CT OF THE HEART

CONTEMPORARY CARDIOLOGY
CHRISTOPHER P. CANNON, MD
SERIES EDITOR
CT of the Heart: Principles and Applications, edited by Platelet Glycoprotein IIb/IIIa Inhibitors in Cardiovascular
U. Joseph Schoepf, MD, 2005 Disease, Second Edition, edited by A. Michael
Principles of Molecular Cardiology, edited by Marschall S. Lincoff, MD, 2003
Runge, MD and Cam Patterson, MD, 2005 Heart Failure: A Clinicians Guide to Ambulatory
Heart Disease Diagnosis and Therapy: A Practical Diagnosis and Treatment, edited by Mariell L. Jessup,
Approach, Second Edition, edited by Gabriel M. MD and Evan Loh, MD, 2003
Khan, MD, FRCP, FRCP (C), FACC, 2005 Management of Acute Coronary Syndromes, Second
Cardiovascular Genomics: Gene Mining for Edition, edited by Christopher P. Cannon, MD 2003
Pharmacogenomics and Gene Therapy, edited by Aging, Heart Disease, and Its Management: Facts and
Mohan K. Raizada, PhD, Julian F. R. Paton, PhD, Controversies, edited by Niloo M. Edwards, MD,
Michael J. Katovich, PhD, and Sergey Kasparov, MD, Mathew S. Maurer, MD, and Rachel B. Wellner, MPH,
PhD, 2005 2003
Surgical Management of Congestive Heart Failure, edited Peripheral Arterial Disease: Diagnosis and Treatment,
by James C. Fang, MD and Gregory S. Couper, MD, edited by Jay D. Coffman, MD and Robert T.
2005 Eberhardt, MD, 2003
Cardiopulmonary Resuscitation, edited by Joseph P. Cardiac Repolarization: Bridging Basic and Clinical
Ornato, MD, FACP, FACC, FACEP and Mary Ann Peberdy, Science, edited by Ihor Gussak, MD, phd, Charles
MD, FACC, 2005 Antzelevitch, phd, Stephen C. Hammill, MD, Win K.
Cardiac Transplantation: The Columbia University Shen, MD, and Preben Bjerregaard, MD, DMSC, 2003
Medical Center/New York-Presbyterian Hospital Essentials of Bedside Cardiology: With a Complete Course
Manual, edited by Niloo M. Edwards, MD, Jonathan in Heart Sounds and Murmurs on CD, Second Edition,
M. Chen, MD, and Pamela A. Mazzeo, 2004 by Jules Constant, MD, 2003
Heart Disease and Erectile Dysfunction, edited by Robert Primary Angioplasty in Acute Myocardial Infarction,
A. Kloner, MD, PhD, 2004 edited by James E. Tcheng, MD, 2002
Coronary Disease in Women: Evidence-Based Diagnosis Cardiogenic Shock: Diagnosis and Treatment, edited by
and Treatment, edited by Leslee J. Shaw, PhD and Rita David Hasdai, MD, Peter B. Berger, MD, Alexander
F. Redberg, MD, FACC, 2004 Battler, MD, and David R. Holmes, Jr., MD, 2002
Complementary and Alternative Cardiovascular Medicine, Management of Cardiac Arrhythmias, edited by Leonard I.
edited by Richard A. Stein, MD and Mehmet C. Oz, MD, Ganz, MD, 2002
2004 Diabetes and Cardiovascular Disease, edited by Michael T.
Nuclear Cardiology, The Basics: How to Set Up and Johnstone, md and Aristidis Veves, MD, DSC, 2001
Maintain a Laboratory, by Frans J. Th. Wackers, MD, Blood Pressure Monitoring in Cardiovascular Medicine and
PhD, Wendy Bruni, BS, CNMT, and Barry L. Zaret, MD, Therapeutics, edited by William B. White, MD, 2001
2004 Vascular Disease and Injury: Preclinical Research, edited
Minimally Invasive Cardiac Surgery, Second Edition, by Daniel I. Simon, MD, and Campbell Rogers, MD,
edited by Daniel J. Goldstein, md, and Mehmet C. Oz, 2001
MD 2004 Preventive Cardiology: Strategies for the Prevention and
Cardiovascular Health Care Economics, edited by William Treatment of Coronary Artery Disease, edited by
S. Weintraub, MD, 2003 JoAnne Micale Foody, MD, 2001
CT OF THE HEART
PRINCIPLES AND APPLICATIONS

EDITED BY

U. JOSEPH SCHOEPF, MD
DEPARTMENT OF RADIOLOGY,
MEDICAL UNIVERSITY OF SOUTH CAROLINA,
CHARLESTON, SC

FOREWORD BY
ALEXANDER R. MARGULIS, MD, DSC (HON)
CLINICAL PROFESSOR OF RADIOLOGY,
WEILL MEDICAL COLLEGE OF CORNELL UNIVERSITY,
NEW YORK, NY
 2005 Humana Press Inc.
999 Riverview Drive, Suite 208
Totowa, New Jersey 07512
www.humanapress.com

For additional copies, pricing for bulk purchases, and/or information about other Humana titles, contact Humana at the above address or at any of the following
numbers: Tel.: 973-256-1699; Fax: 973-256-8341; E-mail:humana@humanapr.com; Website: humanapress.com

All rights reserved.

No part of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying,
Microfilming, recording, or otherwise without written permission from the Publisher.

All articles, comments, opinions, conclusions, or recommendations are those of the author(s), and do not necessarily reflect the views of the publisher.

Due diligence has been taken by the publishers, editors, and authors of this book to ensure the accuracy of the information published and to describe generally
accepted practices. The contributors herein have carefully checked to ensure that the drug selections and dosages set forth in this text are accurate in accord with
the standards accepted at the time of publication. Notwithstanding, as new research, changes in government regulations, and knowledge from clinical experience
relating to drug therapy and drug reactions constantly occurs, the reader is advised to check the product information provided by the manufacturer of each drug
for any change in dosages or for additional warnings and contraindications. This is of utmost importance when the recommended drug herein is a new or
infrequently used drug. It is the responsibility of the health care provider to ascertain the Food and Drug Administration status of each drug or device used in
their clinical practice. The publisher, editors, and authors are not responsible for errors or omissions or for any consequences from the application of the
information presented in this book and make no warranty, express or implied, with respect to the contents in this publication.

Cover illustration: figure 6D from chapter 25, CT Angiography for Assessment of Coronary Artery Anomalies, by Steffen C. Froehner, Matthias Wagner, Juergen
Brunn, and Rainer R. Schmitt; figures 13 and 14 from chapter 3, Scan Techniques for Cardiac and Coronary Artery Imaging With Multislice CT, by Bernd M.
Ohnesorge, Brian R. Westerman, and U. Joseph Schoepf; figure 13 from chapter 4, Image Reconstruction for ECG-Triggered and ECG-Gated Multislice CT,
by Thomas Flohr and Tinsu Pan; figure 13 from chapter 24, Visualization Techniques for Contrast-Enhanced CT Angiography of Coronary Arteries, by Jean-
Louis Sablayrolles and Pascal Giat; and figure 9 from chapter 18, Multidetector-Row CT Assessment of Left-Ventricular Function, by Kai Uwe Juergens and
Roman Fischbach.

Production Editor: Tracy Catanese

Cover design by Patricia F. Cleary

This publication is printed on acid-free paper.

ANSI Z39.48-1984 (American National Standards Institute) Permanence of Paper for Printed Library Materials.

Photocopy Authorization Policy:

Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted by Humana Press Inc., provided that
the base fee of US $25.00 per copy is paid directly to the Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that
have been granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana Press Inc. The fee code for
users of the Transactional Reporting Service is: [1-58829-303-3/05 $25.00].

e-ISBN: 1-59259-818-8

Printed in the United States of America. 10 9 8 7 6 5 4 3 2 1

Library of Congress Cataloging-in-Publication Data

CT of the heart : principles and applications / edited by U. Joseph Schoepf.

p. ; cm. -- (Contemporary cardiology)
Includes bibliographical references and index.
ISBN 1-58829-303-3 (alk. paper)
1. Heart--Tomography.
[DNLM: 1. Heart--radiography. 2. Tomography, X-Ray Computed. WG 141.5.T6 C959 2004] I. Schoepf, U. J. (U. Joseph), 1969- II. Series: Contemporary
cardiology (Totowa, N.J. : Unnumbered)
RC683.5.T66C785 2004
616.1'20757--dc22
2004003443
 Foreword

Radiologic technology has made dramatic advances in The entire contents are meticulous and comprehensive,
the last 25 years, and none have been more impressive than from the introduction about the past, present, and future
those in computed tomography (CT). The progress in the of CT of the heart, through the technical underpinning of
speed of obtaining images, computing, postprocessing, the method and the various clinical, physiologic, and
and spatial resolution has been incredible. The result is pathologic applications of CT in studying the heart.
that CT has moved from displaying purely morphologic This book fills an immense need, particularly at a time
information to providing valuable physiologic data as well. when cardiac screening with CT, whether one agrees with
Whether with electron beam or multidetector-row CT, this practice or not, is a reality. Furthermore, with the
advances are impressive and nowhere have the applica- rapid increase of aging populations in the industrialized
tions been more useful and dramatic than in the heart. world, noninvasive diagnostic approaches are increas-
This multiauthored book, CT of the Heart, edited by U. ingly needed. As technology continues to advance and
Joseph Schoepf, MD, is a splendid rendition of the state-of- applications of CT to heart studies expand, it is my hope
the-art in CT imaging of the heart; however, where appro- that the editor will bring this book up to date with a new
priate, it also features comparisons with other technical edition.
approaches, such as magnetic resonance and ultrasound.
The contributors are leading radiologists, cardiologists, Alexander R. Margulis, MD, DSc (HON)
physicists, engineers, and basic and clinical scientists Clinical Professor of Radiology
from Europe, the United States, Israel, and Japan. Weill Medical College of Cornell University

v
 Preface

Through the ages of exploration and enlightenment the ease burden is a promising and exciting but yet untested
heart has kept its fascination as the metaphor of life. Its concept.
firm entrenchment in human emotion and consciousness Within CT of the Heart, the reader will be exposed to
and the overwhelming socioeconomic importance of its a variety of expert opinions on the respective topics. Some
diseases make noninvasive visualization and diagnosis of authors will assume a more optimistic or a more conser-
the heart and its diseases the coveted holy grail of medi- vative perspective on cardiac CT applications. Because of
cal imaging. the lack of large-scale clinical studies, only future expe-
Imaging of the heart has always been technically chal- rience will show who may be right. It is a declared goal of
lenging, because of the hearts continuous motion. The this book to showcase the full scope of current develop-
introduction and ongoing technical improvement of fast ments, research, and scientific controversy regarding the
ECG-synchronized computed tomography (CT) scanning principles and applications of CT of the heart. Truth is
of the heart has enabled imaging of the elusive but cardi- most likely to be found in the equilibrium of opinions. The
nal cardiac anatomy and pathology with a combination of publisher and I have striven to maintain this equilibrium
speed and spatial resolution that is hitherto unparalleled by providing a platform for differing opinions and for
by other noninvasive imaging modalities. Accordingly, different technical approaches to CT of the heart. To miti-
considerable interest has been directed in recent years at gate commercial overtones and bias, which so often
the beneficial utilization of CT for noninvasive interroga- accompany the first steps of a potentially important new
tion of the coronary arteries and for imaging studies of technology, contributions of users and/or developers of
anatomic and functional sequelae of ischemic heart dis- all cardiac CT manufacturers were included. Scientists
ease, such as cardiac perfusion, motion, and viability. The representing different companies graciously disregarded
current and potential future roles of CT for these and other commercial divisions and agreed to co-author chapters in
applications are the subject matter of this book. order to provide the reader with a truly balanced view on
CT of the Heart, however, does not claim to have all of cardiac CT technology.
the answers. CT of the heart is a nascent but rapidly evolv- Accordingly, CT of the Heart is the work of many. I am
ing field, and the act of condensing expert knowledge and indebted to Dr. Christopher Cannon, the series editor, and
experience in the format of a book can only result in a to Paul Dolgert of Humana Press for entrusting me with
snapshot of the status quo at a certain point in time. Novel the role of editor. I am grateful to my chairman, Dr. Philip
iterations of existing technology and profoundly new con- Costello, for his unfailing guidance, support, and friend-
cepts of medical imaging are already on the horizon. The ship. I feel very, very honored by all the kindness that my
benefits and indications of integrating CT into the diag- many friends in the cardiac imaging community have
nostic algorithm of heart disease is intensely researched shown me by volunteering their time, their knowledge,
and discussed: to date, the diagnostic value of CT coro- their experience, and the vision that went into their respec-
nary calcium measurements and the exact role of this tive contributions. All authors are highly respected
marker for cardiac risk stratification remain unclear and experts in their fields and this book would never have
controversial. We are only beginning to understand the come to pass without their incredible support, for which I
am so grateful. Finally I would like to thank Tracy Catanese
usefulness and potential clinical application of CT
and Craig Adams of Humana Press for so efficiently and
angiography for noninvasive detection of coronary artery
expertly steering the production of CT of the Heart.
stenosis. Cross-sectional assessment of the coronary
artery wall for noninvasive identification, characteriza- U. Joseph Schoepf, MD
tion, and quantification of atherosclerotic lesions and dis- Charleston, South Carolina

vii
 Contents

FOREWORD ..................................................................... v 9 Detection and Quantification of Coronary

PREFACE ...................................................................... vii Calcium With Electron Beam CT .................. 83
Axel Schmermund, Stefan Mhlenkamp,
CONTRIBUTORS ............................................................. xi and Raimund Erbel
PART I INTRODUCTION AND HISTORICAL 10 Detection and Quantification of Coronary
BACKGROUND Calcium With Dual-Slice CT ........................ 91
Joseph Shemesh
1 CT of the Heart: Past, Present, and Future ......... 3 11 Detection and Quantification
William Stanford of Calcified Coronary Plaque
With Multidetector-Row CT........................ 101
PART II TECHNICAL BACKGROUND J. Jeffrey Carr
2 Electron Beam CT of the Heart ......................... 15 12 Coronary Calcium Scoring
David G. Hill With Multidetector-Row CT:
3 Scan Techniques for Cardiac and Coronary Rationale and Scoring Techniques .............. 111
Artery Imaging With Multislice CT .............. 23 Roman Fischbach and David Maintz
Bernd M. Ohnesorge, Brian R. Westerman, 13 Noninvasive Quantification of Coronary
and U. Joseph Schoepf Calcium: Quantification Methods,
4 Image Reconstruction for ECG-Triggered Scanner Types, Scan Protocols,
and ECG-Gated Multislice CT ...................... 45 Accuracy, and Reproducibility .................... 129
Thomas Flohr and Tinsu Pan Stefan Ulzheimer, Kaiss Shanneik,
5 Phase-Correlated Image Reconstruction and Willi A. Kalender
Without ECG .................................................. 55
Marc Kachelriess and Willi A. Kalender PART IV CONTRAST-ENHANCED CT OF THE HEART:
6 Radiation Dose From CT of the Heart ............... 63
MORPHOLOGY AND FUNCTION
Cynthia H. McCollough 14 CT of the Pericardium ...................................... 145
Reinhard Groell
PART III DETECTION AND QUANTIFICATION 15 Multidetector-Row CT for Assessment
OF CORONARY CALCIUM of Cardiac Valves ......................................... 155
7 Coronary Calcium Screening: Jrgen K. Willmann and Dominik Weishaupt
An Epidemiologic Perspective ....................... 71 16 Multislice CT of the Heart and Great Vessels
Christopher J. ODonnell and Udo Hoffmann in Congenital Heart Disease Patients .......... 161
8 Coronary Calcium Scanning: Jean-Franois Paul
Why We Should Perform It ............................ 79 17 Imaging of Cardiac and Paracardiac Masses
Axel Schmermund, Stefan Mhlenkamp, and Pseudotumors ........................................ 171
and Raimund Erbel Bernd J. Wintersperger

ix
x CONTENTS

18 Multidetector-Row CT Assessment 29 Contrast-Enhanced Electron Beam

of Left-Ventricular Function ....................... 183 CT and Multidetector-Row CT
Kai Uwe Juergens and Roman Fischbach in the Evaluation of Coronary
19 Imaging Intramyocardial Microcirculatory Stent Patency ................................................ 311
Function Using Fast Computed Heiko Pump, Stefan Mhlenkamp,
Tomography ................................................. 195 Raimund Erbel, and Rainer Seibel
Stefan Mhlenkamp, Axel Schmermund, 30 CT Angiography for the Detection
Birgit Kantor, Raimund Erbel, of Coronary Artery Stenosis ........................ 321
and Erik L. Ritman Koen Nieman and Filippo Cademartiri
20 Approaches for Assessing Myocardial 31 Complementary Use of Coronary Calcium
Viability With Multidetector-Row CT ........ 207 Scoring and CT Angiography ...................... 333
Yasushi Koyama and Teruhito Mochizuki Alexander W. Leber
32 CT vs Magnetic Resonance for Imaging
PART V CONTRAST-ENHANCED CT OF THE HEART: of the Coronary Arteries .............................. 339
CORONARY ARTERIES Armin Huber
21 Anatomy of the Coronary Arteries
and Veins in CT Imaging ............................. 219 PART VI CONTRAST-ENHANCED CT OF THE HEART:
Robert J. M. van Geuns PRINCIPLES OF ATHEROSCLEROSIS AND VESSEL
and Filippo Cademartiri WALL IMAGING
22 Techniques and Protocols for Acquisition 33 Pathology and Pathophysiology of Coronary
and Display of Contrast-Enhanced CT Atherosclerotic Plaques ............................... 351
Angiography ................................................. 229 Renu Virmani, Allen P. Burke,
Christoph R. Becker Frank D. Kolodgie, Andrew Farb,
23 Contrast Material Injection Techniques for CT Aloke V. Finn, and Herman Gold
Angiography of the Coronary Arteries .......... 237 34 Pathogenesis of the Vulnerable
Filippo Cademartiri and Koen Nieman Atherosclerotic Plaque ................................. 365
24 Visualization Techniques for Contrast- Masanori Aikawa
Enhanced CT Angiography 35 Multidetector-Row CT Imaging of Clinical
of Coronary Arteries .................................... 247 and Preclinical Coronary Atherosclerosis ... 377
Jean-Louis Sablayrolles and Pascal Giat Christoph R. Becker
25 CT Angiography for Assessment of Coronary 36 Multidetector-Row CT vs Intravascular
Artery Anomalies .......................................... 259 Ultrasound for Coronary Plaque
Steffen C. Froehner, Matthias Wagner, Characterization ........................................... 381
Juergen Brunn, and Rainer R. Schmitt Axel Kuettner
26 Multidetector-Row CT for Assessment 37 Multidetector-Row CT vs Magnetic
of Kawasaki Disease .................................... 279 Resonance Imaging for Coronary
Toru Sakuma and Kunihiko Fukuda Plaque Characterization ............................... 389
27 Multidetector-Row CT of the Coronary Konstantin Nikolaou, Christoph R. Becker,
Arteries for Planning of Minimally and Zahi Fayad
Invasive Bypass Surgery .............................. 287 38 Multidetector-Row CT for Detection
Christopher Herzog, Selami Dogan, and of Noncalcified and Calcified Coronary
Thomas J. Vogl Lesions: Clinical Significance ..................... 399
28 CT Angiography for Assessment of Coronary Stephen Schroeder
Bypass Grafts .................................................. 301
Marcello De Santis INDEX ......................................................................... 401
 Contributors

MASANORI AIKAWA, MD, PhD, Cardiovascular STEFFEN C. FROEHNER, MD, Department of

Division, Department of Medicine, Brigham and Diagnostic and Interventional Radiology, Herz- und
Womens Hospital and Harvard Medical School, Gefaess-Klinik GmbH, Bad Neustadt an der Saale,
Boston, MA Germany
CHRISTOPH R. BECKER, MD, Department of Clinical KUNIHIKO FUKUDA, MD, Department of Radiology,
Radiology, University of Munich, Munich, Germany The Jikei University School of Medicine, Tokyo,
JUERGEN BRUNN, MD, Department of Diagnostic and Japan
Interventional Radiology, Herz- und Gefaess-Klinik PASCAL GIAT, PhD, Division of Computed
GmbH, Bad Neustadt an der Saale, Germany Tomography, General Electric Medical Systems,
ALLEN P. BURKE, MD, Department of Cardiovascular Buc, France
Pathology, Armed Forces Institute of Pathology, HERMAN GOLD, MD, Cardiac Unit, Department of
Washington, DC Internal Medicine, Massachusetts General Hospital,
FILIPPO CADEMARTIRI, MD, Department of Boston, MA
Radiology, Erasmus Medical Center, Rotterdam, REINHARD GROELL, MD, Department of Radiology,
The Netherlands Medical University Graz, Graz, Austria
J. JEFFREY CARR, MD, MSCE, Division of CHRISTOPHER HERZOG, MD, Institute for Diagnostic
Radiological Sciences and Public Health Sciences, and Interventional Radiology, J. W. Goethe
Wake Forest University School of Medicine, University, Frankfurt, Germany
Winston-Salem, NC DAVID G. HILL, PhD, General Electric Medical
MARCELLO DE SANTIS, MD, Department of Systems, South San Francisco, CA
Radiology, S. Andrea Hospital, Rome, Italy UDO HOFFMANN, MD, Department of Radiology,
SELAMI DOGAN, MD, Department of Thoracic and Massachusetts General Hospital and Harvard
Cardiovascular Surgery, J. W. Goethe University, Medical School, Boston, MA
Frankfurt, Germany ARMIN HUBER, MD, Department of Clinical
RAIMUND ERBEL, MD, Department of Cardiology, Radiology, University of Munich, Munich, Germany
University Clinic Essen, Essen, Germany KAI UWE JUERGENS, MD, Department of Clinical
ANDREW FARB, MD, Department of Cardiovascular Radiology, University of Muenster, Muenster,
Pathology, Armed Forces Institute of Pathology, Germany
Washington, DC MARC KACHELRIESS, PhD, Institute of Medical
ZAHI FAYAD, MD, Zena and Michael A. Wiener Physics, University of Erlangen-Nrnberg,
Cardiovascular Institute, Mount Sinai School of Nrnberg, Germany
Medicine, New York, NY WILLI A. KALENDER, PhD, Institute of Medical
ROMAN FISCHBACH, MD, Department of Clinical Physics, University of Erlangen-Nrnberg,
Radiology, University of Muenster, Muenster, Nrnberg, Germany
Germany BIRGIT KANTOR, MD, Division of Cardiovascular
ALOKE V. FINN, MD, Cardiac Unit, Department of Diseases and Internal Medicine, Mayo Clinic
Internal Medicine, Massachusetts General Hospital, College of Medicine, Rochester, MN
Boston, MA FRANK D. KOLODGIE, PhD, Department of
THOMAS FLOHR, Division of Computed Tomography, Cardiovascular Pathology, Armed Forces Institute
Siemens Medical Solutions, Forchheim, Germany of Pathology, Washington, DC

xi
xii CONTRIBUTORS

YASUSHI KOYAMA, MD, Department of Cardiology, TORU SAKUMA, MD, Department of Radiology, The
Ehime Prefectural Imabari Hospital, Ehime, Japan Jikei University School of Medicine, Tokyo, Japan
AXEL KUETTNER, MD, Department of Radiology, AXEL SCHMERMUND, MD, Department of Cardiology,
University of Tbingen, Tbingen, Germany University Clinic Essen, Essen, Germany
ALEXANDER W. LEBER, MD, Medizinische Klinik I, RAINER R. SCHMITT, MD, Department of Diagnostic
University Hospital Clinic Munich-Grosshadern, and Interventional Radiology, Herz- und Gefaess-
Ludwig Maximilians University, Munich, Germany Klinik GmbH, Bad Neustadt an der Saale, Germany
CYNTHIA H. MCCOLLOUGH, PhD, Department of U. JOSEPH SCHOEPF, MD, Department of Radiology,
Radiology, Mayo Clinic College of Medicine, Medical University of South Carolina, Charleston, SC
Rochester, MN STEPHEN SCHROEDER, MD, PhD Division of
DAVID MAINTZ, MD, Department of Clinical Cardiology, Eberhard-Karls-Universitt, Tbingen,
Radiology, University of Muenster, Muenster, Germany
Germany RAINER SEIBEL, MD, Institute of Diagnostic and
ALEXANDER R. MARGULIS, MD, DSc (HON), Interventional Radiology, University of Witten-
Department of Radiology, Weill Medical College of Herdecke, Witten-Herdecke, Germany
Cornell University, New York, NY KAISS SHANNEIK, MSc, Institute of Medical Physics,
TERUHITO MOCHIZUKI, MD, Department of University of Erlangen-Nrnberg, Nrnberg,
Radiology, Ehime University School of Medicine, Germany
Ehime, Japan JOSEPH SHEMESH, MD, Department of Cardiology,
STEFAN MHLENKAMP, MD, Department of The Grace Ballas Cardiac Research Unit, Sheba
Cardiology, University Clinic Essen, Essen, Medical Center, Sackler School of Medicine, Tel
Germany Aviv, Israel
KOEN NIEMAN, MD, Department of Cardiology WILLIAM STANFORD, MD, Division of Chest and
(Thoraxcenter) and Department of Radiology, Cardiovascular Radiology, Department of
Erasmus Medical Center, Rotterdam, The Radiology, Ray J. and Lucille A. Carver University
Netherlands of Iowa College of Medicine, Iowa City, IA
KONSTANTIN NIKOLAOU, MD, Department of Clinical STEFAN ULZHEIMER, PhD, Division of Computed
Radiology, University of Munich, Munich, Germany, Tomography, Siemens Medical Solutions,
and Zena and Michael A. Wiener Cardiovascular Forchheim, Germany
Institute, Mount Sinai School of Medicine, New ROBERT J. M. VAN GEUNS, MD, Department of
York, NY Radiology, Erasmus Medical Center, Rotterdam,
CHRISTOPHER J. ODONNELL, MD, MPH, National The Netherlands
Heart, Lung, and Blood Institutes Framingham RENU VIRMANI, MD, Department of Cardiovascular
Heart Study, Framingham, MA, and Cardiology Pathology, Armed Forces Institute of Pathology,
Division, Department of Medicine, Massachusetts Washington, DC
General Hospital and Harvard Medical School, THOMAS J. VOGL, MD, Institute for Diagnostic and
Boston, MA Interventional Radiology, J. W. Goethe University,
BERND M. OHNESORGE, PhD, Division of Computed Frankfurt, Germany
Tomography, Siemens Medical Solutions, MATTHIAS WAGNER, MD, Department of Diagnostic
Forchheim, Germany and Interventional Radiology, Herz- und Gefaess-
TINSU PAN, General Electric Medical Systems, Klinik GmbH, Bad Neustadt an der Saale, Germany
Waukesha, WI DOMINIK WEISHAUPT, MD, Institute of Diagnostic
JEAN-FRANOIS PAUL, MD, Radiology Unit, Marie Radiology, University Hospital Zurich, Zurich,
Lannelongue Hospital, Plessis Robinson, France Switzerland
HEIKO PUMP, MD, Institute of Diagnostic and BRIAN R. WESTERMAN, PhD, Division of Computed
Interventional Radiology, University of Witten- Tomography, Toshiba Medical Solutions, Los
Herdecke, Witten-Herdecke, Germany Angeles, CA
ERIK L. RITMAN, MD, PhD Physiological Imaging JRGEN K. WILLMANN, MD, Institute of Diagnostic
Research Laboratory, Department of Physiology and Radiology, University Hospital Zurich, Zurich,
Biomedical Engineering, Mayo Clinic College of Switzerland
Medicine, Rochester, MN BERND J. WINTERSPERGER, MD, Department of
JEAN-LOUIS SABLAYROLLES, MD, Centre Clinical Radiology, University of Munich, Munich,
Cardiologique du Nord, Saint Denis, France Germany
 CHAPTER 1 / PAST, PRESENT, FUTURE OF CT 1

INTRODUCTION
I
AND HISTORICAL BACKGROUND
 CHAPTER 1 / PAST, PRESENT, FUTURE OF CT 3

1 CT of the Heart
Past, Present, and Future

WILLIAM STANFORD, MD

INTRODUCTION patient was scanned (Fig. 2A). Scan times of 4.5 min per image
Imaging of the heart and great vessels has previously been were required.
done with plain film, cardiac catheterization, nuclear medicine, The technology was then advanced with the addition of fan
and echocardiography as the primary imaging modalities. The beam architecture to the translate/rotate process. With this
recent newer advances in the computed tomography (CT) and refinement, each radiation beam activated multiple detectors
magnetic resonance imaging (MRI) technologies, however, rather than a single detector. Thus increased numbers of images
have dramatically changed our approach to imaging cardiac were possible from each tube activation. With this upgrade,
disease. CT and MRI, supplemented by CT angiography and imaging times were reduced to approx 2.5 min per image (Fig. 2B).
MRI angiography, are increasingly replacing the chest film, as The next advance was the introduction of a continuously
well as nuclear andto some extentecho imaging as the rotating X-ray tube coupled with a continuously rotating detec-
primary modalities in evaluating heart disease. tor array. This further decreased scan times to approx 18 s
CT images, which initially took over 4 min to generate, can (Fig. 2C).
now be obtained in 50100 ms with electron beam imaging Around 1978, the rotating detector array concept was changed
and in 125500 ms with helical imaging. Importantly, these to that of a fixed detector array, and this further decreased scan
advances in temporal resolution are rapid enough to essentially times to approx 2 s per image, and this is the configuration
stop cardiac motion, and thus visualization of extremely small present in many of our conventional CT scanners in operation
structures such as calcium deposits within the walls of the coro- today (Fig. 2D). As a consequence of these advances, CT was
nary arteries is now possible. Along with this increased tempo- able to become a major workhorse for whole-body imaging,
ral resolution has come a concomitant increase in spatial which included many cardiac applications. However, although
resolution, and isotropic voxels as small as 0.5 mm3 are now the excellent resolution and absence of overlying structures
identifiable. These advances, along with the increasing use of allowed visualization of the pericardium and many of the rela-
3D reconstruction techniques, have revolutionized cardiac tively static abnormalities such as intracardiac filling defects
imaging and have moved CT from only anatomic visualization from thrombi and tumors, the contracting heart was not well
into the arena of functional and perfusion imaging. visualized.
An additional problem with this conventional technology
HISTORICAL was that the scanner cabling restricted tube movement and al-
CT was first introduced by Sir Godfrey Hounsfield in the lowed for only a single tube rotation to a fixed point before the
1970sa short 30 yr ago (1,2) (Fig. 1). Hounsfield was an tube had to be returned to its original position. Thus a continu-
electrical engineer working for EMI, an electronics firm in ous tube movement was not possible and scan times were rela-
England. While there, he conceived the idea of taking cross- tively fixed. Because of this, cardiac applications were
sectional X-ray data and reformatting these data into images. restricted to a 2-s-per-image acquisition time. Yet, the ability
For this, he and Alan M. Cormack, a Tufts professor of math- to visualize not only the vessel lumina and cardiac chamber
ematics working independently, received the Nobel Prize in endocardium but also the vessel wall and the surrounding
medicine in 1979. myocardium were important applications not heretofore pos-
Hounsfields first-generation scanner used a translate/rotate sible. Thus CT provided a more comprehensive view of cardio-
technology. In this methodology, an X-ray source moving lat- vascular pathology.
erally activated a series of single detectors before moving to In the early 1980s, an important advance moved CT into the
another position and repeating the process. This translate/rotate realm of cardiac imaging. This was the introduction of the elec-
process was repeated until the entire circumference of the tron beam technology concept by Dr. Douglas Boyd of the Uni-
versity of California, San Francisco (3). The electron beam
From: Contemporary Cardiology: CT of the Heart: scanner, while having an appearance similar to a conventional
Principles and Applications CT scanner, did not have an X-ray tube rotating around the
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ patient. Instead, electrons were generated by a source and then

3
4 STANFORD

Fig. 1. Godfrey N. Hounsfield, the father of computed tomography.

For their achievements, Hounsfield and Alan M. Cormack were
awarded the Nobel Prize in medicine in 1979.

Fig. 2. (A) Diagram of the first-generation translate/rotate CT scan-

ner. To create the image, a series of exposures were taken as the
X-ray source moved a short distance laterally. The tube then rotated
to a different position and the sequence was repeated. A single image
required 4.5 min exposure time. (B) A later modification of the trans-
late/rotate sequence introduced fan beam architecture to activate
multiple detectors before moving to a new position. With this modi-
fication, scan times were reduced to 2.5 min/image. (C) The next
generation of scanners used a rotating X-ray source coupled to rotat-
ing detectors. Scan times were reduced to 18 s/image. (D) Subsequent
modifications used a rotating X-ray source with a fixed detector sarray.
Scan times were 2 s/image.
 CHAPTER 1 / PAST, PRESENT, FUTURE OF CT 5

Fig. 3. Cross-sectional diagram of the GE-Imatron C150 XP electron beam scanner. Electron sweeps of tungsten target rings generate X-rays
that traverse the patient to activate detectors in the gantry over the patient. With this development, imaging times were reduced to 50100 ms.
(Courtesy GE-Imatron Inc.)

bent electromagnetically to sweep tungsten target rings located THE PRESENT

in the gantry beneath the patient. The X-rays produced by the CONVENTIONAL CT
electron sweep traversed the patient and were collected by The design of our current conventional CT scanners is that
fixed solid-state detectors located in the gantry above the patient of a rotating X-ray source activating fixed detectors in the gan-
(Fig. 3). This technology decreased scan times to 50100 ms, try surrounding the patient. The tube movement is limited by
which essentially froze cardiac motion and thus dramatically cables so that once the tube rotates around the patient, it has to
changed our ability to image the beating heart. For the first stop and rewind. This constrained tube travel is a major limita-
time it was possible to view cardiac contractions and to visualize tion of conventional CT; however, the cross-sectional image
small structures such as calcium deposits within the walls of the detail is excellent and spatial resolutions of 912 line pairs (lp)/
coronary artery. cm are possible.
An additional major advance in CT imaging came in the early ELECTRON BEAM CT
1990s with the introduction of helical/spiral CT imaging and its The electron beam CT (EBCT) scanner (GE-Imatron) works
slip-ring technology (4) (Fig. 4). With these advances, the X-ray on a different principle than does conventional CT. With EBCT,
beam was able to continuously rotate around the patient as the electrons sweep tungsten target rings to produce X-rays that
patient moved through the scanner gantry. These innovations enter from beneath the patient, and are attenuated and collected
decreased scan times to approx 5001000 ms and ultimately pro- by solid-state detectors in the gantry above the patient. The
duced data sets with spatial resolutions as small as 0.5 mm3. The technology, because of the absence of moving parts, dramati-
initial platform for helical CT scanners consisted of a single X-ray cally decreases scan times to 50100 ms, which is rapid enough
source and a single detector ring; however, subsequent develop- to essentially freeze cardiac motion.
ments in helical CT technology have enlarged imaging platforms There are several different EBCT scanner models in
from a single detector to 4 rows and now 16 rows, and in the future operation. The older C150 and C300 scanners operate at 130
32-, 64-, and 256-row detectors are expected. Along with the kV, 625 mA, 83 kW to produce 100-ms, 1.5-, 3-, and 6-mm
increase in numbers of detector rows were advances in the tem- slice thickness images at resolutions of 9.5 lp/cm. The technol-
poral resolution, and now imaging times now as fast as 420 ms are ogy also makes it possible to generate flow and movie mode
possible. These and other innovations in multislice helical CT 50-ms, 8-mm dual slice thickness images from four target rings.
have allowed the entire heart to be scanned within a 2030 s The latter ability is important in functional imaging; its spatial
breath-hold, and do so with excellent spatial resolution (5). resolution is limited at 4.5 lp/cm.
6 STANFORD

Fig. 4. A modern helical CT imaging suite.

The newer EBCT e-Speed version operates at 140 kV, 1012 same-level images, is triggered off the R wave of the
1000 mA, 140 kW. It can generate dual 1.5-, 3-, and 8-mm ECG. Once completed, it then takes 8 ms for the scanner to
slices, at 50 ms temporal resolution. The spatial resolution is 10 reset, and another 1012 image data set at the same or usually
lp/cm. This version also has a 100-ms mode for higher spatial a different level can be obtained. The spatial resolution in the
resolution and a 33-ms mode for higher temporal resolution. flow and movie modes is moderate at 4.5 lp/cm; however, with
Both are significant advances in EBCT technology. the new e-Speed model, resolutions of 10 lp/cm are possible.
Flow Mode Movie sequences are important in quantifying cardiac function
EBCT uses several sequences to image the heart. In the flow and in evaluating abnormal contraction patterns, such as those
mode sequence, the C150 and C300 scanners generate eight seen post myocardial infarction. Also, by completely opacify-
8-mm images in 224 ms. Then, following an 8-ms delay to reset ing the cardiac chambers, filling defects such as thrombi or
the beam, an additional eight 8-mm slices can be generated in tumors are readily identified.
another 224 ms. The advantage of this configuration is that a Step Volume and Continuous Volume Scans
stack of multilevel images can be obtained almost simulta- Lastly, in the step volume scan (SVS) and continuous vol-
neously, thus allowing visualization of a contrast bolus as it ume scan (CVS) modes, 1.5-, 3-, or 6-mm single-slice thick-
enters, peaks, and washes out of a region of interest (Fig. 5). nesses are obtainable. In the SVS sequence, the temporal
This sequence is similar to the first-pass studies used in nuclear resolution of the C150 and C300 platforms is 100 ms, which is
medicine. Flow mode sequences are useful in evaluating coro- 9 images per second, with the pixel sizes varying from 0.06 to
nary artery bypass graft patency, intracardiac shunts, and arte- 1 mm2. With the e-Speed, 50-ms volume scan sequences at 17
riovenous malformations, as well as identifying bolus arrival or 34 images per second are possible, depending upon whether
times in aberrant vessels. one or two detectors are used. The SVS images can be triggered
Movie Mode at preset intervals from the ECG signal. The maximal resolu-
A more commonly used EBCT sequence is the movie (cine) tion is 10 lp/cm, depending upon the field of view and the
sequence. In this configuration, the C150 and C300 scanners reconstructive algorithm selected. In the SVS mode, images are
can generate images of the contracting heart every 58 ms (17 acquired in a manner similar to that of conventional CT scan-
images per second). These images can be acquired during sys- nersi.e., a single target ring is swept to produce an image and
tole and diastole, which allows visualization during a cardiac then the table moves and a second image is generated. Multiple
contraction (Fig. 6). This sequence, usually consisting of sweeps of the same target can be taken and volume averaged if
 CHAPTER 1 / PAST, PRESENT, FUTURE OF CT 7

Fig. 5. An electron beam CT flow mode sequence showing a contrast bolus opacifying the right atrium and ventricle, and then washing out
and later appearing in the left ventricle. This sequence is useful in evaluating bypass graft patency, intracardiac shunts, and arteriovenous
malformations, as well as identifying contrast arrival in aberrant vessels.

additional resolution is required; however, this comes at the reduced gantry rotation times from 1 s to 420 ms, and with
price of increased radiation exposure. segmented reconstruction, image times approximating 100 ms
The continuous volume mode (CVS) is similar to that of are possible. An additional advantage of the helical technology
spiral or helical CT methodologies, with up to 140 images on is that the detector information is generated as a volumetric
the C150, up to 280 images on the C300, and 400 images on the data set, and this permits later reformations of different slice
e-Speed scanners possible during an acquisition period of about thickness that allows 3D reconstructions. Helical CT images
33 s. The minimum exposure time is 100 ms (33 ms on e-Speed) are commonly ECG gated, and this further decreases motion
and the images can be contiguous or overlapped. On the C150 artifacts by being able to restrict image acquisition to the quiet
or C300 scanners, this mode allows a complete 140-slice, 3- phase of the cardiac cycle. The latter is particularly important
mm thick data set to be acquired during a 17-second breath- in coronary calcification screening.
hold. Scan widths of 1.510 mm are possible. Helical/Spiral CT Sequence
HELICAL CT In cardiac imaging, helical CT scanners are operated in two
Current-generation helical/spiral CT scanners have multirow basic modes, both gated from the patients ECG. The first mode,
detectors capable of generating 1 to 16 slices of varying thick- designated prospective gated scanning, activates the X-ray tube
ness with each gantry rotation. Technological advances have for only the time needed to acquire a partial image. This is
8 STANFORD

Fig. 7. Electron beam CT long axis (four-chamber) view image show-

ing the left and right atria, left and right ventricles, and left-ventricular
outflow tract.

Retrospective gated scanning is the other operating mode of

helical scanners. This sequence involves the helical scanning
of the entire heart while at the same time recording the patients
ECG. After the scan, images are reconstructed at a preselected
phase of the cardiac cycle. This mode is commonly used for
coronary artery angiography and for cardiac functional analy-
sis. In order to avoid anatomical gaps in the data set, the helical
pitch is set lower than that used for general body imaging. Thus
the dose to the patient is higher. Nevertheless, the ability to
reconstruct images in multiple phases from the same high-reso-
lution data set can provide important information. Submillime-
ter slices, as small as 0.5 mm, can be used to achieve a high
spatial resolution, and cardiac data collection can be accom-
plished in about 30 s. However, the patients heart rate is again
a major factor in maintaining high image quality. If the heart
rate exceeds 70 beats per minute (bpm), segmented image
reconstruction can be used. This algorithm uses data from two,
three, or four consecutive cardiac cycles to reconstruct a single
image, thus improving the temporal resolution significantly
Fig. 6. Electron-beam short axis movie mode images at the mid-left-
ventricular level in diastole (A) and systole (B). This sequence is and extending the range of heart rates that can be scanned with
useful in quantifying cardiac function and evaluating abnormalities of high-quality images.
wall motion.
ADDITIONAL CONSIDERATIONS
IMPORTANCE OF TEMPORAL RESOLUTION
Although conventional, helical, and EBCT are able to iden-
roughly half the gantry rotation time, or as short as 200250 ms. tify cardiac anatomy, motion artifacts can still remain a problem.
The time of data collection is measured from the R wave, and With contrast enhancement, conventional CT can image cardiac
the operator selects this delay. This mode is frequently used for chambers and great vessel anatomy extremely well. However,
coronary artery calcium imaging, because the radiation dose to helical/spiral CT and EBCT, because of their faster scan times,
the patient is kept to a minimum. However, rapid or irregular are suited better for imaging moving structures. In a typical car-
patient heart rates can affect image quality and the reproduc- diac sequence, the superior and inferior vena cavae, pulmonary
ibility of the study. Data collection periods that extend beyond arteries, and aorta are routinely visualized, as are the right and left
diastole may exhibit motion artifacts that can reduce accuracy. atria and ventricles, and their outflow tracts (Fig. 7). With CT, the
A 4-detector ring scanner rotating at 0.5 s and programmed to interfaces between the contrast-enhanced cardiac chambers and
provide 3-mm slices can acquire a data set in about 20 s, a myocardium are usually well defined, especially if the image is
reasonable breath-hold for most patients. Sixteen-row detector acquired during diastole; in systole this interface may be partially
scanners in the same study reduced scanning times to roughly degraded by motion. Overall, both helical/spiral CT and EBCT,
812 s. because of their faster scan times, generally better define struc-
 CHAPTER 1 / PAST, PRESENT, FUTURE OF CT 9

Fig. 8. Heart model showing the cross-sectional planes used in cardiac imaging.

tural detail (5). Although it is important that the patient remain dye, an earlobe densitometer records the arrival of the bolus.
quiet and suspend respiration, often it is still possible with the Other techniques are the sure start technologies, where the
EBCT and helical/spiral CT scanners to obtain satisfactory scanner monitors the rise in contrast attenuation over the area
images, even if the individual is unable to hold his or her breath. of interest. When the attenuation approaches the preselected
CONTRAST CONSIDERATIONS threshold, the scanner is triggered automatically. Alternatively,
In cardiac imaging, contrast administration is often neces- a fixed delay of 1530 s may be used.
sary. The contrast material is usually administered as a bolus or POSITIONAL CONSIDERATIONS
as a continuous infusion, but because of the speed of the scan- Cardiac CT scanning, especially in evaluating cardiac func-
ners, contrast arrival times become critical. Commonly, volumes tion, requires imaging in planes less familiar to radiologists.
of 80150 mL of contrast material, varying from 240 to 370 mg The oblique position of the heart requires modifications to our
of iodine per mL, are administered. Generally, the contrast traditional scanning planes. The two configurations commonly
material is infused at a rate of 14 mL/s using a power injector. used in cardiac imaging are the short axis view, wherein the
For contrast administration, timing methodologies become ventricle is bread-loafed (Fig. 8), and the four-chamber
important to ensure contrast optimization. One method is to view, which slices the heart longitudinally from top to bottom
administer a test bolus of 10 mL of contrast material and per- and produces images that are similar to those familiar to our
form repeated imaging over the area of interest to determine echocardiology colleagues (Fig. 7).
contrast bolus arrival time. Another is to perform a circulation RADIATION DOSAGE
time using either cardiogreen dye or a solution of 0.5% magne- The effective dose is the overall radiation exposure to the
sium sulfate. Both are injected at approx 4 mL/s. With magne- patient. This is designated in mSv, and is frequently equated to
sium sulfate, approx 10 mL of 0.5% solution is commonly months of background exposure. For an EBCT calcium study,
administered, with bolus arrival being manifested by a warm the effective dose would be 1.0 mSv in males and 1.3 mSv in
sensation in the back of the tongue or throat. With cardiogreen women (4.0 and 5.2 mo of background radiation, respectively) (6).
10 STANFORD

Fig. 9. Electron beam CT image showing an apical thrombus in the tip Fig. 10. Helical CT image of calcifications in the left anterior de-
of an akinetic left ventricle in a postmyocardial-infarction patient. scending coronary artery.

For a multislice helical/spiral CT prospectively triggered have excellent temporal and spatial resolution; however, one
coronary calcification study, the effective dose would be 1.5 must recognize that because some organs such as the heart do
mSv in males and 1.8 mSv in women (6 and 7.2 mo of back- not lie in conventional imaging planes, true short and longitu-
ground radiation). If a multislice helical/spiral CT coronary cal- dinal axis anatomical images may not be possible without
cium study were retrospectively triggered, this would increase reformatting. To partially overcome this problem, some scanner
the radiation exposure to 3.0 mSv in males and 3.6 mSv in couches and/or gantries can be tilted to position the heart into
women, which would amount to 12 and 14.4 mo of background a more appropriate scanning plane. Generally, cardiac images
radiation (6). are acquired in a short axis plane similar to bread loafing, or
For a 50-ms EBCT cine sequence, the exposure would be in the longitudinal axis projection similar to the four-chamber
0.58 mSv per image; for a 10-image sequence, this would give view of echocardiography (Fig. 8). Because of the excellent
an exposure of 5.8 mSv (23.2 mo of background exposure). resolution of CT, imaging of structures as small as 0.5 mm is
One additional advantage of the electron beam technology is possible. Additionally, tissues with different CT attenuations,
that the radiation beam enters from beneath the patient, which
such as fat, myocardium, and calcium, can be readily identi-
reduces exposure to breast and thyroid tissue.
fied. An example is in imaging of the pericardium, where the
If a CT angiographic study were done using retrospective
low-attenuation epicardial and mediastinal fat lie adjacent to
gating, the radiation dose could reach as high as 10.9 mSv in
the higher-attenuation, fibrous pericardium. Commonly, the
men and 13.0 mSv in women (43.6 and 52 mo) (6). However,
tube current modulation techniques have now been introduced addition of iodinated contrast material allows further definition
that can ramp down the power during noncritical times, and this of the myocardium, and functional imaging of the heart is
can decrease radiation exposure by up to 80% (5). possible.
CARDIAC APPLICATIONS DISADVANTAGES OF CT
Both EBCT and helical/spiral CT are useful in defining car-
The disadvantages of even the faster CT technologies are the
diac and great vessel anatomy. The excellent spatial and tempo-
ral resolutions allow identification of intracardiac filling defects still somewhat long scan times which, unfortunately, do not
such as thrombi and tumors (7) (Fig. 9), and the identification of completely eliminate cardiac motion. Other disadvantages
defects of the pericardium (8). Functional imaging can quantify inherent in all scanners are that they are stationary and the
stroke volumes and ejection fractions, as well as evaluate patient has to be transported to the scanner and be subjected to
abnormalities of wall motion and perfusion, especially in post- radiation. Also, contrast administration is often necessary, and
infarct patients (9,10,11). More recently, imaging of coronary weight restrictions may prevent some patients being scanned.
calcification (12,13,14) (Fig. 10) and CT angiography for iden- THE FUTURE
tifying coronary stenoses (15,16,17) (Fig. 11) and soft,
noncalcified coronary atherosclerotic plaque (18) (Fig. 12) have The ideal X-ray imaging system would be to combine the
become extremely important applications. attributes of conventional radiography to produce images with
spatial resolution in the order of 5 1p/mm and acquisition times
ADVANTAGES OF CT of tens to hundredths ms that would be sufficient to freeze
The main advantage of CT imaging is that cross-sectional physiologic motion. In addition, the ideal system would respond
images without overlying structures are generated. The images linearly over a wide range, exclude scatter, and provide 3D
 CHAPTER 1 / PAST, PRESENT, FUTURE OF CT 11

Fig. 11. Electron beam CT shaded surface display image showing a stenosis in the left anterior descending coronary artery pre- (A) and post-
(B) balloon angioplasty. (Courtesy Moshage et al. [15].)

Fig. 12. Helical CT angiography image demonstrating areas of noncalcified plaque in the wall of an left anterior descending coronary artery.

information free of superimposing tissues with isotropic reso- expected. Already, prototype platforms of 256 detector rows
lution throughout (19). Also, it would provide the ability to are being developed and tested. Isotropic voxel resolutions of
image in transverse sections and thus avoid the problem of 0.5 0.5 0.5 mm are now possible, and from these 3D data
superimposed tissues. CT still has a long way to go, but contin- sets it is possible to replicate true anatomy. Sector reconstruc-
ued improvements in temporal and spatial resolution are tion can decrease scan times to below 100 ms, and EBCT scan-
12 STANFORD

ners that may reduce scan times to under 33 ms are under 9. Stanford W. Electron-beam computed tomography in the evaluation
development. Techniques to decrease scatter and noise are of ischemic heart disease. In: Taveras JM, Ferrucci JT (eds), Radi-
ology on CD ROM: Diagnosis-Imaging-Intervention. Lippincott,
being evaluated, and these should further improve image qual- Williams and Wilkins, Philadelphia: 2002; Chapter 19:2.
ity and decrease motion blur. These and other refinements have 10. Mochizuki T, Murase K, Higashino H, et al. Two- and three-dimen-
made possible functional and perfusion imaging. sional CT ventriculography: a new application of helical CT. AJR
Thus, the future of CT imaging of the heart is very exciting, Am J Roentgen 2000;174:203208.
11. Wang T, Ritman EL. Regional myocardial perfusion-quantitation
and expectations are that CT will play an ever-increasing role with high-speed, volume scanning CT (Abstract). Circulation
in cardiac imaging. 1987;76 (Part 2, Suppl.):IV5.
12. Stanford W. Coronary artery calcification as an indicator of preclini-
REFERENCES cal coronary artery disease. Radiographics 1999;19:14091419.
1. Hounsfield GN. Computerized transverse axial scanning (tomogra- 13. Becker CR, Jakobs TF, Aydemir S, et al. Helical and single-slice
phy): Part I. Description of system. Brit J Radiol 1973;46:10161022. conventional CT versus electron beam CT for the quantification of
2. Godfrey N. HounsfieldAutobiography. Nobel e-Museum. Avail- coronary artery calcification. AJR Am J Roentgen 2000;174:
able at: http://www.nobel.se/medicine/laureates/1979/hounsfield- 543547.
autobio.html. 14. Nallamothu BK, Saint S, Bielak LF, et al. Electron-beam computed
3. Boyd DP, Lipton MJ. Cardiac computed tomography. Proc IEEE tomography in the diagnosis of coronary artery disease: a meta-
1983;71:298307. analysis. Arch Intern Med 2001;161:833838.
4. Kalendar WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT 15. Moshage WE, Achenbach S, Seese B, Bachmann K, Kirchgeorg M.
with single breath-hold technique: continuous transport and con- Coronary artery stenoses: three-dimensional imaging with electro-
tinuous scanner rotation. Radiology 1990;176:181183. cardiographically triggered, contrast agent-enhanced, electron-beam
5. Cody DD, Moxley DM, Davros W, Silverman PM. Principles of CT. Radiology 1995;196:707714.
multislice computed tomographic technology. In: Silverman PM 16. Achenbach S, Moshage W, Ropers D, Nossen J, Daniel WG. Value
(ed), Multislice Computed Tomography. Lippincott, Williams, & of electron-beam computed tomography for the noninvasive detec-
Wilkins, Philadelphia: 2002;129. tion of high-grade coronary-artery stenoses and occlusions. N Engl
6. Hunold P, Vogt FM, Schmermund A, et al. Radiation exposure dur- J Med 1998;339:19641971.
ing cardiac CT: effective doses at multi-detector row CT and elec- 17. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary
tron-beam CT. Radiology 2003;226:145152. artery stenoses by contrast-enhanced, retrospectively electrocardio-
7. Stanford W, Rooholamini SA, Galvin JR. Ultrafast computed to- graphically-gated, multislice spiral computed tomography. Circula-
mography for detection of intracardiac thrombi and tumors. In: tion 2001;103:25352538.
Elliott LP (ed), Cardiac Imaging Imaging in Infants, Children, and 18. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection
Adults. JB Lippincott, Philadelphia: 2001;494500. and evaluation of atherosclerotic coronary plaques with multislice
8. Stanford W. Computed tomography and ultrafast computed tomog- computed tomography. J Am Coll Cardiol 2001;37-14301435
raphy in pericardial disease. In: Elliott LP (ed), Cardiac Imaging in 19. Mahesh M. Search for isotropic resolution in CT from conven-
Infants, Children, and Adults. JB Lippincott, Philadelphia: tional through multiple-row detector. Radiographics 2002;22:
1991;415420. 949962.
 CHAPTER 2 / ELECTRON BEAM TOMOGRAPHY 13

TECHNICAL BACKGROUND II
 2 Electron Beam CT of the Heart

DAVID G. HILL, PhD

INTRODUCTION a group of magnets used to control and shape the beam. Two
Electron beam tomography (EBT)* was developed by Dou- dipoles deflect the beam onto one of four fixed targets and
glas Boyd, PhD, and his associates at Imatron, starting in 1977; sweep it along that target. A solenoid and two quadrupoles
the first clinical installation was at University of California at focus the beam and ensure a proper elliptical shape aligned so
San Francisco in 1984. The challenge was to design a CT scan- that the minor diameter of the ellipse is in the scan direction.
ner capable of imaging the heart without motion artifacts; this The tungsten targets form a 210 arc and are backed by a water-
required data acquisition times of 100 ms or less. In order to cooling system, allowing at least 20 s of continuous scanning
study perfusion, covering most of the heart without moving the at full power without reaching thermal limits. X-rays are
patient table was also necessary. Acquisition times of 100 ms generated as the electrons strike the tungsten and are collimated
or less required a design with no moving parts except for the into a thin fan that irradiates the patient. The entire beam system,
patient table. The result is EBT, where an electron beam gen- from electron source to target ring, is under vacuum and can be
erates the X-ray views needed for CT by sweeping across fixed considered a very large, water-cooled, stationary X-ray tube.
targets. Reconstruction of a cross-sectional image from CT data
Since 1983, through several models, EBT has been continu- requires that at least one ray in every direction pass through
ously improved in spatial resolution, temporal resolution, and each pixel; a ray connects a source point and a detector (180
software capability, and has been distributed worldwide. plus the X-ray fan angle). Current EBT geometry meets this
In addition to the original anatomy, wall motion, and perfusion requirement with a 216 detector and a 210 target, providing
applications, applications such as measurement of coronary a 47.5-cm central region for imaging.
calcium, CT coronary angiography, lung studies, and standard X-rays from any of the four targets can reach the dual slice
CT radiological examinations have been developed and intro- detector. If the targets are used in series, 76 mm of a patient can
duced. be seen without moving the patient. Brass collimation rings
Unless otherwise stated, this article quotes specifications minimize the number of X-rays that could pass through the
for the e-Speed model. patient but not strike a detector. Collimation and detector size
give dual 7-mm slices with any of the targets. For narrower
TECHNOLOGY slices, one target is used, and an additional collimator is
EBT eliminates moving parts in the X-ray generation sys- employed to limit the X-ray beam to dual 1.5-, single 3-, or dual
tem by employing an electron beam that sweeps around a fixed 3-mm slices as measured at the isocenter. This additional col-
tungsten-coated target to generate a fan of X-rays. The detector limation is also prepatient, to give maximum dose utilization.
is mounted in a fixed position above the target ring, providing During image reconstruction, the data for dual 1.5-mm slices
a fourth generation (fixed detector, moving beam) CT design. can be combined to make a single 3-mm slice, and the data for
Sufficient data to reconstruct an image can then be acquired in dual 3-mm slices can be combined to make a single 6-mm slice.
one sweep of the beam around the target, which in the current As in any CT scanner, X-rays are detected and digitized by
model of the scanner (e-Speed) can be as short as 33 ms. The solid-state detectors. Special EBT challenges are the high data
speed and position of the beam and data acquisition system are rates and short integration times necessary to acquire sufficient
under computer control, allowing flexibility in the choice of data for an image in 33 ms or less.
temporal resolution, spatial resolution, and X-ray signal. Reconstruction of these images requires that the data be
The schematic shown in Fig. 1 is representative of an EBT corrected for beam hardening, beam and detector position, scat-
system. The electron source accelerates electrons at potentials tered X-rays, and incoming flux. The result is a cross-sectional
of up to 140 kV, producing a 1000-mA beam. Under high CT image with very high temporal resolution and high spatial
vacuum, the beam expands down a beam pipe until it reaches resolution.
*Also referred to as EBCT in the published literature. A single sweep is the passage of the electron beam along a
target from beginning to end (210), producing sufficient data
From: Contemporary Cardiology: CT of the Heart: to reconstruct an image. Most sweeps used for imaging move
Principles and Applications at constant angular velocity. A reset lasting 46 ms, during
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
which no X-rays are generated, brings the beam back to the

15
16 HILL

Fig. 1. Electron beam CT schematic. Fig. 2. Typical W wire signal.

starting position for the next sweep. If the next sweep does not a phantom containing a set of pins. Analysis of the raw data
immediately follow the preceding sweep, the beam is either from the pin scans gives the reconstruction system the location
turned off or deflected to a beam stop downstream of the deflec- of the beam corresponding to the data as acquired, enabling
tion coils, where it continues to move until the next sweep is interpolation of the raw data to the positions required by the
commanded. Because of its high power, the electron beam can reconstruction algorithm.
never be allowed to stop on the target without damaging the Control of the scanner requires numerous computers net-
target. Protections against such damage are built into the hard- worked together, dealing with control of the beam, collimation,
ware and software controlling the scanner. table, power supply, data acquisition, reconstruction, and oper-
The geometry of the electron source requires the perveance, ator interaction.
defined as V1.5/I, to be a constant; there is only one mA value
for a given kV setting. Thus, an e-Speed with nominally 1000 SCAN MODES
mA at 140 kV has 895 mA at 130 kV. Standard sweep speeds with the e-Speed are 33 ms, 50 ms,
EBT calibration and data correction involve scanning of air and 100 ms. Scan modes consist of different combinations of a
to get incoming flux, measurement of water to set the CT num- series of sweeps on one or more targets, with patient-table
ber of water to 0, and correction of beam hardening and scatter- motion or with no patient-table motion, and/or with an electro-
ing to ensure uniformity of CT numbers in imaging of an organ cardiogram (ECG) or timed trigger. If the patient table is not
or object with constant density. programmed to move while a sweep is being executed, the
An additional EBT calibration is tuning, which is used to result is an axial scan mode. Axial scans, triggered from an
determine with high precision both the exact location of the ECG signal, are typical for cardiac studies. If the patient table
beam with time, and the size, shape, and orientation of the beam moves while the X-rays are on, the result is a continuous vol-
spot. The beam condition at any given time is controlled by a ume scan, analogous to a spiral or helical scan with a mechani-
sequence of coil currents vs time given to the magnets that steer cal CT. Continuous volume modes are typically used for CT
and shape the beam. angiographic studies of the great vessels, for example.
An EBT scanner has a fifth target, located beyond those An external ECG monitor generates triggers at the R wave.
used for patient studies; a series of three-wire bundles in a W Sweeps begin at a user-specified time after the R wave, which
shape are located in front of this target. As the beam sweeps can be set either as a time interval in seconds or as a percentage
over this target and crosses one of the W wires, the location of of the RR interval. If triggering is based on percentages, the
the peak on the central wire gives the time of the beam with time is estimated from the previous seven heart RR intervals
respect to data acquisition; the width of the peak on the central and thus adapts to changes in the heart rate. Extensive clinical
wire gives the beam spot width; the shape and similarity of studies (1) show, for an EBT scanner with its high temporal
the pulses from the angled side wires give the orientation of the resolution, that the point in the RR interval with minimum
elliptical beam spot and its length; the distance between the motion is about 300 ms after the R wave, near the end of the T
side peaks is related to the beam radius from isocenter. An wave. Since the length of systole (starting at the R wave) tends
example of a W wire signal is shown in Fig. 2. to be more constant with heart-rate change than the length of
The first part of tuning consists of adjusting the coil currents diastole, triggering at the end of the T wave (end systole) seems
in time to give the desired parameters. Using equations that to have fewer problems with irregular heart rates than does
characterize the coil currents as a function of deflection angle, triggering by percentage of the RR interval.
the tuned values are transferred to the sweep descriptions that EBT scanners use prospective triggering. Before the scan
will be used for scanning on the X-ray targets. The second part begins, a point (or points) in the heart cycle is chosen for the
of tuning is a scan, with the transformed coil current values, of sweep to occur, and only those data are acquired and recon-
 CHAPTER 2 / ELECTRON BEAM TOMOGRAPHY 17

Fig. 3. (A) ECG tracesingle trigger at end systole (represented by
). (B) ECG tracetriggering eight phases of the heart. (C) ECG trace
triggering at two positions within the heart cycle.

structed. Therefore, all exposure of the patient contributes to studies of the heart at multiple phases in the same study time as
the study images; high temporal resolution and accurate trig- a single-phase study, and allows the clinician to appreciate
gering algorithms make this both feasible and efficient. changes in the heart over the heart cycle. Multiple phases is a
The simplest ECG-triggered scan mode has one sweep after cine mode including patient-table motion, and is used with all
the trigger at a prespecified time after the R wave; patient-table collimations. The high temporal resolution gives minimal
motion is after the sweep, so the patient table is at rest in motion artifact at any phase.
the new position in time for the next scan. This is the scan mode A further generalization of this mode allows the user to
used for a coronary calcium study or for an electron beam position sweeps at any point throughout the heart cycle. For
angiography (EBA) study (contrast in the arteries) with dual example, this might be used to scan at approximately end sys-
1.5-mm, single 3-mm, or dual 3-mm collimation. Table incre- tole and end diastole to obtain a measurement of ejection fraction.
ment is usually equal to the collimation, although smaller Figure 3 shows a representative ECG trace with several
increments giving overlapping images are used in some cases. possible triggering modes indicated.
Instead of just one sweep after the trigger, more than one For any ECG-triggered mode with patient-table motion, if
sweep can be used, with each sweep immediately following its the time required for the number of sweeps requested plus the
predecessor. The patient table moves to its next position after patient-table motion is less than the RR interval, the table is
the set number of sweeps has been performed. This allows positioned for the next location when the time to start the next
18 HILL

Table 1
Selected Specifications

Specification C150 C300 / C150 High Resolution e-Speed

Sweep speeds-thin slices 100 ms 100 ms 33, 50, and 100 ms

Sweep speeds-multitarget 50 ms 50 ms 33, 50, and 100 ms
Thin slice collimation 1.5, 3, and 6 mm 1.5, 3, and 6 mm dual 1.5 mm, single 3 mm, dual 3 mm
Multitarget collimation dual 7 mm dual 7 mm dual 7 mm
In-plane spatial resolution 7 line pairs (lp)/cm (100 ms) 9.5 lp/cm (100 ms) 10 lp/cm (50 ms)
13 + lp/cm (100 ms)
7 lp/cm (33 ms)
Power 83 kW (130 kV, 625 mA) 83 kW (130 kV, 625 mA) 140 kW (140 kV, 1000 mA)
116 kW (130 kV, 895 mA)
Trigger times (%RR) 4080 (all) 4080 (all) 099 (all)
0 (50 ms only) 0 (50 ms only)
Trigger times (seconds after R) 0.2730.999 0.2730.999 0.0601.500

set of sweeps arrives; if the patient table is not in place for the volume scanning but without patient-table motion during data
next heart cycle, one heart cycle is skipped. acquisition.
A multitarget cine study, typically covering the entire heart A PreView (General Electric, formerly Imatron) (scout or
cycle, can also be performed. Sweeps every 33, 50, or 100 ms localization scan) is created using a special sweep while the
from the first target cover the time between first and second patient table moves continuously. The beam is swept very rap-
R waves; the second target is used to cover the third to fourth idly along the tuning target; near three oclock and again near
R-wave interval, the third target with the fifth to sixth R-wave six oclock, the beam moves onto the X-ray target for a short arc
interval, and the last target with the seventh to eighth R-wave at the slower 50-ms speed. X-rays reaching the patient are gen-
interval. If more coverage is needed, the patient table is then erated only when the beam is on the X-ray target. A transmis-
moved 40 mm or more and additional scans taken. sion image is synthesized from the source points along these
A flow study is used to follow a bolus of contrast. One or arcs, giving views in both anterior/posterior and lateral direc-
more sweeps is taken each heart cycle without patient-table tions. From these views the user not only can set the location in
motion. If all four targets are used, then one sweep is taken on the patient direction to start and stop acquiring data, but also
each of the targets during one heart cycle, allowing the mea- can set the center point and field of view for the reconstructed
surement of perfusion over 76 mm of the body using the same images in the studies to follow. Because PreView images are
contrast bolus. Some heart cycles are usually skipped to made from a number of source positions through a process of
sample the bolus curve but minimize dose. This is the standard tomosynthesis, the plane of focus can be changed from the
mode for the measurement of myocardial perfusion. default value of the isocenter, or multiple planes can be com-
All of these modes can be triggered manually or by time puted and viewed. The depth of field is sufficiently large that
instead of by ECG; however, cardiac studies are almost always isocenter is normally chosen.
ECG-triggered. IMAGE QUALITY
If the table moves while the beam sweeps continually, the EBT is designed for high temporal resolution: 33, 50, or
result is a continuous volume scan, similar to a helical or spiral 100 ms in a single image without requiring data from more
scan with a mechanical CT. Data can be acquired with any of than one heart cycle. This temporal resolution is short enough
the available collimations. Images are reconstructed at any that images can be acquired at any point in the heart cycle with
location along the scan; the slice width in the image is deter- no or almost no motion artifact.
mined from the collimation and the table travel corresponding Maximum in-plane resolution ranges from 7 line pairs (lp)/
to the sweeps included in the reconstruction. As an EBT scan- cm at 33 ms to 13 lp/cm at 100 ms. An extensive set of recon-
ner has a fixed mA for a given kV, in order to change the struction kernels allow the optimization of resolution to the
exposure in an image without changing the slice width, the clinical study being performed. Narrowest slices are 1.5 mm.
speed of the patient table must also change. For example, in a Because the arteries are typically moving from 20 to 100 mm/
study using 3-mm collimation and reconstructing a 3-mm slice, sec (2), resolving the arteries clearly requires a high temporal
twice the exposure in the 3-mm slice requires the patient table resolution imaging system.
to move half as fast.
A continuous volume scan is the normal scan mode for study- SPECIFICATIONS
ing the noncardiac vessels, the lungs, abdomen, or any other As of this writing there are three EBT models in clinical use:
organ not requiring ECG triggering. Of course, the axial modes C150, C300, and e-Speed. Some critical parameters such as
described previously with a timed or manual trigger may also sweep speed, slice width, and resolution are summarized in
be used; however, the study time is longer than with continuous Table 1.
 CHAPTER 2 / ELECTRON BEAM TOMOGRAPHY 19

DOSE Table 2
Dose Information
As an EBT scanner irradiates a patient over a 210 arc (essen-
tially from two oclock to ten oclock), dose near the surface of Parameter C150/C300 e-Speed
the patient is not uniform. For a patient lying supine, dose is a
CT dose index (CTDI) at A (center) 1.8 mGy 1.9 mGy
factor of five lower anterior than it is posterior. The result is that CTDI at B (12 oclock) 1.2 mGy 1.2 mGy
dose to the breast is minimized during a cardiac examination. CTDI at C (6 oclock) 7.9 mGy 6.3 mGy
Prospective triggering also means that a patient receives CTDI at D,E (3 or 9 o-clock) 5.6 mGy 5.1 mGy
only the dose necessary for the images desired for the study. CTDIw (weighted average) 3.7 mGy 3.6 mGy
The user can choose to acquire data over the entire heart cycle Typical cardiac studies covering entire heart with one sweep each
or over only a part of it, and the scanner acquires just those data. heart cycle:
Table 2 gives the CT dose index (CTDI) (3) information for
kV 130 140
a scan that would be typical of a calcium or coronary mA 625 1000
angiographic single-phase study for an EBT scanner, as well Sweep time 100 ms 50 ms
as an estimate of the effective dose (E) if the entire heart is Collimation 3 mm 3 mm
covered. The 32-cm body CTDI is quoted with a 3-mm slice (or Table increment/heart cycle 3 mm 3 mm
dual 1.5-mm with e-Speed) and a single sweep. The CTDI as Effective dose 0.7 mSv 0.7 mSv
defined is an estimate of the single-slice dose taking into Sweep time 100 ms 50 ms
account the tails of the dose profiles from the adjacent slices. Collimation 1.5 mm dual 1.5 mm
E is estimated (4) by multiplying the CTDIvol (weighted CTDI Table increment/heart cycle 1.5 mm 3 mm
adjusted for table speed) times the distance covered (which is Effective dose 1.4 mSv 0.7 mSv
the dose-length product) 0.017.
TYPICAL APPLICATIONS
Typical cardiac studies with EBT include coronary calcium
(5), EBA of the coronary arteries (6,7), cardiac anatomy (8), without contrast); some 610 heart cycles later a sequence of
wall motion studies (9), and perfusion studies (10). EBT scan- one sweep every other heart cycle begins through the expected
ners are also excellent vehicles for CT angiography of the non- peak of the contrast (18 s or so) followed by a sweep every 3
cardiac vessels, as well as for lung studies (11) . The references heart cycles or so to follow the contrast wash out. The circula-
given are to published studies as examples; no attempt has been tion time (the time between injection and arrival at the coronary
made to make this a complete bibliography. Figure 4 is an arteries) is then determined from the change in the CT number
example of an axial image showing coronary calcium; Fig. 5 in the aorta or left ventricle with time. EBA data acquisition
shows a 3D volume rendering of an EBA study, and Fig. 6 shows starts after a delay from the start of injection that is approxi-
a maximum intensity projection image of the same data set. mately equal to the circulation time. If the circulation time
Studies of coronary calcium use an ECG-triggered axial study uses all four targets and the dual 7-mm collimation, it
mode with one phase/heart cycle. Most recently the preferred becomes a perfusion study.
trigger point is near end systole, a time of minimum motion of The length of the EBA injection should be approximately
the right coronary (1) with a 3-mm slice, 130 kV, and no con- equal to the time required to complete the series, estimated
trast. Historically, most coronary calcium scanning has been from the distance to cover (which when divided by the table
done with triggering either near 80% or more recently near 40% increment gives the number of heart cycles needed) and the start-
of the RR interval. All EBT scanner models have a calcium ing heart rate. For most patients, the heart rate will increase
scanning mode that matches slice width, reconstruction param- during the study. Optimization of the contrast is very impor-
eters, and kV to the early data on C100/C150, so that the tant; in addition to correct circulation time and injection dura-
Agatston (12) score for calcium is independent of EBT model; tion, it may involve a saline chaser and variation of the contrast
thus, historical data can be used for clinical comparison. Other flow rate during the injection. The best studies come with the
scoring methods such as volume score (13) and mass score (14) contrast signal uniform in time and the ratio of the average CT
have been proposed and implemented that may be more inde- number of the contrasted vessels to the X-ray noise in the images
pendent of the scan mode. as large as possible.
Coronary angiographic studies (electron beam angiography, Most EBA studies are analyzed on a workstation making use
or EBA) use ECG-triggered axial modes with one or more of reformats, maximum intensity projections, and volume ren-
phase/heart cycle and intravenous contrast injection. If a single derings to appreciate the details of the arteries. Multiple-phase
phase is used, then the scanner is triggered at the point of mini- EBAs make arterial motion obvious, although any individual
mum motion; if multiple phases are being used, then the point phase can be analyzed. Clinical evaluation of the data frequently
of minimum motion is usually included as one of the phases. includes observing 2D and 3D renderings beat as the display
In order to optimize the contrast, circulation time is deter- cycles through the phases. With an appropriate choice of scan
mined before an EBA study, using a short bolus of contrast and parameters, the study may cover a sufficient portion of the heart
a flow study looking at a given level. Typically a scan is trig- cycle to allow the estimation of ejection fraction. Valve motion
gered at the start of contrast injection (for a baseline value can also be visualized.
20 HILL

Fig. 4. Calcium example; 3-mm slice; 130 kV; 895 mA; 50 ms.

Fig. 5. Three-dimensional rendering of an electron beam angiography Fig. 6. Maximum intensity projection image of left anterior descend-
study; dual 1.5-mm slice; 140 kV; 1000 mA; 50 ms. ing artery; same data set as Fig. 5.
 CHAPTER 2 / ELECTRON BEAM TOMOGRAPHY 21

Sufficient cardiac functional information is often gathered the number of detectors in the patient direction for increased
from a multiphase study. In addition, there are two ways to coverage with each sweep. As cardiac imaging continues to
perform a dedicated wall-motion or function study. One is essen- grow in clinical importance, EBT is an ideal modality for im-
tially an EBA using dual 7-mm collimation covering the entire proved applications and efficient usage.
heart cycle, with a 15-mm patient-table motion every other
heart cycle. An alternative, avoiding patient-table motion, is to REFERENCES
move the beam sequentially through the different targets instead 1. Mao S, Budoff MJ, Bakhsheshi H, Liu SC. Improved reproducibility
of moving the patient table. of coronary artery calcium scoring by electron beam tomography
with a new electrocardiographic trigger method. Invest Radiol
A perfusion study is a circulation time study with 76-mm 2001;36(7):363367.
coverage, using all four targets and the dual 7-mm collimation. 2. Achenbach S, Ropers D, Holle J, Muschiol G, Daniel WG, Moshage
This mode allows the measurement of perfusion throughout the W. In-plane coronary arterial motion velocity: measurement with
myocardium using one bolus of contrast. There are three gaps electron-beam CT. Radiology 2000;216(2):457463.
of approx 5 mm each (between the targets) that are not fully 3. Title 21 Code of Federal Regulations Part 1020 Section 33(c)(2)(i).
4. McCullough C. Patient dose in computed tomography of the heart.
illuminated. Herz 2003;28:16.
A complete EBCT cardiac exam consists of a PreView for 5. Budoff MJ, Raggi P. Coronary artery disease progression assessed
localization, a calcium study, a flow study used for both perfu- by electron-beam computed tomography (review). Am J Cardiol
sion and circulation time determination, a multiphase EBA, 2001;88(2A):46E50E.
and possibly a wall-motion study using residual contrast from 6. Achenbach S, Moshage W, Ropers D, Nossen J, Daniel W. Value of
electron beam computed tomography for the noninvasive detection
the EBA. of high-grade coronary-artery stenoses and occlusions. N Engl J
The EBCT patient table can tilt 25 and slew 25 (side to Med 1998;22:19641971.
side in the horizontal plane). Tilt and slew allow the acquisition 7. Lu B, Zhuang N, Mao S, Bakhsheshhi H, Liu S, Budoff M. Image
of wall-motion or perfusion data in approximately the short quality of three-dimentional electron beam coronary angiography. J
Comput Assist Tomogr 2002;26:202209.
axis position (tilt 15, slew 15 to 25) or the long axis po-
8. Westra S, Hurteau J, Galindo A, McNitt-Gray M, Boechat M, Laks
sition (tilt 0, slew +25), approximating traditional cardiac H. Cardiac electron-beam CT in children undergoing surgical repair
views. for pulmonary atresia. Radiology 1999;213:502512.
Continuous volume is the mode of choice to analyze the 9. Rumberger JA. Use of electron beam tomography to quantify car-
noncardiac vessels, yielding clear motion-free images. For diac diastolic function (review). Cardiol Clin 2000;18(3):547556.
10. Bell MR, Lerman LO, Rumberger JA. Validation of minimally in-
example, using on the e-Speed dual 1.5-mm collimation , 50-ms
vasive measurement of myocardial perfusion using electron beam
sweeps, and a 3-mm table increment, the patient table moves at computed tomography and application in human volunteers. Heart
54 mm/s, allowing most areas of interest to be covered in 10 s 1999;81:628635.
or less. 11. Robinson T, Leung A, Moss R, Blankenberg F, Al-Dabbagh H,
Northway W. Standardized high-resolution CT of the lung using
FUTURE DIRECTIONS spirometer-triggered electron beam CT scanner. AJR Am J
Roentgenol 1999;172:16361638.
EBCT scanners are designed for excellent cardiac imaging. 12. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M
There is no inherent limit on how fast an EBCT scanner can Jr, Detrano R. Quantification of coronary artery calcium using
sweep the beam, as there are no moving parts, so sweep times ultrafast computed tomography. J Am Coll Cardiol 1990;15:827832.
can become even shorter, improving temporal resolution. 13. Callister TQ, Cooil B, Raya SP, Lippolis NJ, Russo DJ, Raggi P.
Potential temporal resolution is limited only by available X-ray Coronary artery disease: improved reproducibility of calcium scor-
ing with an electron-beam CT volumetric method. Radiology
power (signal-to-noise in the images) and data acquisition 1998;208:807814.
speeds. Future directions are likely to include using a detector 14. Utzheimer S, Kalender W. Assessment of calcium scoring perfor-
with smaller elements for thinner slices, as well as increasing mance in cardiac computed tomography. Eur Radiol 2003;13:484497.
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 23

3 Scan Techniques for Cardiac and Coronary

Artery Imaging With Multislice CT

BERND M. OHNESORGE, PhD, BRIAN R. WESTERMAN, PhD,

AND U. JOSEPH SCHOEPF, MD

INTRODUCTION provided much more accurate CT imaging of the heart and the
Cardiac imaging is a demanding application for any coronary arteries (1114). Recent clinical studies have demon-
noninvasive imaging modality. On the one hand, high temporal strated the potential of MSCT to differentiate and classify lipid,
resolution is needed to virtually freeze cardiac motion and thus fibrous, and calcified coronary plaques (15). Despite these
avoid motion artifacts in the images. On the other hand, suffi- promising advances, the 4-slice CT scanner technology still
cient spatial resolutionat best submillimeteris required to faces some challenges and limitations with respect to motion
adequately visualize small and complex anatomical structures artifacts in patients with higher heart rates, limited spatial reso-
like the coronary arteries. The complete heart volume has to be lution, and long breath-hold times (12). In 2001, a new genera-
examined within a single short breath-hold time to avoid breath- tion of MSCT systems with simultaneous acquisition of up to
ing artifacts and to limit the amount of contrast agent, if neces- 16 slices was introduced (16,17). With submillimeter slice
sary. The motion of the heart is both complex and very fast. acquisition and gantry rotation times shorter than 0.5 s, both
Some estimates of the temporal resolution needed to freeze spatial and temporal resolution are improved, while examina-
cardiac motion in any phase of the cardiac cycle are as low as tion times are considerably reduced.
10 ms. In 1984, electron beam computed tomography (EBCT) In this chapter, we present the basic technology of MSCT
was introduced as a noninvasive imaging modality for the cardiac scanning with a special focus on recommended scan
diagnosis of coronary artery disease (14). Its temporal resolu- techniques for different clinical applications. We will also dis-
tion of 100 ms allows for relatively motion-free imaging of the cuss the technology advances and improved clinical perfor-
cardiac anatomy in the diastolic phase, even at higher heart mance of state-of-the-art 16-slice CT equipment compared with
rates. Because the EBCT at that time was limited to axial scan- 4-slice CT scanners.
ning for electrocardiogram (ECG)-synchronized cardiac inves-
TECHNOLOGY PRINCIPLES
tigations, a single breath-hold scan of the heart required slice
widths of at least 3 mm. The resulting axial resolution was TECHNOLOGY OVERVIEW
therefore limited and not adequate for 3D visualization of the CT examinations of the heart should be performed in a single,
coronary arteries. With the advent of subsecond rotation, com- short breath-hold scan with high temporal resolution to elimi-
bined with prospective and retrospective ECG-gating, nate cardiac motion and high, preferably isotropic, spatial reso-
mechanical single-slice helical or spiral CT systems with supe- lution. It has become increasingly apparent that submillimeter
rior general image quality entered the realm of cardiac imaging slices are necessary to adequately visualize small and complex
(4,5). Since 1999, 4-slice CT systems, which have the potential cardio-thoracic anatomy and the coronary arteries.
to overcome some of the limitations of single-slice cardiac CT In 1984, EBCT was introduced as the first cross-sectional
scanning, have been used to establish ECG-triggered or ECG- noninvasive imaging modality that could visualize the cardiac
gated multislice CT (MSCT) examinations of the heart and anatomy and the coronary arteries (1). During this period,
the coronary arteries in clinical use (610). As a result of the mechanical scanners typically had single-slice detectors and
increased scan speed with four simultaneously acquired slices, a minimum gantry rotation time of 0.751.0 s, and were not
coverage of the entire heart volume with thin slices within one considered of value for strictly cardiac imaging. With presently
breath-hold became feasible. The improved axial resolution available EBCT scanners, temporal resolution of 100 ms pro-
vides motion-free images of the cardiac anatomy in the dias-
From: Contemporary Cardiology: CT of the Heart:
tolic phase of the cardiac cycle even at higher heart rates (2).
Principles and Applications Cardiac anatomy can be covered in a single breath-hold of 30
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ 40 s with slice widths of 3 mm, but this limits the diagnostic

23
24 OHNESORGE, WESTERMAN, AND SCHOEPF

accuracy of coronary artery visualization (3). In 1998, spiral/ ECG-TRIGGERED MSCT IMAGING
helical CT systems with simultaneous acquisition of four Prospective ECG triggering has long been used in conjunc-
detector slices and rotation time of 500 ms were introduced tion with EBCT and single-slice spiral CT (14). A trigger
(6,7,10), which provided a substantial performance increase signal is derived from the patients ECG based on a prospective
over the single- and dual-slice spiral CT systems that had been estimation of the present RR interval, and the scan is started at
available until then. These MSCT scanners can cover larger a user-defined time after a detected R wave, usually during
scan volumes with slice collimation down to 0.5 mm and thus diastole. MSCT allows simultaneous acquisition of several
provide higher spatial resolution for improved visualization of slices in one heartbeat with a cycle time that usually allows
small and complex anatomy. scanning in every other heartbeat (Fig. 2A). Thus shorter breath-
Detector configurations in use today enable simultaneous hold times are required compared to single-slice scanners, and
collimation of four slices with different slice widths. In the respiratory artifacts can be virtually eliminated. To achieve the
fixed array design, detector rows with equal spacing are used, best possible temporal resolution, scan data are acquired only
while the adaptive array and hybrid detector designs con- during a partial gantry rotation (approximately two thirds of a
sist of fewer detector rows, containing elements that become rotation with 240260 projection data) that covers the mini-
wider away from the center of the detector (Fig. 1A,B). In all mum amount of data required for image reconstruction. Con-
designs the thinnest slices result from collimation of the inner ventional partial image reconstruction based on fan beam
four detector rows, and thicker slices are generated by elec- projection data results in a temporal resolution equal to the
tronic combination of adjacent detector rows. acquisition time of the partial scan. Optimized temporal reso-
Regardless of detector design, it was apparent from the lution can be achieved with parallel-beam-based half-scan
operation of the 4-slice scanners that no fundamental barrier reconstruction algorithms that provide a temporal resolution of
prevented the construction of CT scanners that could acquire half the rotation time in a central area of the scan field of view
more slices simultaneously, and indeed this quickly came to (6,7,10,16). Thus, prospective ECG triggering is also the most
pass. Systems capable of eight slices became available in 2000, dose-efficient method of ECG-synchronized scanning, as only
with commensurate increases in exam speed. Meanwhile, the very minimum of scan data needed for image reconstruction
8- and 10-slice CT scanners are being used that provide further are acquired. However, usually only relatively thick slices (3
improved volume coverage with about 1-mm slice width and mm with EBCT, 2.53 mm with 4-, 8-, and 16-slice CT) are
500 ms rotation time (18). The first 16-slice CT scanners, intro- used for prospectively triggered acquisition, to maintain a rea-
duced in early 2002, provide faster rotation time (down to 400 sonably short single breath-hold. Thus, resulting data sets are
ms) and submillimeter detector collimation for routine volume often not suitable for 3D or multiplanar reformation (MPR)
imaging (16,17). The 16-slice systems are sufficiently fast to reconstruction of small cardiac anatomy. In addition, prospec-
cover the entire heart, scanning with submillimeter slices, in a tive ECG-triggered scans are sensitive to changes in heart rate
reasonable breath-hold. All of the 16-slice scanners have during acquisition, so significant fluctuation or arrhythmia can
adopted a detector design with elements of two sizes. In the have a severe effect on image quality.
examples given (SOMATOM Sensation 16, Siemens, and ECG-GATED MSCT SCANNING
Aquillion 16, Toshiba Medical) (Fig. 1C,D) the 16 central rows AND IMAGE RECONSTRUCTION
define submillimeter detector acquisition. By adding outer Retrospective ECG gating overcomes the limitations of pro-
detector rows on both sides, wider slices of 1 mm and above can spective ECG triggering with regard to scan time and spatial
be achieved. resolution, and can provide more consistent image quality for
Higher temporal resolution than the older mechanical CT examination of patients with changing heart rate during the
scanners is provided by a combination of faster gantry rotation scan. This technique requires multislice spiral scanning with
speed, with rotation times down to 500 ms, and specialized slow table speed and simultaneous recording of the ECG trace
reconstruction algorithms (10). As of 2002, the shortest gantry that is used for retrospective assignment of image reconstruc-
rotation time was reduced to 400 ms, and more recently even to tion (810). Phase-consistent coverage of the heart requires a
370 ms, thus further improving temporal resolution (16,17). highly overlapping spiral/helical scan with a table feed adapted
The combination of fast rotation time and multislice acquisi- to the heart rate in order to avoid gaps between image stacks
tion with submillimeter spatial resolution has proved to be of that are reconstructed in consecutive heart cycles. The image
particular importance for improved cardiac image quality stacks are reconstructed at exactly the same phase of the heart
(17,1921). cycle and cover the entire heart and adjacent anatomy (Fig. 2B).
Motion artifacts caused by cardiac pulsation can be mini- To achieve gapless coverage of the entire heart over a wide
mized in high-resolution CT studies by limiting image recon- range of heart rates, pitch values between 0.2 and 0.4 are
struction to those parts of the cardiac cycle associated with the employed (pitch equals table feed per rotation, divided by the
least motion, typically during diastole. The heart phases can be nominal width of the X-ray beamequal to all slices com-
determined from a simultaneously recorded ECG signal. Two bined). Images are reconstructed during every heart beat, and
different ECG synchronization techniques are most commonly somewhat faster scan coverage is possible as compared to
employed for cardiac CT scanningprospective ECG trigger- prospective ECG triggering. Moreover, the continuous spiral
ing and retrospective ECG gating. acquisition enables reconstruction of overlapping image slices,
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 25

Fig. 1. (A) Equally spaced detector elements build the fixed array detector (FAD) for a 4-slice system. Different 4-slice collimation settings
(4 1.25, 4 2.5, 4 3.75, 4 5 mm) are produced by electronic combination of adjacent elements. (B) Differently sized detector elements
build the adaptive array detector (AAD) for a 4-slice system. Different 4-slice collimation settings (4 2.5, 4 5 mm) are produced by electronic
combination of adjacent elements. For 4 1 mm collimation, partial shielding of the elements with 1.5-mm width is required. Partial
illumination of the inner two detector elements allows for 2 0.5-mm collimation for high-resolution scanning. (C) The latest 16-slice CT
scanners are also based on the AAD design with differently sized detector elements. 16-slice collimation settings 16 0.75 mm and 16 1.5
mm are produced by equally spaced 16 0.75-mm detector elements in the center, eight additional 1.5-mm detector elements (four on each
side), and electronic combination of adjacent elements. (D) Other 16-slice CT designs allow for illumination of 16 0.5-mm slices in the
detector center and up to 16 2.0-mm slices via adding detector elements with 1.0-mm detector spacing on each side that are also electronically
combined.
26 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 2. (A) Illustration of sequential volume coverage with prospec-

tively ECG-triggered multislice scanning. Multiple images (one image
per detector slice) are acquired at a time with a certain delay after a
detected R wave. The slice thickness is equivalent to the collimation
(hatched blocks), and the temporal resolution equals half the rotation
time. Owing to the limitation of the scan cycle time to about 1 s, a scan
can be acquired every other heart cycle for usual heart rates. (B) A
scan with continuous table feed and continuous exposure is acquired
for retrospectively ECG-gated multislice spiral scanning. Stacks of
overlapping images can be reconstructed with a temporal resolution
of half the rotation time in every cardiac cycle. Continuous 3D images Fig. 3. (A) Schematic illustration of an adaptive segmented image
can be reconstructed in different phases of the cardiac cycle by selec- reconstruction approach for electrocardiogram (ECG)-gated 16-slice
tion of the data ranges with certain phase relations to the R waves. spiral CT scanning. Dashed lines are used to indicate the z positions
of the detector slices, which continuously and linearly change posi-
tion relative to the patient with constant spiral feed. The ECG signal
is simultaneously recorded during scan acquisition and is displayed at
the bottom of the diagram. At heart rates below a certain threshold,
and thus a longitudinal spatial resolution about 20% below the one segment of consecutive multislice spiral data is used for image
slice width can be achieved (e.g., 1.0 mm for 1.25-mm slices, reconstruction. At higher heart rates, two sub-segments from adjacent
0.8 mm for 1.0-mm slices, 0.6 mm for 0.75-mm slices, 0.5 mm heart cycles contribute to the partial scan data segment. In each car-
for 0.6-mm slices). For these reasons, retrospective ECG-gated diac cycle, a stack of images is reconstructed at different z positions
covering a small sub-volume of the heart, which is shown as a box.
reconstruction is the preferred method for imaging cardiac
The combination of sub-volumes from all heart cycles during the scan
anatomy and the coronary arteries. Thin slices provide the nec- provides a continuous 3D data set of the entire heart. (B) Temporal
essary spatial resolution for diagnosis within a breath-hold time resolution as a function of the heart rate for the adaptive segmented
that can be managed by most patients. Reconstructing images reconstruction approach, using 0.5-s, 0.42s, and 0.37-s gantry rota-
from fan beam data of a partial rotation (usually 240260) tion time. If data from two consecutive heart cycles are used for image
reconstruction, the temporal resolution strongly depends on heart rate.
within every cardiac cycle yields a temporal resolution equal to For 0.42-s rotation time, the temporal resolution reaches its optimum
half of the gantry rotation time. of 105 ms at 81 bpm; for 0.37 s the optimum of 92 ms is present at 71
A multislice spiral interpolation between the projections of bpm. Thus, clinically robust image quality can be achieved also in
adjacent detector rows is used in order to compensate for table patients with higher heart rates, i.e., in patients with stable sinus
movement and to provide a well-defined slice sensitivity pro- rhythm and heart rates in the range 7585 bpm.
file and images free of spiral artifacts. The temporal resolution
can be improved by using scan data from more than one heart
cycle for reconstruction of an image (segmented reconstruc- tion between (rotation time)/2 and (rotation time)/2 M is
tion) (9,22,24) (Figs. 3 and 4). While the number of fan beam possible (where M equals the number of segments or heart
projections needed to reconstruct the image remains the same, cycles used). Most recent cardiac reconstruction algorithms
collecting them from multiple (up to four) consecutive cardiac automatically adapt the number of heart cycles used with chang-
cycles reduces the data collection time in each cycle and thereby ing heart rate during the scan, in order to optimize temporal
improves temporal resolution. Depending on the relationship resolution if higher heart rates are present (Figs. 3 and 4). While
between rotation time and patient heart rate, a temporal resolu- the use of segmented reconstruction may extend cardiac imag-
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 27

and the center of rotation. Cone-beam artifacts are most pro-

nounced at high-contrast boundaries. Typical sources of cone-
beam artifacts are the ribs or the pelvic bones. Because the heart
is usually well centered and does not contain large high-con-
trast structures, cone-beam artifacts are negligible for todays
16-slice CT scanners (20,21). Since cone-beam reconstruction
algorithms are computationally intensive, they tend to be time
consuming. If they are not needed for ECG-gated cardiac scan-
ning, then the time can be better utilized for more rapid recon-
struction using less complex algorithms. Cardiac CT scanning
with more than 16 slices, however, will require advanced car-
diac cone-beam algorithms in future CT systems (25).
Usually the diastolic phase of the cardiac cycle is chosen for
image reconstruction of cardiac and coronary morphology;
however, owing to the highly overlapping scan acquisition,
image data can be reconstructed for each x, y, and z position
within the scanned volume over the entire course of the cardiac
cycle. This allows for retrospective selection of phases for
reconstruction that provide the best image quality for anatomy
with special motion patterns in individual patients (26,27). To
improve image quality in the presence of arrhythmia, selected
image stacks can either be discarded or arbitrarily shifted within
the cardiac cycle, so that reconstruction always coincides with
the same interval during diastole at each level of the cardiac
volume. Besides the morphological information that is derived
from images reconstructed in diastole, additional reconstruc-
tions of the same scan data set in different phases of the cardiac
cycle (Fig. 2B) enable analysis of basic cardiac function param-
eters such as end-diastolic volume, end-systolic volume, ejec-
tion fraction, and regional wall motion.
Relatively high radiation exposure is associated with retro-
spectively ECG-gated imaging of the heart because of continu-
ous X-ray exposure and overlapping data acquisition at low
table speeds. All data can be used for image reconstruction in
Fig. 4. (A) Schematic illustration of an alternative adaptive segmented different cardiac phases, but if only a very limited interval (i.e.,
image reconstruction technique for electrocardiogram-gated 16-slice
spiral CT scanning with use of up to three segments per slice. (B) diastolic phase) in the cardiac cycle is targeted during recon-
Temporal resolution as a function of the heart rate for this approach struction, a significant portion of the acquired data and radia-
for 0.5-s and 0.4-s gantry rotation time. By using up to three segments tion exposure is redundant. A high potential for exposure
per slice, temporal resolution can be further increased in patients with reduction during ECG-gated spiral scanning is offered by an
certain and very stable heart rate at the expense of slower table speed,
on-line reduction of the tube output within each cardiac cycle
thus resulting in higher radiation exposure and longer breath-hold
times. When heart rate is 95105 bpm, 0.5-s gantry rotation provides during phases that are of less importance for ECG-gated recon-
theoretically better temporal resolution than 0.4-s gantry rotation. struction (10). In this approach, the nominal tube output is
Temporal resolution is a complex function of heart rate, gantry rota- applied only during the diastolic phases of the cardiac cycle
tion time, and helical pitch. that are likely to be reconstructed. In the rest of the cardiac
cycle, the tube output is reduced. Depending on the heart rate,
an overall exposure saving of 3050% can be achieved without
ing with CT to higher heart rates, the associated drawback is compromising on image quality (28). This technique is particu-
slower table speed, resulting in increased patient dose and larly useful for examination of coronary artery calcification.
longer breath-holds. As with nonsegmented reconstruction For coronary CT angiography in patients with higher heart rates,
algorithms, an irregular heart rate may severely compromise optimal images of the different coronary vessels may need to be
the study. reconstructed in wider ranges of the RR interval, leaving little
Recent publications (16,23) have demonstrated that cone- time in which to reduce tube current.
beam reconstruction algorithms become mandatory for general A special technique for ECG-gated spiral scanning with
purpose CT scanning with eight and more slices to avoid severe increased volume coverage for cardio-thoracic applications has
image artifacts. The severity of cone-beam-induced artifacts been introduced for 4-slice CT scanners (29). It allows for sup-
depends on the number of simultaneously acquired slices, on pression of cardiac pulsation by means of reconstruction with
the width of each slice, and on the distance between an object a temporal resolution that equals half the rotation time, and to
28 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 5. Investigation of the spatial resolution in the scan direction for cardiac CT examinations with 4- and 16-slice CT with a resting
longitudinal resolution phantom. The phantom includes air-filled spheres 0.4 to 3.0 mm in diameter that can be examined with multiplanar
reformation cuts along the scan direction. A maximum longitudinal resolution of 0.81.0 mm can be achieved with 4-slice CT using 4 1-mm
collimation, 1.3-mm slice width, 0.6-mm image increment, and 3-mm/s table feed. 16-slice CT provides up to 0.6-mm longitudinal resolution
based on 16 0.75-mm collimation, 0.8-mm slice width, 0.4-mm image increment, and 6.6-mm/s table feed.

ECG gating that eliminates scan data acquired during the sys- includes contrast-enhanced coronary arteries containing stents
tolic phase, with its rapid heart motion. The latest 16-slice CT and atherosclerotic plaques of known dimensions and catego-
scanners equipped with this protocol can provide motion-free ries. Data sets from 16-slice scanners can be reconstructed with
coverage of the entire thoracic anatomy with submillimeter different image widths for optimization of the trade-off between
slices within a single breath-hold. However, cone-beam recon- spatial resolution and signal-to-noise ratio for specific clinical
struction techniques may be required for this application, as applications. For 16 0.75-mm collimation, slices with a
high-contrast thoracic anatomy located in the periphery of the thinnest possible image of 0.8 mm can be generated, thus
scan field of view is also of interest for diagnosis (30). allowing for substantially improved visualization of coronary
PERFORMANCE EVALUATION artery lumen compared to 4-slice CT.
WITH PHANTOM STUDIES
The performance of 4-slice CT scanners in comparison to a CLINICAL IMAGING PROTOCOLS
new 16-slice CT scanner with respect to cardiac imaging has The optimal scan technique for a given clinical application
been evaluated with computer simulation studies and phantom very much depends on the desired spatial and contrast resolu-
measurements. The results demonstrate the advancements of tion for that application. Prospective ECG triggering can be
16-slice CT scanners in scan speed and spatial resolution. readily used for contrast-enhanced imaging of great-vessel
The achievable spatial resolution with high-resolution car- morphology within limited scan ranges and for non-contrast-
diac scan protocols is demonstrated with a z resolution phan- enhanced detection of coronary calcification with low radia-
tom (Fig. 5). The z resolution phantom consists of a Lucite plate tion exposure. Nevertheless, retrospective ECG gating has been
with rows of cylindrical holes of different diameters in the shown to be useful for quantification of coronary calcium with
transverse direction. The 4-slice CT scanner with 4 1 mm improved reproducibility owing to reconstruction with over-
collimation, 0.5 s rotation speed and 3 mm/s table feed (pitch lapping images and shorter scan times (31). Contrast-enhanced
0.375) can resolve structures of 1.0 mm in size using 1.3 mm visualization of small cardiac morphology and the coronary
reconstructed slice width and 0.5 mm image increment. With arteries with the best possible spatial and temporal resolution,
16-slice CT technology and submillimeter collimation the spa- as well as complete coverage of the cardio-thoracic vascula-
tial resolution is improved even at higher scan speed. Based on ture, is feasible only with retrospectively ECG-gated scan
16 0.75 mm collimation, 0.42 s rotation speed and 6.6 mm/s acquisition. With 4-slice CT scanners and retrospectively ECG-
table feed (pitch 0.31), 0.6 mm sized objects can be delineated gated spiral scanning, a true 3D data set of the cardiac and
using 0.8 mm reconstructed images and 0.4 mm image increment. coronary anatomy can be acquired with 0.6 0.6-mm in-plane
A computer model of an anthropomorphic heart phantom resolution and 1.0-mm longitudinal resolution based on 4
demonstrates the clinical relevance of the increased spatial 1-mm slice collimation. A 1012-cm scan range can be cov-
resolution with 16-slice CT acquisition (Fig. 6). The model ered in a 3040-s breath-hold time using 1-mm slices. The
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 29

Fig. 6. An anthropomorphic numerical heart and coronary artery phantom was used to evaluate the influence of spatial resolution and slice width
on visualization of coronary artery lumen and coronary plaque. The contrast enhanced left coronary artery segments (enhancement 250 HU)
contain plaques with different properties and a stent in the proximal left anterior descending (LAD) coronary artery: lipid plaque 30HU, fibrous
plaque 80 HU, calcified plaque 500 HU, and a stent with 50% in-stent lumen narrowing caused by a lesion with 30 HU (A). 16-slice CT with
16 0.75-mm collimation allows for retrospective reconstruction of different slice thicknesses between 0.8 mm and 3.0 mm (B). The phantom
was reconstructed with 1.3-mm (C), 1.0-mm (D), and 0.8-mm (E) slice width, and the proximal LAD was displayed with multiplanar
reformations. Differentiation of the lesions and visualization of in-stent lumen is possible with slice width 1.0 mm.

subsequent introduction of 8-slice CT scanners has reduced different institutions and manufacturers, may have widely
breath-hold times by about a factor of two, but this is still too varying dose consequences. The estimations of the effective
long for routine submillimeter slice scanning. Most recently, patient dose values in Table 1 (34) are based on calculations
16-slice CT scanners provide improved longitudinal resolution of the generally available computer program WinDose (35)
with submillimeter collimation (0.50.75 mm) and improved that have been validated with phantom measurements (36). It
in-plane spatial resolution of 0.5 0.5 mm. Despite thinner should be kept in mind that programs such as WinDose calcu-
slices, breath-hold times can be reduced to 1520 s for a 12-cm late effective dose to a mathematical phantom, and the dose to
scan range owing to simultaneous acquisition of 1216 detec- individual patients will vary depending on their size and
tor slices (24). weight, as well as the specific protocols used.
Diagnostically adequate image quality should be provided For the detection and quantification of coronary artery
at the minimum radiation exposure possible. Therefore, scan calcium, contrast enhancement is generally not used, although
protocols have to be developed for different applications with accurate measurement of coronary calcium is also feasible
optimized image quality, spatial resolution, and radiation using the thin-slice CT angiography protocol (37). For coro-
exposure. Various publications are available that discuss the nary CT angiography, the contrast medium injection must be
radiation exposure of MSCT in cardiac applications (32,33). carefully tailored by either using a test bolus or automatic bolus
The calculation of patient dose from CT examinations has triggering technique. Since scan times for imaging of the heart
become more complex since the introduction of multislice on modern 4-, 8-, or 16-slice CT scanners range from 15 to
scanners. Not only is dose related to tube current and voltage 40 s, 80 to 150 mL of contrast medium at injection rates between
and exposure time, but also to slice thickness and helical/ 3 and 5 mL/s is needed to maintain homogenous vascular opaci-
spiral pitch. Thus, different exam protocols, as developed by fication throughout the scan. Saline chasing has proven manda-
30 OHNESORGE, WESTERMAN, AND SCHOEPF

Table 1
Cardiac CT Scan Protocols With 4-Slice and 16-Slice CT

ECG-trigger (Ca-Score) ECG-gating (Ca-Score) ECG-gating (High-Res)

4-Slice CT 16-Slice CT 4-Slice CT 16-Slice CT 4-Slice CT 16-Slice CT

Scan range [mm] 120 120 120 120 100 100

Collimation per slice [mm] 2.03.0 1.03.0 2.02.5 1.02.0 1.01.25 0.51.0
Image thickness [mm] 3.0 3.0 3.0 3.0 1.31.5 0.61.25
Rotation Time [ms] 500 370500 500 370500 500 370500
Table Feed 812 mm/scan 1624 mm/scan 67.5 mm/s 1216 mm/s 34 mm/s 58 mm/s
kv/mA 120/100 120/100 120/100 120/100 120/300 120/400
Eff. Dose [mSv],
Male/Female
(No ECG Modulation) approx 0.6/1.0 approx 0.5/0.8 approx 2.0/3.0 approx 1.8/2.7 approx 6.0/9.0 approx 8.0/10.0
Eff. Dose [mSv],
Male/Female
(ECG-Modulation) n.a. n.a. approx 1.2/1.6 approx 1.2/1.6 approx 3.5/5.5 approx 5.0/6.5

tory for reduction of the volume of contrast medium needed for developed, and Ca-scoring data acquired with MSCT need care-
consistently high vascular enhancement. This also helps in ful clinical evaluation and comparison to EBCT results.
avoiding streak artifacts, which frequently arise from dense Both ECG-triggered sequential and ECG-gated helical/spi-
contrast material in the superior vena cava and the right atrium ral scanning can be used to acquire calcium-scoring data. Scan-
and sometimes interfere with the interpretation, especially of ning is performed without contrast enhancement in a
the right coronary artery. Techniques for contrast bolus optimi- cranio-caudal direction from the top of the pulmonary artery
zation have been developed in the past (38,39) but have not trunk to the apex of the heart (roughly a 12-cm scan range). A
been widely used, since reasonable results could be obtained by 3-mm slice width is routinely used for EBCT scanning, as it
adapting single-slice CT strategies for contrast administration represents the thinnest slice width that allows for a single
to dual- and 4-slice scanners. However, the introduction of breath-hold scan. 4-, 8-, and 16-slice CT scanners generally use
ever-faster CT acquisition techniques now requires careful slice widths of 2.0 to 3.0 mm as the closest match to the stan-
custom tailoring of the bolus for achieving adequate and con- dard EBCT protocols. A tube voltage of 120 kV is recom-
sistent contrast medium attenuation within the cardiovascular mended, as it provides the best relation of contrast-to-noise
system. ratio and radiation exposure. For most MSCT scanners,
100 mA tube current is adequate to achieve sufficiently high
CLINICAL APPLICATION signal-to-noise levels to detect small calcified lesions. The tube
QUANTIFICATION OF CORONARY CALCIFICATION current may need to be increased for large patients (e.g., to
EBCT has been established as a noninvasive imaging 150 mA) to maintain a diagnostic signal-to-noise level at the
modality for the detection and quantification of coronary cal- expense of increased radiation exposure.
cium by using the Agatston-scoring algorithm (2). With EBCT ECG-triggered scanning allows the scan volume to be cov-
scanning, typically 3-mm thick slices are acquired contigu- ered in 2025 s with 4-slice CT scanners and in 1015 s with
ously with prospective ECG triggering in mid-diastole and an 8- and 16-slice CT scanners. The effective patient dose is about
exposure time of 100 ms per slice. An effective dose of about or less than 1 mSv (see also Table 1). Since the newer scanners
0.9 mSv was reported for this protocol (40). A known limitation acquire multiple contiguous slices simultaneously, the likeli-
of coronary artery calcium scoring with EBCT is the high hood of slice misregistration in the z-axis is less than in the case
interscan variability associated with the technique (41). This of single-slice scanners. This, together with the reduced breath-
high variability has limited the usefulness of coronary artery hold time, suggests that multislice scanners will be the pre-
calcium measurements for tracking the progression of athero- ferred modality for this examination. Comparative studies of
sclerosis with time or for patients undergoing statin (lipid-low- EBCT and prospectively ECG-triggered MSCT were able to
ering) therapy, which may become a potentially powerful future demonstrate good measurement agreement in phantom experi-
application of this technique (42). ments (31) and high correlation in patient studies (40,44,45).
Since the introduction of sub-second gantry rotation, imag- However, a high mean interexamination variability of 2232%,
ing of coronary artery calcium has also been evaluated with comparable to the interexamination variability of two consecu-
single-slice (4) and dual-slice mechanical CT (43). Owing to tive EBCT examinations, is common with prospectively ECG-
increased imaging performance in terms of temporal resolution triggered MSCT, owing to the presence of motion artifacts and
and volume coverage, MSCT can be expected to provide coro- misregistration between the sequential acquisitions.
nary calcium quantification with substantially higher accuracy ECG-gated spiral scanning provides shorter breath-hold
and better reproducibility than mechanical single- and dual- times as compared to ECG-triggered scanning, as well as more
slice CT (31,40). Optimal acquisition techniques have to be consistent volume coverage based on overlapping images. A
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 31

Fig. 7. Different scan protocols for 4- and 16-slice CT were compared in a phantom study to investigate the sensitivity for the detection of small
amounts of coronary calcification. The phantom consisted of a Lucite cylinder embedded in an anthropomorphic chest phantom with inserts
of different concentrations of calcium hydroxylapatite (Ca HA) and different diameters (200, 400, and 800 mg/cm3 Ca HA) (A). Slices with
3-mm slice width were reconstructed based on scan data acquired with 133 mAs and 4 2.5-mm, 16 1.5-mm, and 16 0.75-mm collimation
(SOMATOM Sensation 4 and SOMATOM Sensation 16, Siemens). The smallest calcifications (1 mg mass, arrows) can be detected only with
the 16 0.75-mm protocol. A case study performed with 16-slice CT and 16 1.5-mm collimation reveals small calcifications in the middle
segment (B) and in the distal segment (C) of the right coronary artery (case study courtesy of the Department of Radiology, Grosshadern Clinic,
University of Munich, Germany).

12-cm scan range can be completed in 1520 s with 4-slice CT in the context of screening. Especially in this population it is
and in 610 s with 8- and 16-slice CT. The spiral interpolation imperative to limit radiation dose to a minimum. This can be
algorithms generate 3-mm thick images (full width at half achieved by properly adapting scan protocols (49), or by using
maximum) based on a slice thickness of 2.0 to 2.5 mm for 4- and sophisticated technical developments such as ECG-based tube
8-slice CT scanners and 1.252.0 mm for 16-slice CT scanners current modulation (28), which can decrease effective radia-
(see also Table 1). Optimal detection accuracy of small calcified tion exposure of the patient by as much as 50% (28).
lesions might be achieved with 16-slice scanners via The modulation transfer function of the convolution kernel
reconstruction of 3-mm thick images from scan acquisition that is used for image reconstruction has considerable influ-
with submillimeter collimation (46). Retrospectively ECG-gated ence on in-plane spatial resolution and signal-to-noise ratio,
spiral/helical scanning represents the preferred scan technique and thus can seriously affect quantitative measurements. For
for minimized inter-examination variability. Recent indepen- coronary calcium scanning, a medium-sharp convolution ker-
dent studies found interscan variability of about 10% or less for nel is used that provides moderate image noise for low radiation
repeat 4-slice CT scanning (31,47,48), which may be accurate exposure and about 0.6 0.6-mm in-plane resolution (approx
enough to detect significant changes in the total atherosclerotic 4 line pairs (lp)/cm at 50% and 9 lp/cm at 2% MTF). Edge
disease burden in patients with and without specific therapy. enhancement from a sharp kernel should be avoided, as it may
Interscan variability may be further improved with the latest lead to an overestimation of scores and to misleading artifac-
16-slice CT scanners, primarily based on faster rotation time tual lesions at the pericardium close to the coronary arteries.
and increased scan speed (Fig. 7). In comparison to prospec- The traditional method for calculating calcium scores is that
tively ECG-triggered technique, the drawback to using retro- described by Agatston (2). Recent studies describing better
spective ECG gating is higher patient dose. Frequently, healthy, results for interscan and for inter- and intra-observer variability
asymptomatic individuals undergo coronary calcium scoring strongly suggest that truly quantitative approaches may offer
32 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 8. Presentation of the platform used for quantification of coronary calcification (syngo Calcium Scoring, Siemens). Lesions exceeding
the calcium threshold of 130 HU are identified with 3D-based picking and viewing tools, and are assigned to the different coronary arteries
(left main, left anterior descending, circumflex, and right coronary arteries). Coronary calcifications are quantified my means of Agatston score,
calcium volume, and calcium mass. Calibration factors that are predetermined with phantom measurements and that depend on the scan protocol
are the basis for calculation of calcium mass. The quantitative measurements are displayed and reported in table format.

significant advantages over the Agatston scoring method Cardiac and coronary CT imaging with MSCT requires
(31,47,48). The new methods include both a measurement of ECG-gated thin-slice spiral scan protocols with table speed
lesion volume (but without the arbitrary weighting factor used determined by the patients heart rate in order to ensure com-
in Agatston scoring) and calculation of calcium mass. plete phase-consistent coverage of the heart with overlapping
Advanced software platforms (Fig. 8) allow for assessment of images. Most MSCT scanners provide scan protocols with
equivalent volume and total calcified plaque burden in terms of overlapping helical/spiral pitches between 0.2 and 0.4 that
absolute calcium mass, based on actual scanner-specific cali- enable gapless volume coverage for heart rate higher than
bration factors (47,48). The latter technique probably has the approx 40/min. The helical/spiral pitch is defined as the table
greatest potential to increase accuracy, consistency, and feed per gantry rotation divided by the X-ray beam width (equal
reproducibility of coronary calcium assessment (50), and thus to slice width multiplied by the number of slices).
may replace traditional scoring methods in the near future (51). Usually, the cranio-caudal size of the heart to be covered by
CARDIAC AND CORONARY CT ANGIOGRAPHY the scan is in the range 1012 cm. Four-slice CT scanners with
The image requirements for noninvasive coronary CT an- 500 ms rotation time and slice thickness of 1.01.25 mm can
giography are high spatial resolution, low-contrast detectabil- cover the entire heart during a 3040-s breath-hold. With 8-slice
ity, and temporal resolution. In addition to scanner CT scanners, the breath-hold time can be reduced to 2025 s with
performance, image quality depends on various patient param- the same spatial and temporal resolution. Recent 16-slice CT
eters, and optimization of examination protocols is critical for scanners provide high-resolution cardiac scan protocols with
reliable diagnostic studies. submillimeter slice collimation (0.50.75 mm) and in-plane
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 33

Fig. 9. Coronary CT angiography examination of a patient with suspected coronary artery disease (A) using a 16-slice CT scanner in correlation
with conventional coronary angiography (B). Scan protocol: 16 0.75-mm detector collimation, 1.0-mm reconstructed slice width, 420-ms
rotation speed, 12-cm scan range covered in 16 s (SOMATOM Sensation 16, Siemens). Coronary CT angiography reveals high-grade stenosis
in the proximal left descending coronary artery near the bifurcation of the first diagonal branch (arrow, 90% stenosis) and in the middle segment
of the left descending coronary artery next to the bifurcation of the second diagonal branch (double arrow, 70% stenosis). Both lesions are
confirmed by coronary angiography (case courtesy of Dr. C. S. Soo, HSC Medical Center, Kuala Lumpur, Malaysia).

Fig. 10. Coronary CT angiography examination of a patient who presented with chest pain in the emergency room, using a 16-slice CT scanner.
Scan protocol: 16 0.5-mm detector collimation, 0.6-mm reconstructed slice width, 400-ms rotation speed, 12-cm scan range covered in 32
s (Aquillion 16, Toshiba). Visualization with maximum intensity projection reveals a significant amount of noncalcified plaque and moderate
lumen narrowing in the proximal left descending coronary artery (A, arrow) and a high-grade lumen obstruction in the right coronary artery
(B, double arrow) (case courtesy of Dr. H. Anno, Fujita Health University, Japan).

spatial resolution up to 0.5 0.5 mm (52,53). The image recon- their breath-hold period. These precautions are particularly
struction protocols are adapted to the clinical priorities. As a important when scanning with systems of 4 or 8 slices. With 16-
result of the factors described above, accurate imaging of small slice CT scanners and submillimeter collimation, usually slices
coronary vessels with 4-slice scanners is limited. An important not thinner than 1.0 mm are reconstructed, with 0.6 0.6-mm
part of the scan protocol involves patient preparation. Patients in-plane resolution to achieve optimal contrast resolution for
should practice breath-holding for the expected duration of the assessment of coronary narrowing and intraluminal lesions
scan so that they are well prepared. A short period of oxygen (Figs. 9 and 10). Special algorithms with image width of less
breathing prior to the procedure can also help patients extend than 1 mm and in-plane resolution of 0.5 0.5 mm can be
34 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 11. Coronary CT angiography examination of a patient after percutaneous transluminal coronary angioplasty with stent in the distal right
coronary artery, using a 16-slice CT scanner. Scan protocol: 16 0.75-mm detector collimation, 0.75-mm reconstructed slice width, 420-ms
rotation speed, 12-cm scan range covered in 16 s (SOMATOM Sensation 16, Siemens). The patient presented for follow-up examination after
the intervention and was examined to rule out re-stenosis. Coronary CT angiography reveals a patent stent lumen (arrow) but calcified and
noncalcified atherosclerotic lesions and high-grade lumenal narrowing of about 70% proximal to the stent (double arrow). (A) The lumenal
narrowing can be readily displayed with 3D volume-rendering technique in combination with cut planes that remove overlying anatomy. (B)
Multiplanar reformation allows for visualization of the open in-stent lumen and of the narrowed lumen proximal to the stent. (C) Additional
display of the same anatomy with maximum intensity projection provides clear assessment of the calcified and noncalcified atherosclerotic
lesions related to the stenotic lesion. (D) Conventional angiography confirms the patent stent and the 70% stenosis, which was successfully
dilated in the same session (case courtesy of Dr. C. S. Soo, HSC Medical Center, Kuala Lumpur, Malaysia).

applied for high-resolution reconstruction of small high-con- sition with highly overlapping helical/spiral pitch (i.e., pitch
trast structures (e.g., coronary stents and calcified coronary much less than 1) in a 10-cm scan range results in an effective
segments) in a limited range (Figs. 11 and 12). dose of approx 78 mSv for males and approx 911 mSv for
Rotation times down to 370 ms and the extended number of female patients (based on protocol examples in Table 1).
slices, up to 16, result in a reduced scan time of 1520 s (see Despite increased tube current and spatial resolution with
also Table 1). Thus, 16-slice CT can also cover larger scan 16-slice CT scanners, radiation exposure does not considerably
ranges of 1820 cm with ECG-gated thin-slice spiral scan pro- increase, due to better dose utilization of 16-slice detector
tocols in a reasonably short breath-hold of 2530 s that enables geometry (16,17). However, radiation exposure increases with
high-resolution imaging of most parts of the great thoracic reduced spiral pitch and with extension of the scan range. Using
vasculature and coronary bypass grafts over their entire course ECG-gated dose modulation, radiation exposure can be reduced
(Fig. 13). by 3050% depending on heart rate, provided that images do
Optimization of scan protocols in terms of radiation expo- not need to be reconstructed over a wide range of the RR inter-
sure is particularly important for contrast-enhanced CT imag- val. ECG-gated dose modulation can reduce radiation exposure
ing of the coronary arteries. Sufficiently high spatial resolution to approx 3.56.0 mSv for male and to approx 4.58.0 mSv for
and low-contrast resolution has to be achieved in patients of all female patients, and works best in patients with a reasonably
sizes at the lowest possible radiation exposure. For tube voltage stable heart rate during the scan.
of 120 kV and 500 ms rotation time, tube current of approx 300 The overall diagnostic quality of coronary CT angiography
mA should be used for imaging 4- and 8-slice CT with slice largely depends on choice of the appropriate reconstruction
width of 1 to 1.25 mm. The tube current may need to be time point within the cardiac cycle, the rate and stability of the
increased to approx 350400 mA for 16-slice CT using submil- patients heart rate during the examination, and the effective-
limeter slices and faster rotation time. ECG-gated scan acqui- ness of contrast enhancement. The motion pattern of the left
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 35

Fig. 12. Coronary CT angiography examination of a patient after percutaneous transluminal coronary angioplasty with two stents in the left
descending coronary artery using a 16-slice CT scanner. Scan protocol: 16 0.5-mm detector collimation, 0.6-mm reconstructed slice width,
400-ms rotation speed, 12-cm scan range covered in 32 s (Aquillion 16, Toshiba Medical). Cross-sectional reconstruction of orthogonal
multiplanar reformations and longitudinal visualization of the vessel based on automatic center-line detection and data segmentation shows
both stents open (Vessel Probe, Vital Images) (case courtesy of Dr. Strobie, Florida Institute for Advanced Diagnostic Imaging).

heart and the left anterior descending and circumflex coronary ogy, such as the cardiac chambers and the great vessels, is
arteries follows the left-ventricular contraction, whereas the possible in patients who present with higher heart rates without
right coronary artery moves synchronously with the right heart, resorting to drugs, although some image artifacts may be
i.e., the right atrium. Because of these different motion pat- present.
terns, different reconstruction time points over the cardiac cycle The reliability of multislice cardiac and coronary CT
are frequently necessary in order to display the different coro- angiography in patients with arrhythmias is limited. However,
nary vessels with a minimum of motion artifact (26,27). Most misinterpretations of the ECG signal can be partially compen-
studies agree that patient heart rate is inversely related to image sated by retrospective editing of the ECG trace. Persistent
quality of cardiac and coronary CT angiography (1214,26,27) irregular heart rates, such as in patients with atrial fibrillation,
because of motion artifacts. It has been shown that MSCT pro- are contraindications for coronary CTA, but the assessment of
vides better diagnostic accuracy and reliability of results in the greater cardiac morphology such as the ventricles and atria
patients with slow heart rates. Study data based on 4-slice CT usually remains possible.
with 500-ms rotation time indicates that an upper limit to The presence of heavy coronary calcification may limit the
achieve consistently appropriate image quality lies between 65 value of CT coronary imaging, because beam hardening and
and 75 beats per minute (bpm) (26,27). At higher heart rates, partial volume effects, and an inability to distinguish between
adequate image quality can be achieved by using segmented calcium and contrast, can completely obscure the coronary
image reconstruction, but overall results may be less consistent lumen. Metal objects such as stents, surgical clips, and sternal
and reproducible. Even with the recent reduction of gantry ro- wires can also obscure the evaluation of underlying structures.
tation times to as little as 370 ms (Fig. 14), heart rate still remains Use of the thinnest possible slice width reduces partial-volume
the crucial factor in determining motion artifact and thus image artifacts and improves visualization of calcified coronary seg-
quality. In those patients with heart rates above the 7580 bpm ments. Additionally, dedicated filtering could be beneficial to
threshold, the practical options are use of segmented recon- the imaging of calcified vessels.
struction to improve temporal resolution where possible, or Optimization of contrast media injection protocols for car-
reduction of the heart rate pharmacologically by administering diac and coronary CT angiography is aimed at providing
beta-blockers. Reliable evaluation of larger cardiac morphol- homogenous enhancement within the entire course of the coro-
36 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 13. Coronary CT angiography examination of two patients using a 16-slice CT scanner. with 16 0.75-mm detector collimation,
420-ms rotation speed, and extended volume coverage of 15 cm in order to visualize the heart, the ascending aorta, and the pulmonary vessels
(SOMATOM Sensation 16, Siemens). (A) The first case demonstrates abnormal anatomy of the left descending coronary artery, including
multiple fistula. Real-time 3D rendering is of advantage for assessment of the complex anatomy conditions. (B) In the second case, three
bypasses to right coronary artery, left anterior descending coronary artery, and marginal branch are visualized with 3D volume-rendering
technique. The bypass to the left anterior descending coronary artery (arrow) shows a patent proximal and distal anastomosis. However, a 50%
stenosis (arrow head) is present in the proximal part of the bypass. The other two bypasses reveal open lumen over the entire course (cases
courtesy of Dr. G. Lo, Department of Radiology, Hongkong Sanatorium Hospital [A] and the departments of Radiology and Cardiology, Rhn-
Klinikum Bad Neustadt, Germany [B]).

nary arteries in order to facilitate density-threshold-dependent

2 and 3D visualization. Contrast attenuation within the vessel
should be high enough to allow for lesion detection, but not
high enough to obscure calcified coronary artery wall lesions.
Adequate opacification can be achieved in the majority of pa-
tients over the course of a 40-s scan with 140 mL of 300 mg/mL
iodinated contrast material injected at a flow rate of 3.5 mL/s.
Owing to increased acquisition speed with 16-slice CT (approx
20 s scan time) the amount of contrast media can be reduced to
80100 mL, but the injection rate will need to be increased to
4 to 5 mL/s. The use of reliable contrast-tracking software
remains the best option for achieving the desired opacification

Fig. 14. (left) Coronary CT angiography examination to rule out

coronary artery disease using a 16-slice CT scanner. Scan protocol:
16 0.75-mm detector collimation, 1.0-mm reconstructed slice width,
370-ms rotation speed, 12-cm scan range covered in 16 s (SOMATOM
Sensation 16 Cardiac, Siemens, Germany). Display of the coronary
tree with 3D volume-rendering technique in combination reveals all
main coronary segments normal. Also the rapidly moving right coro-
nary artery can be displayed free of motion owing to the fast rotation
speed (case courtesy of Department of Cardiology, University of
Erlangen, Germany).
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 37

Fig. 15. CT angiographic examination of a patient with an occlusion of the left anterior descending coronary artery using a 16-slice CT scanner.
Scan protocol: 16 0.75-mm detector collimation, 1.0-mm reconstructed slice width, 420-ms rotation speed, 12-cm scan range covered in 16
s (SOMATOM Sensation 16, Siemens). In addition to reconstruction of the coronary artery tree, ECG-gated spiral image reconstruction was
performed in 10 different time points during the cardiac cycle with a distance of 10% of the RR interval to provide input data for cardiac function
evaluation. End-diastolic volume and end-systolic volume of the left ventricle as well as ejection fraction can be readily assessed based on short
axis and long axis multiplanar reformations with 5-mm thickness that are loaded into a dedicated functional evaluation software (syngo Argus,
Siemens) (case courtesy of the Department of Radiology, Tbingen University, Germany).

in patients with a wide range of cardiac output. Use of saline permits retrospective reconstruction of images at all phases of
chasing technique (e.g., with a bolus of 50 mL of saline injected the cardiac cycle, it is a relatively simple matter to automati-
immediately after the iodine bolus) may be helpful for better cally segment the chamber and generate left-ventricular ejec-
contrast bolus utilization and for reducing streak artifacts aris- tion fractions and wall motion using the same data set collected
ing from dense contrast material in the superior vena cava and for coronary CTA. Thus, the diagnosis of cardiac and coronary
the right heart. In order to minimize the likelihood of the latter morphology and also of basic cardiac function parameters can
effect, the injection site should be considered carefully. Thor- be derived from a single contrast-enhanced ECG-gated spiral
ough patient preparation remains an integral part of achieving examination with thin-section acquisition. Standard analysis
consistent results. techniques yield most of the metrics associated with cardiac
EVALUATION OF CARDIAC FUNCTION function, including ejection fraction, end-diastolic and end-
In addition to the diagnosis of cardiac and coronary mor- systolic volumes, stroke volume, and cardiac output. Left-ven-
phology, evaluation and quantification of cardiac function pro- tricular regional function can also be quickly displayed on polar
vides important information for the assessment of cardiac and maps of percent wall thickness, wall motion, and ejection frac-
coronary diseases. Because ECG-gated helical/spiral scanning tion (Figs. 15 and 16).
38 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 16. Cardiac Function Analysis software (Toshiba Medical Systems) reconstructs short- and long-axis images of the heart at multiple phases
of the cardiac cycle. From this data set, ejection fraction, cardiac output, wall motion, and other parameters are calculated and displayed. This
provides additional diagnostic information without any additional radiation to the patient (case courtesy of Dr. Shapiro, Johns Hopkins
University, Baltimore).

First study results (54) show that basic cardiac function Rapid cardiac motion during the systolic phase of the car-
parameters derived with 4-slice CT correlate well with the gold- diac cycle can cause motion artifacts in the end-systolic recon-
standard techniques MRI and coronary angiography, based on structions. However, correct delineation of the ventricle walls
a standardized heart phase selection for end-diastolic and end- is usually still possible with sufficient accuracy. The latest
systolic CT reconstruction and semi-automated evaluation tools 16-slice CT scanners have the potential to further improve the
(Figs. 15 and 16). accuracy of cardiac function measurement as compared to
Although dose is significantly reduced during end diastole 4- and 8-slice CT scanners, based on increased gantry rotation
and systole with ECG-controlled tube current modulation, suf- speed.
ficient contrast resolution can still be obtained for functional CARDIO-THORACIC IMAGING
evaluation. Use of MPRs in short and long heart axis with thick- Thoracic CT studies are frequently degraded by motion
ness of 58 mm enables appropriate delineation of the ventricle artifacts caused by transmitted cardiac pulsation. Typical diag-
wall in both end-diastolic and end-systolic reconstruction. nostic pitfalls related to this effect are false-positive findings of
Automated direct 3D reconstruction of oblique planes in pre- aortic dissection and distortion of paracardiac lung segments.
defined views such as short and long heart axis and in multiple Suppression of cardiac pulsation artifacts improves image qual-
phases of the cardiac cycle will enable a more efficient ity in CT studies of the thorax, including the heart. Important
workflow for cardiac function analysis in the future. indications are the planning and follow-up of surgical proce-
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 39

Fig. 17. Electrocardiogram (ECG)-gated examination of the thoracic aorta in a 30-s breath-hold using a 16-slice CT scanner with 16 0.75-mm
detector collimation and 420-ms rotation speed. The thoracic vasculature is displayed with 3D volume-rendering technique. Based on ECG-gated
reconstruction that suppresses cardiac pulsation, the entire thoracic aorta can be evaluated free of pulsation artifacts (case courtesy of the
Department of Radiology, Cleveland Clinic Foundation, USA).

dures, diagnosis or exclusion of aortic dissection and aortic nation with specialized ECG-gated cone-beam reconstruction
aneurisms, detection of pulmonary embolism, assessment of techniques that enable faster volume coverage (30).
congenital heart and vessel disease, and early detection and
reliable diagnosis of paracardiac lung nodules. FUTURE PERSPECTIVE
Prospectively ECG-triggered and retrospectively ECG- Further CT technology enhancements are required for accu-
gated protocols have been successfully applied to suppress rate and consistent diagnosis of cardiac and coronary diseases,
cardiac-related motion artifacts in thoracic studies. With 4-slice including the detection and quantification of coronary artery
CT, ECG-triggered or ECG-gated thin-slice scanning of the stenosis and coronary atherosclerotic lesions. Detection and
thorax is usually not possible within a single breath-hold, but quantification of the degree of lumen narrowing with the ability
can be used for re-scanning that part of the thorax where most to differentiate a 1020% change in vessel cross-sectional area
pulsation is present. With the substantial performance increase represents a future goal for cardiac CT imaging. To meet this
of 16-slice CT, retrospectively ECG-gated coverage of the tho- goal, future CT systems need to achieve a spatial resolution in
rax is feasible in a single breath-hold and with a single contrast all three dimensions (isotropic voxels) of 0.40.5 mm for visu-
injection (Fig. 17). Thus, this technique now has the potential alization of the main coronary vessels and 0.3 mm or better for
to be used as a standard examination for multislice spiral smaller branches. A recent study indicates that a heart-rate-
examination of the lung and thoracic vasculature. As a future independent temporal resolution of 100 ms or less allows for
step, ECG-gated thoracic imaging may become possible even virtual elimination of motion artifacts in phases of the cardiac
with submillimeter resolution owing to 16-slice CT in combi- cycle with limited motion, with heart rates up to approx 100
40 OHNESORGE, WESTERMAN, AND SCHOEPF

Fig. 18. (A) Illustration of a prototype system with large-area detector technology, capable of acquiring 256 0.5-mm slices simultaneously
(Toshiba Medical Systems). Experimental ex vivo and human studies demonstrate the potential of such future CT scanners to cover large patient
volumes with a single gantry rotation. (B) The 3D scan of the chest demonstrates the potential for capturing the heart in one rotation (images
courtesy of Y. Saito, Toshiba Medical Systems).

Fig. 19. The anthropomorphic heart and coronary artery phantom displayed in Fig. 6 was used to compare the maximum possible spatial
resolution of 16-slice CT with future area detector CT systems (A) that may provide detector pixel sizes below 0.5 mm (CT prototype with flat-
panel detector system, Siemens). Such systems may be able to cover the heart in a single scanner rotation with isotropic resolution of 0.25 mm.
Extremely high spatial resolution can improve visualization of in-stent lumen and small coronary segments compared to 16-slice CT (B), but
do not necessarily provide improved delineation of noncalcified coronary lesions because of compromised signal-to-noise ratio (C).

bpm (55). For motion-free imaging at very high heart rates and substantial increase of radiation exposure, which should not
in phases with rapid cardiac motion for analysis of cardiac exceed the amount of invasive diagnostic coronary angiography.
function, a temporal resolution of 50 ms or less might be The very short acquisition times of EBCT, down to 50 ms,
required (56). Scan acquisition within a single, short breath- combined with prospectively ECG-triggered scanning, enable
hold is mandatory for minimized contrast-medium injection motion-free imaging of the coronary arteries for patients with
and to avoid respiratory artifacts. Breath-hold times of 15 s or moderate and higher heart rates and stable sinus rhythm. How-
less are appropriate for stable patients, but breath-hold times of ever, the restrictions in spatial resolution and contrast-to-noise
10 s or less are advisable for patients who are ill or uncoopera- ratio as well as rather long breath-hold times limit the ability of
tive. Ideally, all data of the complete heart anatomy would be EBCT today to reliably visualize all main coronary artery seg-
acquired within a single heart cycle or less without patient ments and noncalcified atherosclerotic plaques. New EBCT
movement. All these requirements related to spatial resolution, detectors are under evaluation that allow for simultaneous
temporal resolution, and scan time have to be achieved without acquisition of two 1.5-mm slices and increased in-plane spatial
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 41

Fig. 20. CT/positron emission tomography (PET) examination of a patient with a known occlusion of the left anterior descending coronary
artery and history of myocardial infarction (biograph Sensation 16, Siemens). The 3D volume-rendering reconstruction of the ECG-gated
16-slice CT scan reveals the occluded coronary vessel and a related infarct scar. The PET scan demonstrates a perfusion defect and necrotic
myocardium at the anterior wall (case courtesy of the Department of Radiology of Tbingen University, Germany, and the Department of
Nuclear Medicine of the Technical University of Munich, Germany).

resolution via finer structuring of the elements of the fixed these algorithms to date is restricted to cerebral and bone
detector ring. With these detectors, the heart can be scanned with imaging. Practical cardiac imaging has also benefited tremen-
1.5-mm slices in a 3040-s breath-hold. However, limited sig- dously from the development of advanced software, particu-
nal-to-noise ratio will remain a major challenge for EBCT tech- larly with regard to image reconstruction. Another approach to
nology for high-resolution cardiac and coronary artery imaging. improved temporal resolution may lie with advances in seg-
The temporal resolution of current multislice mechanical mented image reconstruction. While current algorithms are
CT scanners needs to be improved to provide motion-free and robust, using as many as four cardiac cycles, these and related
robust coronary imaging also for moderate and high heart rates. techniques may develop to further extend the range of patient
Since the introduction of CT in 1972, there has been a persistent heart rates that can reliably be imaged.
and successful effort to reduce gantry rotation times, reaching Area detector technology and related new cone beam recon-
a remarkable 0.4 . Further increased rotation speed is still a very struction techniques are being researched that may allow the
real goal; however, the rotational forces on the gantry compo- entire coronary anatomy to be covered in a single heartbeat
nents (tube, generator, and detector) increase as the square of without movement of the table. Prototype systems with large
the rotation speed. area detectors exist, capable of acquiring as many as 256 slices
Thinner slices, leading to improved z-axis resolution, are of 0.5 mm simultaneously. Alternative designs can provide
technically feasible, but require a higher radiation dose to in-plane and through-plane spatial resolution even up to 0.2
maintain a high contrast-to-noise ratio, and more detector-slices mm using flat-panel detector technology (Figs. 18 and 19).
to maintain a short scan time (<20s). As a complement to fur- With these CT scanners, imaging of high-resolution morphol-
ther increased spatial resolution, advanced beam-hardening and ogy as well as dynamic and functional information via repeated
metal artifact reduction algorithms can be developed that im- scanning of the same scan range may become possible. The
prove imaging of calcified and stented coronary vessels; use of application potential of such technology is being evaluated with
42 OHNESORGE, WESTERMAN, AND SCHOEPF

experimental systems using phantom models and post mortem ter 16-slice scanning and increased gantry rotation speed for cardiac
imaging. RFo, Fortschr Rntgenstr 2002;174:10221027.
hearts. Initial experience shows that todays area detector tech-
18. Funabashi N, Komiyama N, Yanagawa N, et al. Coronary artery
nology is still too limited in contrast resolution, and the high patency after metallic stent implantation evaluated by multislice
radiation dose that is needed to provide adequate signal-to- computed tomography. Circulation 2003;107:147148.
noise ratio even for high-contrast studies is unacceptable for 19. Kopp AF, Kttner A, Heuschmid M, Schrder S, Ohnesorge B,
use in human subjects. Claussen CD. Multidetector-row CT cardiac imaging with 4 and 16
slices for coronary CTA and imaging of atherosclerotic plaques. Eu
The latest MSCT scanners also allow combining 16-slice
Radiol 2002;12(Suppl 2): S17S24.
CT with PET cameras. These systems may allow for a clinical 20. Flohr TG, Kttner A, Bruder H, et al. Performance evaluation of a
combination of cardiac CT imaging and cardiac PET scans in multi-slice CT system with 16-slice detector and increased gantry
a single examination and subsequent fusion of the information rotation speed for isotropic submillimeter imaging of the heart. Herz
on cardiac morphology and myocardial function and metabo- 2003;28:719.
21. Flohr TG, Schoepf UJ, Kttner A, et al. Advances in cardiac imaging
lism (Fig. 20). The clinical potential of these scanners is cur-
with 16-slice CT-systems. Acad Radiol 2003;10:386401.
rently under evaluation, and initial study data can be expected 22. Flohr T, Ohnesorge B. Heart-rate adaptive optimization of spatial
in the near future. and temporal resolution for ECG-gated multi-slice spiral CT of the
heart. J Comp Assist T 2001;25:907923.
REFERENCES 23. Flohr T, Stierstorfer K, Bruder H, et al. New technical developments
1. Boyd DP, Lipton MJ. Cardiac computed tomography. Proc IEEE in multislice CTpart 1: approaching isotropic resolution with sub-
1982;71:298307 millimeter 16-slice scanning. Fortschr Rntgenstr (RFo) 2002;
2. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, 174:10221027.
Detrano R. Quantification of coronary artery calcium using ultrafast 24. Flohr T, Ohnesorge B, Bruder H, et al. Image reconstruction and
computed tomography. JACC 1990;15:827832. performance evaluation for ECG-gated spiral scanning with a
3. Achenbach S, Moshage W, Ropers D, Nssen J, Daniel WG. Value 16-slice system. Med Phys 2003;30(10):26502662.
of electron-beam computed tomography for the non-invasive detec- 25. Bruder H, Flohr TG, Stierstorfer K, Rauscher A, Hlzel A, Schaller
tion of high-grade coronary artery stenoses and occlusions. N Engl S. ECG-gated dynamic cardiac volume imaging with CT area detec-
J Med 1998;339:19641971. tors (abstract). Radiology 2002;225(P):310.
4. Becker CR, Jakobs TF, Aydemir S, et al. Helical and single-slice 26. Hong C, Becker CR, Huber A, et al. ECG-gated reconstructed multi-
conventional CT versus electron beam CT for the quantification of detector row CT coronary angiography: effect of varying trigger
coronary artery calcification. AJR 2000;174:543547. delay on image quality. Radiology 2001;220:712717.
5. Bahner ML, Bse J, Lutz A, Wallschlger H, Regn J, van Kaick G. 27. Kopp AF, Schrder S, Kttner A, et al. Coronary arteries: retrospec-
Retrospectively ECG-gated spiral CT of the heart and lung. Eur tively ECG-gated multi-detector row CT angiography with selective
Radiol 1999;9:106109. optimization of the reconstruction window. Radiology 2001;221:
6. Ohnesorge B, Flohr T, Schaller S, et al. Technische Grundlagen und 683688.
Anwendungen der Mehrschicht CT. Radiologe 1999;39:923931. 28. Jakobs T, Becker CR, Ohnesorge B, Flohr T, Schoepf UJ, Reiser
7. Taguchi K, Aradate H. Algorithm for image reconstruction in multi- MF. Reduction of radiation exposure with ECG-controlled tube
slice helical CT. Med Phys 1998;25(4):550561. current modulation for retrospectively ECG-gated helical scans of
8. Ohnesorge B, Flohr T, Becker CR, et al. Cardiac imaging by means the heart. Eur Radiol 2002;12:10811086.
of electrocardiographically gated multisection spiral CT: initial 29. Flohr T, Prokop M, Schpf, et al. A new ECG-gated multislice spiral
experience. Radiology 2000;217:564571. CT scan and reconstruction technique with extended volume
9. Kachelriess M, Ulzheimer S, Kalender WA. ECG-correlated image coverage for cardio-thoracic applications. Eur Radiol 2002;12:
reconstruction from subsecond multi-row spiral CT scans of the 15271532.
heart. Med Phys 2000;27:18811902. 30. Flohr T, Bruder H, Kttner A, Heuschmid M, Schaller S, Ohnesorge
10. Ohnesorge B, Becker CR, Flohr T, Reiser MF. Multi-Slice CT in BM. ECG-gated spiral scanning of the lung and mediastinal vessels
Cardiac ImagingTechnical Principles, Clinical Application and with optimized temporal resolution and cone-correction on a
Future Developments. Springer, New York: 2002. 16-slice CT system: performance evaluation and initial clinical
11. Knez A, Becker CR, Leber A, Ohnesorge B, Reiser MF, Haberl R. results (abstract). Radiology 2002;225(P):449.
Non-invasive assessment of coronary artery stenoses with 31. Kopp AF, Ohnesorge B, Becker C, et al. Reproducibility and accu-
multidetector helical computed tomography. Circulation 2000;101: racy of coronary calcium measurement with multidetector-row ver-
e221e222. sus electron beam CT. Radiology 2002;225:113119.
12. Niemann K, Oudkerk M, Rensing BJ, et al. Coronary angiography 32. Ulzheimer S, Halliburton SS, McCollough CH, Becker CR, White
with multi-slice computed tomography. Lancet 2001;357:599603. RD, Kalender WA. Evaluation of image quality and calcium scoring
13. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral performance in multislice cardiac computed tomography (abstract).
computed tomography angiography for determination of coronary Radiology 2002;221(P):458.
artery stenoses. Am J Cardiol 2002;88:11911194. 33. Hunold P, Vogt FM, Schmermund A, Kerkhoff G, Debatin JF,
14. Kopp AF, Schrder S, Kttner A, et al. High resolution multi-slice Barkhausen J. Radiation exposure during noninvasive coronary
computed tomography with retrospective gating for angiography artery imaging: comparison of multislice CT and electron beam CT.
in coronary arteries: results in 102 patients. Eur Heart J 2002;23: Radiology 2002;221(P):503504.
17141725. 34. Schoepf UJ, Becker CR, Ohnesorge BM, Yucel EK (2003). CT of
15. Schrder S, Kopp AF, Baumbach A, et al. Non-invasive detection coronary artery disease. Radiology, in press.
and evaluation of atherosclerotic plaque with multi-slice computed 35. Kalender WA, Schmidt B, Zankl M, Schmidt M. A PC program for
tomography. JACC 2001;37:14301435. estimating organ dose and effective dose values in computed tomog-
16. Flohr T, Stierstorfer K, Bruder H, Simon J, Schaller S. New techni- raphy. Eur Radiol 1999;9:555562.
cal developments in multislice CT, part 1: approaching isotropic 36. Schmidt B, Ulzheimer S, Kalender WA. Dose in multi-slice cardiac
resolution with sub-mm 16-slice scanning. Fortschr Rntgenstr CT: assessment of organ effective dose values with Monte Carlo
2002;174:839845. methods (abstract). Radiology 2001;221(P):414.
17. Flohr T, Stierstorfer K, Bruder H, Simon J, Schaller S, Ohnesorge B. 37. Hong C, Becker CR, Schoepf UJ, Ohnesorge B, Brning R, Reiser
New technical developments in multislice CT, part 2: sub-millime- MF. Absolute quantification of coronary calcification in non-con-
 CHAPTER 3 / MULTISLICE SCAN TECHNIQUES 43

trast and contrast-enhanced multislice CT studies. Radiology 47. Ohnesorge B, Kopp AF, Fischbach R, et al. Reproducibility of coronary
2002;223:474480. calcium quantification in repeat examinations with retrospectively
38. Fleischmann D, Rubin GD, Bankier AA, Hittmair K. Improved ECG-gated multislice spiral CT. Eur Radiol 2002;12:15321540.
uniformity of aortic enhancement with customized contrast medium 48. Moser K, Bateman T, Case J, et al. The influence of acquisition
injection protocols at CT angiography. Radiology 2000;214:363371. mode on the reproducibility of coronary artery calcium scores using
39. Schweiger GD, Chang PJ, Brown BP. Optimizing contrast enhance- multi-detector computed tomography (abstract). Circulation
ment during helical CT of the liver: a comparison of two bolus track- 2002(Suppl. II);106(19):II479.
ing techniques. Am J Roentgenol 1998;171:15511558. 49. Hong C, Bae KT, Pilgram TK, et al. Coronary artery calcium mea-
40. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium surement with multi-detector row CT: in vitro assessment of effect
measurement: agreement of multirow detector and electron beam of radiation dose. Radiology 2002;225:901906.
CT. Am J Roentgenol 2001;176:12951298. 50. Ulzheimer S, Kalender WA. Assessment of calcium scoring perfor-
41. Achenbach S, Ropers D, Mohlenkamp S, et al. Variability of repeated mance in cardiac computed tomography. Eur Radiol 2003;13:484497.
coronary artery calcium measurements by electron beam tomogra- 51. Becker CR, Schpf UJ, Reiser MF. Coronary calcium scoring: medi-
phy. Am J Cardiol 2001;87:210213. cine and politics. Eur Radiol 2003;13:445447.
42. Achenbach S, Ropers D, Pohle K, et al. Influence of lipid-lowering 52. Nieman K, Cademartiri F, Lemos PA, et al. Reliable noninvasive
therapy on the progression of coronary artery calcificationa pro- coronary angiography with fast submillimeter multislice spiral com-
spective study. Circulation 2002;106:10771082. puted tomography. Circulation 2002;106:20512054.
43. Shemesh J, Apter S, Stroh CI, et al. Tracking coronary calcification 53. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery
by using dual-section spiral CT: a 3-year follow-up. Radiology stenoses with thin-slice multi-detector row spiral computed tomogra-
2000;217:461465. phy and multiplanar reconstruction. Circulation 2003;107:664666.
44. Daniell A, Friedman J, Berman D, et al. Concordance of coronary 54. Juergens KU, Grude M, Fallenberg EM, Heindel W, Fischbach R.
calcium estimation between multi-detector and electron beam CT Using ECG-gated multidetector CT to evaluate global left ventricu-
(abstract). Circulation 2002(Suppl. II);106(19):II479. lar myocardial function in patients with coronary artery disease.
45. Stanford W, Thompson B, Burns TL, Heery S, Burr M. Multi-detec- AJR Am J Roentgenol 2002;179:15451550.
tor helical CT versus electron beam CT in the quantification of coro- 55. Achenbach S, Ropers D, Holle J, et al. In-plane coronary arterial
nary artery calcification: emphasis on lower calcium scores. motion velocity: measurement with electron beam CT. Radiology
Radiology 2004;230:397402. 2000;216:457463.
46. Ohnesorge B, Flohr T, Heuschmid M, Becker C. Evaluation of dif- 56. Wang Y, Watts R, Mitchell I, et al. Coronary MR angiography:
ferent examination protocols for coronary artery calcium quantifi- selection of acquisition window of minimal cardiac motion with
cation with ECG-gated 16-slice spiral CT (abstract). Radiology electrocardiography-triggered navigator cardiac motion prescann-
2002;225(P):239. inginitial results. Radiology 2001;218:580585.
 CHAPTER 4 / ECG-TRIGGERED MSCT 45

4 Image Reconstruction for ECG-Triggered

and ECG-Gated Multislice CT

THOMAS FLOHR AND TINSU PAN

INTRODUCTION the limitations of established 4-slice/8-slice CT scanners, and

Computed tomography (CT) imaging of the heart and the first clinical studies have already demonstrated enhanced clini-
coronary anatomy requires high temporal resolution to avoid cal performance (15,16). While system properties such as gan-
motion artifacts and to achieve sufficient spatial resolution try rotation time and detector slice width determine the intrinsic
at best submillimeterto adequately visualize small anatomi- temporal and spatial resolution of the data, dedicated scan and
cal structures such as the coronary arteries. Furthermore, the image reconstruction techniques are needed to optimize the
complete heart volume has to be examined within the time of outcome of cardiac CT examinations. In this chapter, we present
one breath-hold. First attempts to use single-slice spiral CT the basics of ECG-triggered and ECG-gated MSCT scanning.
systems for cardiac scanning were not convincing because of We give an overview on image reconstruction techniques, start-
poor temporal resolution and insufficient volume coverage with ing with single-segment partial scan reconstruction and ending
thin slices (1,2). Since 1999, 4-slice CT systems with higher with multisegment approaches. We discuss the pros and cons of
volume coverage speed and improved temporal resolution single- and multisegment reconstruction. We demonstrate the
thanks to faster gantry rotation (rotation time down to 0.5 s) properties of reconstruction algorithms with patient scans, and
have been clinically used for electrocardiogram (ECG)-trig- end with a short summary and discussion.
gered or ECG-gated multislice CT (MSCT) examinations of
SCAN TECHNIQUES FOR ECG-CONTROLLED
the cardiac anatomy (39). Coverage of the entire heart volume
with thin slices (4 1-mm/4 1.25-mm collimation) within one
MULTISLICE CT: ECG-TRIGGERED AXIAL AND
breath-hold period became feasible, allowing for new applica- ECG-GATED SPIRAL SCANNING
tions such as high-resolution CT angiographies of the coronary For ECG-synchronized examinations of the cardio-thoracic
arteries (69). First clinical studies have demonstrated the abil- anatomy, either ECG-triggered axial scanning or ECG-gated
ity of MSCT to characterize lipid, fibrous, and calcified coro- spiral scanning can be used. The most basic approach is pro-
nary plaques (10). Despite all promising advances, challenges spectively ECG-triggered axial scanning, which was previously
and limitations remain for cardiac MSCT with 4-slice detec- introduced with electron beam CT. The patients ECG signal is
tors. Spatial resolution is still not sufficient to clearly depict monitored during examination, and axial scans are started with
stents or severely calcified coronary arteries (6,7). Temporal a predefined temporal offset relative to the R waves, which can
resolution is not yet adequate for patients with higher heart be either relative (given as a certain percentage of the RR inter-
rates, and a diagnostic outcome cannot be guaranteed in these val time) or absolute (given in ms), and either forward or reverse
cases despite careful selection of the reconstruction interval (5). Data acquisition is therefore triggered by the R waves of
(11,12). The scan time of about 40 s required to cover the entire the patients ECG signal. A schematic illustration of absolute
heart volume (approx 12 cm) with 4 1-mm, 4 1.25-mm and relative phase setting is given in Fig. 1. The principle of
collimation is at the limit for a single breath-hold scan. In 2000, a ECG-triggered multislice axial scanning is illustrated in Fig. 2.
shorter scan time was realized with 8 1.25-mm collimation, The coordinate system in Fig. 2 shows the patients ECG-signal
cardiac MSCT, which enables scan times of about 20 s. In 2001, as a function of time on the horizontal axis and the position of
a new generation of MSCT systems was introduced. With the detector slices relative to the patient on the vertical axis (in
simultaneous acquisition of up to 16 submillimeter slices and this example, four detector slices are indicated). Usually, par-
gantry rotation times down to 0.4 s, spatial resolution in the tial scan data intervals (180 + detector fan angle) are acquired.
transverse direction and temporal resolution are further Thanks to optimized half-scan reconstruction algorithms with
improved, while examination times are considerably reduced adequate data weighting, a temporal resolution up to half the
(13,14). Sixteen-slice systems have the potential to overcome gantry rotation time per image (250 ms for 0.5 s gantry rotation)
can be achieved in a sufficiently centered region of interest.
From: Contemporary Cardiology: CT of the Heart: The number of images acquired with every scan corresponds to
Principles and Applications
the number of active detector slices. In between the individual
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
axial scans, the table moves to the next z position; the heart

45
46 FLOHR AND PAN

Fig. 3. Principle of retrospectively electrocardiogram (ECG)-gated

spiral scanning. The patients ECG signal is indicated as a function of
time on the horizontal axis, and the position of the detector slices
relative to the patient is shown on the vertical axis. The table moves
Fig. 1. Schematic illustration of absolute and relative phase setting for continuously, and continuous spiral scan data of the heart volume are
electrocardiogram (ECG)-controlled CT examinations of the cardio- acquired. Only scan data acquired in a pre-defined cardiac phase,
thoracic anatomy. usually the diastolic phase, are used for image reconstruction, which
are marked as boxes.

reconstruction. Different from ECG-triggered axial scanning,

scan data from every heart cycle can be used for image recon-
struction; therefore, the entire heart volume can be covered
with a 4-slice system within one breath-hold period using 1-mm
or 1.25-mm slices. Image reconstruction during different heart
phases is feasible by shifting the start points of the data seg-
ments used for image reconstruction relative to the R waves.
For a given start position, a stack of images at different z posi-
tions covering a small subvolume of the heart can be recon-
Fig. 2. Principle of electrocardiogram (ECG)-triggered multislice structed thanks to the multislice data acquisition (4,5,17).
axial scanning. The patients ECG signal is indicated as a function of Prospective ECG triggering combined with step and shoot
time on the horizontal axis. The position of the detector slices relative
to the patient (four slices in this example) is shown on the vertical
acquisition of axial slices has the benefit of smaller patient dose
axis. Usually, partial scan data intervals (180 + detector fan angle) than ECG-gated spiral scanning, since scan data are acquired in
are acquired, which are marked as boxes. the previously selected heart phases only. It does not, however,
provide continuous volume coverage with overlapping slices,
and misregistration of anatomical details cannot be avoided.
volume is therefore covered in a step and shoot technique. As Furthermore, reconstruction of images in different phases of
a result of the time necessary for table motion, only every sec- the cardiac cycle for functional evaluation is not possible.
ond heart beat can be used for data acquisition, which limits the Because ECG-triggered axial scanning depends on a reliable
minimum slice width to 2.5 mm with 4-slice or 1.25 mm with prediction of the patients next RR interval by using the mean
8-slice CT systems if the whole heart volume has to be covered of the preceding RR intervals, the method encounters its limi-
within one breath-hold period. tations for patients with severe arrhythmia. To maintain the
With retrospective ECG gating, the heart volume is covered benefits of ECG-gated spiral CT but reduce patient dose, ECG-
continuously by a spiral scan. The patients ECG signal is controlled dose-modulation has been developed (18). During
recorded simultaneously to allow for a retrospective selection the spiral scan, the output of the X-ray tube is modulated
of the data segments used for image reconstruction. The prin- according to the patients ECG. It is kept at its nominal value
ciple of retrospectively ECG-gated spiral scanning is illustrated during a user-defined phase of the cardiac cyclein general,
in Fig. 3. Similar to Fig. 2, the patients ECG signal is indicated the mid- to end-diastolic phase. During the rest of the cardiac
as a function of time on the horizontal axis, and the position of cycle, the tube output is reduced. Clinical studies with 4-slice
the detector slices relative to the patient is shown on the vertical CT systems have demonstrated dose reduction by 3050%,
axis. The table moves continuously, and continuous spiral scan depending on the patients heart rate, using ECG-controlled
data of the heart volume are acquired. Only scan data acquired dose modulation (18).
in a predefined cardiac phase, usually the diastolic phase, are
used for image reconstruction (5,17). These scan data segments IMAGE RECONSTRUCTION FOR ECG-TRIGGERED
are indicated as boxes in Fig. 3. A variety of dedicated recon- MULTISLICE AXIAL SCANNING
struction approaches for ECG-gated spiral MSCT have been Using prospective ECG-triggering, axial scan data are
introduced with 4-slice CT scanners (35,17), resulting in a acquired. Since the table is stationary for each individual scan
temporal resolution of half the gantry rotation time (250 ms for and moves only between scans, no multislice spiral interpola-
0.5 s gantry rotation time) or better, thanks to multisegment tion is necessary. In general, partial scans are performed, with
 CHAPTER 4 / ECG-TRIGGERED MSCT 47

Fig. 4. Geometry of CT data acquisition. A modern CT scanner

acquires data in fan beam geometry, characterized by the projection
angle and by the fan angle within a projection. Another set of
variables serving the same purpose is and b. b denotes the distance
of a ray from the iso-center. and b are the coordinates of a ray in Fig. 5. Sinogram illustrating parallel rebinning according to Eq. 1 for
parallel geometry. the partial scan data segment. The sinogram curves for three image
points a, b, and c at different positions within the scan field of view
are indicated.
a scan data segment covering 180 plus the detector fan angle
(about 5060, depending on system geometry). Including
Using this equation, the measured fan beam data can be
some additional data for smooth transition weighting of
transformed to parallel data, a procedure called rebinning. In
complementary rays to avoid streaking artifacts resulting from
parallel geometry, 180 of scan data are necessary for image
data inconsistencies at the beginning and at the end of the scan
reconstruction. For data acquisition in fan beam geometry, a
interval, the total scan interval of a partial scan is p = 240
partial scan interval larger than 180, namely 180 plus the
260. This is the minimum necessary for image reconstruction
detector fan angle, is necessary to provide 180 of parallel data
throughout the entire scan field of view (SFOV) of usually 50
for any image point within the SFOV. In the center of rotation,
cm diameter. The temporal resolution at a certain point in the
for = 0, 180 of the acquired fan beam data are sufficient to
SFOV is determined by the acquisition time window of the data
provide 180 of parallel data; see Eq. (1). If all redundant data
contributing to the reconstruction of that particular image point.
are neglected, temporal resolution in the center of rotation can
Similar to slice sensitivity profiles (SSP), temporal resolution
be as good as 180/360 times the rotation time of the scanner;
may be characterized by time sensitivity profiles (TSP). The
for 0.5 s rotation, Tima = 0.25 s. As a consequence of the
temporal resolution Tima assigned to an image is the full width
rebinning Eq. (1), data with different fan beam projection angles
at half maximum (FWHM) of the TSP. In a conventional ap-
and hence different acquisition times contribute to a parallel
proach, the entire partial scan data segment is used for image
projection at projection angle . Fig. 5 is a sinogram and illus-
reconstruction in any point of the SFOV. Redundant data are
trates the rebinning procedure. Obviously, the temporal resolu-
weighted using algorithms such as the one described by Parker
tion is not constant, but depends on the position of the image
(19). The temporal resolution Tima in this case is p/360
point in the SFOV. This is illustrated in Fig. 5 for the example
times the rotation time of the scanner; for 0.5 s rotation, Tima
of three different image points a, b, and c. Measurement values
= 0.330.36 s. To improve temporal resolution, modified
contributing to point a in the iso-center are located on a straight
reconstruction approaches for partial scan data have been pro-
vertical line in the sinogram ( = 0). Hence, the temporal reso-
posed (5,17), which are best explained in parallel geometry. A
lution for this particular point is half the rotation time of the
modern CT scanner acquires data in fan beam geometry, char-
scanner. Data contributing to points b and c are located on
acterized by the projection angle and by the fan angle within
sinosoidal curves in the sinogram. A measure for temporal reso-
a projection. Another set of variables serving the same purpose
lution is given by the path lengths of the sinosoidal curves
is and b. is the azimuthal angle and b denotes the distance
within the gray shaded parallel sinogram. Obviously, the total
of a ray from the iso-center (see Fig. 4). and b are used to label
acquisition time of the data for point b is longer than for the
rays when projection data are in the form of parallel projec-
central point a, whereas it is even shorter for point c. For clini-
tions. A simple coordinate transformation relates the two sets
cal applications, the heart should be sufficiently centered within
of variables:
the SFOV to maintain a stable temporal resolution of half the
= + , b = RFsin. (1) gantry rotation time. It is not possible to make use of the areas
48 FLOHR AND PAN

Fig. 6. Spiral interpolation scheme for a four-slice scanner using one segment of multislice data for image reconstruction. The z position of
the four detector slices changes linearly relative to the patient due to the constant spiral feed. The spiral interpolation is indicated for some
representative projection angles . Please note that in general no interpolation between projections measured at different projection angles,
i.e., different acquisition times, is performed, even in cases where this would be possible (180-type interpolation). On the bottom, the
electrocardiogram signal is shown schematically.

with better temporal resolution, as start and end position of the images, every z position of the heart has to be seen by a detector
X-ray source during ECG-triggered acquisition depend on the slice at every time during the N heart cycles. As a consequence,
patients heart rate and cannot be fixed. the larger the N and the lower the patients heart rate, the more
the spiral pitch has to be reduced. If the pitch is too high, there
IMAGE RECONSTRUCTION FOR ECG-GATED
will be z positions which are not covered by a detector slice in
MULTISLICE SPIRAL SCANNING
the desired phase of the cardiac cycle. To obtain images at these
Using ECG-gated multislice spiral scanning, the heart vol- z positions, far-reaching spiral interpolations have to be per-
ume is covered continuously by a spiral scan. Image recon- formed, which degrade SSPs and reduce transverse resolution.
struction for ECG-gated multislice spiral scanning therefore
consists of two parts: multislice spiral interpolation to compen- SINGLE-SEGMENT RECONSTRUCTION
sate for the continuous table movement and to obtain scan data At low heart rates, a single-segment reconstruction (N = 1)
at the desired image z position, followed by a partial scan yields the best compromise between sufficient temporal reso-
reconstruction of the axial data segments as described above. A lution on the one hand and adequate volume coverage with thin
single-slice partial scan data segment is generated for each slices on the other. For N = 1, consecutive multislice spiral data
image using a partial rotation of the multislice spiral scan that from the same heart period are used to generate the single-slice
covers the given z position. For each projection angle within partial scan data segment for an image. The spiral interpolation
the multislice data segment, a linear interpolation is performed scheme for a 4-slice scanner using one segment of multislice
between the data of those two detector slices that are in closest data is illustrated in Fig. 6, together with the calculation of the
proximity to the desired image plane zima. The temporal reso- spiral interpolation weights for some representative projection
lution, which is limited to half the gantry rotation time for pro- angles. In general, no interpolation between projections mea-
spective ECG triggering, can be improved up to 1/(2N) times sured at different projection angles, i.e., different acquisition
the rotation time by using scan data of N subsequent heart cycles times, is performed, even in cases where this would be possible
for image reconstruction (17). With increased N, better tempo- (180-type interpolation). The temporal resolution is constant
ral resolution is achieved, but at the expense of reduced volume and equals half the gantry rotation time of the scanner using
coverage within one breath-hold time or loss of transverse reso- optimized partial scan reconstruction techniques as described
lution. To maintain good transverse resolution and thin-slice above. For 0.5-s gantry rotation time, temporal resolution is
 CHAPTER 4 / ECG-TRIGGERED MSCT 49

Fig. 7. Spiral interpolation scheme for a four-slice scanner using N = 2 subsegments of multislice data from consecutive heart periods for image
reconstruction. Both subsegments have to fit together to build up a partial scan data interval. The spiral interpolation is indicated for some
representative projection angles .

Tima = 0.25 s. Some of the recently introduced 16-slice CT same relative phase of the patients heart cycle to reduce the
systems offer gantry rotation times even shorter than 0.5 s, such total time interval contributing to an image. If the patients
as 0.42 s or 0.4 s. In this case, temporal resolution can be as heart cycle and the rotation of the scanner are completely syn-
good as Tima = 0.21 s or 0.2 s. chronous, the two requirements are contradictory. For instance,
for a heart rate of 60 beats per minute (bpm) and a 360 rotation
MULTISEGMENT RECONSTRUCTION time of 0.5 s, the same heart phase always corresponds to the
At higher heart rates, temporal resolution can be improved same projection angle segment, and a partial scan interval can-
by dividing the partial scan data segment used for image recon- not be divided into smaller subsegments acquired in successive
struction into N = 24 subsegments acquired in subsequent heart periods. Then no better temporal resolution than half
heart cycles. Each subsegment is generated by using data from the gantry rotation time is achieved. In the best case, when the
one heart period only, and there are temporal gaps between the patients heart cycle and the rotation of the scanner are opti-
multislice data segments used for image reconstruction. Simi- mally desynchronized, the entire partial scan interval may be
lar to the case N = 1, for each projection angle within divided into N subsegments of equal length, and each
subsegment j, a linear interpolation is performed between the subsegment is restricted to a data time interval of 1/(2N) times
data of those two detector slices that are in closest proximity to the rotation time within the same relative heart phase. Gener-
the desired image plane. The result are N single-slice partial ally, depending on the relation of rotation time and patient heart
scan subsegments located at the given image z position zima (see rate, temporal resolution is not constant but varies between one
Fig. 7 for the example N = 2). half and 1/(2N) times the gantry rotation time. There are sweet
With this technique, the patients heart rate and the gantry spots, heart rates with optimum temporal resolution, and heart
rotation time of the scanner have to be properly desynchronized rates where temporal resolution cannot be improved beyond
to allow for an improved temporal resolution. Two requirements half the gantry rotation time. Multisegment approaches rely on
have to be met: firstly, start- and end-projection angles of the a complete periodicity of the heart motion, and they encounter
subsegments have to fit together to build up a full partial scan their limitations for patients with arrhythmia or patients with
interval (see Fig. 8). As a consequence, the start projections of changing heart rates during examination. In general, clinical
subsequent subsegments have to be shifted relative to each practice suggests the use of N = 1 segment at lower heart rates
other. Secondly, all subsegments have to be acquired in the and N 2 segments at higher heart rates. In some CT scanners,
50 FLOHR AND PAN

Fig. 8. Schematic illustration of the two-segment cardio reconstruction approach. The circular graph represents the full partial scan interval
in parallel geometry (180 of parallel data), which is composed of two subsegments acquired in subsequent heart cycles. These two subsegments
have to be measured in the same phase of the cardiac cycle, and start- and end-projections have to fit together. To fulfill both requirements,
the patients heart cycle and the rotation of the scanner have to be asynchronous.

the partial scan data segment is automatically divided into one segment reconstruction (ACV algorithm). The 0.5-s two-seg-
or two subsegments depending on the patients heart rate dur- ment reconstruction covers 6473 bpm, whereas the 0.42-s two-
ing examination (Adaptive Cardio Volume ACV algorithm segment reconstruction is preferable for 7390 bpm. Figures
[17]). At heart rates below a certain threshold, one subsegment 11 and 12 show examples of patient scans at higher heart rates,
of consecutive multislice spiral data from the same heart period to illustrate the potential of multisegment reconstruction to
is used. At higher heart rates, two subsegments from adjacent improve image quality in selected cases. Again, prospectively
heart cycles contribute to the partial scan data segment. In some adapting the rotation time of the scanner and exploiting
other CT scanners, the single-segment partial scan images are multisegment reconstruction requires a stable and predictable
reconstructed prospectively as baseline images, followed by a heart rate during examination and complete periodicity of the
two-segment reconstruction retrospectively for a potential gain heart motion.
of temporal resolution for higher heart rates. Alternatively, a
two-segment prospective reconstruction can be prescribed with DISCUSSION AND SUMMARY
a slower table feed for scanning the same cardiac anatomy in We have reviewed the scanning and reconstruction tech-
two heart cycles, followed by a retrospective single-segment niques for ECG-controlled MSCT. Two basic scanning tech-
reconstruction for comparison with the two-segment recon- niques are discussed: ECG-triggered axial scanning and
struction. A single-segment reconstruction is usually preferred ECG-gated spiral scanning. For ECG-triggered axial scanning,
for its reliability, even though a two-segment reconstruction the X-ray will be turned on only for the duration of the data for
may yield a better image quality. image reconstruction; therefore, the dose to the patient can be
Another approach is to prospectively adjust the rotation time kept to a minimum. For ECG-gated spiral scanning, the dose is
of the scanner to the heart rate of the patient to obtain best higher, since the X-ray is turned on for the whole duration of the
possible temporal resolution for a multisegment reconstruc- scanning. Ways of reducing the X-ray dose have been pro-
tion. The range of heart rates is preferred to cover 15 bpm for posed, such as ECG-controlled dose modulation (18). There
each selected gantry cycle, to account for the patients heart are two major reconstruction techniques associated with spiral
rate variation during a scan. Figure 9 shows an example of scanning: one is single-segment reconstruction, where each
using 0.5-s and 0.6-s gantry rotation time for this implementa- image is reconstructed with the data from a single cardiac cycle;
tion. The 0.5-s two-segment reconstruction covers low 60s to the other is N-segment reconstruction, where each image is
75 bpm, followed by 0.6-s three- to four-segment reconstruc- reconstructed with the data from N contiguous cardiac cycles.
tion for 7590 bpm, and 0.5-s three- to four-segment recon- The two-segment reconstruction has the advantage of extend-
struction for the heart rates of over 90 bpm. Figure 10 shows ing from single-segment reconstruction without changing the
another example of using 0.42-s and 0.5-s gantry rotation time gantry rotation cycle. The three- and four-segment reconstruc-
in combination with an automatic selection of one- and two- tions can be used when the heart rate is becoming too high (>75
 CHAPTER 4 / ECG-TRIGGERED MSCT 51

Fig. 9. A design of N-segment reconstruction with gantry cycles of 0.5 and 0.6 s to improve temporal resolution for higher heart rates. The
0.5-s two-segment reconstruction covers from the low 60s to 75 bpm, followed by 0.6-s three- to four-segment reconstruction for 7590 bpm,
and 0.5-s three- to four-segment reconstruction for heart rates of over 90 bpm.

Fig. 10. A design of automatic selection of single- and two-segment reconstruction (ACV algorithm) with gantry cycles of 0.42 and 0.5 s.
0.42-s gantry rotation is preferable for low heart rates (single-segment reconstruction) and for 7390 bpm (two-segment reconstruction).
The 0.5-s two-segment reconstruction yields better temporal resolution for 6473 bpm.

bpm), and a slower gantry cycle of 0.6 s can be used to optimize in the cranio-caudal direction (or Z coordinate) to submillime-
the temporal resolution. Single-segment reconstruction is the ter, with already submillimeter resolution in the in-plane direc-
clinically most robust reconstruction technique. The reliability tion (or X and Y coordinates); this is critical for coronary artery
of obtaining good quality images with N-segment reconstruc- imaging, and has shortened the scan time of 40 s (4-slice) to
tion generally goes down when N becomes larger from 2, to 3, 20 s (8- or 16-slice) or even 10 s (16-slice), which is well under
to 4. Multisegment reconstruction requires a stable heart rate a single breath-hold. The gantry rotation cycle of 0.5 s to 0.42
during examination and complete periodicity of the heart motion. or 0.4 s has improved the temporal resolution of 0.25 s to 0.21
The cardiac MSCT has also benefited significantly from the and 0.2 s, which will help the MSCT become a scanner for most
advancements of 4 1-mm/4 1.25-mm to 16 0.75-mm/ patients. The image quality that can be obtained in clinical
16 0.625-mm collimation and faster gantry speeds of 0.5 s routine is demonstrated in Fig. 13. Meanwhile, first clinical
to 0.42 s or 0.4 s. The wider coverage with thinner slices (16 experience has demonstrated the potential of 16-slice technol-
0.75 mm/16 0.625 mm) has improved the spatial resolution ogy for cardiac imaging (15,16,20).
52 FLOHR AND PAN

Fig. 11. A patient study with heart rates of 97101 bpm with three- to four-segment reconstruction. Shown here are the volume-rendered image,
and the curved reformatted images of the left anterior descending (LAD), left circumflex (LCX), and right coronary artery (RCA). It demon-
strates the potential for three- to four-segment reconstruction to improve the temporal resolution for higher heart rates.

Fig. 12. A patient study of heart rates of around 92 bpm with two-segment reconstruction. Shown here are the volume-rendered image, and
a spider view maximum intensity projection showing the origins of left anterior descending, left circumflex, and right coronary artery
(courtesy of Dr. Ropers, Erlangen University, Germany).
 CHAPTER 4 / ECG-TRIGGERED MSCT 53

Fig. 13. Case of a 68-yr-old male patient with a history of stroke secondary to a bilateral carotid obstruction. The CT scan was obtained on a
16-slice CT system. The images show severe calcification in the left anterior descending (LAD) and right coronary artery (RCA), as well as
large pericardial calcifications. The calcifications are displayed in endoscopic and maximum intensity projection viewing techniques. Corre-
sponding to the coronary angiogram, a 40% stenosis of the left main (LM) and 80% stenosis of the LAD and RCA were found (courtesy of Prof.
Oudkerk, Groningen University, Groningen, the Netherlands).
54 FLOHR AND PAN

REFERENCES 12. Kopp A, Schrder S, Kttner A, et al. Coronary arteries: retrospec-

tively ECG-gated multi-detector row CT angiography with selective
1. Bahner M, Boese J, Lutz A, et al. Retrospectively ECG-gated spiral optimization of the image reconstruction window. Radiology
CT of the heart and lung. Eur Radiol 1999;9:106109. 2001;221:683688.
2. Kachelriess M, Kalender W. Electrocardiogram-correlated image 13. Flohr T, Stierstorfer K, Bruder H, Simon J, Schaller S. New techni-
reconstruction from subsecond spiral computed tomography scans cal developments in multislice CT, part 1: approaching isotropic
of the heart. Med Phys 1998;25:24172431. resolution with sub-mm 16-slice scanning. Fortschr Rntgenstr
3. Kachelriess M, Ulzheimer S, Kalender W. ECG-correlated image 2002;174:839845.
reconstruction from subsecond multi-slice spiral CT scans of the 14. Flohr T, Bruder H, Stierstorfer K, Simon J, Schaller S, Ohnesorge B.
heart. Med Phys 2000;27:18811902. New technical developments in multislice CT, part 2: sub-millime-
4. Taguchi K, Anno H. High temporal resolution for multi-slice helical ter 16-slice scanning and increased gantry rotation speed for cardiac
computed tomography. Med Phys 2000,27(5):861872. imaging. Fortschr Rntgenstr 2002;174:10221027.
5. Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of 15. Kopp AF, Kttner A, Heuschmid M, Schrder S, Ohnesorge B,
electro-cardiographically gated multisection spiral CTinitial Claussen CD. Multidetector-row CT cardiac imaging with 4 and 16
experience. Radiology 2000;217:564571. slices for coronary CTA and imaging of atherosclerotic plaques. Eu
6. Achenbach S, Ulzheimer S, Baum U, et al. Noninvasive coronary Radiol 2002;12(Suppl 2):S17S24.
angiography by retrospectively ECG-gated multi-slice spiral CT. 16. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama
Circulation 2000;102:28232828. PMT, de Feyter PJ. Reliable noninvasive coronary angiography with
7. Becker C, Knez A, Ohnesorge B, Schpf U, Reiser M. Imaging of fast submillimeter multislice spiral computed tomography. Circula-
non calcified coronary plaques using helical CT with retrospective tion 2002;106:20512054.
EKG gating. AJR 2000;175:423424. 17. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial
8. Knez A, Becker C, Leber A, Ohnesorge B, Reiser M, Haberl R. Non- and temporal resolution for ECG-gated multi-slice spiral CT of the
invasive assessment of coronary artery stenoses with multidetector heart. JCAT 2001;25(6):907923.
helical computed tomography. Circulation 2000;101:e221e222. 18. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of
9. Nieman K, Oudkerk M, Rensing B, et al. Coronary angiography with the heart with retrospective ECG gating: reduction of radiation
multi-slice computed tomography. Lancet 2001;357:599603 exposure by ECG-controlled tube current modulation. Eur Radiol
10. Schroeder S, Kopp A, Baumbach A, et al. Noninvasive detection and 2002;12:10811086.
evaluation of atherosclerotic coronary plaques with multi-slice com- 19. Parker D. Optimal short scan convolution reconstruction for fanbeam
puted tomography. JACC 2001;37(5):14301435. CT. Med Phys 1982;9(2):254257.
11. Hong C, Becker C, Huber A, et al. ECG-gated reconstructed multi- 20. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery
detector row CT coronary angiography: effect of varying trigger stenosis with thin-slice multi-detector row spiral computed tomogra-
delay on image quality. Radiology 2001;220:712717. phy and multiplanar reconstruction. Circulation 2003;107:664666.
 CHAPTER 5 / NON-ECG IMAGE RECONSTRUCTION 55

5 Phase-Correlated Image
Reconstruction Without ECG

MARC KACHELRIESS, PhD AND WILLI A. KALENDER, PhD

INTRODUCTION the origin of the sync signal is not a property of the phase-
Improvements in computed tomography (CT) technology correlated reconstruction algorithm itself; it can be easily sub-
such as the introduction of spiral CT, subsecond rotation times, stituted by any other appropriate synchronization measure.
and multislice data acquisition have stimulated cardiac CT There are several reasons to look into alternative synchroni-
imaging within the last decade (Fig. 1). Cardiac spiral CT started zation approaches. Recording the ECG requires additional
with the introduction of dedicated phase-correlated reconstruc- hardware and additional patient-handling effort. Further, extra
tion algorithms for single-slice spiral CT in 1997 (13). These systoles that do not necessarily correspond to cardiac motion
approaches have been generalized to the case of multislice spiral are interpreted as valid signals and may therefore impair image
CT (MSCT) acquisition (2,46) to process data of 4-slice scan- quality in the reconstructions. Sometimes the ECG signal is
ners. Since then, the algorithms have been extended to the case corrupted due to hardware failure, is unavailable, or totally
of cone-beam scanning with 16 slices (710) and to scanners useless. A fallback solution is desired in such cases. Addition-
with far more than 16 slices (11). Vendor-specific implemen- ally, for some applications that may profit from cardiac motion
tations that take into account the cone angle have not been reduction, such as thoracic imaging, applying the ECG leads is
announced yet. inconvenient or not desired.
What all cardiac reconstruction algorithms have in common KYMOGRAM DETECTION
is the need to synchronize the reconstruction with heart motion.
Recently, a hardware-independent synchronization method,
With this synchronization, the algorithms seek to use projec-
the kymogram, has become available (1214). The method in
tion data from a temporal window (allowed data ranges) aligned
its present implementation consists of a raw data-based center
to the synchronization points. The size of these allowed data
of mass (COM) detection of the patient cross-section currently
ranges determines the temporal resolution of the reconstruction
scanned. Since the heart is in permanent motion, the detected
and thus the level of motion-artifact reduction. The position of
COM will vary with the tube position, and the synchronization
the windows determines the motion phase that is visible in the
information can be derived from these variations.
reconstruction. The most elaborate phase-weighting strategy
The COM detection is illustrated in Fig. 2. The mathemati-
currently available is the cardio-interpolation (CI) method,
cal relations between object data and projection data state that
which is an adaptive multiphase weighting: minimal-width
the COM of the projection (at angle ) of an object is equal to
allowed data ranges of all heart cycles contributing to a given
the projection (at angle ) of the COM of the object. In our case
z position are determined to form a 180 complete high tempo-
this is illustrated with two small disks (Fig. 2A). The disk in the
ral resolution data set ready for reconstruction. Inter-segment
center of the heart is the patients COM. The second disk is
combination artifacts are eliminated by applying a multi-
the COM of the projection data. The positions of the disks are
triangular weighting function to each data segment followed by
related by projecting along the ray direction. The kymogram
proper normalization. Other known methods such as single-
algorithm uses the acquired raw data and computes the projec-
and bi-phase weighting restrict the image contributions to one
tion COM for each view angle . Then, adjacent projections are
or two heart cycles, respectively, depending on the local heart
used to compute the intersection of the corresponding lines of
rate, and are a subset of CI.
COM projection and the current patient COM results (Fig. 2B).
The synchronization signal traditionally stems from the
So far, the algorithm results in a COM curve similar to the
patients electrocardiogram (ECG), which is simultaneously
plots on the right in Fig. 3. The curves are obscured by a domi-
recorded during the scan. In those cases, the reconstruction
nant slowly varying bias; heart motion is not visible yet. The
procedure is termed ECG-correlated reconstruction. Note that
multiplanar reformations (MPRs) attached to the COM plots
clarify the origin of the bias: the patient COM is dominated by
From: Contemporary Cardiology: CT of the Heart: large structures such as the overall thorax shape or the liver.
Principles and Applications As soon as the bias is removed (e.g., by subtracting a running
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ mean of appropriate length), the cardiac motion is restored and

55
56 KACHELRIESS AND KALENDER

Fig. 1. Four generations of spiral CT scanners. Collimation and rotation times (bottom) are given for typical CT coronary angiography
protocols.

Fig. 2. (A) The center of mass (COM) of the current cross-section is projected onto the COM of the projection data. (B) Adjacent projection
COMs can be used to find the object COM by simple intersection.

becomes visible. Figure 4 shows a plot of the unbiased heart and synchronization is performed with respect to the K peaks,
COM motion. The dots hue values are chosen as the cardiac just as it is done with the ECGs R peaks.
ECG phase (relative to RR). The fact that dots of similar color
are located close together indicates a correlation of the hearts HEART IMAGING
COM and the ECG phase. For example, the upper right area Examples of kymogram-based reconstructions are given in
corresponds to systolic, the lower left area to diastolic motion. Figs. 6, 7, and 8. The images clearly indicate the high quality
The two-dimensional signal can now be reduced to a one- of kymogram-based reconstruction. The patient shown in Fig. 8
dimensional function by projecting it onto the principal axis of is an example for an ECG-based reconstruction that failed. In
cardiac in-plane motion (diagonal line in Fig. 4). An example this case, the ECG monitor reported twice as many R peaks as
of such a resulting kymogram function is given in Fig. 5. Visual a result of a misinterpretation of the T waves. The kymogram-
inspection indicates a good correlation to the patients ECG correlated reconstructions of this patient are therefore of higher
except for regions at scan start and end that correspond to ana- quality than the ECG-correlated images.
tomical levels above and below the heart. In analogy to the It must be pointed out that the kymogram phase and the ECG
ECGs R peaks, the kymograms maxima are called K peaks, phase are correlated up to an unknown scan- and patient-spe-
 CHAPTER 5 / NON-ECG IMAGE RECONSTRUCTION 57

Fig. 3. The detected center of mass (COM) is a function of the z position (and thus of time). The plots show that the COM is dominated
by global patient structures, and the heart motion cannot be readily deduced therefrom.

Fig. 4. Scatter plot showing the unbiased cardiac motion for a complete scan (about 30 s of cardiac motion). The dots are colored according
to a simultaneously acquired electrocardiogram signal. Dots of similar hue are located close together.
58 KACHELRIESS AND KALENDER

Fig. 5. The kymograms K peaks are highly correlated to the electrocardiograms R peaks. Correlation is less optimal at the start and end of
the scan, i.e., above and below the heart, where hardly any cardiac motion can be detected.

Fig. 6. Patient example reconstructed with the standard reconstruction algorithm extended parallel backprojection (EPBP) Std and with the
phase-correlated reconstruction algorithm EPBP cardio-interpolation (CI) at 0% and 50% of KK, respectively. Parameters: 12 0.75-mm
collimation, 0.42 s per rotation, 3-mm table increment.
 CHAPTER 5 / NON-ECG IMAGE RECONSTRUCTION 59

Fig. 7. Volume rendering of a kymogram-based volume reconstruction. The coronary arteries are depicted at great length.

Fig. 8. Case where two R peaks per heart cycle were assumed by the electrocardiogram (ECG)-based reconstruction as a result of a misinter-
pretation of the ECGs T waves. The heart rate reported by the ECG monitor was 130 bpm, but the true heart rate was 65 bpm, and was detected
by the kymogram algorithm. The images are reconstructed at a fixed z position (dashed lines in the plots above) at increments of 20% of RR
and KK, respectively. Only the kymogram-based reconstructions show temporal contiguity.

cific constant, the so-called phase lag: Reconstructions at 0% THORACIC IMAGING

of KK therefore do not necessarily correspond to reconstruc- Imaging of pericardial lung areas usually suffers from
tions at 0% of RR. In cases with defect ECG signals, the phase motion artifacts and blurring as a result of cardiac motion. The
lag is no longer constant (see Fig. 8). use of phase-correlated reconstruction algorithms promises to
60 KACHELRIESS AND KALENDER

Fig. 9. Improved lung imaging using kymogram-correlated reconstruction for scans of the thorax. The motion artifacts at the borders of the
heart and the double contours in the lung vanish using the kymogram.

Fig. 10. Pericardial motion can be greatly reduced using kymogram-correlated reconstruction.
 CHAPTER 5 / NON-ECG IMAGE RECONSTRUCTION 61

Fig. 11. The kymogram detection algorithm and complete reconstruction pipelines are available for clinical use, e.g., on the PC-based VAMP
syngo Explorer workstation (17).

improve the image quality in lung imaging. In general, no ECG One particular alternative to ECG correlation is the kymo-
is acquired for standard thorax scans, and the use of kymogram- gram algorithm (Fig. 11). It is based on a COM tracking of the
based reconstruction approaches applied to lung images is the cardiac motion. Kymogram-gated reconstruction has proven to
method of choice (15,16). Figures 9 and 10 demonstrate results be adequate to image the heart and adjacent lung areas with
obtained from two patients scanned with a 16 0.75-mm col- high quality. The results are comparable to those of ECG-based
limation, a rotation time of 0.5 s, and a table increment of 6 mm synchronization. Kymogram correlation also seems to be espe-
per rotation (pitch 0.5). The scan mode allows us to cover a scan cially useful to improve thoracic imaging in general. Poten-
range of 40 cm during a single breath- hold. The potential for tially, it may replace ECG-based approaches in general and
motion artifact reduction in the lung region can be seen in the thereby eliminate the need for ECG hardware and efforts by
axial displays and in the multiplanar reformations (inserts in personnel to record high-quality ECG signals.
Fig. 9). Blurring and double contours are greatly reduced by the
kymogram-based approach. REFERENCES
1. Kachelriess M, Kalender WA. ECG-based phase-oriented recon-
CONCLUSIONS struction from subsecond spiral CT scans of the heart. Radiology
Synchronization alternatives can be used for those imaging 1997;205(P):215.
2. Kachelriess M, Kalender WA, Karakaya S, et al. Imaging of the
cases where applying the ECG leads is inconvenient, and they heart by ECG-oriented reconstruction from subsecond spiral CT
may further serve as a fallback solution in cases of corrupted scans. In: Glazer G, Krestin G (eds), Advances in CT IV. Springer
ECG signals. Verlag, New York: 1998;137143.
62 KACHELRIESS AND KALENDER

3. Kachelriess M, Kalender WA. Electrocardiogram-correlated image 11. Kachelriess M, Knaup M, Kalender WA. Extended parallel back-
reconstruction from subsecond spiral CT scans of the heart. Med projection for standard 3D and phase-correlated 4D axial and spiral
Phys 1998;25(12):24172431. cone-beam CT with arbitrary pitch and 100% dose usage. Med Phys
4. Kachelriess M, Ulzheimer S, Kalender WA. ECG-correlated imag- 2004;31(6), in press .
ing of the heart with subsecond multi-slice spiral CT. IEEE Trans- 12. Kalender WA, Kachelriess M. Computertomograph mit objek-
actions on Medical Imaging 2000;19(9):888901. tbezogener Bewegungsartefaktreduktion und Extraktion der Objekt-
5. Taguchi K, Anno H. High temporal resolution for multislice helical bewegungsinformation (Kymogramm). European Patent Office
computed tomography. Med Phys 2000;27(5):861872. (Patent pending). 1999.
6. Flohr T, Ohnesorge B, Kopp AF, Becker C, Halliburton SS, Knez A. 13. Kachelriess M, Kalender WA. Kymogram-correlated image recon-
A reconstruction concept for ECG-gated multi-slice spiral CT of the struction from subsecond multi-slice spiral CT scans of the heart.
heart with pulse-rate adaptive optimization of spatial and temporal Radiology 2000;217(P):439.
resolution. Radiology 2000;217(P):438. 14. Kachelriess M, Sennst D-A, Maxlmoser W, Kalender WA. Kymo-
7. Kachelriess M, Fuchs T, Lapp R, Sennst D-A, Schaller S, Kalender gram detection and kymogram-correlated image reconstruction from
WA. Image to volume weighting generalized ASSR for arbitrary sub-second spiral computed tomography scans of the heart. Med
pitch 3D and phase-correlated 4D spiral cone-beam CT reconstruc- Phys 2002;29(7):14891503.
tion. Proceedings of the 2001 Int. Meeting on Fully 3D Image 15. Kachelriess M, Sennst D-A, Kalender WA. Reconstruction of
Reconstruction 2001;179182. motion-free pericardial lung images from standard spiral CT scans
8. Kachelriess M, Sennst D-A, Kalender WA. 4D phase-correlated using kymogram correlation. Radiology 2002;225(P):403.
spiral cardiac reconstruction using image to volume weighting gen- 16. Lell M, Dassel M, Kalender WA, Bautz WA, Kachelriess M.
eralized ASSR for a 16-slice cone-beam CT. Radiology 2001; Improvement of image quality in thoracic CT comparing standard
221(P):457. reconstruction with kymogram-based reconstruction. Radiology
9. Kachelriess M, Kalender WA. Extended parallel backprojection for 2002;225(P):567.
cardiac cone-beam CT for up to 128 slices. Radiology 2002;225(P):310. 17. Sennst D-A, Kachelriess M, Leidecker C, Schmidt B, Watzke O,
10. Sourbelle K, Kachelriess M, Kalender WA. Feldkamp-type recon- Kalender WA. Syngo explorer: an extensible software-based plat-
struction algorithm for spiral cone-beam (CB) computed tomogra- form for reconstruction and evaluation fo CT images. RadioGraphics
phy (CT). Radiology 2002;225(P):451. 2004;24:601612.
 CHAPTER 6 / RADIATION DOSE FROM CT 63

6 Radiation Dose From CT of the Heart

CYNTHIA H. MCCOLLOUGH, PhD

INTRODUCTION To represent dose for a specific scan protocol, which almost

The issue of radiation dose from X-ray computed tomogra- always involves a series of scans, it is essential to take into
phy (CT) has received much attention recently in both the popu- account any gaps or overlaps between the radiation dose pro-
lar media and scientific literature (15). This is in part due to the files from consecutive rotations of the X-ray source. This is
fact that the dose levels from CT typically exceed those from accomplished with use of a dose descriptor know as the Vol-
conventional radiography and fluoroscopy, and that the use of ume CTDIw (CTDIvol), where
CT continues to grow. Thus, CT contributes a significant por-
CTDIvol = [(N T)/ I ] CTDIw
tion of the total collective dose from ionizing radiation deliv-
ered to the public from medical procedures. It is important, and
therefore, that physicians ordering or performing these exami-
N = the number of simultaneous axial scans per X-ray
nations have an understanding of the dose delivered from a source rotation
cardiac CT, as well as how that amount of radiation compares T = the thickness of one axial scan (mm)
to those from other imaging procedures that use ionizing radia- I = the table increment per axial scan (mm) (10).
tion.
In helical CT, the ratio of the table travel per rotation (I) to
HOW TO DESCRIBE THE DOSE FROM A CT the total nominal beam width (N T) is referred to as pitch
EXAMINATION: CTDI AND DLP (1011). Hence
CT dose descriptors, the basic tools required for understand- CTDIvol = (1/pitch) CTDIw.
ing radiation dose in CT, have been in existence for many years,
yet continue to be refined as multidetector-row CT (MDCT) So, whereas CTDIw represents the average radiation dose
evolves. The primary measured value is known as the CT Dose over the x and y directions, CTDIvol represents the average
Index (CTDI) and represents the integrated dose, along the z radiation dose over the x, y, and z directions. This provides a
axis, from one axial CT scan (one rotation of the X-ray tube) single CT dose parameter, based on a directly and easily mea-
(68) (Fig. 1). Typically, a 100-mm long ionization chamber sured quantity, which represents the average dose within the
is used for routine measurements. Thus, the subscript 100 is scan volume for a standardized (CTDI) phantom (10). CTDIvol
used to denote the measurement length. All other CT dose is a useful indicator of the dose for a specific exam protocol,
descriptors are derived from this primary measured value. It is because it takes into account protocol-specific information such
important to note that the CTDI is always measured in the axial as pitch. Its value may be displayed prospectively on the con-
scan mode, and that doses for helical scan modes are calculated sole of newer CT scanners, although it may be mislabeled on
from the axial information. The equipment used to measure some systems as CTDIw. Recent consensus agreement on these
CTDI is shown in Fig. 2. definitions is reflected in newer scanner software releases (10).
The CTDI varies across the field of view. For example, for Thus, CTDIvol estimates the average radiation dose within
body CT imaging, the CTDI is typically a factor or two higher the irradiated volume of a CT acquisition. The SI units are
at the surface than at the center of the field of view. The average milliGray (mGy). It does not indicate, however, the total energy
CTDI across the field of view is given by the weighted CTDI deposited into the scan volume. Its value remains unchanged
(CTDIw), where CTDIw = 2/3 CTDI(edge) + 1/3 CTDI(center) whether there are 20 or 40 scans acquired.
(910). Figure 3 gives the typical relative distribution of dose To better represent the overall energy (or dose) delivered by
in the head and body phantoms. CTDIw is a useful indicator of a given scan protocol, the dose can be integrated along the scan
scanner radiation output for a specific kVp and mAs. CTDIw length to compute the dose-length product (DLP), where
is reported in terms of absorbed dose to air (910). DLP (mGy-cm) = CTDIvol (mGy) scan length (cm) (9).
The DLP reflects the total energy absorbed (and thus the
From: Contemporary Cardiology: CT of the Heart: potential biological effect) attributable to the complete scan
Principles and Applications acquisition. Thus, a limited abdomen CT might have the same
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

63
64 MCCOLLOUGH

Fig. 1. Computed Tomography Dose Index (CTDI) is the integral under the radiation dose profile from a single axial scan.

Fig. 2. Typical equipment used for measuring CT dose index.

Fig. 3. Typical dose distributions (%) across the image field-of-view.

 CHAPTER 6 / RADIATION DOSE FROM CT 65

Table 1
Sample Volume CT Dose Index (CTDIvol)
and Dose Length Product (DLP) Values for Common CT Exams

Chest Abdomen Abdomen and pelvis

Peak kilovotage (kVp) 120 120 120

Tube current (mA) 200 300 300
Exposure time (s) 0.5 0.5 0.5
Detector configuration (N T) 4 5 mm 4 5 mm 4 5 mm
Table index per rotation (I) 15 mm 15 mm 15 mm
Pitch (I/N T) 0.75 0.75 0.75
Reconstructed scan width (mm) 5 5 5
Scan length (cm) 40 20 40
CTDIvol (mGy) 12.0 19.1 19.1
DLP (mGy cm) 480 382 764

Table 2
Scan Acquisition Parameters, Volume CT Dose Index (CTDIvol),
and Dose Length Product (DLP) for Coronary Calcification Imaging (12-cm scan length)

EBCT MDCT1 MDCT1 MDCT2 MDCT2

Data acquisition method Prospective Prospective Retrospective Prospective Retrospective

triggering triggering gating triggering gating
Peak kilovotage (kVp) 130 120 120 120 120
Tube current (mA)a 630 140 100 150 150
Exposure time (s) 0.1 0.36 0.5 0.33 0.5
Detector configuration (N T) 1 3 mm 4 2.5 mm 4 2.5 mm 4 2.5 mm 4 2.5 mm
Table index per rotation (I) 3 mm 10 mm 3.75 mm 10 mm 3.75 mm
Pitch (I/N T) 1 1 0.375 1 0.375
Reconstructed scan width (mm) 3 2.5 3 2.5 2.5
CTDIvol (mGy) 3.5 4.6 12.5 4.7 20.3
DLP (mGy cm) 42 55 150 56 243
Effective doseb (mSv) 0.7 0.9 2.6 1.0 4.1
a mA can be increased in multidetector-row CT (MDCT) scanners for larger patients to avoid an increase in image noise. The mA
values for MDCT 1 and MDCT 2 were provided by the respective manufacturers, and do not necessarily produce an identical level of
image noise.
b Effective Dose estimate, with k = 0.017 mSv (mGy cm)1. This value is averaged between male and female models (see text and
ref. 5).
EBCT, electron beam CT.

CTDIvol as an abdomen and pelvis CT, but the latter exam would AUTOMATIC EXPOSURE CONTROL
have a greater DLP, proportional to the greater z extent of the It is technologically feasible for CT systems to adjust the
scan volume. X-ray tube current (mA) in response to variations in X-ray
Table 1 demonstrates the differences in CTDIvol and DLP intensity at the detector (1213), much as fluoroscopic X-ray
for typical body CT exams. The values are for demonstration systems adjust exposure automatically. This capability, in vari-
only; they can vary by scanner model, vendor, and image qual- ous implementations, is now available commercially on MDCT
ity requirements. Note that a change in technique (mAs/rota- systems in response to wide interest from the radiology com-
tion) affects the CTDIvol, while a change in acquisition length munity. Some systems adapt the tube current based on changes
(at the same technique) is reflected by the DLP. in attenuation along the z axis, others adapt to changes in
In cardiac CT, the anatomic scan length is relatively con- attenuation as the X-ray tube travels around the patient. The
stant (typically 12 cm); thus, the variability in CTDIvol and DLP ideal is to combine both approaches with an algorithm that
is primarily a result of differences in scanner output and scan chooses the correct tube current to achieve a predetermined
acquisition parameters. Tables 2 and 3 provide the scan acqui- level of image noise.
sition parameters, CTDIvol, and DLP for coronary calcification With regard to cardiac CT, the radiation dose for a retrospec-
imaging and coronary angiography. Data are provided for an tively gated exam, where the X-ray tube is kept continuously
electron beam CT (EBCT) system as well as for MDCT sys- on throughout the acquisition, can be dramatically reduced if
tems from two different manufacturers. the tube current is reduced during portions of the cardiac cycle
66 MCCOLLOUGH

Table 3
Scan Acquisition Parameters, Volume CT Dose Index (CTDIvol)
and Dose Length Product (DLP) for Coronary Angiography (12-cm Scan Length)

EBCT MDCT1 MDCT2

Data acquisition method Prospective Retrospective Retrospective

triggering gating gating
Peak kilovotage (kVp) 130 120 120
Tube current (mA)a 630 300 300
Exposure time (s) 0.1 0.5 0.5
Detector configuration (N T) 1 3 mm 4 1 mm 4 1.25 mm
Table index per rotation (I) 2 mm 1.5 mm 1.9 mm
Pitch (I/N T) 0.66 0.375 0.375
Reconstructed scan width (mm) 3 1.25 1.25
CTDIvol (mGy) 5.3 46 55
DLP (mGy cm) 64 547 662
Effective doseb (mSv) 1.1 9.3 11.3
a mA can be increased in multidetector-row CT (MDCT) scanners for larger patients to avoid
an increase in image noise. The mA values for MDCT 1 and MDCT 2 were provided by the
respective manufacturers, and do not necessarily produce an identical level of image noise.
b Effective Dose estimate, with k = 0.017 mSv (mGy cm)1. This value is averaged between
male and female models (see text and ref. 5).
EBCT, electron beam CT.

that are not likely to be of interest for the reconstructed data. fundamental, but perhaps unspoken, question, What is the
Thus, in addition to modulation of the tube current based on likelihood that I will be harmed from this exam. Characteriz-
patient attenuation, the tube current can be modulated by the ing the radiation dose in terms of E and comparing that value
ECG signal. Since cardiac motion is least during diastole and to some meaningful levelfor instance, one years E from
greatest during systole, the projection data are least likely to be naturally occurring background radiationbetter conveys to
corrupted by motion artifact for diastolic-phase reconstruc- the patient the relative potential for harm from the medical
tions. Accordingly, the tube current is reduced during systole. exam. Table 4 provides typical values of E for several common
Dose reductions of approx 50% have been reported using such imaging exams, as well as the annual level of background
a strategy (14). The implementation of these and other dose- radiation in the US (approx 3.6 mSv).
reduction strategies is expected industry-wide over the next It is important to remember, however, that E describes the
several years, in response to the strong concern about the radia- relative whole-body dose for a particular exam and scanner, but
tion dose from CT. is not the dose for any one individual, as E calculations use
many assumptions including a mathematical model of a stan-
EFFECTIVE DOSE dard human body that does not accurately reflect any one
It is important to recognize that the potential biological individual. Effective dose is best used to optimize exams and
effects from ionizing radiation depend not only on the radiation compare risks between proposed exams. It is a broad measure
dose, but also on the biological sensitivity of the tissue or organ of risk, and as such should not be quoted with more than one or
system irradiated. A 100-mGy dose to an extremity would two significant digits.
not have the same potential biological effect (detriment) as Specific values of E can be calculated using several different
a 100-mGy dose to the pelvis (15). Effective dose (E) is a dose software packages (17), which are based on the use of data from
descriptor that reflects this difference in biologic sensitivity one of two sources: the National Radiological Protection Board
(1617). It is a single dose parameter that reflects the risk of a (NRPB) in the UK (18) or the Institute of Radiation Protection
nonuniform exposure in terms of an equivalent whole-body (GSF) in Germany (19). To minimize controversy over differ-
exposure. The units of E are milliSieverts (mSv). ences in E values that are purely the result of calculation meth-
Although the concept of effective dose has some limitations odology and data sources, a generic estimation method was
when applied to medical populations, it does facilitate the com- proposed by the European Working Group for Guidelines on
parison of biological effect between diagnostic exams of dif- Quality Criteria in CT (9), where E is estimated from the non-
ferent types (16,17). Published values of E per DLP (9) allow controversial value of DLP: E = k DLP, where the values of
convenient estimates of E based on the DLP value provided at k are dependent only on the region of the body being scanned
the CT scanner console. The use of E facilitates communication (head, neck, thorax, abdomen, or pelvis) (Table 5). The values
with patients regarding the potential harm of a medical exam of E predicted by DLP and the values of E estimated using more
that uses ionizing radiation. For example, when a patient rigorous calculation methods are remarkably consistent, with a
inquires, What dose will I receive from this exam, an answer maximum deviation from the mean of approx 1015%. Hence,
in the units of mGy or mGy cm will not likely answer the more the use of DLP to estimate E appears to be a reasonably robust
 CHAPTER 6 / RADIATION DOSE FROM CT 67

Table 4
Effective Dose Values for Common Imaging Examinations

Examination Effective dose (mSv)

Head CT 12
Chest CT 57
Abdomen and pelvis CT 811
Selective coronary angiogram 35
Posterior-anterior and lateral chest X-ray 0.040.06
Average annual background radiation in the US 3.6

Table 5
Values of the Conversion Coefficient k for Use in Estimating
Effective Dose (E) (mSv) From Dose Length Product (DLP)
(in mGy cm) According to the Formula E = k DLP (9)

Anatomic region k (mSv mGy1 cm1)

more publications regarding the clinical efficacy of cardiac CT
Head 0.0023 as compared to alternate imaging modalities. Discussions of
Neck 0.0054 patient safety and societal cost will include discussions of the
Chest 0.017
radiation dose from the various procedures. To ensure that these
Abdomen 0.015
Pelvis 0.019 discussions are accurate, it is essential that the dose informa-
tion associated with specific techniques be reported in an accu-
rate and complete fashion using standardized terminology.
REFERENCES
method for estimating E. The effective doses for several car- 1. Brenner DJ, Elliston CD, Hall EJ, Berdon WE. Estimated risks of
diac CT examinations, based upon values of DLP, are given in radiation-induced fatal cancer from pediatric CT. AJR 2001;176:
Tables 2 and 3. 289296.
2. Donnelly LF, Emery KH, Brody AS, et al. Minimizing radiation
SUMMARY dose for pediatric body applications of single-detector helical CT:
strategies at a large childrens hospital. AJR 2001;176:303306.
The fundamental dose parameter in CT, the CTDI, is mea- 3. Haaga JR. Radiation dose management. AJR 2001;177:289291.
sured at edge and center locations in an acrylic phantom for a 4. Nickoloff EL, Alderson PO. Radiation exposure to patients from
given kVp, mAs, and scan width, and reported for a given exam CT: reality, public perception, and policy. AJR 2001;177:285287.
5. Pierce DA, Preston DL. Radiation-related cancer risks at low dose
protocol as the volume CTDI (CTDIvol) in units of mGy. This
among atomic bomb survivors. Radiation Research 2000;154:178186.
value represents the average dose in a standard acrylic phan- 6. Shope TB, Gagne RM, Johnson GC. A method for describing the
tom for a given exam protocol. Another relevant dose param- doses delivered by transmission x-ray computed tomography. Med
eter is that of effective dose, which is given in units of mSv. The Phys 1981;8:488495.
effective dose is a single dose parameter that best represents the 7. American Association of Physicists in Medicine. Standardized
methods for measuring diagnostic x-ray exposures. Report no. 31.
radiation detriment corresponding to a given exam protocol. AAPM, New York, 1990.
Neither of these parameters is an estimate of radiation dose to 8. Nagel HD. Radiation exposure in computed tomography. Frankfurt:
any one individual, but rather should be used to optimize and COCIR, 2000.
compare exam protocols. E can be estimated with good accu- 9. European guidelines for quality criteria for computed tomography.
Luxembourg: European Commission, 2000.
racy from the DLP, which is equal to the CTDIvol multiplied by
10. International Electrotechnical Commission. Medical Electrical
the total scan length (in cm). Equipment. Part 244: Particular Requirements for the Safety of
Techniques to modulate the tube current as a function of X-ray Equipment for Computed Tomography. IEC publication No.
patient attenuation or the time within the cardiac cycle are 60601-2-44 Amendment 1.
important innovations that will reduce the dose from cardiac 11. McCollough CH, Zink FE. Performance evaluation of a multi-slice
CT system. Medical Physics 1999;26:22232230.
CT by at about a factor of two. Hence, coronary artery calcium 12. Gies M, Kalender WA, Wolf H, Suess C, Madsen M. Dose reduction
examinations, which currently have E values between 1 and 4 in CT by anatomically adapted tube current modulation I: simulation
mSv, may be able to be conducted using retrospective gating studies. Medical Physics 1999;26:22352247.
techniques and an E of less than 2 mSv. CT coronary angiog- 13. Kalender WA, Wolf H, Suess C. Dose reduction in CT by anatomi-
cally adapted tube current modulation II: phantom measurements.
raphy, which currently requires an E of approx 10 mSv with
Medical Physics 1999;26:22482253.
retrospectively gated MDCT may be performed with approx 5 14. Jakobs TF, Becker CR, Ohnesorge B, et al. Eur Rad 2002;12:
mSv if ECG tube current modulation is applied. These effective 10811086.
dose values are of similar magnitude to those from a chest CT 15. Committee on the Biological Effects of Ionizing Radiation. Health
examination (57 mSv) or a conventional (diagnostic) coro- effects of exposure to low levels of ionizing radiation, BEIR V.
Washington, DC: National Academy, 1990.
nary angiogram (35 mSv). 16. International Commission on Radiological Protection (ICRP). 1990
Finally, as cardiac CT image quality, diagnostic accuracy, Recommendations of the ICRP. Publication 60. ICRP, New York,
and availability all continue to improve, there will be more and NY, 1991.
68 MCCOLLOUGH

17. McCollough CH, Schueler BA. Calculation of effective dose. Medi- 19. Zankl M, Panzer W, Drexler G. The calculation of dose from exter-
cal Physics 2000;27:828837. nal photon exposures using reference human phantoms and Monte
18. Jones DG, Shrimpton PC. Survey of CT practice in the UK. Part 3: Carlo methods. Part IV: Organ dose from computed tomographic
Normalised organ doses calculated using Monte Carlo techniques, examinations, GSF-Bericht 30/91. Neuherberg, Germany: GSF
NRPB-250. Oxon, United Kingdom: National Radiological Protec- Forschungszentrum fur Umwelt und Gesundtheit, Institut fur
tion Board, 1991. Strahlenschutz, 1991.
 CHAPTER 7 / CORONARY CALCIUM SCREENING 69

DETECTION AND QUANTIFICATION

III
OF CORONARY CALCIUM
 CHAPTER 7 / CORONARY CALCIUM SCREENING 71

7 Coronary Calcium Screening

An Epidemiologic Perspective

CHRISTOPHER J. ODONNELL, MD, MPH AND UDO HOFFMANN, MD

INTRODUCTION dynamics of coronary calcification under drug treatment. In

Because coronary artery disease (CAD) is the most frequent this chapter, we review the strengths and limitations of avail-
cause of death in industrialized nations (1) and the available able original research and discuss consensus statements of the
tools for prediction of CAD onset are imperfect, there is a need American College of Cardiology/American Heart Association
for new methods to screen apparently healthy individuals to (ACC/AHA) regarding the use of CAC as an adjunct to estab-
identify those at increased risk. Current risk prediction of CAD lished risk factors. We also provide insight into ongoing
is based on the patients age and sex as well as on the presence research and propose studies that could help to facilitate evi-
and extent of established, modifiable coronary risk factors such dence-based practice guidelines.
as hypertension, hyperlipidemia, diabetes mellitus, and ciga- CORONARY CALCIFICATION AND CORONARY
rette smoking. Risk prediction algorithms such as the ARTERY DISEASE
Framingham Heart Study coronary risk score have been shown
Experts agree that CT testing can determine whether calci-
feasible and valid for CAD risk prediction in the United States
fications are present in the walls of the coronary arteries. If
population (2). As a consequence, those traditional risk factors
calcification of any amount is present, it follows that athero-
and/or risk-factor algorithms have been incorporated into treat-
sclerosis is present in the coronary artery. Hence, patients with
ment guidelines for hyperlipidemia (3) and hypertension.(4).
no symptoms but with detectable calcification can be said to
However, coronary risk scores can explain only 70% of the
have coronary artery disease not detectable by the usual clinical
overall risk for CAD and are more sensitive than specific.
tests (subclinical atherosclerosis). However, there is contro-
Consequently, there is need to develop new strategies to iden-
versy regarding the relation of the calcification score to the
tify patients at high risk, specifically among patients who appear
prevalence of CAD and the incidence of cardiac events such as
to be at intermediate (i.e., 620% 10-yr risk) or low risk accord-
unstable angina, myocardial infarction, and sudden cardiac death.
ing to traditional risk factors (5,6).
Early studies by Rumberger (6a) showed that there is a lin-
Recently developed fast computed tomography (CT) tech-
ear relationship between the amount of calcium and the overall
niques such as multidetector-row CT (MDCT) and electron
amount of atherosclerotic plaque. Logically, a number of stud-
beam CT (EBCT) now offer the opportunity to noninvasively
ies have shown that high Agatston scores are associated with
detect and quantify coronary artery plaque burden (Figs. 14).
the presence and number of obstructive coronary artery lesions.
The conduct of both techniques and their similarities and dif-
While the technique appears to be quite sensitive, there is no
ferences are described in detail elsewhere in this book. These
one-to-one correlation, and the location of calcification and
CT scanners are capable of nearly freezing the motion of the
stenosis may be different. Two meta-analyses evaluated the
heart as a result of very fast image acquisition and synchroni-
diagnostic accuracy of CAC (EBCT) for the detection of sig-
zation to ECG signals. Both have been shown to be highly
nificant coronary artery stenosis as compared to coronary
sensitive for the detection of calcified coronary atherosclerotic
angiography. In both studies, the overall sensitivity and speci-
plaques. A typical report from either MDCT or EBCT contains
ficity of CAC were 8090% and 4050%, with a maximum
the Agatston score, a semiquantitative measure that is based on
joint value of 75%, respectively (7,8). Moreover, the summary
the area and a weighted density factor for coronary calcium.
odds ratios were increased 20-fold (95% confidence interval
In the past decade, a large number of studies have been
[CI], 5 to 88) indicating that the odds of having a significant
conducted to examine the predictive value of CT coronary
coronary artery stenosis are 20 times higher if calcium was
artery calcium (CAC) as a predictor for coronary risk. In addi-
detected. However, the generalizability of these studies is lim-
tion, this technique has been shown to be capable of following
itedmost of the patients were middle-aged men. Apart from
the natural history of coronary atherosclerosis and tracking the
a small number of studies where most patients were referred to
angiography as a result of chest pain, the indication for coro-
From: Contemporary Cardiology: CT of the Heart: nary angiography was not specified. However, most likely
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ almost all patients were symptomatic for CAD or had known

71
72 ODONNELL AND HOFFMANN

Fig. 1. A cross-sectional image through the aorta and the origin of the
left coronary artery (dashed arrow). A moderate amount of calcifica- Fig. 2. A cross-sectional image through the aorta; a moderate amount
tion can be easily identified as bright signals (solid arrows). of calcification can be easily identified as bright signals (arrows).

Fig. 3. A cross-sectional image through the aorta (Ao), the left ventricle (LV) and the right atrium (RA). Left image: high image quality, no
image artifacts, aortic valve calcification. Right image: image artifacts in the right coronary artery (solid arrow) and mitral valve calcification
(dashed arrow).

CAD. Consequently, the prevalence of disease (CAD) was very dictive value/negative predictive value) is difficult to answer.
high in the study populations (with one exception >50%, in In addition, most studies were designed only as observational
about a third of the studies 7095%), and the clinically impor- studies, and blinding was not mentioned as a quality criterion
tant question What is the probability of having/not having in all but two of the analyzed studies. In summary, the available
significant CAD given a positive/negative test? (positive pre- information is still limited, and the results cannot be general-
 CHAPTER 7 / CORONARY CALCIUM SCREENING 73

Fig. 4. A cross-sectional image through the aorta and the left atrium. Left image: high image quality, no image artifacts, arrow pointing at the
circumflex and the left coronary artery. Right image: image artifacts in the circumflex, the left coronary artery, and the aortic valve (arrows).

ized. As a consequence, the use of CAC for the accurate predic- or death) after 37 12 mo of follow-up. After adjustment for
tion of prevalent significant CAD has not been implemented risk factors ascertained by questionnaire, there were statisti-
into clinical guidelines. cally increased risks associated with the presence of CAC for
hard events in men (95% CI, 3.9 [1.213]) and no significantly
PREDICTIVE VALUE OF CORONARY CALCIFICATION
increased risks in women (95% CI, 1.5 [0.210]) (10). In a
FOR CARDIOVASCULAR EVENTS second prospective study of all-cause mortality in 10,377 self-
Clinicians are currently interested in the role of CAC in referred men and women, there were 249 deaths in approx
prediction of cardiovascular events. Since a large proportion of 60 mo of follow-up. Increasing CAC scores were associated
all cardiovascular events occur in asymptomatic patients, it is with significantly increased relative risks for death, and receiver
justified to differentiate asymptomatic persons from symptom- operating characteristics (ROC) curves were significantly
atic patients with known CAD. greater for CAC score added to risk factors compared with risk
factors alone (p < 0.001) (11). In a third prospective study of
CORONARY CALCIFICATION
5585 men and women recruits in whom an EBCT test was
IN ASYMPTOMATIC PERSONS
obtained to decide upon eligibility for inclusion in a random-
A small number of prospective studies have evaluated the ized controlled lipid lowering trial, there were 122 events, of
prediction of cardiac events by coronary calcification. In a which 43 were hard events after over 50 mo of follow-up.
recent meta-analysis of prospective studies, there was an Overall, there were statistically increased risks for hard events
increased risk for a combined outcome (nonfatal myocardial (9.9 [5.218.9]) as well as all endpoints associated with a high
infarction [MI], congenital heart disease [CHD], death, or CAC score. In an analysis of the incremental predictive ability
revascularization) associated with a calcium score elevated of CAC over Framingham risk score in a subset of 1817 sub-
above a median value (risk ratio 8.7 [95% CI, 2.728.1]). The jects in whom risk factors was measured, there was a several-
relative risk of hard endpoints (MI or death) was also increased fold increase in risk for cardiovascular endpoints in the highest
(risk ratio 4.2 [95% CI 1.611.3]) (9). Three more recent compared with the lowest tertile of CAC score (12).
prospective studies have added data from more than 20,000
patients to the meta-analysis. In 5635 asymptomatic, low-to- LIMITATIONS OF THE AVAILABLE DATA
medium-risk men and women self-referred for EBCT, there While there is consistency across the available studies for an
were 224 events, of which only 61 were hard endpoints (MI association of CAC scores with the presence of significant CAD
74 ODONNELL AND HOFFMANN

and the risk of cardiovascular events, a number of method- sclerosis are recording risk factor data prospectively. A final
ological concerns make it difficult to interpret study results and set of concerns pertains to study follow-up and outcomes. First,
apply them to clinical practice. Importantly, there are concerns the level of completion of follow-up is inconsistent across stud-
regarding study generalizability/external validity, validity of ies, with some studies having less than 90% follow-up. For
the risk factor measures, and resultant multivariable models example, in one of the largest recent studies, only a 64% fol-
used in the studies. In addition, the conclusions are drawn from low-up was achieved (10). Losses to follow-up can raise ques-
relatively small numbers of hard cardiovascular outcomes (all tions about the validity of study findings. A second concern is
together numbering only several hundred). the focus of the available studies on either combined hard (i.e.,
A major limiting factor of most studies has been the inclu- MI and death) and/or soft (i.e., revascularization using percu-
sion of self-referred patients. None of the available evidence taneous coronary interventions or coronary artery bypass sur-
for CAC screening has been drawn from community-based gery) outcomes. Soft intervention outcomes are often used, but
cohorts representative of typical populations which might be may not be appropriate because neither the study patients nor
considered for screening. Subjects studied to date have a high their physicians were blinded to the CAC score. Thus, given the
prevalence of subclinical CAD and CAD risk factors, often fact that many patients were self-referred and not part of a
making them different from persons in the community of a blinded study, there is a substantial risk that soft outcomes may
similar age and sex. The baseline characteristics of many of be contaminated by intervention bias introduced by knowledge
these cohorts suggest that patient populations already at sub- of the CAC test results. No study with cardiovascular mortality
stantially elevated risk have been studied. Thus, data from these as a primary outcome has been reported to date, and in virtually
studies may not be representative of apparently healthy low- all available studies, there have been too few cardiovascular
and medium-risk groups for whom screening may be more deaths to reliably examine the association of CAC. The recent
relevant. Moreover, the self-referred subjects to date may dif- study of Shaw et al. shares many of the limitations of the other
fer significantly in other unmeasured characteristics that self-referral studies, but it did report a positive association of
accompany health-seeking behaviors. CAC score with all-cause mortality (11). One strength of the
In the available studies, data are limited for young persons, Shaw study was its demonstration of consistent, statistically
women, and non-Caucasians. Much of the data to date have significant increases in mortality across increasing CAC scores
been derived largely from middle-aged Caucasian men. As a in a large cohort. Nevertheless, mortality data were drawn from
consequence of this selection bias, age- and gender-specific the National Death Index, and no further data on cause of death
CAC thresholds for younger individuals for subjects below or on non-fatal outcomes were ascertained (11).
40 yr of age, women, and non-Caucasian populations may be Interestingly, there is a small and growing body of evidence
relatively unreliable due to lack of statistical power (small in population-based cohorts that calcification of the aorta pre-
sample sizes, insufficient numbers of events) or wholly dicts risk for CAD and other cardiovascular diseases indepen-
unavailable. Potentially important racial differences in preva- dent of measured risk factors. The prevalence of calcified
lence of CAC have been noted in the Multiethnic Study of lesions detected on thoracic and abdominal radiographs like
Atherosclerosis (MESA) (13); in particular, the prevalence of that of coronary calcium, increases steeply with age (17). A
CAC appears to be lower in African Americans and other major number of recent studies have demonstrated that thoracic and
ethnic groups in the United States (1416). abdominal aortic calcifications are independent, prospective
Another significant concern pertains to the method of ascer- predictors of CVD. Calcification of the aortic arch detected by
tainment of risk factor data. It has been suggested that a high plain antero-posterior radiographs predicts an increased risk of
CAC score predicts coronary risk independently and probably cardiovascular diseases in men and women (18). In the
incrementally to traditional coronary risk factors and coronary Framingham Heart Study, the presence and extent of abdomi-
risk scores. ROC curve analyses are often provided to bolster nal aortic calcification detected by lateral lumbar radiographs
the case for CAC screening over and above risk factors. How- is a predictor of CVD, CVD mortality, and CHD in middle-
ever, in virtually every available study, blood pressure determi- aged men and women (mean age 60 yr), independent of age and
nations, lipid levels, and other risk factors are obtained actually measured cardiovascular risk factors (17). Abdominal
retrospectively by questionnaire, self-report from the study aortic calcification would likely be detected quite reliably and
patient, and/or review of the medical record. There is a substan- accurately by CT imaging, and further study of the prognostic
tial risk of misclassification in self-reported risk-factor data. value of such CT findings in addition to or instead of coronary
Therefore, calculations of Framingham coronary risk or other calcium is clearly warranted.
multivariable risk models using self-reported risk factors algo-
rithms would tend to underestimate the risk conferred by risk ADDITIONAL QUESTIONS SURROUNDING
factors (bias towards the null) and lead to a more favorable CAC SCREENING
comparison of CAC scores with risk factors. Standardized A number of additional concerns have arisen surrounding
research methods that mirror office-based methods for mea- CAC screening and are summarized in Table 1. First, the CT
surement of blood pressure, lipid levels, and diabetes have not test does lead to significant exposure to radiation. The test
been employed in the available studies. By way of contrast, exposes the patient to a limited amount of ionizing radiation
prospective epidemiologic cohort studies such as the (0.7 to 3 mSv) that is equivalent to 25100% of the natural
Framingham Heart Study and the Multiethnic Study of Athero- background radiation exposure that an individual in the United
 CHAPTER 7 / CORONARY CALCIUM SCREENING 75

Table 1
Suggested Guidelines and Implications for Coronary Artery Calcium Scoring
(Adapted From Partners HealthCare System 2002 Guidelines)

Guidelines for ordering a CT test for coronary calcium

Coronary calcium screening is not recommended for asymptomatic individuals at low risk.
Coronary calcium screening will be performed only upon a physicians request.
Patients should be informed that most insurers do not reimburse for coronary screening. The out-of-pocket expense for a multidetector-row
CT test for coronary calcium is $350$450.
A positive calcium score might be valuable in determining whether a patient who appears to be at intermediate congenital heard disease (CHD)
risk is actually at high risk. However, we do not currently recommend coronary calcium scoring for this group.
Ongoing research will better define the clinical subgroups, such as elderly asymptomatic persons or those with a strong family history, in whom
the management of other risk factors might be modified according to the calcium score.

Guidelines for Interpreting a CT Test for Coronary Calcium

Some patients and physicians will choose to obtain coronary calcium studies despite the lack of consensus for these studies. We recognize that
clinicians who might not have ordered the test themselves are increasingly faced with the test results and asked for their advice. Therefore
the following comments are provided from the recent ACC/AHA consensus document:
In the absence of coronary calcium (i.e., a negative CT test)
Atherosclerotic plaque, including unstable plaque, is very unlikely, although it is possible that significant coronary atherosclerosis may not
be detected.
Significant luminal obstructive disease is highly unlikely.
Angiographically normal coronary arteries occur in the majority of such patients.
The absence of calcium may be consistent with a low risk of CHD events over the next 2 to 5 yr.

For a test that is positive for coronary calcium:

The presence of calcium confirms the presence of a coronary atherosclerotic plaque.
The greater the amount of calcium, the greater the likelihood of occlusive coronary artery disease, although there is not a one-to-one relationship,
and findings may not be site specific.
The total amount of calcium correlates best with the total amount of atherosclerotic plaque, although the true plaque burden is underestimated.
A high calcium score (e.g., Agatston score > 400) may be consistent with moderatehigh risk of CHD events over the next 25 yr.

Fig. 5. 90th Percentile Agatston scores derived from large numbers of men and women screened with electron beam CT. Data from ref. 22.

States receives per year (2.5 to 3 mSv). This is less than the dose additional exposure to radiation or even invasive testing in order
received during a diagnostic cardiac catheterization (approx to confirm or exclude a potentially serious diagnosis such as an
4.5 mSv) and only a fraction of the occupational exposure limit aortic aneurysm or malignancy. Third, as mentioned above,
for a radiation worker in the United States (50 mSv/y). Although unbiased data on age-adjusted coronary calcium thresholds are
the available data suggest that these risks are extremely low not yet available, because the available data are derived from
from such exposure (19), for some individuals even this low cohorts that may or may not be representative of general screen-
risk may be considered unacceptably high. Second, several ing subjects. Moreover, simple thresholds (such as an Agatston
recent studies have demonstrated that there is a high prevalence score of 400) that do not adjust for age may overestimate risk
of potentially clinically important findings that are incidentally in older persons and underestimate risk in younger persons
detected on CT scans (20,21). Such incidental findings may (Fig. 5) (22). A fourth and related concern is the fact that many
provoke significant anxiety in the patient and may require different CT scan machines are now available (including both
76 ODONNELL AND HOFFMANN

EBCT and MDCT machines) and standardized scanning and nations of lipid-lowering agents, anti-platelet therapies, and
scoring algorithms have not been implemented across machine angiotensin-converting enzyme inhibitors. While the optimal
types. Moreover, there is a small but significant level of mea- indications for the use of noninvasive tests for clinical moni-
surement variability consistently seen across all scanner types, toring of the response to therapy remain to be determined,
and the degree of variability may differ. Fifth, all CAC scoring existing expert consensus statements such as that provided by
to date has been conducted using the Agatston score. While the the ACC/AHA in 2000 provide cogent standards for the
Agatston score has provided an invaluable reference score for present time. CAC testing has not yet been reliably shown to
research to date, newer mass-based scoring techniques may add incremental information for risk stratification in general-
substantially reduced the misclassification introduced by image izable patient cohorts, and ongoing studies should address the
noise. A final concern regarding CT scanning is its currently important outstanding questions regarding the appropriate
high cost. The cost for the test ranges from $375$450, and the types of men and women in whom to consider coronary cal-
test is not currently covered by most health insurance providers. cium screening.

CRITICAL EVIDENCE NEEDED REFERENCES

FOR CLINICAL GUIDELINES 1. American Heart Association. 2000 Heart and Stroke Statistical
Both risks and benefits must be considered for any potential Update. 2000. Dallas, Texas, American Heart Association.
2. Wilson PW, DAgostino RB, Levy D, et al. Prediction of coronary
screening test. The benefits would include potential for reduc- heart disease using risk factor categories. Circulation 1998;97:
tion of risk provided by an accurate diagnosis of increased risk 18371847.
and the potential benefits from early detection of incidental 3. Third Report of the National Cholesterol Education Program (NCEP)
findings (e.g., small lung cancers). The risks include the mor- Expert Panel on Detection, Evaluation, and Treatment of High Blood
bidity and mortality risk of an incorrect diagnosis (negative or Cholesterol in Adults (Adult Treatment Panel III) final report.
Circulation 2002;106:31433421.
positive), risks of radiation exposure, morbidity and mortality 4. Chobanian AV, Bakris GL, Black HR, et al. The Seventh Report of
from detection of benign incidental findings, and high costs of the Joint National Committee on Prevention, Detection, Evaluation,
potentially unnecessary testing. The current body of studies and Treatment of High Blood Pressure: the JNC 7 report. JAMA
suggests that there may be a role for CAC screening, but to date 2003;289:25602572.
there are no strong unbiased data for or against CAC screening 5. Greenland P, Smith JS, Jr., Grundy SM. Improving coronary heart
disease risk assessment in asymptomatic people: role of traditional
over and above traditional risk-factor assessments in the general risk factors and noninvasive cardiovascular tests. Circulation
population. The limitations of existing studies can be addressed 2001;104:18631867.
only by prospective, observational studies conducted in 6. Wilson PW, Smith SC, Blumenthal RS, et al. Task force #4how
unbiased, population-based cohorts of both Caucasian and non- do we select patients for atherosclerosis imaging? J Am Coll Cardiol
Caucasian men and women. It has been suggested that coronary 2003;41:18981906.
6a. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz RS.
calcium screening may be most beneficial in persons at Coronary artery calcium area by electron-beam computed tomogra-
intermediate coronary risk (5,6). At least 40% of the population phy and coronary atherosclerotic plaque area. A histopathologic cor-
of middle-aged men and women may fall into this category (5). relative study. Circulation 1995;92:21572162.
In the proposed studies, risk-factor measurements should be 7. Nallamothu BK, Saint S, Bielak LF, et al. Electron-beam computed
office-based rather than self-reported. Ideally, measures of risk tomography in the diagnosis of coronary artery disease: a meta-
analysis. Arch Intern Med 2001;161:833838.
factors as well as coronary calcification would be conducted in 8. ORourke RA, Brundage BH, Froelicher VF, et al. American Col-
as blinded a manner as possible. Several such population-based lege of Cardiology/American Heart Association expert consensus
studies are currently underway. These studies include the document on electron-beam computed tomography for the diagnosis
Framingham Heart Study and the Multiethnic Study of Athero- and prognosis of coronary artery disease. J Am Coll Cardiol 2000;36:
sclerosis. Final recommendations regarding use of CT calcium 326340.
9. OMalley PG, Taylor AJ, Jackson JL, et al. Prognostic value of
awaits the completion of these studies. coronary electron-beam computed tomography for coronary heart
disease events in asymptomatic populations. Am J Cardiol 2000;85:
CONCLUSION 945948.
10. Kondos GT, Hoff JA, Serukov A, et al. Electron-beam tomography
Coronary atherosclerosis underlies CHD and is influenced coronary artery calcium and cardiac events: a 37-month follow-up
by the interaction of multiple risk factors over time in a suscep- of 5635 initially asymptomatic low- to intermediate-risk adults.
tible host. Traditional risk-prediction models, such as the Circulation 2003;107:25712576.
Framingham risk score, are reasonable predictors of CHD risk. 11. Shaw LJ, Raggi P, Schisterman E, Berman DS, Callister TQ.
Interest in CAC screening is aimed at improving the prediction Prognostic value of cadiac risk factors and coronary artery calcium
screening for all-cause mortality. Radiology 2003;228:826833.
of risk for future clinical events. Future studies will need to 12. Arad Y, Roth M, Newstein D, Guerci AD. Coronary calcification,
ascertain the incremental benefit of CAC measures over tradi- coronary disease risk factors, and atherosclerotic cardiovascular
tional risk factors and the cost effectiveness of incorporating disease events: the St. Francis Heart Study. J Am Coll Cardiol
such measures into global risk-assessment algorithms. Abstract 2003.
In the future, CAC tests may play an important role in 13. Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of
atherosclerosis: objectives and design. Am J Epidemiol 2002;
assessment of patients at intermediate risk for CHD events. 156:871881.
Higher-risk individuals may be candidates for more aggres- 14. Hatwalkar A, Agrawal N, Reiss DS, et al. Comparison of prevalence
sive pharmacological risk-factor modification using combi- and severity of coronary calcium determined by electron beam
 CHAPTER 7 / CORONARY CALCIUM SCREENING 77

tomography among various ethnic groups. Am J Cardiol 2003; 19. Morin RL, Gerber TC, McCollough CH. Radiation dose in com-
91:12251227. puted tomography of the heart. Circulation 2003;107:917922.
15. Budoff MJ, Yang TP, Shavelle RM, et al. Ethnic differences in coro- 20. Horton KM, Post WS, Blumenthal RS, et al. Prevalence of signifi-
nary atherosclerosis. J Am Coll Cardiol 2002;39:408412. cant noncardiac findings on electron-beam computed tomography
16. Lee TC, OMalley PG, Feuerstein I, et al. The prevalence and sever- coronary artery calcium screening examinations. Circulation
ity of coronary artery calcification on coronary artery computed 2002;106:532534.
tomography in black and white subjects. J Am Coll Cardiol 2003; 21. Hunold P, Schmermund A, Seibel RM, et al. Prevalence and clinical
41:3944. significance of accidental findings in electron-beam tomographic
17. Wilson PW, Kauppila LI, ODonnell CJ, et al. Lumbar aortic calci- scans for coronary artery calcification. Eur Heart J 2001;22:
fication is an important predictor of vascular morbidity and mortal- 17481758.
ity. Circulation 2001;103:15291534. 22. Hoff JA, Chomka EV, Krainik AJ, Daviglus M, Rich S, Kondos GT.
18. Witteman JC, Kannel WB, Wolf PA, et al. Aortic calcified plaques Age and gender distributions of coronary artery calcium detected by
and cardiovascular disease (the Framingham Study). Am J Cardiol electron beam tomography in 35,246 adults. Am J Cardiol
1990;66(15):10601064. 2001;87(12):13351339.
 CHAPTER 8 / CORONARY CALCIUM SCANNING: WHY 79

8 Coronary Calcium Scanning

Why We Should Perform It

AXEL SCHMERMUND, MD, STEFAN MHLENKAMP, MD,

AND RAIMUND ERBEL, MD

It is increasingly recognized that only by preventing ische- factor exposure in the actual individual has really led to the
mic heart disease can morbidity and mortality associated with development of coronary atherosclerosis. The cumulative
cardiovascular diseases be substantially reduced. There are two effect of risk-factor exposure over many years can be assessed
main preventive strategies: one that focuses on community-wide by its impact on coronary plaque development (12).
measures (population approach) and one that focuses on There is a relationship between the extent of calcified plaque
high-risk individuals (high-risk approach) (1). Certainly, both burden and that of total plaque burden (13,14). Thus, the pri-
approaches have strengths and limitations, and should work best mary aim is not to diagnose coronary stenoses, but rather to
if applied in a complementary fashion. Although clear goals can detect and quantify coronary plaque burden. Importantly, coro-
be set for community-wide measures (such as promoting healthy nary calcium in many instances seems to indicate coronary
diet, increasing physical activity, and marginalizing smoking), disease activity. Calcium is a frequent feature of plaque rupture
strategies for successful implementation depend on political, (found in 7080% of cases) (1517). Among all types of plaques
cultural, and economic surroundings. Trials in Central and West- that can be defined histologically, the extent of calcium is great-
ern Europe have failed to fulfil the expectations (2,3). By con- est in healed plaque rupture, which is frequently observed in
trast, the high-risk approach has been demonstrated to be sudden coronary death (16,18). Further, calcium is associated
effective regarding both lifestyle and pharmacological inter- with positive arterial remodeling (19). The mechanisms lead-
ventions (1). The problem with this approach is the need to ing to positive arterial remodeling appear to share common
reliably identify high-risk subjects. aspects with those ultimately leading to plaque rupture, and
Risk prediction algorithms derived from large epidemio- plaques displaying positive remodeling of the arterial segment
logical studiesin particular the Framingham study and the are prone to rupture.
Mnster (Germany) Procam study (4,5)serve as the corner- All prospective studies in seemingly healthy older adults
stone of office-based risk assessment. Absolute 10-yr risk of have found substantial increases in relative risk of a cardiovas-
myocardial infarction or cardiac death is calculated. However, cular hard event in the presence of increased amounts of coro-
risk factors identify adverse outcomes in a population. Because nary calcium (2025). Consistently, a very high negative
of the wide variability of atherosclerosis in subjects with for- predictive value (99%) regarding cardiovascular hard endpoints
mally identical risk-factor exposure (6), individual discrimina- has been reported. Conversely, high-risk persons were identi-
tion is often unsatisfactory (7). Further, there is considerable fied whose risk exceeded an event rate of 2% per year (2025).
uncertainty as to the application in external populations with These studies have prompted the authors of the American Heart
different characteristics (811). The Framingham algorithm Association Prevention Conference V and the National Choles-
appears to overestimate absolute risk in a German population terol Education Program Adult Treatment Panel III to list elec-
by approx 50% (11). tron beam computed tomography coronary calcium scanning
To estimate the risk of a coronary event, risk-factor assess- as a possible diagnostic modality in adults with several risk
ment is the most common procedure, and this is justified by factors, in particular older adults (26,27). The actual guidelines
overwhelming evidence. Risk factors lead to events by the for the prevention of CAD issued by the European Society of
patho-physiological intermediary of atherosclerotic plaque Cardiology and other leading European societies (European
disease. In this sense, risk factors yield a statistical probability Heart Journal) expand on this. It is stated that the calcium score
that coronary plaques have developed. Coronary calcium scan- is an important parameter to detect asymptomatic individuals
ning allows for direct visualization of coronary atherosclerotic at high risk for future CVD events, independent of the tradi-
plaque disease. It can answer the question of whether risk- tional risk factors (28).
The favorable data from the above-mentioned trials were all
From: Contemporary Cardiology: CT of the Heart: obtained in selected cohorts, which has been criticized.
Principles and Applications Recently, however, results from population-based studies with
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ unselected participants have confirmed these results. The

79
 80

80
SCHMERMUND, MHLENKAMP, AND ERBEL

Fig. 1. Sixty-one-year-old male who agreed to participate in the Heinz Nixdorf Recall Study, an epidemiological project in the general population, 3 d after having run the Cologne marathon
in 4:28 h. The man was in very good health and feeling well. His only cardiovascular risk factor was systemic hypertension, yielding a Framingham risk of 12% (intermediate risk category)
and a Procam risk of 5.7% (low risk category). The electrocardiograph recording in the setting of the Heinz Nixdorf Recall Study and further laboratory analyses revealed subacute anterior
myocardial infarction. The left panel shows his electrocardiogram with T inversion in I, aVL and V4-V6 (arrows) and slow R progression. His total coronary calcium score determined by
electron-beam CT was 1105 (right panel). The figure shows a section at the base of the heart with massive calcification of the left main stem and proximal portion of the left anterior descending
coronary artery. He had more coronary calcium than approx 94% of males his age (94th percentile). Coronary angiography showed severe three-vessel disease with left main stem
involvement. He underwent successful aorto-coronary bypass surgery with complete revascularization and had an uneventful course.
 CHAPTER 8 / CORONARY CALCIUM SCANNING: WHY 81

Rotterdam Coronary Calcification Study and the St. Francis 7. Wald NJ, Law M, Watt HC, et al. Apolipoproteins and ischaemic
Heart Study, with a total of almost 7600 healthy subjects, have heart disease: implications for screening. Lancet 1994;343:7579.
8. DAgostino RB Sr, Grundy S, Sullivan LM, Wilson P. CHD Risk
consistently observed relative risks associated with elevated Prediction Group. Validation of the Framingham coronary heart
coronary calcium scores on the order of 811 (2931). In both disease prediction scores: results of a multiple ethnic groups inves-
population-based studies, the predictive power of coronary tigation. JAMA 2001;286:180187.
calcium was independent of and additive to risk-factor infor- 9. Laurier D, Chau NP, Cazelles B, Segond P. PCV METRA Group.
mation. In the Rotterdam study, subjects with a Framingham Estimation of CHD risk in a French working population using
a modified Framingham model. J Clin Epidemiol 1994;47:13531364.
risk score below the median who had a calcium score in the 10. Menotti A, Puddu PE, Lanti M. Comparison of the Framingham risk
upper tertile had a relative risk of 8.9 compared with subjects function-based coronary chart with risk function from an Italian
with the same Framingham risk score whose calcium score was population study. Eur Heart J 2000;21:365370.
in the lower tertile (30). 11. Hense HW, Schulte H, Lowel H, Assmann G, Keil U. Framingham
Imaging of coronary calcium allows direct detection of coro- risk function overestimates risk of coronary heart disease in men and
women from Germanyresults from the MONICA Augsburg and
nary atherosclerotic plaques and estimation of the extent of the PROCAM cohorts. Eur Heart J 2003;24:937945.
disease and its distribution in the coronary tree (32). As opposed 12. Hoeg JM, Feuerstein IM, Tucker EE. Detection and quantitation of
to other methods of direct vessel wall diagnostics such as calcific atherosclerosis by ultrafast computed tomography in chil-
B-mode ultrasound analysis of carotid artery intima-media dren and young adults with homozygous familial hypercholester-
thickness and plaques, coronary atherosclerosis can be directly olemia. Arterioscler Thromb 1994;14:10661074.
13. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz
assessed. As opposed to noninvasive exercise tests such as RS. Coronary artery calcium area by electron-beam computed
exercise stress testing or stress echocardiography, actual tomography and coronary atherosclerotic plaque area. A histopatho-
coronary plaques are detected, with all the consequences logic correlative study. Circulation 1995;92:21572162.
regarding comparative predictive abilities. Coronary calcium 14. Schmermund A, Denktas AE, Rumberger JA, et al. Independent and
scanning adds previously unachievable prognostic power to incremental value of coronary artery calcium for predicting the extent
of angiographic coronary artery disease: comparison with cardiac risk
the information derived from risk-factor analysis. An example factors and radionuclide perfusion imaging. J Am Coll Cardiol
is shown in Fig. 1. 1999;34:777786.
Coronary calcium scanning is helpful in particular in healthy 15. Farb A, Burke AP, Tang AL, et al. Coronary plaque erosion without
subjects with an undetermined cardiovascular risk whose man- rupture into a lipid core. A frequent cause of coronary thrombosis in
agement is unclear. In these subjects, the coronary calcium sudden coronary death. Circulation 1996;93:13541364.
16. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary cal-
score can direct the treating physician towards general advice cification: insights from sudden coronary death victims. Z Kardiol
in the case of a low or negative score, or towards intensive 2000;89 Suppl 2:4953.
treatment in the case of a high score. If the score is intermediate, 17. Schmermund A, Erbel R. Current perspective: unstable coronary
this is also helpful. The physician has much more security that plaque and its relation to coronary calcium. Circulation 2001; 104:
indeed, an intermediate risk is present and the patient can be 16821687.
18. Burke AP, Kolodgie FD, Farb A, et al. Healed plaque ruptures and
treated accordingly. Finally, selected symptomatic patients sudden coronary death: evidence that subclinical rupture has a role
with known coronary artery disease can benefit from coronary in plaque progression. Circulation 2001;103:934940.
calcium scanning if detailed information on the extent of ath- 19. Burke AP, Kolodgie FD, Farb A, et al. Morphological predictors of
erosclerotic plaque disease is needed and prognostication is arterial remodeling in coronary atherosclerosis. Circulation 2002;
especially important. 105:297303.
20. Raggi P, Callister TQ, Cooil B, et al. Identification of patients at
REFERENCES increased risk of first unheralded acute myocardial infarction by elec-
tron-beam computed tomography. Circulation 2000;101:850855.
1. Yusuf S, Reddy S, unpuu S, Anand S. Global burden of cardio- 21. Arad Y, Spadaro LA, Goodman K, Newstein D, Guerci AD. Predic-
vascular disease. Part II: variations in cardiovascular disease by tion of coronary events with electron beam computed tomography.
specific ethnic groups and geographic regions and prevention strat- J Am Coll Cardiol 2000;36:12531260.
egies. Circulation 2001;104:28552864. 22. Wong ND, Hsu JC, Detrano RC, Diamond G, Eisenberg H, Gardin
2. World Health Organization European Collaborative Group. Euro- JM. Coronary artery calcium evaluation by electron beam computed
pean Collaborative Trial of Multifactorial Prevention of Coronary tomography and its relation to new cardiovascular events. Am J
Heart Disease. Final report on the 6-year results. Lancet 1986;1: Cardiol 2000;86:495498.
869872. 23. Park R, Detrano R, Xiang M, et al. Combined use of computed
3. Thelle D. Prevention of cardiovascular diseases: a scientific tomography coronary calcium scores and C-reactive protein levels
dilemma. Scand Cardiovasc J 2000;34:103105. in predicting cardiovascular events in nondiabetic individuals. Cir-
4. Wilson PWF, DAgostino RB, Levy D, Belanger AM, Silbershatz culation 2002;106:20732077.
H, Kannel WB. Prediction of coronary heart disease using risk factor 24. Kondos GT, Hoff JA, Sevrukov A, et al. Electron-beam tomography
categories. Circulation 1998;97:18371847. coronary artery calcium and cardiac events: a 37-month follow-up
5. Assmann G, Cullen P, Schulte H. Simple scoring scheme for cal- of 5635 initially asymptomatic low- to intermediate-risk adults.
culating the risk of acute coronary events based on the 10-year Circulation 2003;107:25712576.
follow-up of the prospective cardiovascular Munster (PROCAM) 25. Shaw LJ, Raggi P, Schisterman E, Berman DS, Callister TQ. Prog-
study [erratum: Circulation 2002;105:900]. Circulation. 2002;105: nostic value of cardiac risk factors and coronary artery calcium
310315. screening for all-cause mortality. Radiology 2003;228:826833.
6. Greenland P, Abrams J, Aurigemma GP, et al. Prevention Confer- 26. Greenland P, Smith SC, Grundy SN. Current perspective: improving
ence V. Beyond secondary prevention: identifying the high-risk coronary heart disease risk assessment in asymptomatic people. Role
patient for primary prevention. Noninvasive tests of atherosclerotic of traditional risk factors and noninvasive cardiovascular tests. Cir-
burden. Writing Group III. Circulation 2000;101:e16e22. culation 2001;104:18631867.
82 SCHMERMUND, MHLENKAMP, AND ERBEL

27. Third Report of the National Cholesterol Education Program (NCEP) 30. Vliegenthart R. Coronary calcification and the risk of cardiovascu-
Expert Panel on Detection, Evaluation, and Treatment of High Blood lar disease. An epidemiologic study. PhD Thesis. Thoraxcentre
Cholesterol in Adults (Adult Treatment Panel III) Final Report. Rotterdam, Netherlands, March 19, 2003.
Circulation 2002;106:31433421. 31. Arad Y, Roth M, Newstein D, Guerci A. Coronary calcification,
28. De Backer G, Ambrosioni E, Borch-Johnsen K, et al. Executive coronary disease risk factors, and atherosclerotic cardiovascular
Summary. European guidelines on cardiovascular disease preven- disease events: the St. Francis Heart Study. Hotline Session, ACC
tion in clinical practice. Third Joint Task Force of European and 2003.
Other Societies on Cardiovascular Disease Prevention in Clinical 32. Schmermund A, Baumgart D, Mhlenkamp S, et al. Natural history
Practice. Eur Heart J 2003;24:16011610. and topographic pattern of progression of coronary calcification in
29. Vliegenthart R, Oei HH, Breteler MM, et al. Coronary calcification symptomatic patients: an electron-beam CT study. Arterioscler
is a strong predictor of all-cause and cardiovascular mortality in Thromb Vasc Biol 2001;21:421426.
elderly. Circulation 2002;106(Suppl):II743.
 CHAPTER 9 / CALCIUM DETECTION WITH EBCT 83

9 Detection and Quantification of Coronary

Calcium With Electron Beam CT

AXEL SCHMERMUND, MD, STEFAN MHLENKAMP, MD,

AND RAIMUND ERBEL, MD

TECHNIQUE this scoring algorithm, it should not be completely abandoned

Electron beam computed tomography (EBCT) employs in favor of more recent algorithms.
well-known computed tomography technology. However, in
CORONARY CALCIUMWHAT DOES IT MEAN?
distinction to other CT machines, no mechanical parts are
moved. Whereas in usual CT machines, the distance between Non-contrast-enhanced cardiac EBCT allows for direct,
cathode and anode is very short, it measures approx 9 ft in noninvasive visualization of calcified coronary plaques (Fig. 1).
EBCT. The electron beam, which produces the X-rays by strik- The primary aim is not to diagnose coronary stenoses, but rather
ing the anode, is steered over this distance by an electromag- to detect and quantify coronary plaque burden. Coronary
netic deflection system. The latest generation of EBCT plaques, usually not highly stenotic, are the underlying sub-
machines (e-Speed, GE Imatron) achieves an image acquisi- strate of the acute coronary syndromes (5). Coronary calcium
tion time of only 30 ms, which is sufficient to freeze the motion is a specific expression of coronary atherosclerotic plaque dis-
of the heart. ease (6,7). There is a relationship between the extent of calci-
Standardized methods for imaging, identification, and quan- fied plaque burden and that of total plaque burden (69). Total
tification of coronary calcium using EBCT have been estab- plaque burden is one of the most important predictors of coro-
lished (1,2). The current generation of EBCT scanners is usually nary risk (10,11). The more calcium detected, the more plaque
operated in the high-resolution mode with continuous, non- there is. This carries direct implications for an individuals
overlapping slices of 3-mm thickness and an acquisition time coronary risk (5).
of 100 ms. Patients are positioned supine, and a sufficient num- It may be argued that calcium is associated with an overall
ber of slices is obtained to cover the complete heart through the increased activity of coronary atherosclerotic disease. Histo-
apex (usually 3640 slices). Electrocardiographic triggering is pathologic studies have demonstrated that coronary calcium is
performed at a fixed point within the RR interval, generally a frequent feature of plaque rupture (5,12,13) and is found even
40% or 80%. Coronary calcium is defined as a hyperattenuating in subjects who die of sudden coronary death as the first mani-
lesion above the threshold of a CT density of 130 Hounsfield festation of ischemic heart disease under the age of 50 yr (14).
units (HU) in an area of two or more adjacent pixels. The cal- Among all types of histologically defined types of plaques
cium score is a product of the area of calcium and a factor rated detected in young victims of sudden coronary death, acute rup-
14, dictated by the maximum CT density (1). The calcium tures contain calcium most frequently (80%), while healed
score can be calculated for a given coronary segment, a specific ruptures contain the greatest amount of calcium (13,15). Plaque
coronary artery, or for the entire coronary system (Fig. 1). erosions, on the other hand, are associated with little calcium
More recently, a volumetric score has been introduced, (12,13). Calcium is found preferentially in plaques with expan-
which uses isotropic interpolation and may thus be more repro- sive (positive) arterial remodeling (16). The mechanisms
ducible (3). Further, calcium mass measurements have been leading to expansive arterial remodeling appear to share com-
suggested, which would render measurements obtained with mon aspects with those ultimately leading to plaque rupture,
various CT technologies (EBCT or spiral CT with different and plaques displaying positive remodeling of the arterial seg-
slice thickness) comparable (4). Such measurements require ment are prone to rupture (17). Finally, lesions characterized by
simultaneous phantom calibration and are currently not widely ultrasound radiofrequency analysis as unstable frequently con-
available. Because the above-mentioned calcium score has tain calcium (18).
been employed since 1990, and all long-term data are based on The findings in studies using conventional intravascular
ultrasound have been less unequivocal, perhaps as a result of
differences in the ability to define calcium and lesion charac-
From: Contemporary Cardiology: CT of the Heart:
teristics (1923). Whereas some studies have detected less
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ calcium in the culprit lesion in patients with unstable angina

83
84 SCHMERMUND, MHLENKAMP, AND ERBEL

Fig. 1. Coronary calcium scans at the level of the left anterior descending coronary artery. Left panel: No coronary calcium. Mid panel:
Moderate coronary calcium with typical localization in the area of bifurcation with first diagonal branch, total calcium score 108. Right panel:
Extensive coronary calcium, total calcium score 3.532.

pectoris or acute myocardial infarction than in patients with determined by questionnaire. In a multivariate analysis in a
stable symptoms (1921), others did not find a difference subgroup of patients with all information available (n = 787),
(22,23), or, in a prospective study, detected more calcium in the calcium score cutpoints were associated with odds ratios for
patients who later sustained an event (24). In summary, coro- suffering any event in the range of 1420 and the risk factors,
nary calcium indicates the presence and extent of coronary including age, with odds ratios in the range of 36 (31).
atherosclerotic plaque disease. The weight of evidence sug- Raggi et al. found that subjects with calcium scores in the
gests that coronary calcium indicates atherosclerotic disease highest quartile had an odds ratio of 21.5 (95% CI, 2.8162.4)
activity and is associated with healed or acute plaque rupture for suffering acute myocardial infarction or cardiac death (32).
and positive arterial remodeling. Risk factor data were obtained by questionnaire. Subjects in the
highest quartile of cardiovascular risk factor distribution had
CORONARY CALCIUMWHAT DOES IT PREDICT? an odds ratio of 7.0 (95% CI, 1.631.5).
In patients presenting to the emergency room with chest Wong et al. assessed risk factors (by questionnaire) and
pain and no initial objective signs of myocardial ischemia, a coronary calcium scores in 881 subjects. After adjustment for
negative EBCT indicated an excellent prognosis with regard to risk factors, coronary calcium scores in the third or fourth
major cardiac events over the subsequent 14 mo (2527). quartile were associated with a relative risk of 4.5 or 8.8 (33).
EBCT yielded negative predictive values in the range of Accordingly, these three studies suggest that EBCT was inde-
98100%. In symptomatic patients undergoing coronary pendent of risk-factor information for predicting events and
angiography, increased amounts of coronary calcium detected was clearly superior compared with risk-factor analysis alone.
by EBCT were highly predictive of subsequent events over 30 mo However, risk factors were determined only by questionnaire.
(2830). In direct comparison, EBCT performed better than Detrano et al. reported an odds ratio of 2.7 associated with
coronary angiography (that is, number of stenotic major coro- a calcium score above the median (34). Receiver operating
nary arteries) in this respect (2830). curve analysis of calcium scores for separating subjects with vs
A number of studies have examined the predictive value of without acute myocardial infarction or coronary death yielded
EBCT coronary calcium quantification in selected asymptom- an area under curve ( SE) of 0.64 0.05. For comparison, this
atic populations. All of these studies have demonstrated that value was 0.86 0.07 in the study by Arad et al. (31). Risk
the coronary calcium score predicts coronary events, defined as factors were assessed by questionnaire and direct measure-
an aggregate of coronary death, nonfatal myocardial infarction, ments of laboratory values. An ECG was recorded in all sub-
and revascularization (Table 1). Arad et al. observed that the jects. The combined analysis of risk factors and ECG yielded
prespecified calcium score cutpoints of 80 and 160 both were an area under the curve of 0.69 0.05, so that for predicting
associated with odds ratios of approx 22 for suffering coronary hard events, EBCT was not superior in this report. However, in
death or acute myocardial infarction, with wide confidence a later analysis of the same data set, but excluding diabetic
intervals (CIs) (95% CI, 5.197.4 and 6.477.4, respectively) subjects, coronary calcium was a strong and independent
(31). There was a tendency for higher scores in subjects with predictor of coronary death and nonfatal myocardial infarction
hard events compared to subjects who underwent revasculari- as well as revascularization (35). After adjustment for risk
zation. In this study, the status of established risk factors was factors, the relative risk of sustaining a hard event was 4.96.1
 CHAPTER 9 / CALCIUM DETECTION WITH EBCT 85

Table 1
Prognostic Studies in Selected Patient Populations

Patients Patients
with event: without event:
calcium calcium
Mean Median No. of score score
Author No. of Mean Women calcium calcium Follow-up events (mean SD, (mean SD, Definition
(reference) patients age (%) score score (mo) (%) median) median) RR of events

Arad (31) 1177 53 11 29 ~ 156 4 43 39 764 935 135 432 14.3 Coronary
(3.3) (511) (3) death,
nonfatal
MI, revasc.
Raggi (32) 632 52 9 50 101 254 3.1 32 27 303 441 92 240 12.5 Coronary
(4.3) (141) (2.4) death,
nonfatal MI
Wong (33) 926 54 10 21 143 5 40 28 n/a n/a 4.5a MI, revasc.,
(3.0) cerebrovasc.
event
Detrano (34) 1196 66 8 11 452 457 44 41 46 n/a n/a 2.3 Coronary
(3.8) death,
nonfatal
MI, revasc.
Park (35) 967 66 8 10 n/a n/a 77 104 395 571 195 378 4.4 Coronary
(203) (37) 7.5a death,
nonfatal
MI, revasc.
aAdjusted for risk factors. SD, standard deviation; MI, myocardial infarction.

in the highest vs the lowest tertile of calcium scores (depending

on the level of C-reactive protein), and the risk for all endpoints
(including revascularization) was 4.47.5. In this analysis,
coronary calcium was superior to combined risk-factor analy-
sis (on the basis of actual laboratory data) and ECG and also to
C-reacting protein (35).
These findings were corroborated by the analysis of large
databases with a total of >15,000 patients (36,37). Kondos et
al. examined subjects who were all self-referred and who were
followed over a mean of 3.1 yr. They observed a relative risk
of myocardial infarction or cardiac death of 7.2 in men with a
coronary calcium score in the upper quartile (36). In women,
the event rate was only 0.4%, so that no meaningful analysis
was possible. Shaw et al. documented only total mortality (37).
A calcium score in the range between 400 and 1000 was asso- Fig. 2. Rate of coronary artery disease (CAD)-related deaths associ-
ated with different coronary calcium scores in the Rotterdam Coro-
ciated with a 6.2-fold increase in death rate. A calcium score nary Calcification Study, a population-based study. Mean follow-up
>1000 was associated with a 12.3-fold increase in death rate. was 2.7 yr (38).
Perhaps the most important aspect of these studies was that
high-risk persons could be identified whose risk exceeded a
rate of 2% per year (hard events) (36,37) or 1% per year for
overall mortality. cluding 73 coronary heart disease deaths. There was a strong
Along the same lines, Wahys et al. reported retrospective and graded relationship between the EBCT-derived calcium
data (38). In analogy with a study in symptomatic patients (30), score and cardiovascular disease events (39). The 15% of par-
asymptomatic patients with very high calcium scores (>1000) ticipants with the highest scores had a relative risk of 11 of
had a greatly elevated risk of sustaining an event (38). dying from ischemic heart disease. The yearly death rate
The Rotterdam Coronary Calcification Study was the first regarding only ischemic heart disease was almost 2% (Fig. 2).
population-based study to have yielded data in unselected par- In the approx 50% of participants with low or negative calcium
ticipants (39,40). In this project, 2013 persons from a suburb of scores, this rate was only 0.07% (39). After adjustment for risk
Rotterdam, the Netherlands, were recruited. After a mean fol- factors, the association between coronary calcium scores and
low-up time of 2.7 yr, 116 cardiovascular events occurred, in- cardiovascular as well as coronary morbidity and mortality
86 SCHMERMUND, MHLENKAMP, AND ERBEL

Table 2
Indications for the Use of CT Coronary Calcium Scanning

Asymptomatic persons:
1. Persons with risk factors who cannot be determined by office-based risk assessment to have either a low or a high cardiovascular risk
2. Older persons in whom the established risk factors lose some of their predictive value and whose risk remains indeterminate
Symptomatic persons:
3. Patients in whom advanced risk stratification is useful, for example if extensive coronary plaque disease is suspected
4. Patients presenting to the emergency room with nonspecific chest pain (rule out myocardial infarction)

remained strong (40). The relative risk of coronary heart dis- magnitude of the problem warrants tools for improved risk
ease increased gradually in patients with higher coronary cal- prediction, and published data suggest a prominent role for
cium scores compared with those in the reference category with EBCT. However, it is currently unresolved who benefits
calcium scores of 0100, which comprised 50% of all partici- from an EBCT scan in terms of future coronary events
pants (40). It was 3.2 in subjects with scores of 101500, 5.1 in which can be prevented on the basis of the calcium study.
subjects with 5011000, and 8.3 in subjects with scores >1000. Clearly, the indications for a scan need to be strictly defined
The coronary event rates in 2.7 yr were 0.8% in the lowest score to avoid overdiagnosis and disproportionate increases in
group and increased to 3.3%, 5.0%, and 8.7% in the other cal- health care expenditure.
cium score groups. Indeed, the authors of the American Heart Association
Prevention Conference V and the National Cholesterol
In summary, the EBCT-derived coronary calcium score has
Education Program Adult Treatment Panel III list coro-
demonstrated its ability to identify subjects with a very low risk
nary calcium scanning as a useful diagnostic modality in
of cardiovascular and coronary disease as well as subjects with
adults with several risk factors, in particular older adults
a very high rate of events (>2% per year). This ability goes
(43,44). Fig. 3 shows how coronary calcium scoring might
beyond that of conventional risk-factor analysis, is indepen- be used to improve individual risk stratification.
dent of it, and adds incremental predictive value. The most recent statement is provided in the actual guide-
PROGRESSION OF CORONARY CALCIUM lines for the prevention of CAD issued by the European
Society of Cardiology and other leading European societ-
Follow-up coronary calcium scans appear to be useful, as ies (45). The European guidelines state that the calcium
the progression of calcification is influenced by low-density score is an important parameter to detect asymptomatic
lipoprotein values and can be attenuated or even stopped by individuals at high risk for future CVD events, indepen-
lipid-lowering (statin) medication (3,41). In patients dent of the traditional risk factors. This statement expands
followed in an outpatient setting, the progression of coronary on the previous statements (43,44).
calcium burden is approx 25% per year (42). However, there is The interpretation of the EBCT scan is straightforward
great inter-individual variability, probably as a result of the with regard to prognostication, and it indeed applies to any
interplay of numerous factors which influence the process. person regardless of symptomatic status. Because it is
Also, baseline plaque burden determines the rate of progres- internationally comparable, the Agatston calcium score
sion (42). Progression is observed at typical predilection sites remains the standard algorithm, but an increasing role of
of atherosclerosis in the coronary tree. Changes in overall coro- calcium area and/or volume-scoring algorithms can be
nary calcium result from uniform changes at these sites. This expected. The absence of identifiable coronary calcium is
uniform pattern of change suggests that the development of consistent with no or at most minimal coronary atheroscle-
calcified plaque disease is a coronary systemic process (42). rosis. Individuals in this group fall into a favorable prog-
In patients with elevated cholesterol values, progression of nostic group. On the other hand, Agatston calcium scores
coronary calcium appears to be accelerated and on the order of exceeding 80100 are highly consistent with at least non-
50% per yr (3). In patients treated with statin therapy, the annual obstructive CAD (46,47). The presence of moderate cal-
progression is substantially reduced. Values between 10 and cium scores in an otherwise asymptomatic individual
20% per year have been reported (3,41). If low-density lipopro- suggests that further evaluations may be prudent, includ-
ing careful identification and aggressive treatment of modi-
tein (LDL)-cholesterol can be reduced to levels <100 mg/dL, it
fiable risk factors. It cannot be overemphasized that
appears possible to stop the progression of coronary calcium
calcium scores need to be interpreted with regard to age
(41). The value of coronary calcium scanning for assessing the
and sex distribution. According to the natural history of
treatment effect in individual patients remains to be clarified. coronary atherosclerosis, calcium scores increase in higher
CORONARY CALCIUMHOW CAN IT BE USED? age groups. At the same time, the prognostic significance
is shifted, because clinically overt coronary artery disease
There are four major areas where EBCT scanning for coro-
occurs in some patients with increasing age, but not in
nary calcium can offer clinical value (Table 2). others. With this in mind, individuals with significant coro-
1. Application of coronary calcium scoring to define the nary calcium (calcium score >400500) clearly have
extent of coronary atherosclerosis in asymptomatic indi- advanced coronary atherosclerosis and require strict mea-
viduals at risk for coronary disease is of great interest. The sures regarding modifiable risk factors. Additionally,
 CHAPTER 9 / CALCIUM DETECTION WITH EBCT 87

Fig. 3. Algorithm for the use of electron beam CT coronary calcium scanning for advanced risk stratification (41,42). In patients in whom office-
based risk assessment determines an intermediate risk level (approx 40% of patients), calcium scanning can be used for furtheradvanced
risk stratification (box). Even in the presence of a rather intermediate calcium score, the scan is useful. In this case, the probability of a truly
intermediate risk is much greater than if determined by office-based risk assessment alone. Of note, the initial office-based risk estimate might
be less accurate than the aggregate risk estimate including coronary calcium scanning. It is open to question whether one should concentrate
on examining intermediate-risk patients in order to define their management more precisely or whether some patients with a presumably low
or high risk on the basis of office-based risk assessment may also profit from undergoing a coronary calcium scan.

Fig. 4. Algorithm used by the authors for interpreting electron-beam computed tomography coronary calcium scans (2). In a first step, the total
calcium score is considered. Because of the relationship between coronary calcified and coronary total plaque burden, a high calcium score
indicates a high coronary plaque burden. In a second step, the total calcium score is put in perspective regarding age and sex of the individual
person. If the calcium score is greater than expected (above the median value), cardiovascular risk is increasingly elevated.

depending on all available clinical information, further cation of data from one system to another. On the basis of
evaluations should probably be considered to determine these considerations, the German Working Group on
the potential for myocardial ischemia (47). Cardio-CT has suggested a scheme for interpretation of
Apart from the calcium score, age and sex of the patient EBCT coronary calcium scans (Fig. 4) (2).
should be considered by using specific percentile values 2. The issues explained above also pertain to symptomatic
(normogram) such as those provided for >35,000 adults patients with known coronary artery disease in that the
in ref. 48. These percentile values were calculated on the extent of coronary arteriosclerosis represents an important
basis of EBCT studies. They appear to be very similar to predictor of events (2830). Therefore, coronary calcium
values derived with four-slice spiral CT (49). But, dynamic scanning may be helpful in selected individuals with coro-
changes in imaging parameters need to be considered that nary artery disease where it appears particularly important
may pose problems regarding the comparability and appli- to obtain complete prognostic information.
88 SCHMERMUND, MHLENKAMP, AND ERBEL

3. The absence of coronary calcium by EBCT has negative 3. Callister TQ, Raggi P, Cooil B, Lippolis NJ, Russo DJ. Effect of
predictive values between 90% and 100% for obstructive HMG-CoA reductase inhibitors on coronary artery disease as
assessed by electron-beam computed tomography. New Engl J Med
CAD (2527). This appears of value in the patient with no 1998;339:19721978.
prior clinical coronary artery disease who presents to the 4. Hong C, Becker CR, Schoepf UJ, et al. Coronary artery calcium:
emergency department with chest pain or a questionable absolute quantification in nonenhanced and contrast-enhanced
ischemic syndrome, but a negative or nondiagnostic ECG multi-detector row CT studies. Radiology 2002;223:474480.
and negative cardiac enzymes. Recent studies in this set- 5. Schmermund A, Erbel R. Current perspective: unstable coronary
plaque and its relation to coronary calcium. Circulation
ting have demonstrated that the negative predictive value 2001;104:16821687.
is 99100% (2527). Especially in a patient with a low to 6. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcium:
intermediate likelihood of obstructive coronary artery dis- pathophysiology, epidemiology, imaging methods, and clinical
ease, a negative EBCT could thus facilitate early dismissal implications. A statement for health professionals from the Ameri-
can Heart Association. Circulation 1996;94:11751192.
from the emergency department or chest pain unit. Con-
7. ORourke RA, Brundage BH, Froelicher VF, et al. American Col-
versely, if the calcium score is consistent with moderate or lege of Cardiology/American Heart Association Expert Consensus
severe coronary atherosclerosis, then additional cardiac Document on electron-beam computed tomography for the diagno-
evaluation might be appropriate. sis and prognosis of coronary artery disease. J Am Coll Cardiol
Finally, in a patient with an indeterminate or otherwise 2000;36:326340.
8. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz
questionable stress test result, EBCT can be valuable by RS. Coronary artery calcium area by electron-beam computed
adding anatomic information on the extent of coronary tomography and coronary atherosclerotic plaque area. A histopatho-
atherosclerosis to the information on the functional sig- logic correlative study. Circulation 1995;92:21572162.
nificance of the disease sought with the stress test. Other- 9. Schmermund A, Denktas AE, Rumberger JA, et al. Independent and
wise, in this group of patients, the same issues apply as incremental value of coronary artery calcium for predicting the
extent of angiographic coronary artery disease: comparison with
explained above. cardiac risk factors and radionuclide perfusion imaging. J Am Coll
4. The progression of coronary atherosclerosis is an impor- Cardiol 1999;34:777786.
tant area of research and has considerable implications 10. Proudfit WL, Bruschke VG, Sones FM, Jr. Clinical course of patients
with regard to an adverse prognosis in the presence of with normal or slightly or moderately abnormal coronary
accelerated disease progression. Recent investigations arteriograms: 10-year follow-up of 521 patients. Circulation 1980;62:
712717.
have documented substantial progress in the reliability of 11. Emond M, Mock MB, Davis KB, et al. Long-term survival of medi-
EBCT scanning (50,51). The median variability of an cally treated patients in the Coronary Artery Surgery Study (CASS)
EBCT scan repeated in the same patient within a few min- Registry. Circulation 1994;90:26452657.
utes is much less than the annual rate of progression found 12. Farb A, Burke AP, Tang AL, et al. Coronary plaque erosion without
in symptomatic or high-risk patients (3,41,42). Indeed, rupture into a lipid core. A frequent cause of coronary thrombosis in
sudden coronary death. Circulation 1996;93:13541364.
EBCT enables tracking the progression of calcified plaque 13. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary
disease, which appears to parallel the mode of progression calcification: insights from sudden coronary death victims. Z Kardiol
in angiographic studies and predicts future myocardial 2000;89(Suppl 2):4953.
infarction (52). Recent data suggest that the influence of 14. Schmermund A, Schwartz RS, Adamzik M, et al. Coronary athero-
risk-factor modification on disease progression can be sclerosis in unheralded sudden coronary death under age fifty: histo-
pathologic comparison with healthy subjects dying out of hospital.
evaluated non-invasively using EBCT. This offers a new
Atherosclerosis 2001;155:499508.
avenue for following patients both with regard to the prog- 15. Burke AP, Kolodgie FD, Farb A, et al. Healed plaque ruptures and
nosis related to differential progression of coronary ath- sudden coronary death: evidence that subclinical rupture has a role
erosclerosis and with regard to the effect of various in plaque progression. Circulation 2001;103:934940.
treatment strategies. At present, the ability of EBCT to 16. Burke AP, Kolodgie FD, Farb A, et al. Morphological predictors of
arterial remodeling in coronary atherosclerosis. Circulation 2002;
monitor the efficacy of therapeutic interventions appears
105:297303.
established only for small groups of patients and not nec- 17. Ward MR, Pasterkamp G, Yeung AC, Borst C. Arterial remodeling.
essarily for individuals. There is substantial variability of Mechanisms and clinical implications. Circulation 2000;102:
progression in patients with apparently similar risk-factor 11861191.
profiles under treatment. However, this may only reflect 18. Nair A, Kuban BD, Tuzcu EM, Schoenhagen P, Nissen SE, Vince
DG. Coronary plaque classification with intravascular ultrasound
the complex, multifactorial nature of the disease. The use
radiofrequency data analysis. Circulation 2002;106:22002206.
of EBCT to assess therapeutic interventions by direct visu- 19. Rasheed Q, Nair R, Sheehan H, Hodgson J McB. Correlation of
alization of the coronary arteries holds great promise. Fur- intracoronary ultrasound plaque characteristics in atherosclerotic
ther, the first available data suggest that subjects with coronary artery disease patients with clinical variables. Am J Cardiol
increased progression of coronary calcium are much more 1994;73:753758.
20. Mintz GS, Pichard AD, Popma JJ, et al. Determinants and correlates
likely to sustain a clinical event than subjects with less
of target lesion calcium in coronary artery disease: a clinical,
progression (52,53). angiographic and intravascular ultrasound study. J Am Coll Cardiol
1997;29:268274.
REFERENCES 21. Beckman JA, Ganz J, Creager MA, Ganz P, Kinlay S. Relationship
1. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M of clinical presentation and calcification of culprit coronary artery
Jr, Detrano R. Quantification of coronary artery calcium using stenoses. Arterioscler Thromb Vasc Biol 2001;21:16181622.
ultrafast computed tomography. J Am Coll Cardiol 1990;15: 22. Bocksch WG, Schartl M, Beckmann SH, Dreysse S, Paeprer H.
827832. Intravascular ultrasound imaging in patients with acute myocardial
2. Erbel R, Moshage W. Ttigkeitsbericht Arbeitsgruppe Elektro- infarction: comparison with chronic stable angina pectoris. Coron
nenstrahltomographie. Z Kardiol 1999;88:459465. Artery Dis 1994;5:727735.
 CHAPTER 9 / CALCIUM DETECTION WITH EBCT 89

23. de Feyter PJ, Ozaki Y, Baptista J, et al. Ischemia-related lesion 39. Vliegenthart R, Oei HH, Breteler MM, et al. Coronary calcification
characteristics in patients with stable or unstable angina. A study is a strong predictor of all-cause and cardiovascular mortality in
with intracoronary angioscopy and ultrasound. Circulation 1995;92: elderly. Am Heart Association Scientific Sessions 2002 (Abstract
14081413. no. 3651).
24. Abizaid AS, Mintz GS, Abizaid A, et al. One-year follow-up after 40. Vliegenthart R. Coronary calcification and the risk of cardiovascu-
intravascular-ultrasound assessment of moderate left main coronary lar disease. An epidemiologic study. PhD Thesis. Thoraxcentre
artery disease in patients with ambiguous angiograms. J Am Coll Rotterdam, Netherlands, March 19, 2003.
Cardiol 1999;34:707715. 41. Achenbach S, Ropers D, Pohle K, et al. Influence of lipid-lowering
25. Laudon DA, Vukov LF, Breen JF, Rumberger JA, Wollan PC, therapy on the progression of coronary artery calcification: a pro-
Sheedy PF II. Use of electron-beam computed tomography in the spective evaluation. Circulation 2002;106:10771082.
evaluation of chest pain patients in the emergency department. Ann 42. Schmermund A, Baumgart D, Mhlenkamp S, et al. Natural history
Emerg Med 1999;33:1521. and topographic pattern of progression of coronary calcification in
26. McLaughlin VV, Balogh T, Rich S. Utility of electron beam com- symptomatic patients: an electron-beam CT study. Arterioscler
puted tomography to stratify patients presenting to the emergency Thromb Vasc Biol 2001;21:421426.
room with chest pain. Am J Cardiol 1999;84:327328. 43. Greenland P, Smith SC, Grundy SN. Current perspective: Improv-
27. Georgiou D, Budoff MJ, Kaufer E, Kennedy JM, Lu B, Brundage ing coronary heart disease risk assessment in asymptomatic people.
BH. Screening patients with chest pain in the emergency department Role of traditional risk factors and noninvasive cardiovascular tests.
using electron beam tomography: a follow-up study. J Am Coll Circulation 2001;104:18631867.
Cardiol 2001;38:105110. 44. Third Report of the National Cholesterol Education Program (NCEP)
28. Detrano R, Hsiai T, Wang S, et al. Prognostic value of coronary Expert Panel on Detection, Evaluation, and Treatment of High Blood
calcification and angiographic stenoses in patients undergoing coro- Cholesterol in Adults (Adult Treatment Panel III) Final Report. Cir-
nary angiography. J Am Coll Cardiol 1996;27:285290. culation 2002;106:31433421.
29. Keelan PC, Bielak LF, Ashai K, et al. Long-term prognostic value 45. De Backer G, Ambrosioni E, Borch-Johnsen K, et al. Executive
of coronary calcification detected by EBCT in patients undergoing Summary. European guidelines on cardiovascular disease preven-
coronary angiography. Circulation 2001;104:412417. tion in clinical practice. Third Joint Task Force of European and
30. Mhlenkamp S, Lehman N, Schmermund A, Grnemeyer D, Seibel Other Societies on Cardiovascular Disease Prevention in Clinical
R, Erbel R. Electron beam CT-based coronary calcium quantities Practice. Eur Heart J 2003;24:16011610.
predict future hard events in symptomatic males with advanced coro- 46. Rumberger JA, Sheedy PF, Breen JF, Schwartz RS. Electron beam
nary artery diseasea 5 year follow-up study. Eur Heart J computed tomographic calcium score cutpoints and severity of
2003;92:845854. associated angiographic lumen stenosis. J Am Coll Cardiol 1997;29:
31. Arad Y, Spadaro LA, Goodman K, Newstein D, Guerci AD. Predic- 15421548.
tion of coronary events with electron beam computed tomography. 47. Rumberger JA, Brundage BH, Rader DJ, Kondos G. Electron beam
J Am Coll Cardiol 2000;36:12531260. computed tomographic coronary calcium scanning: a review and
32. Raggi P, Callister TQ, Cooil B, et al. Identification of patients at guidelines for use in asymptomatic persons. Mayo Clin Proc 1999;
increased risk of first unheralded acute myocardial infarction by elec- 74:243252.
tron-beam computed tomography. Circulation 2000;101:850855. 48. Hoff JA, Chomka EV, Krainik AJ, et al. Age and gender distribu-
33. Wong ND, Hsu JC, Detrano RC, et al. Coronary artery calcium tions of coronary artery calcium detected by electron beam tomog-
evaluation by EBCT and its relation to new cardiovascular events. raphy in 35,246 adults. Am J Cardiol 2001;87:13351339.
Am J Cardiol 2000;86:495498. 49. Schmermund A, Erbel R, Silber S. Age and gender distribution of
34. Detrano RC, Wong ND, Doherty TM, et al. Coronary calcium does coronary artery calcium measured by four-slice computed tomogra-
not accurately predict near-term future coronary events in high-risk phy in 2,030 persons with no symptoms of coronary artery disease.
adults [published errata in Circulation 2000;101:697 and Circula- Am J Cardiol 2002;90:168173.
tion 2000;101:1355]. Circulation 1999;99:26332638. 50. Adamzik M, Schmermund A, Reed JE, Adamzik S, Behrenbeck T,
35. Park R, Detrano R, Xiang M, et al. Combined use of computed Sheedy PF 2nd. Comparison of two different software systems for
tomography coronary calcium scores and C-reactive protein levels electron-beam CT-derived quantification of coronary calcification.
in predicting cardiovascular events in nondiabetic individuals. Cir- Invest Radiol 1999;34:767773.
culation 2002;106:20732077. 51. Achenbach S, Ropers D, Mhlenkamp S, et al. Variability of repeated
36. Kondos GT, Hoff JA, Sevrukov A, et al. Electron-beam tomography coronary artery calcification measurements by electron beam
coronary artery calcium and cardiac events: a 37-month follow-up tomography. Am J Cardiol 2001;87:210213.
of 5635 initially asymptomatic low- to intermediate-risk adults. 52. Raggi P, Cooil B, Shaw LJ, et al. Progression of coronary calcium
Circulation 2003;107:25712576. on serial electron beam tomographic scanning is greater in pa-
37. Shaw LJ, Raggi P, Schisterman E, Berman DS, Callister TQ. Prog- tients with future myocardial infarction. Am J Cardiol 2003;92:
nostic value of cardiac risk factors and coronary artery calcium 827829.
screening for all-cause mortality. Radiology 2003;228:826833. 53. Arad Y, Roth M, Newstein D, Guerci A. Coronary calcification,
38. Wayhs R, Zelinger AB, Raggi P. High coronary artery calcium scores coronary disease risk factors, and atherosclerotic cardiovascular
pose an extremely elevated risk for hard events. J Am Coll Cardiol disease events: the St. Francis Heart Study. Hotline Session,
2002;9:225230. ACC 2003.
 CHAPTER 10 / CALCIUM DETECTION WITH DUAL-SLICE CT 91

10 Detection and Quantification of Coronary

Calcium With Dual-Slice CT

JOSEPH SHEMESH, MD

INTRODUCTION rarely contain a few calcium granules. Atheroma, type IV

The introduction of dual-slice spiral technique by Elscint lesions, emerge primarily in the decade after puberty at the
in 1992 allowed the imaging of the heart with 3-mm slices highly susceptible artery sites, and immediately contain cal-
within 30 s (a breath-hold). The advantage of double-helical cium granules within some smooth muscle cells and extracel-
computed tomography (DHCT) over single-slice helical CT lularly among the vast mass of lipidic particles (the lipid core)
is its ability to acquire data twice as fast as single-slice helical that now forms, replaces, and expands an extensive region in
CT owing to its two parallel arcs of detectors that are simul- the deep intima. According to Stary (1), cell organelles were
taneously irradiated by a single node. Both single- and dual- damaged and sequestered calcium. Intracellular calcium gran-
slice techniques acquire a volume of spatial information over ules become extracellular when dead cells break up, but cal-
several seconds without electrocardiographic gating. These cification of cell components apparently also occurs only after
data can be reconstructed into thin overlapping axial images, their release into the extracellular space after cell disintegra-
with an acquisition time for each image of approx 0.61.0 s. tion. Lipidic cell remnants can calcify, including matrix
The use of overlapping methods compensates for the absence vesicles, cell membranes, secondary and tertiary lysosomes,
of electrocardiogram (ECG) triggering, enabling the detec- macrophage foam cells, and smooth muscle cells. The athero-
tion of minimal spotty calcific lesions with an area threshold mas of people up to early middle age do not often obstruct the
of 0.5 mm2. This technique was further improved by the 4- arterial lumen much, but are vulnerable to rupture or fissur-
and 16-slice multidetector spiral techniques, which enable ing, leading to acute coronary syndrome (2). These poten-
the use of ECG triggering with subsecond slice time. The tially dangerous lesions cannot be detected by the current
dual-slice technique of the Twin system (Philips LTD) is a noninvasive diagnostic methods, which are all based on flow-
contribution to practical cardiology. The noninvasive detec- limiting protruding lesions that obstruct at least 50% of the
tion and measurement of the early stages of atherosclerosis vessel lumen. This preclinical stage of CAD can be detected
within the wall of the coronary arteries provides the clinician by the high resolution of the spiral technique, providing a
with a new perspective on coronary artery disease (CAD) and potential means for population screening and an anatomical
adds important information to the current noninvasive meth- basis for primary prevention measures.
ods. This chapter will summarize our experience of using the The proportion of lipid-to-calcium deposit is much higher
coronary calcium score in clinical practice. in young than old people. From about the fifth decade of life,
advanced lesions contain lumps and plates of calcium rather
CORONARY CALCIUM AND ATHEROSCLEROSIS than, or in addition to, the much smaller granules. Granules
In order to recommend the most appropriate diagnostic test increase in size by encrustation, and adjacent granules fuse to
for each individual, it is essential to recognize the underlying form the larger structures. Type V and VI lesions contain
process of the disease. For CAD, it is atherosclerosis, which large-sized deposits but continue to have a core of extracellu-
starts early in life. Atherosclerotic lesions (types I and II) are lar lipid. In high-grade calcifications, designated as type VII
found in 69% of children by the time of puberty (1). The lesions, mineral dominates over all other lesion components.
changes are minimal in such lesionslipids have not yet ac- Lamellar bone (osseous metaplasia) may be present in such
cumulated exracellularly, the matrix between cells remains lesions. Type VVII lesions are more often associated with
unchanged, and they are not associated with the emergence of lumen obstruction, forming the basis of the chronic manifes-
calcium deposits. Type III lesions are the link between the tations of CAD, stable angina pectoris (SAP) with generally
minimal changes of children and the atheroma of adults, and ischemic changes upon stress tests diagnostic modalities.
These symptomatic obstructive lesions require revascular-
From: Contemporary Cardiology: CT of the Heart: ization by angioplasty or bypass operationthe chronic,
Principles and Applications soft, or interventional events.
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

91
92 SHEMESH

Fig. 1. The main indications for coronary calcium screening are summarized in relation to stages of the atherosclerotic process.
Screening: asymptomatic individuals may have mild to moderate amounts of coronary calcifications (CC) (Stary stages IIIIV), or may have
earlier noncalcific plaques (Stary III). At these stages the current noninvasive tests are normal, as well as the coronary angiogram. The goal
of the coronary scan is to provide a more sensitive screening tool, in order to introduce early primary preventive measures to treat the
atherosclerosis systematically.
Diagnostic: the diagnosis of heavy calcifications (Stary VVII) helps to confirm the presence of obstructive disease in clinically equivocal
patients. The goal is to identify obstructive vessels for revascularization in order to treat ischemia and anginal symptoms. Coronary CT for this
indication should not be confused with its use as a screening tool.

CLINICAL MANIFESTATION OF CAD: and only 7 (10%) had extensive lesions. We demonstrated in
CHRONIC VS ACUTE this study that extensive calcium characterizes the coronary
CAD is clinically manifested by two different main syn- arteries of patients with chronic SAP, whereas first AMI most
dromes. The acute syndrome strikes suddenly as sudden death, frequently occurs in mildly calcified or noncalcified culprit
acute myocardial infarction (AMI), or unstable angina. This arteries. We recently confirmed this observation in a prospec-
syndrome is related to fissure or rupture of one or more vulner- tive study (4), which investigated the underlying calcific ath-
able plaques, generally characterized by a lipid-rich core and erosclerotic lesions in a cohort of 50 hypertensive patients who
thin, fibrous cap. The chronic CAD syndrome manifests when sustained acute vs chronic coronary events. Their mean age
obstructive lesions develop over many years, resulting in was 66 6 yr, and 78% were male. They were all participants
ischemic changes upon stress tests, with angina pectoris (AP), in the INSIGHT calcification study and sustained a coronary
or without pain (silent ischemia) (Fig. 1). Pathologic, angio- event during a 3-yr follow-up. Twenty-nine patients had an
graphic, and intracoronary ultrasound studies reveal sharp dif- acute and 21 had a chronic event. High prevalence of coronary
ferences in the distribution of coronary atherosclerosis and artery calcification (CAC) (total calcium score [TCS] >0) was
plaque morphology between these two coronary syndromes. observed in both groups: 93% in the acute and 96% in the
Utilizing DHCT, we evaluated the coronary calcium pat- chronic group. Three-vessel calcification was found in 82%
terns in 149 patients: 47 with chronic SAP and 102 patients of the chronic as compared to only 31% in the acute group.
surviving a first AMI (3). Prevalence of coronary calcium was Median TCS was 906 in the chronic and 63 in the acute group.
81% among the AMI patients and 100% in the SAP patients. The main clinical application of these two studies (3,4) is the
The 547 calcific lesions identified in the AMI patients and the contribution of DHCT to the better understanding of the CAC
1242 lesions in the SAP patients were categorized into three characteristics and their potential as predictor of ischemic coro-
groups according to their extent: mild, intermediate, and exten- nary events. While interventional procedures are related to
sive. The age-adjusted percentages of the highest level of cal- higher amounts of CAC, many of the acute events occur against
cification, among AMI vs SAP patients, were as follows: mild a background of minor or mild calcifications. This should draw
18 vs 3%, intermediate 49 vs 18%, and extensive lesions 33 vs attention to the malignant potential of the underlying athero-
79% respectively (p < 0.01). In the AMI group, 73 culprit arter- mas presented by minor CAC. Furthermore it is essential to
ies were identified; 16 (22%) had no calcium detected, while 30 distinguish between acute and interventional events when the
(41%) showed mild lesions, 20 (27%) had intermediate forms, prognostic significance of CAC is assessed.
 CHAPTER 10 / CALCIUM DETECTION WITH DUAL-SLICE CT 93

PREDICTION OF ANGIOGRAPHIC AP had angiographically obstructive CAD, compared to 70%

OBSTRUCTIVE CAD in men and 40% in women with atypical chest pain (10). We
When compared to coronary angiography, the presence of could conclude that using DHCT for excluding coronary cal-
CAC determined by DHCT (measured as total coronary score cium yields the best results in middle-aged and elderly patients
>0) indicates the presence of obstructive CAD with a sensitiv- with chronic atypical chest pain.
ity of 91% and specificity of 52% in one study (5) and 88 and CHEST PAIN IN WOMEN
52% respectively in another study (6). The relatively low speci- Women have higher rates of false-positive ECG and thal-
ficity reflects the presence of CAC in non- or mildly protruding lium stress tests. They also show poor predictive value of
atheromas within the vessel wall. In fact, the specificity of CT anginal pains for the prediction of angiographically obstruc-
is low only when angiography is considered the gold standard. tive coronary disease. Moreover, women have a higher likeli-
The truth is that patients who have CAC detected by CT and hood of alternative mechanisms of chest pain, making the
normal angiograms have false-negative angiograms rather than diagnosis of CAD even more difficult than in men. Therefore,
false-positive CT scans. Sekiya and colleagues showed that chest pain in women still poses a diagnostic challenge. The
these patients have coronary atherosclerosis with reduced ven- value of exercise electrocardiography as well as thallium scin-
tricular function and coronary reserve (7). The accuracy of tigraphy have been questioned in women as a result of the
DHCT for the diagnosis of angiographically obstructive CAD higher frequency of false-positive results (11). The demon-
is similar to that obtained by electron beam CT (EBCT), as stration of normal perfusion by radionuclide test, however, is
indicated in a meta-analysis by Brahmajee and colleagues (8). not always predictive of the absence of CAD or future coro-
These authors pooled 1662 patients from nine studies and found nary events (12). We studied 48 symptomatic women (mean
a pooled sensitivity of 92.3% and pooled specificity of 51.2%. age 65 5 yr) referred for coronary angiography because of
CONTRIBUTION OF DHCT typical angina, or atypical chest pain with a positive exercise
or thalium scintigraphy test (13). Women with angiographi-
IN CHEST PAIN EVALUATION
cally normal coronary arteries had lower TCS than those with
Chest pain is one of the most frequent reasons for referring CAD, 5.7 11 vs 580 634, respectively. Seven women with
patients to cardiac evaluation. The diagnostic procedures in angiographically normal CAD demonstrated mild coronary
many of these patients are directed at excluding the presence of calcification (TCS <50). Of the 11 women without CAC, none
obstructive CAD. We found that for this indication, DHCT is had CAD. The absence of CAC on DHCT scans in women >60
particularly helpful as a first-line diagnostic procedure in yr is highly predictive of normal coronary arteries with 61%
equivocal clinical situations. However, in order to maximize sensitivity, 100% specificity, and 85% accuracy. DHCT pro-
the clinical benefits of this diagnostic tool and to avoid unnec- vides a rapid and reliable noninvasive tool for the diagnosis of
essary costs and irradiation, it is essential to define the proper normal coronary arteries and may spare unnecessary invasive
criteria for referring patients for coronary calcium CT. Our intervention in a great number of elderly female patients, as
experience in this field can be summarized as follows: well as significant cost savings.
CT for Chest Pain Evaluation Highly Recommended: In another study we further evaluated the contribution of
When CAD is not yet diagnosed DHCT for the detection of diseased coronary arteries in women
Chronic chest pain with anginal pain, positive exercise test, and angiographically
Atypical chest pain normal coronary arteriesSyndrome X (14). Eighty-one
Equivocal results of stress tests, thallium, or echo stress consecutive women (mean age 64 yr) referred to coronary
In patients unable to undergo stress test angiography for chest pain evaluation underwent DHCT. They
In patients older than 50 yr, particularly women were divided into three groups according to stress test and
Less Useful in Patients With: angiographic results: Group 1, the Normal, had a normal
Known CAD exercise test and angiographically normal coronary arteries;
Acute chest pain syndrome Group 2, Syndrome X, were those with abnormal exercise
Typical AP in men test and normal coronary arteries; Group 3, the CAD group,
Unequivocal positive stress test or thallium scintigraphy comprised women with at least one angiographically diseased
results vessel. The prevalence of coronary calcification among Syn-
We evaluated the prevalence and score of CAC in patients drome X patients was 63% compared to 96% in women with
categorized by chest pain type and by gender, and their value CAD and 22% in the normal group. The lowest values of TCS
in determining which patients with chronic chest pain are were obtained for the normal group. Significantly higher values
optimal candidates for coronary artery calcification scoring were found for Syndrome X group, and the highest for the CAD
with DHCT (9). We found CAC in 87% of men with typical AP group. This study demonstrates that the majority of women
compared to only 52% of those with atypical chest pain. This over 50 yr with Syndrome X exhibit calcific coronary athero-
difference was less pronounced in women54% in women sclerosis on DHCT. These changes of the coronary artery wall
with typical AP vs 44% in those with atypical chest pain. This may lead to altered endothelial-dependent vasomotion and
prevalence of CAC is in accordance with the estimated preva- reduced coronary flow reserve (15). We believe, therefore, that
lence of angiographic CAD in patients with AP and atypical women with Syndrome X having CAC at DHCT should receive
chest pain: 90% of the men and 60% of the women with typical anti-atherosclerotic and probably anti-ischemic therapy.
94 SHEMESH

In many of our patients, DHCT helped the clinician to

achieve the accurate diagnosis, as in the following examples:
Case 1A 58-yr-old asymptomatic subject. He used to
perform an ECG stress test each year as a part of an annual
check-up program. He has no other risk factors. During the last
6 yr his stress test revealed ST changes, which were interpreted
by the cardiologists who performed the tests as false positive.
The patient was referred for coronary scan, revealing extensive
calcification of the left circumflex artery (Fig. 2). A thallium
stress test was than performed, revealing ischemic perfusion
defect at the corresponding site. Coronary angiography was
recommended. A 90% obstruction of proximal left circumflex
was found, which has been successfully opened (Fig. 3A,B).
Comments: This case demonstrates the potential of DHCT
in the differentiation between false-positive ECG stress test
and true silent ischemia: The target patients are those with or
without chest pain who have a positive stress test but a normal
thallium scan, or inconclusive tests. The differential diagnosis
between false and true ischemia is easier if the clinician knows Fig. 2. Heavily calcific left circumflex (LCX) coronary artery.
the coronary calcium score of the patient. The presence of
extensive calcification, as shown in this case, favors the pres-
ence of obstructive CAD and strengthens the indication for
coronary angiography. Comments: This case demonstrates a failure of the stress
Case 2A 63-yr-old woman. She has hypercholesterolemia ECG and thallium SPECT to correctly estimate the severity of
and positive family history for CAD. The patient complained of CAD. The role of a high coronary calcium score as an indicator
retrosternal chest burning pain and discomfort for 6 yr. She has of obstructive CAD is emphasized: the DHCT results prompted
diaphragmatic hernia. The clinical manifestations raise two the cardiologist in this case to perform coronary angiography,
main possible origins for her pains: gastric and coronary. She which led to the diagnosis of multiple obstructive lesions and
was referred for stress ECG, revealing unequivocal results with resulted in a coronary bypass operation.
1.5-mm ST depression at high rate. At this point the patient was It should be remembered, however, that most of the patients
referred for coronary scanning, upon which no calcium was with obstructive plaques are symptomatic. According to our
detected (TCS = 0, Fig. 4AC). No further cardiac test was done, experience, the estimated prevalence of extensive CAC in
and no cardiac events were recorded during 5 yr of follow-up. asymptomatic subjects ages 4060 yr is approx 5%.
Comments: Accurate diagnosis of angiographically normal Case 4A 55-yr-old male. Asymptomatic, current smoker.
coronary arteries in women over 60 yr old can be achieved by He walks 10 km daily. He performs an annual stress ECG test
DHCT. This clinical application of the test is based on the well- during a routine check-up. He completed 17 min of the Bruce
accepted concept that angiographically obstructive disease is protocol up to the maximal predictive heart rate without any
very unlikely in the absence of CAC. Absence of CAC indi- limiting symptoms or ECG ischemic changes. He consented to
cates a false-positive ECG stress test and saves the patient from undergo DHCT for one of our investigational protocols. Exten-
having to undergo coronary angiography. sive calcification was found, with TCS = 1350 (Fig. 6AC).
Case 362-yr-old male, asymptomatic, without any other Consequently, he was referred for thallium SPECT, which was
risk factors. He underwent a stress test during his routine annual completely normal; further diagnostic procedures were not
check-up. (Bruce protocol, 10 min, maximal heart rate 156/ recommended. During 2 yr of follow-up, the patient remained
min, 12.9 METS). Up-sloping ST depression of 2.5 mm was asymptomatic and has not sustained a coronary event.
observed in leads II, III, aVF, V4V6 without pain. The patient Comments: This case demonstrates that extensive CAC may
was referred for thallium single photon emission CT (SPECT) be present without any clinical signs of obstructive CAD. An
by his cardiologist, revealing a suspected small, mild, revers- angiographically patent lumen may be observed despite the
ible defect at the distribution of marginal left circumflex. At presence of extensive CAC within the vessel wall. This can be
this point, the cardiologist decided not to order coronary angio- explained by the Glagov remodeling phenomenon (16).
graphy. The patient then consented to undergo DHCT for one The presence of extensive calcification (above the 90th per-
of our investigational protocols, revealing very high TCS of centile) in asymptomatic subjects requires thallium stress test.
1725 (>95th percentile) with four-vessel calcification, suggest- If thallium is normal, coronary angiography is not recom-
ing the presence of obstructive CAD (Fig. 5). Consequently, he mended.
was referred for coronary angiography, revealing severe mul- Case 5A 75-yr-old woman, hypertensive and hypercho-
tiple-vessel obstructions with main Lt obstruction of 65%, lesterolemic. She complained of chest pain for 15 yr and has
proximal LAD of 70%, and mid RCA with 85%, leading to a deep negative T waves in the precordial leads of her ECG.
coronary bypass operation. Echocardiography revealed concentric LVH. Her cardiologist
 CHAPTER 10 / CALCIUM DETECTION WITH DUAL-SLICE CT 95

Fig. 3. Heavily calcific left circumflex coronary artery (see Fig. 2) led to angiography, which revealed >90% obstruction at the corresponding
portion of the vessel (A). Successful angioplasty was performed (B).

Fig. 4. Coronary vessel without calcifications. (A) Main left coronary

(main LCA) and left anterior descending (LAD) arteries. (B) Proxi-
mal right coronary artery (RCA). (C) Distal RCA.

suspected a concomitant obstructive CAD with previous silent

non-q-wave MI. He recommended thallium SPECT, which was
nonconclusive with a small perfusion defect, interpreted as
breast attenuation artifact. At this point the cardiologist consid-
ered coronary angiography. We suggested performing a DHCT.
The scan was done and revealed no CAC (Fig. 7); further diag-
nostic tests were not recommended. The patient has now com-
pleted 6 yr of follow-up and has not sustained a coronary event.
Comments: This case emphasizes the negative predictive
value of DHCT. The absence of CAC in a 75-yr-old woman
made the diagnosis of angiographically obstructive CAD very
96 SHEMESH

Fig. 5. Three-dimensional reconstruction of the heart with extensive

coronary calcification. MLCA, main left coronary artery; LAD, left
anterior descending artery; LCX, left circumflex artery; RCA, right
coronary artery.

Fig. 6. Extensive coronary calcifications may be present without symptoms, as in this 55-yr-old male patient. (A) Left anterior descending
(LAD). (B) Left circumflex coronary artery (LCX). (C) Distal right coronary artery (RCA).

unlikely. The deep negative T waves of the woman reflects her are in patients younger than 50 yr who might have noncalcific
left-ventricular hypertrophy. CAD, and in cases of acute chest pain, which may be the con-
Coronary CT may also help the diagnosis of co-existing sequence of soft-plaque rupture. (In our studies, up to 19% of
CAD in patients with atrial fibrillation (17) or CLBBB. The patients with first AMI had no CAC on DHCT; the younger the
main limitations of the use of DHCT for chest-pain evaluation patient, the higher this percentage may be.)
 CHAPTER 10 / CALCIUM DETECTION WITH DUAL-SLICE CT 97

TRACKING THE PROGRESSION OF THE

CORONARY ATHEROSCLEROTIC PROCESS
DHCT CAN ACCURATELY TRACK
THE CALCIUM PROGRESSION
Studies of the progression of coronary atherosclerosis have
hitherto been limited to invasive angiographic techniques
among symptomatic patients. The ability to quantify coronary
atherosclerosis by a simple noninvasive technique provides
clinicians and researchers with a simple method to monitor the
effect of anti-atherosclerotic treatment. We have demonstrated
the accuracy of DHCT in tracking the progression of CC as
measured during a 3-yr follow-up (20) (Fig. 8). Ninety-four of
the 246 patients who entered the study had no calcium (TCS = 0).
Among the 152 with calcium (TCS >0), 33 had minimal
amounts of calcium (TCS 19), and 119 had a total score >9; 60
of them (50%) had baseline TCS <100 and 75% had TCS <250.
The mean TCS of the patients with calcium increased signifi-
cantly each year from 245 at baseline to 288, 322, and 349 at the
Fig. 7. Absence of calcium excludes the presence of concomitant end of the first, second, and third year of follow-up (p for trend
obstructive coronary artery disease in a 75-yr-old female with hyper- <0.01). This corresponds to a percent change from baseline
trophic cardiomyopathy. LAD, left anterior descending artery. TCS of 33, 71, and 117% respectively. This progression of
calcium was not related to age, sex, or any other risk factor
by univariant analysis. The incidence of 3 yr progression of
calcium was almost three times higher among the patients
with calcium at baseline compared to those without calcium:
70% (106/152) compared to 28% (26/94). All the 33 patients
In summary: Coronary calcium scoring can help the clini-
with baseline minimal amounts of calcium (TCS <9) had
cian to confirm the presence of CAD and to inform him about
detectable calcium at follow-up, with increasing mean TCS
its extent. Further diagnostic procedures can be used more
from 3.8 0.45 to 14.9 2.7. We found that baseline TCS is a
efficiently, reducing costs by saving unnecessary procedures.
very strong predictor of the follow-up TCS. This was found by
The test is useful to exclude obstructive CAD in patients
other researchers who used EBCT (21) and described by Yoon
older than 50 yr with chronic atypical chest pain, particularly
and colleagues (22) as calcium begets calcium.
in women, due to their higher prevalence of false-positive stress
tests. CLINICAL SIGNIFICANCE OF CAC PROGRESSION

DIAGNOSIS OF ISCHEMIC ETIOLOGY We investigated the hypothesis that a higher progression

OF DILATIVE CARDIOMYOPATHY rate of CAC is related to cardiac events, in a cohort of hyper-
tensive patients who were followed during a period of 3 yr (23).
OF UNKNOWN ORIGIN
One hundred sixty-eight consecutive hypertensive patients who
The etiology of dilative cardiomyopathy (DCMP) is impor- underwent annual CT scans over a 3-yr period were selected
tant for both therapeutic and prognostic implications, but the among the 547 participants of the INSIGHT calcification side
differential diagnosis may still be difficult after evaluation of arm study (24). Patients were divided into three groups:
the clinical data and generally requires coronary angiography.
We have demonstrated that the presence or absence of CAC on 1. Asymptomatic with prominent atherosclerotic risk factors
DHCT provides a useful tool in differentiating between (n = 116).
ischemic and nonischemic DCMP in middle-aged patients (18). 2. Patients with clinical CAD at baseline (n = 28) (Stable
CAD).
Thirty-two consecutive patients with DCMP who underwent
3. Patients who sustained a coronary event (n = 24) during
coronary angiography were scanned. Thirteen patients (mean
follow-up (Event).
age 60 8) had nonischemic DCMP with normal coronary
arteries, whereas 19 (mean age 60 6) exhibited ischemic Clinical CAD was defined as previous myocardial infarc-
DCMP with at least two obstructed vessels. CAC was found in tion, typical AP with positive stress test, or positive coronary
all patients with ischemic DCMP and was absent in 12 of the 13 angiography (>50% lumen obstruction in at least one epicar-
with nonischemic DCMP. The presence of CAC on DHCT can dial vessel). Coronary events were AMI, unstable AP, sudden
thus identify ischemic DCMP with a sensitivity of 100%, speci- death, hospitalization for increased preexisting angina, coro-
ficity of 92%, and total accuracy of 97%, providing a reliable nary angiography, coronary angioplasty, or coronary bypass
noninvasive tool for the differential diagnosis of ischemic from operation. The progression of CAC was calculated as percent
nonischemic DCMP. The results of this study were reproduced increase from baseline total coronary calcium score. A signifi-
by Budoff et al. on EBCT (19). cant annual progression of total coronary calcium score was
98 SHEMESH

Fig. 8. Progression of coronary calcium. Two spotty lesions at baseline in the proximal right coronary artery (RCA) (A), with significant
increase after 3 yr (B).

observed within each of the three study groups (p < 0.01). Both PROGRESSION UNDER INTERVENTION
patients with stable CAD and asymptomatic patients with a Using DHCT we have been able to demonstrate different
stable rate of CAC developed a coronary event. There was a progression rates of CAC over a 3-yr period, in hypertensive
higher progression rate at the end of follow-up of patients who patients treated by nifedipine once daily vs co-amilozide fol-
sustained a coronary event compared to those who did not: lowing the protocol of the INSIGHT study (24). A total of 201
180% vs 124 in asymptomatic and 118% in the group of stable patients with a TCS > 9 at the onset of the study underwent an
CAD (p < 0.05). annual DSCT for 3 yr. Coronary calcium progression was sig-
Prevalence, extent and annual progression of CAC in nificantly lower in the group treated by nifedipine as compared
patients with documented CAD are significantly higher than in to the co-amilozide group, 40% vs 78% at the third year.
asymptomatic subjects. The mean annualized percent of pro- PROGRESSION IN TRANSPLANTED HEARTS
gression in asymptomatics is 15%, and three times higher (47%) Tracking the progression of allograft atherosclerosis in heart
in patients with CAD. These results sharpen the importance of transplant recipients is currently accomplished using invasive
excluding patients with documented CAD from studies techniques. We used DHCT to track CAC progression in 24
designed to evaluate interventional effects and prognostic value consecutive heart transplant patients (25). The first scan was
of CAC in asymptomatic populations. performed 1.9 1.3 yr after transplantation. After 2 yr of fol-
 CHAPTER 10 / CALCIUM DETECTION WITH DUAL-SLICE CT 99

low-up, 4 died and the remaining 20 patients underwent a sec- nosis of coronary artery disease: a meta-analysis. Arch Intern Med
ond scan. The incidence of CAC at the first scan was 4.2% and 2001;161:833838.
9. Shemesh J, Weg N, Tenenbaum A et al. Usefulness of spiral com-
increased to 40% at the second scan. We could identify very puted tomography (dual-slice mode) for the detection of coronary
mild new spotty lesions with a mean TCS of 6.7. Our findings artery calcium in patient with chronic atypical chest pain, in typical
confirm the results of previous studies done by Barbir and col- angina pectoris, and in asymptomatic subjects with prominent ath-
leagues using EBCT on 102 heart transplant recipients. They erosclerotic risk factors. Am J Cardiol 2001;87:226228.
found calcific lesions in 45% of the patients with a similar 10. DeSanctis RW. Clinical manifestations of coronary artery disease:
chest pain in women. In: Wenger NK, Speroff L, and Packard B
minimal amount of CAC (26). (eds), Cardiovascular Health and Disease in Women. Le Jack Com-
Spiral CT can be a noninvasive alternative for the detection munications, Greenwich, CT: 1993;67.
of newly developed atherosclerotic changes in heart transplant 11. Hung J, Chaitman BR, Lam J, et al. Non-invasive diagnostic test
recipients. The delayed appearance of CAC in heart transplants choices for the evaluation of coronary artery disease in women: a
suggests that CT should be done no less than 2 yr after trans- multivariate comparison of cardiac fluoroscopy, exercise electro-
cardiography, and exercise thallium myocardial scintigraphy. J Am
plantation. Coll Cardiol 1984;4:816.
SUMMARY 12. Niemeyer MG, Van Der Wall EE, Kuyper AF, et al. Discordance of
visual and quantitative analysis regarding false negative and false
Evaluation of chest pain, diagnosis of false-positive ST positive test results in thallium-201 myocardial perfusion scintigra-
changes upon stress test, diagnosis of obstructive CAD, and phy. Am J Physiol Imaging 1991;6:3443.
evaluation of the effect of anti-atherosclerotic treatment con- 13. Shemesh J, Tenenbaum A, Fisman EZ, et al. Absence of coronary
calcification on double helical CT scans: predictor of angiographic
stitute a major part of the daily clinical practice. We have dem- normal coronary arteries in elderly women? Radiology 1996;199:
onstrated that a comprehensive use of DHCT substantially 665668.
contributes to these topics. The accordance between data 14. Shemesh J, Fisman EZ, Tenenbaum A, et al. Coronary artery calci-
obtained from DHCT and EBCT devices in many of these fication in women with syndrome X: usefulness of double helical CT
subjects indicates that for clinical practice, coronary calcium for detection. Radiology 1997;205:697700.
15. Gage JE, Hess OM, Murakami T, et al. Vasoconstriction of stenotic
detection and measurement is relevant regardless of the device
coronary arteries during dynamic exercise in patients with classic
or the technique used. We strongly believe that our and others angina pectoris: reversibility by nitroglycerin. Circulation 1986;73:
results demonstrate the usefulness of the new multislice spiral 865876.
CT devices. The differentiation between minimal, mild, mod- 16. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis GJ.
erate, and extensive CAC is universal. The numerical scale of Compensatory enlargement of human atherosclerotic coronary
arteries. New Engl J Med 1996:316:13711375
each category should be individualized to each specific device
17. Stroh CI, Shemesh J, Motro M. Using fast CT to exclude CAD in
after the most accurate and reproducible protocol has been elderly women (60 and above). N Engl J Med 1996;335 (8):595.
defined. 18. Shemesh J, Tenenbaum A, Fisman EZ, et al. Coronary calcium as a
Essential questions such as whom to refer for this test, for reliable tool for differentiating ischemic from nonischemic
which diagnostic question, how to interpret the results, and the cardiomyopathy. Am J Cardiol 1995;77:191194.
contribution of the test for better treatment of the patients should 19. Budoff MJ, Shavelle DM, Lamont DH, Kim HT, Akinwale P,
Kennedy JM, Brundage BH.Usefulness of electron beam computed
be further discussed. tomography scanning for distinguishing ischemic from nonischemic
cardiomyopathy. J Am Coll Cardiol 1998;32(5):11731178.
REFERENCES
20. Shemesh J, Apter S, Stroh CI, Itzchak Y, Motro M. Tracking coro-
1. Stary HC. The development of calcium deposits in atherosclerotic nary calcification by using dual-section spiral CT: a 3-year follow-
lesions and their persistence after lipid regression. Am J Cardiol up. Radiology 2000;217:461465.
2001;88:1619 21. Schmermund A, Baumgart D, Mhlenkamp S, et al. Natural history
2. Fuster V. Mechanisms leading to myocardial infarction: insights and topographic pattern of progression of coronary calcification in
from studies of vascular biology. Circulation 1994;90:21262146. symptomatic patients: an electron beam CT study. Arterioscler
3. Shemesh J, Stroh CI, Tenenbaum A, et al. Comparison of coronary Thromb Vasc Biol 2001;3:421426.
calcium in stable angina pectoris and in first acute myocardial 22. Yoon HC, Emerick AM, Hill JA, Gjertson DW, Goldin JG. Calcium
infarction utilizing double helical computerized tomography. Am J begets calcium: progression of coronary artery calcification in
Cardiol 1998;81:271275. asymptomatic subjects. Radiology 2002;224:236241.
4. Shemesh J, Apter S, Itzchak Y, Motro M. Coronary calcification 23. Shemesh J, Apter S, Stolero D, Itzchak Y, Motro M. Annual
compared in patients with acute versus in those with chronic coro- progression of coronary artery calcium by spiral computed tomog-
nary events using dual-sector spiral CT. Radiology 2003;226:483488. raphy in hypertensive patients without myocardial ischemia but with
5. Shemesh J, Apter S, Rozenman J, et al. Calcification of coronary prominent atherosclerotic risk facors, in patients with previous
arteries: detection and quantification with double helix CT. Radiol- angina pectoris or acute myocardial infarction which healed, and in
ogy 1995;197:779783. patients with coronary events during follow-up. Am J Cardiol 2001;
6. Broderick LS, Shemesh J, Wilensky RL, et al. Measurement of coro- 87:19351937.
nary artery calcium with double helical CT compared to coronary 24. Motro M, Shemesh J. Calcium channel blocker nifedipine slows
angiography: evaluation of CT scoring methods, interobserver down progression of coronary calcification in hypertensive patients
variation, and reproducibility. AJR Am J Roentgenol 1996;167: compared with diuretics. Hyprtension 2001;37:14101413.
439444. 25. Shemesh J, Tenenbaum A, Stroh CI, et al. Double helical CT as a
7. Sekiya M, Mukai M, Suzuke M, et al. Clinical significance of the new tool for tracking of allograft atherosclerosis in heart transplant
calcification of coronary arteries in patients with angiographically recipients. Invest Radiol 1999;32:503506.
normal coronary arteries. Angiology 1992;43:401407. 26. Barbir M, Lazem F, Bowker T et al. Determinant of transplanted
8. Brahmajee KN, Saint S, Bielak LF, Sonnad SS, Peyser PA, Rubenfire coronary calcium detected by ultrafast computed tomography scan-
M, Fendrick M. Electron-beam computed tomography in the diag- ning. Am J Cardiol 1997;79:16061609.
 CHAPTER 11 / MDCT PLAQUE DETECTION 101

11 Detection and Quantification of Calcified

Coronary Plaque With Multidetector-Row CT

J. JEFFREY CARR, MD, MSCE

INTRODUCTION consistent results, and high patient acceptance. In this chapter

Multidetector-row computed tomography (MDCT) has rap- we will review the history of MDCT for the measurement of
idly developed into a powerful tool for noninvasive measure- CAC, discuss the technical aspects and various implementa-
ment of calcified plaque in the coronary arteries over the past tions of MDCT protocols for measuring CAC, and review sci-
decade. Identification and quantification of coronary artery entific results published and in-progress with cardiac CT and
calcifications (CAC) with X-ray devices is well established in MDCT specifically.
the literature with chest radiographs, fluoroscopy, computed
EVOLUTION OF HELICAL CT
tomography (CT without electrocardiogram [ECG] gating) and
TO CARDIAC-GATED MDCT
cardiac CT (electron beam CT [EBCT], helical CT , and MDCT
with cardiac gating) (16). Calcified plaque is an established The ability to identify the heart and calcifications related to
component of coronary atherosclerosis, and radiographic tech- the coronary arteries was noted even with early-generation CT
niques are highly sensitive to calcified atherosclerotic plaque scanners, despite insufficient temporal resolution to stop car-
(1,7). The presence of calcified plaque documents the presence diac motion (12,13). The motion of the heart blurred anatomic
of subclinical atherosclerosis in the coronary artery. Calcified detail related to the coronary arteries and cardiac chambers.
plaque is an active and regulated process occurring in the vessel The development of slip-ring CT technology, which enabled
wall, with pathways similar to those of bone metabolism (8,9). spiral or helical CT scanning, dramatically improved the tem-
poral resolution of mechanical CT systems. Increased gantry
As of 2003, two consensus documents (1,10) concerning car-
speed provided improved temporal resolution and greater scan
diac CT and the recommendations of the Prevention V Confer-
coverage per unit time, which facilitated protocols incorporat-
ence (11) are available to guide clinical application. The results
ing suspended respiration (14). When CT gantries capable of
of several large epidemiological studies, as well as pharmaceu-
1 s were possible, the nongated ECG scans through the chest
tical trials, will become available during the next five years and clearly demonstrated improved cardiac and coronary morphol-
will provide new information to guide the medical community ogy, and strategies for quantifying CAC were developed
and society at large as to the appropriate utilization of CAC (2,15,16). Cardiac gating of helical CT exams using a helical
screening in the population. acquisition was first coupled with low-pitch overlapping recon-
Cardiac CT is rapidly transitioning from a research tool with struction algorithms designed to maximize temporal resolution
modest clinical application to a diagnostic test integral in our and create multilevel, multiphase images of the entire coronary
management of cardiovascular disease. There is increasing circulation. (17). Synchronized recording of the ECG tracing
evidence that cardiac CT applied to measuring CAC will be during the scan acquisition allowed the images to be aligned
effective in the risk stratification of individuals asymptomatic with the ECG tracing of cardiac activity. This retrospective
for cardiovascular disease (CVD), as will be discussed later in ECG gating was performed after the scan was acquired (i.e.,
this chapter. The rapid development of the cardiac-gated MDCT postexam processing) on a computer workstation, and allowed
techniques was made possible through the pioneering work the user or computer algorithm to select the appropriate dias-
performed with EBCT in the 1980s and 1990s. Technological tolic phase image for measuring CAC (Fig. 1). The introduc-
advances in engineering, manufacturing, and computer sci- tion of the 0.8-s gantry rotation and higher heat unit X-ray tubes
ences made possible the current-generation MDCT systems. made possible a more clinically feasible study with further
The strengths of cardiac CT and, specifically, MDCT, are improved image quality. The 0.8-s gantry rotation resulted in
detailed in the technical chapters, but are based on high spatial cardiac imaging with a temporal resolution of 520 ms per im-
resolution, volumetric coverage, high temporal resolution, age, and the high heat unit tube meant that the scans could be
rapid scan times, high patient productivity, robust protocols, obtained in clinical practice without extended wait periods for
tube cooling. This first-generation cardiac-gated helical CT
From: Contemporary Cardiology: CT of the Heart: technique allowed scanning the entire heart in a single breath-
Principles and Applications hold ranging from 30 to 50 s, depending on heart rate, which
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ determined the helical pitch (4).

101
102 CARR

Fig. 1. Retrospective gating with a single-slice helical CT system. The image is from a retrospectively gated helical scan performed in June
of 1999 (General Electric Medical Systems)CTi, with 0.8-s gantry, SmartScore calcium scoring package on the Advantage Windows
workstation). Overlapping images are reconstructed throughout each cardiac cycle, represented by the gray diamonds. An automated selection
algorithm picks the diastolic phase image at a preprogrammed percentage of the cardiac cycle (red diamond) and is displayed in the image
window below the electrocardiogram (ECG) tracing. As discussed in the text, images are obtained throughout the cardiac cycle, resulting in
significantly higher exposure than in a prospectively triggered sequential protocol. Modern multidetector-row CT systems obtain multiple
channels of data simultaneously (4, 8, 12, or 16, depending on model). However, like the single-slice helical CT system, the helical reconstruc-
tion algorithm requires data throughout the scan, and thus the X-ray tube is on continuously. ECG dose modulation is available and reduces
tube current during systole to improve the dose profile of the study. However, even with tube modulation, prospective triggering results in the
lowest dose profile while maintaining comparable image-quality factors. Calcified plaques are demonstrated in the left anterior descending
coronary artery.
 CHAPTER 11 / MDCT PLAQUE DETECTION 103

The studies comparing the single-slice scanners with ECG The basic operation of a MDCT system for sequential CAC
gating demonstrated high correlation and good agreement study with prospective ECG gating is the same whether it is 4,
between calcium scores obtained with that generation of EBCT 8, or 16 slices. The operator prescribes the start and end loca-
systems (C150) and the single-slice helical CT systems tions based on the scout image of the thorax. The MDCT scan-
(3,4,1820). This new capability for helical CT came with ner moves the CT couch so that the first block of 4, 8, or 16
limitations. First, the overlapping reconstructions resulted in images is appropriately positioned inside the gantry. The scan
a large number of imagestypically 450600 images per scan sequence is started by the breathing instructions, and the first
series. The large number of images required significant com- block of images is obtained at the prescribed offset from the R
putational resources to reconstruct the images, and this pro- wave on the ECG. In the sequential mode, the X-ray beam is on
cess would often have to be completed offline to reduce impact for only a very brief exposure at each level. For a 0.5-s gantry
on patient throughput on the CT scanner. The scans also tended system this time is slightly more than the number of views
to result in relatively long breath-holds, ranging from 35 to 50 required to reconstruct an image, and is typically in the range
s. With the single-slice helical CT, the breath-hold time was of 330360 ms, depending upon manufacturer. For a General
dependent upon the patients heart rate, which was used to set Electric LightSpeed 8 or 16 MDCT with a 0.5-s gantry, the
the speed of the CT couch as it moved the patient through the X-ray exposure time is only 2 s for the entire scan. During
scanner. The single-slice helical CT exam was dose-ineffi- the majority of the 1020-s scan time, the X-ray tube is off
cient in that the patient was irradiated for the entire cardiac because the CT couch is moving to the next location or the CT
cycle, but only a few cardiac phases at most were utilized for is waiting for the appropriate cardiac phase. Upon completion
calcium scoring. Prospective ECG triggering was developed of the first set of images, the CT couch then moves the patient
and tested on these systems, but the application was limited to the next block of scans along the z axis. This movement
by long scan times, which required multiple breath-holds to occurs during the second cardiac cycle, which then allows the
image the entire heart (3). Nonetheless, single-slice helical CT second block of slices to be obtained during the third cardiac
systems operating with effective temporal resolutions of 520 cycle of the breath hold. This sequence of acquiring and skip-
ms demonstrated the capability of mechanical CT scanners to ping a heartbeat continues until all the prescribed scan loca-
accurately and reproducibly measure CAC with comparable tions have been completed.
measurement error to EBCT. Direct comparison of helical CT As of 2003, MDCT systems use various detector arrays with
and EBCT with individuals receiving paired scans on each a length along the z axis of 20 mm. For the MDCT(4) systems,
system demonstrated no statistically significant difference in a 4 2.5-mm slice protocol uses the center 10 mm of the
the interexam or intraobserver variability when using the stan- detector. The introduction of the MDCT(8) and MDCT(16)
dard 130-Hounsfield unit (HU) threshold (21). The foundation systems allowed scans for coronary calcium using an 8 2.5-mm
developed for single-slice helical CT was critical for develop- configuration, thus utilizing the entire 20-mm width of the
ing cardiac applications, reconstruction algorithms, ECG and detector array. A comparison of MDCT systems is presented in
non-ECG gating strategies, and postprocessing for the revolu- Table 1. The direct benefits of this for coronary calcium scan-
tion in CTthe introduction of multislice scanners. ning are the reduction in reduced breath-hold time to 1020 s
as opposed to 2040 with MDCT(4). Only six CT couch trans-
MDCT SYSTEMSAPPLICATION lations and thus only 6 cardiac cyclesas opposed to 12 with
TO CORONARY CALCIUM MEASUREMENT MDCT(4)are needed to image the entire heart, which
The introduction of the four-channel MDCT systems created improves the image quality by reduced variability related to
new possibilities for cardiac imaging. The ability to acquire cardiac beat-to-beat variations, improved slice registration
four 2.5-mm slices results in covering 10 mm along the z axis along the z axis, and reduced patient dose through improved
with each heart beat. The entire heart, typically 120 mm in z dose efficiency.
axis length, could be imaged with only 12 movements of the CT One way to conceptualize prospective ECG gating with
couch. For the first time the entire heart could be covered with MDCT is to mentally construct the ideal CT device, in which
prospective ECG gating in a clinically acceptable breath-hold a detector covers the entire heart. This system might have 48
time. The initial release of the four-channel MDCT systems channels or rows (i.e., capable of 2.5-mm slices 48 levels =
had gantry rotations of 0.8 s, although this was followed by 128-mm detector) through the use of either a panel or dramati-
systems with gantry rotation speeds of 0.5, 0.42, and 0.4 s, cally expanded matrix array detector (22,23). To measure coro-
resulting in temporal resolutions ranging from 520 to 240 ms nary calcium with a similar protocol to todays CT systems, the
when scanning in the cardiac mode (i.e., partial scan recon- hypothetical MDCT(48) scanner would operate in the
struction). The MDCT scanner is operated in the sequential sequential mode and acquire 48 slices, each 2.5 mm thick,
scan mode (on some systems this is called axial or cine scan within a single heart beat. If we assume no further improvement
mode), resulting in significantly less radiation exposure than in gantry rotation speeds beyond 0.4 s and utilize the current
the helical approach utilizing retrospective ECG gating. This partial scan reconstruction algorithm, then scan acquisition and
approach is fundamentally identical to calcium screening pro- breath-hold time would be 250 ms! The entire coronary arterial
tocols implemented with EBCT systems that also utilize a step- tree would be acquired within a 250-ms phase of a single car-
and-shoot sequential scan technique. diac cycle, and no movement of the patient by the CT couch
104 CARR

Table 1
Comparison of MDCT Multislice Capability
on Scan TimeSequential Scan Mode With Prospective ECG Gating

Multislice Slice thickness Slab thickness Slabs per 120 mm Scan time
capability (mm) (mm) coverage (s)

1 2.5 2.5 48 80160

4 2.5 10 12 2040
8 2.5 20 6 1020
8 or 16 2.5 or 1.25 20
20 6 1020
48 2.5 120 1 0.24
Slab thickness roughly corresponds to beam collimation; however, the beam must be slightly wider than the
nominal detector configuration. Current 16-channel systems are limited in their z axis coverage by the 20-mm width
of the detector. These systems can typically acquire in either an 8-slice mode (8 2.5 mm) or 16-slice (16 1.25 mm)
with increased spatial resolution in the slice direction (z axis). The hypothetical 48-slice MDCT system is provided
as an example of how the increasing size of the detector along the z axis will impact future cardiac CT protocols.
Numerous technical challenges remain to be solved before this type of system or some variant is ready for clinical
applications.

would be required. Sequential would be a misnomer, since ment of the reported value. Thus a measurement can have high
only one block or set of 48 images would be obtained. By precision (alternatively, highly reproducible or refined), pro-
acquiring all levels or slices of the coronary tree in a single viding nearly the same value on multiple measurements but
heartbeat, slice misregistration related to beat-to-beat coronary biased from the true value. Likewise a device can be well cali-
motion, breath-holding, and voluntary patient motion would be brated but have significant variability, resulting in decreased
eliminated. A more likely scenario is that this enhanced imag- precision. Most importantly, the degree of accuracy and preci-
ing capability would be coupled with more sophisticated recon- sion must be appropriate for the desired task. Fortunately, car-
struction algorithms, providing either enhanced temporal diac CT can measure plaque burden with high accuracy and
resolution or increased coverage of cardiac motion (multiphase good precision in the range of values for which clinical deci-
imaging) or both to varying degrees. sions are likely to be made, based on existing outcomes data.
MDCT systems will continue to rapidly develop new capa- Specifically, EBCT and MDCT have extremely good measure-
bilities focused on cardiac imaging, based in large part on the ment characteristics for plaque burden above Agatston scores
intense global competition between CT manufacturers. Based of 50. More problematic, as will be discussed later, are the low
on the rapid developments with CT of the past decade, it is scores, where detection and image noise issues have been
foolhardy to speculate what capabilities may be available to shown to negatively impact the ability to measure calcified
cardiac imagers even in the near future from the descendents of plaques. Likewise, changes in calcium scores over short time
todays MDCT systems. It is clear that future MDCT systems intervals can easily be shown in groups of patients, with reason-
will incorporate enhanced resolution both temporal and spatial, able statistical power; however, tracking change in CAC in an
greater coverage of the heart, new methods for cardiac motion individual over a year requires significantly greater precision
compensation, dose reduction techniques, scan reconstruction and accuracy.
and postprocessing applications specifically designed for CT was developed as a medical imaging device, with less
cardiac imaging. What remains unknown is when, and what emphasis on quantification of the X-ray attenuation values
impact these advances in cardiac imaging with CT will have on (a.k.a. CT numbers or HU). The CT attenuation values extend
clinical management. from 1000 to +4000. CT number 0 is calibrated to water, 1000
to air, +1000 to dense bone, and >1000 to metal objects. Limi-
MEASURING CALCIFIED CORONARY tations in the CT scanners ability to measure small calcified
PLAQUE WITH CARDIAC CT objects were established early in the history of CT, during
In quantifying the amount of calcified plaque in the coro- attempts to quantify calcification in pulmonary nodules. Since
nary arteries, the CT system is acting as a measurement device. this effort, there has been significant improvement in the CT
Something common to all measurement techniques is the pres- system stability and quality control over the past decade (26).
ence of measurement error. Comparison of two measurements When comparing CT devices (EBCT vs EBCT, MDCT vs
will by definition incorporate the combined errors of the two MDCT, or EBCT vs MDCT) an understanding of the influ-
measurements. The problems associated with comparing a new ences of accuracy and precision are important to understand
technique to an established technique are well documented in any observed differences. Significant variability in the mea-
medical and statistical literature (24,25). Accuracy is defined surement of coronary calcium between EBCT systems at the
as how well the instrument is calibrated to the true value of the same site maintained and calibrated by the same engineers has
object being measured. Precision indicates the degree of refine- been reported with the 130 HU threshold varying between 77.1
 CHAPTER 11 / MDCT PLAQUE DETECTION 105

and 136.4 mg/cm3 of calcium hydroxyapatite (27). Likewise, STUDIES EVALUATING CALCIFIED PLAQUE
when replicate calcium determinations are performed using MEASUREMENT WITH CARDIAC CT
EBCT, significant measurement error has been documented. In the studies to date, extremely high correlations have been
Mean interscan variability using the Agatston scoring method seen between CT devices for both the amount of calcified plaque
has ranged from 18 to 43%, and there remains debate about the and the rank ordering of patients calcium burdens. In the initial
impact of a 40% vs 80% phase for image acquisition (2832). comparisons of EBCT to single-slice helical CT with ECG gat-
The nature of the calcium score and the range of values from 0 ing, the correlation between scores performed on the same indi-
to greater than 10,000 requires an understanding that the mea- viduals was near perfect (R > 0.97), and mean interscan
surement error changes based on the amount of calcified plaque; variability was 25% (3,4). In addition, the more conservative
a more detailed analysis using methods proposed by Bland to nonparametric Spearmans test of rank order was shown to be
determine limits of agreement is indicated and has been extremely high (0.96), and agreement at grouping patients based
performed by some researchers (24,33). Comparable degrees on Agatston scores of 100, 200, and 400 was 97%, 92%, and 94%
of measurement error, as will be discussed later, have been respectively. The study by Goldin et al. is the only study to date
documented for MDCT. It must be remembered that if one in which a direct comparison of interscan variability between
measurement device is labeled the gold standard, all com- EBCT and helical CT/MDCT has been performed. In this study,
parison instruments will have the error inherent in the gold participants had paired scans and calcium measurements on an
standard included and in some cases erroneously attributed as EBCT (C150) and a GE helical CT system with cardiac gating
error related to the comparison device. The measurement error (0.8-s gantry), and no statistically significant difference in
is related to the physiology of the heart and the size of the interscan or interobserver variability was found when using the
calcified plaques we are measuring with the CT systems. Spe- 130 CT threshold and Agatston scoring method. Differences in
cifically, we are asking the CT system to measure relatively risk stratification were observed between the two methods. The
small calcified plaques located on 3-mm or smaller coronary documented high correlations for both amount of calcium and
arteries using a slice collimation of 2.53 mm, depending on ordering of calcified plaque burden in these studies indicates
CT system. This results in partial volume averaging of the that differences in agreement could be improved through cali-
plaque within or between two slices, which can dramatically bration of calcium plaque measured by CT to an external stan-
alter the CT numbers of the plaque. In addition, motion blurring dard (4,27). This calibration would reduce systematic bias related
is present with EBCT and to a greater extent with MDCT. Blur- to the CT system and further improve agreement.
ring of the plaque secondary to coronary motion is most com- The National Heart, Lung and Blood Institute of the National
mon in the right coronary artery, and alters the values of the Institutes of Health is funding the Multi-Ethnic Study of Ath-
picture elements (pixels) within a given image. Lastly, detec- erosclerosis (MESA), in which 6800 individuals asymptomatic
tion of calcified plaques is determined by the presence of more for CVD have received an extensive CVD exam, including
signal (information about the plaque and surrounding struc- cardiac CT measurement of CAC. MESA exams are performed
tures) than noise. Image noise is random by definition and at six field centers across the United States (39). The partici-
obscures information or signal. The noise in the CT images pants had replicated measures of coronary calcium made dur-
creates false-positive plaques that cannot be differentiated from ing the baseline CT visit in order to measure interscan
true calcified plaques. The impact of image noise on small- variability and to improve the ability to track progression of
lesion detection has been documented with EBCT (34). A limi- calcified plaque in subsequent planned exams of the cohort. In
tation of EBCT is the fixed mAs (approx 60). With mAs fixed, MESA, three field centers utilized EBCT (C150) and three
image noise increases with thoracic size and degrades image MDCT (4-slice). Findings from the baseline exam on cardiac
quality in larger patients. The established relation between body CT interscan variability were presented at the American Heart
size and image noise will creates an erroneous relationship Association Scientific Session in 2002 (40). After adjusting for
between obesity and CAC (35,36). Since both obesity and CAC mean calcium score and body mass index, there was no signifi-
are associated with CVD, this results in confounding the true cant difference between median interscan variability using the
relationship. The ability to alter technique and maintain a con- Agatston score (EBCT 22.7 vs MDCT 24.7, p = 0.15) or volume
stant image quality (i.e., an acceptable level of image noise) is score (EBCT 11.0 vs MDCT 11.9, p = 0.84). The median
a critical benefit of MDCT techniques for CAC (37,38). There interscan percentage differences using the Agatston score were
is by definition a trade-off between improved image quality 18.1 % (EBCT) and 18.7% (MDCT), and are consistent with
(i.e., less image noise means more signal and more X-ray pho- the range of interscan variability published for EBCT of
tons) and increased dose. Thus, image quality should be selected 1843% as well as the result published with MDCT(4) demon-
based on the clinical application of the results. Current and strating interscan variability of 20.4%, 13.9%, and 9.3% for
likely future clinical recommendations indicate changes in Agatston score, volume, and mass respectively (41). The con-
management using significant levels of plaque burden corre- clusions we can draw from these data is that the measurement
sponding to Agatston scores ranging from 100 to 400. How- precision of EBCT and MDCT(4) are very similar. The compo-
ever, if a percentile approach or progression rates are to be nent of the overall observed variability attributable to the CT
utilized in clinical practice, then these issues become critical systems (EBCT vs EBCT, MDCT vs MDCT, EBCT vs MDCT)
and further research into the optimal level of image quality is is small relative to the intrinsic variability present in measuring
required. CAC with existing cardiac CT technology.
106 CARR

MDCTCARDIAC-GATED HELICAL between patient size and image noise with EBCT has been
ACQUISITIONS FOR CAC demonstrated to confound the true relation between CAC and
MDCT systems can scan in a helical mode and can recon- obesity (35). Cardiac imaging requires all aspects of image
quality (SNR, temporal resolution, contrast-to-noise, and
struct images compensated for cardiac motion using either ECG
spatial resolution) to be balanced appropriately for the desired
waveform or the inherent motion of the heart (4244). These
task. In the specific case of detecting and quantifying small
techniques are detailed in other chapters, but evolved directly
calcified plaques in the coronary arteries, the critical influence
from the single-slice helical CT retrospective gating described
of image noise was established early in the literature with EBCT
previously (Fig. 1). Although currently used and developed for
(34). The epidemic of obesity in industrialized societies and the
MDCT coronary angiography, it is clear that these advances in
elevated risk of CVD seen in individuals with the metabolic
coronary imaging can be applied to measuring calcified plaque
syndrome (which includes obesity) means that any device that
in the coronary arteries during a non-contrast-enhanced study.
will be used for screening the general population for CVD
Specifically, the improved spatial and temporal resolution pos-
must be able to handle individuals weighing greater than 100 kg
sible with these techniques has already been shown to reduce
(220 lb) without creating false-positive calcium scores. Main-
inter-exam variability from 35% with EBCT using the standard
taining a consistent level of image quality across various patient
sequential technique and 3-mm slices to 4% with volumetric
sizes will significantly improve image quality and the measure-
MDCT using a helical scan with overlapping 2.5-mm slices in
ment of CAC.
controlled testing with a motion phantom of the coronary arter-
ies (45,46). These methods hold significant potential. The IMPROVED SPATIAL RESOLUTION
increased radiation exposure and limited clinical experience
The continued technological advancement of MDCT sys-
has limited their application to date. Lower-dose methods
tems will allow imaging with increasing spatial resolution in
incorporating mA modulation and low-kV imaging are avail-
the slice or z direction. Current MDCT-16 systems can pro-
able, and initial reports have demonstrated promising results
vide 1.25-mm slices by 16 levels simultaneously with proto-
(4749). At this time, prospective ECG gating with a sequential
cols reconstructing 2.5 mm 8 levels at no increased dose or
scan acquisition is the preferred standard for clinical use, based
exposure to patient. With increasing spatial resolution, issues
on the lower radiation exposure and more extensive clinical
related to dose and SNR become increasingly important. Fun-
and research experience to date. The future application of the
damentally, improved spatial resolution will allow more
helical technique to coronary calcium imaging, although prom-
accurate measurement of small coronary plaques (Fig. 2). If
ising, remains to be fully defined and validated for clinical
the CT technique and radiation exposure is held constant, by
practice.
definition greater image noise will be present on the thinner
IMAGE NOISE AND CARDIAC CT 1.25-mm images, when compared to the 2.5-mm images
Atherosclerosis begins as early as the second decade of life, reconstructed from the same scan data. These improvements
and includes the formation of small calcified lesions (50). in measurement capability may be particularly important in
Detection of calcified plaques in the walls of the coronary measuring the change in plaque burden over time, and this is
arteries requires increasing image quality as the size of the an area of extremely active research. Rapid progression of
plaque becomes smaller. In essence, with CT (EBCT or MDCT) coronary atherosclerosis and thus potentially calcified plaque
the device is measuring those calcified plaques above the pre- may be a key predictor of future clinical events.
defined threshold and size criteria configured in the scoring
STANDARDIZED SCANNING
software. These factors are determined not by the pathobiology
PROTOCOL FOR MDCT
of atherosclerosis but by the imaging specifications of the CT
devices. Individuals with a calcium score of zero represent a The continued technological change in MDCT makes stan-
range of calcified coronary plaque not measurable by existing dardization problematic. For measuring CAC, by far the great-
CT technologies. Previous authors have suggested lowering est experience is using prospective ECG gating (also called
the threshold for CT-measurable calcified plaque from 130 to triggering) and scanning in a sequential scan mode. Gantry
90 CT units (16). This effectively increases the sensitivity of speeds of 0.5 s or greater should be used for cardiac applica-
the CT test for earlier plaque. For helical CT and MDCT sys- tions. 120 kVp with mAs ranging from 50 to 100 are consis-
tems, this is possible through the ability to maintain sufficient tently used in research and clinical applications. The tradeoff
image quality by adjusting the mA or tube current to maintain between 50 and 100 mAs is between dose (effective dose
the signal-to-noise ratio (SNR) at an acceptable level. Without ranges between 1 and 2 mSv, respectively) and image noise.
maintaining the SNR, the image quality is degraded to the point Note that increasing mAs with patient size is desirable to
that image noise obscures the small, calcified plaques. With maintain image quality. In the MESA CT protocol, mA was
EBCT, mAs is fixed. Individuals with increasing thoracic girth increased by 25% for those individuals weighing greater than
have increasing image noise. Image noise can vary on MDCT 100 kg (220 lb). Slice collimation is relatively standardized at
systems depending upon the technique employed, but typically 2.5 mm for MDCT systems when operating in the sequential
range between 8 and 16 noise/HU, compared to 24 noise/HU mode. MDCT scans should be performed at the maximum
for EBCT. Image noise can be scored as CAC resulting in false- multislice capability (i.e., widest nominal beam width) pos-
positive results and inflate the size of true lesions. This relation sible in order to reduce variability related to coronary and
 CHAPTER 11 / MDCT PLAQUE DETECTION 107

Fig. 2. Improvement in spatial resolution using 1.25-mm slice collimation. Images A and B are from the same cardiac-gated scan through the
heart performed on a multidetector-row CT(16) system (General Electric LightSpeed Pro) using 0.4-s gantry rotation and segmented recon-
struction algorithm. (A) The direct axial slice reconstructed with 2.5-mm slice thickness; (B) the same data reconstructed at 1.25-mm slice
thickness. Note how there is increased image noise apparent in the 1.25 slice by the more textured appearance of the homogenous blood in
the ascending aorta and main pulmonary artery. (C,D) Oblique reconstructions positioned to create a cross-section through the left anterior
descending coronary artery using the 2.5-mm and 1.25-mm data respectively. Note how the calcifications in the wall of the left anterior
descending artery demonstrate more detail secondary to the improved spatial resolution with the 1.25-mm dataset. These sets of images
demonstrate how improved technique will be required as slice thickness is reduced if the level of noise is to remain constant.

patient motion. The wider beam will also reduce extraneous density, which may provide important information in relation-
scatter radiation and increase dose efficiency while reducing ship to CVD risk assessment. There are several methods of
patient exposure. determining a calibrated mass, which include calibrating the
CT scanner to an external standard as well as including a cali-
CALIBRATED CALCIUM MASS bration phantom within the scan field of view for each scan
Calibration of any measurement device is central to improv- (Fig. 3). The calibrated calcium mass in milligrams has several
ing accuracy. The Agatston score is a quantitative scoring sys- advantages. Based on imaging principles, it is more robust to
tem with limitations related to the discrete weighting factors the effect of partial volume averaging than either a volume
that increase variability, and it has no straightforward means of score or Agatston score. Data with MDCT in both phantom
calibration (41,51,52). The various volume-scoring methods studies and with human research participants have demon-
allow calculation of the volume of calcified plaque in S.I. units strated less variability with a calibrated mass measurement than
such as milliliters. Volume methods are strongly influenced by either volume or Agatston methods (41,52). Lastly, the cali-
the threshold of CT measurable calcium, typically set at 130 brated mass can be reported in milligrams of calcium and will
HU. It has already been demonstrated that there is variability allow for standardization of measurements across CT systems
between patient and CT system in the determination of this and techniques as well as for currently undiscovered future
threshold (27). Plaque volume does not account for plaque techniques.
108 CARR

Fig. 3. Calibration phantom for calcium mass. One slice from a research study in which a calcium calibration phantom is used. These phantoms
were originally designed for measurement of trabecular bone mineral density with CT, also known as quantitative computed tomography
(QCT). The phantom is positioned underneath the participant during the scan and contains four cylinders with the following concentrations
of calcium hydroxyapatite: 0 mg/mL, 50 mg/mL, 100 mg/mL, and 200 mg/mL. These known quantities of calcium are then used to calibrate
the CT numbers such that the resulting mass determination is reported in mg of calcium hydoxyapatite. Calibration of each patients scan
provides the greatest gain in accuracy.

STUDIES IMPACTING THE CLINICAL individuals whose coronary calcium was measured by either
APPLICATION OF CAC EBCT or MDCT between 2000 and 2002 (39). The results from
Increasing evidence supports the use of CAC as a tool for MESA will provide data on both prevalence and the predictive
better quantifying coronary heart disease risk among asymp- ability of CAC for CVD events in men and women as well as
tomatic individuals. The Saint Francis Heart Study has recently four ethnic groups.
reported data with 4.3 yr on average of follow-up of asymptom-
CONCLUSION
atic individuals aged 5070 yr using EBCT (53). In this study,
participants with CAC scores >100 had a 10-fold increase in Accumulating evidence supports the use of the calcified
relative risk for CVD events. The calcium score had signifi- plaque as measured by cardiac CT (EBCT or MDCT) as a valu-
cantly better prognostic ability for coronary events than the able predictor of future CVD risk. Ongoing population-based
Framingham Risk Index. The National Heart, Lung and Blood research and clinical trials will provide additional information
Institutes MESA is following an asymptomatic cohort of 6800 to determine the appropriate clinical role. Calibration of car-
 CHAPTER 11 / MDCT PLAQUE DETECTION 109

diac CT measurements of coronary calcified plaque burden to 14. Kalender WA, Seissler W, Klotz E, Vock P. Spiral volumetric CT
mass of calcium will improve accuracy and reduce variability. with single-breath-hold technique, continuous transport, and con-
tinuous scanner rotation. Radiology 1990;176:181183.
As both EBCT and MDCT systems continue to improve in 15. Callaway MP, Richards P, Goddard P, Rees M. The incidence of
terms of image quality, tracking the progression of plaque bur- coronary artery calcification on standard thoracic CT scans. Brit J
den and greater understanding of calcified plaque morphology Radiol 1997;70:572574.
will provide important new information on coronary heart dis- 16. Broderick LS, Shemesh J, Wilensky RL, et al. Measurement of
ease. Based on preliminary data, it is likely that coronary cal- coronary artery calcium with dual-slice helical CT compared with
coronary angiography: evaluation of CT scoring methods, inter-
cium provides additive information to traditional CVD risk observer variations, and reproducibility. Am J Roentgenol 1996;167:
factors. The specific results of ongoing epidemiological stud- 439444.
ies will help determine the appropriate at-risk population for 17. Woodhouse CE, Janowitz WR, Viamonte M, Jr. Coronary arteries:
CVD screening with cardiac CT for coronary calcium. Cardiac retrospective cardiac gating technique to reduce cardiac motion
CT with MDCT is well positioned to make screening for sub- artifact at spiral CT. Radiology 1997;204:566569.
18. Becker CR, Knez A, Jakobs TF, et al. Detection and quantification
clinical coronary atherosclerosis a viable and cost-effective of coronary artery calcification with electron-beam and conventional
option for identifying the individual at elevated CVD risk. Once CT. Eur Radiol 1999;9:620624.
these individuals are identified, proven interventions for CVD 19. Knez A, Becker C, Becker A, et al. New generation computed
prevention are currently available and can be employed to tomography scanners are equally effective in determining coronary
reduce the tremendous burden of CVD on the population. calcium compared to electron beam CT in patients with suspected
CAD. Circulation 1998;Suppl I:I-655.
REFERENCES 20. Carr JJ, Burke GL, Goff DC, Crouse JR, DAgostino RA. Coronary
artery calcium scores correlate strongly between fast gated helical
1. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcifica- and electron beam computed tomography. Circulation 1999;99:
tion: pathophysiology, epidemiology, imaging methods, and clini- 1106.
cal implications. A statement for health professionals from the 21. Goldin JG, Yoon HC, Greaser LE, 3rd, et al. Spiral versus electron-
American Heart Association. Writing Group. Circulation 1996;94: beam CT for coronary artery calcium scoring. Radiology 2001;
11751192. 221:213221.
2. Shemesh J, Apter S, Rozenman J, et al. Calcification of coronary 22. Carr JJ. Coronary calcium: the case for helical computed tomogra-
arteries: detection and quantification with double-helix CT. Radiol- phy. J Thorac Imaging 2001;16:1624.
ogy 1995;197:779783. 23. Ning R, Chen B, Yu R, Conover D, Tang X, Ning Y. Flat panel
3. Becker CR, Jakobs TF, Aydemir S, et al. Helical and single-slice detector-based cone-beam volume CT angiography imaging: sys-
conventional CT versus electron beam CT for quantification of coro- tem evaluation. IEEE Trans Med Imaging 2000;19:949963.
nary artery calcification. AJR Am J Roentgenol 2000;174:543547. 24. Bland MJ, Altman DG. Statistical methods for assessing agreement
4. Carr JJ, Crouse JR, 3rd, Goff DC, Jr., DAgostino RB, Jr., Peterson between two methods of clinical measurement. Lancet 1986;1
NP, Burke GL. Evaluation of subsecond gated helical CT for quan- (8476):307310.
tification of coronary artery calcium and comparison with electron 25. Carr JJ, Reed JC, Choplin RH, Case LD. Pneumothorax detection:
beam CT. AJR Am J Roentgenol 2000;174:915921. a problem in experimental design. Radiology 1993;186:2325; dis-
5. Becker CR, Schoepf UJ, Reiser MF. Methods for quantification of cussion 2526.
coronary artery calcifications with electron beam and conventional 26. Zerhouni EA, Spivey JF, Morgan RH, Leo FP, Stitik FP, Siegelman
CT and pushing the spiral CT envelope: new cardiac applications. SS. Factors influencing quantitative CT measurements of solitary
Int J Cardiovasc Imaging 2001;17:203211. pulmonary nodules. J Comput Assist Tomo 1982;6:10751087.
6. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, 27. McCollough CH, Kaufmann RB, Cameron BM, Katz DJ, Sheedy
Jr., Detrano R. Quantification of coronary artery calcium using PF, 2nd, Peyser PA. Electron-beam CT: use of a calibration phantom
ultrafast computed tomography. JACC 1990;15:827832. to reduce variability in calcium quantitation. Radiology 1995;196:
7. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz 159165.
RS. Coronary artery calcium area by electron-beam computed 28. Mao S, Budoff M, Bakhsheshi H, Liu SCK. Improved reproducibil-
tomography and coronary atherosclerotic plaque area. A histopatho- ity of coronary artery calcium scoring by electron beam tomography
logic correlative study. Circulation 1995;92:21572162. with a new electrocardiographic trigger method. Invest Radiol
8. Doherty TM, Detrano RC. Coronary arterial calcification as an active 2001;36:363367.
process: a new perspective on an old idea. Calcif Tissue Int 1994;54: 29. Achenbach S, Ropers D, Mohlenkamp S, et al. Variability of repeated
224230. coronary artery calcium measurements by electron beam tomogra-
9. Wallin R, Wajih N, Greenwood GT, Sane DC. Arterial calcification: phy. Am J Cardiol 2001;87:210213, A8.
a review of mechanisms, animal modes, and the prospects for therapy. 30. Becker CR, Jakobs TF, Aydemir S, et al. Helical and single-slice
Med Res Rev 2001;21:274301. conventional CT versus electron beam CT for the quantification of
10. ORourke RA, Brundage BH, Froelicher VF, et al. American College coronary artery calcification. Am J Roentgenol 2000;174:543547.
of Cardiology/American Heart Association Expert Consensus Docu- 31. Yoon HC, Goldin JG, Greaser LE, 3rd, Sayre J, Fonarow GC.
ment on Electron-Beam Computed Tomography for the Diagnosis Interscan variation in coronary artery calcium quantification in a
and Prognosis of Coronary Artery Disease. JACC 2000:236240. large asymptomatic patient population. Am J Roentgenol 2000;174:
11. Smith SC, Jr., Amsterdam E, Balady GJ, et al. Prevention Confer- 803809.
ence V: Beyond secondary prevention: identifying the high-risk 32. Greaser LE, 3rd, Yoon HC, Mather RT, McNitt-Gray M, Goldin JG.
patient for primary prevention: tests for silent and inducible Electron-beam CT: the effect of using a correction function on coro-
ischemia: Writing Group II. Circulation 2000;101:E126. nary artery calcium quantitation. Acad Radiol 1999;6:4048.
12. Sasaki F, Koga S, Takeuchi A. [Computed tomographic detection of 33. Bielak LF, Sheedy PF, 2nd, Peyser PA. Coronary artery calcifica-
calcification within the heart and the thoracic aorta (authors transl)]. tion measured at electron-beam CT: agreement in dual scan runs and
Nippon Igaku Hoshasen Gakkai ZasshiNippon Acta Radiologica change over time. Radiology 2001;218:224229.
1982;42:123129. 34. Bielak LF, Kaufmann RB, Moll PP, McCollough CH, Schwartz RS,
13. Reinmuller R, Lipton MJ. Detection of coronary artery calcifications Sheedy PF, 2nd. Small lesions in the heart identified at electron
by computed tomography. Cardiovascular Imaging 1987;1:139145. beam CT: calcification or noise?. Radiology 1994;192:631636.
110 CARR

35. Sevrukov A, Pratap A, Doss C, Jelnin V, Hoff JA, Kondos GT. 44. Kachelriess M, Sennst DA, Maxlmoser W, Kalender WA. Kymo-
Electron beam tomography imaging of coronary calcium: the ef- gram detection and kymogram-correlated image reconstruction from
fect of body mass index on radiologic noise. J Comput Assist subsecond spiral computed tomography scans of the heart. Med Phys
Tomogr 2002;26:592597. 2002;29:14891503.
36. Sevrukov A, Jelnin V, Hoff JA, Kondos G. Electron-beam tomog- 45. Kopp AF, Ohnesorge B, Becker C, et al. Reproducibility and accu-
raphy coronary artery calcium scanning: effect of image noise on racy of coronary calcium measurements with multi-detector row
calibration phantom computed tomography values. Am J Cardiol versus electron-beam CT. Radiology 2002; 225:113119.
2001; 88:85E. 46. Hong C, Becker CR, Schoepf UJ, Ohnesorge B, Bruening R, Reiser
37. Van Hoe LR, De Meerleer KG, Leyman PP, Vanhoenacker PK. MF. Coronary artery calcium: absolute quantification in
Coronary artery calcium scoring using ECG-gated multidetector nonenhanced and contrast-enhanced multi-detector row CT studies.
CT: effect of individually optimized image-reconstruction win- Radiology 2002;223:474480.
dows on image quality and measurement reproducibility. Am J 47. Ohnesorge B, Becker C, Flohr T, et al. Coronary calcium scoring with
Roentgenol 2003;181:10931100. electrocardiographically pulsed multislice spiral computed tomo-
38. Mahnken AH, Wildberger JE, Simon J, et al. Detection of coronary graphy and reduced radiation exposure. Am J Cardiol 2001;88:82E.
calcifications: feasibility of dose reduction with a body weight- 48. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of
adapted examination protocol. Am J Roentgenol 2003; 181:533 the heart with retrospective ECG gating: reduction of radiation
538. exposure by ECG-controlled tube current modulation. Eur Radiol
39. Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of 2002;12:10811086.
atherosclerosis: objectives and design. Am J Epidemiol 2002;156: 49. Jakobs TF, Wintersperger BJ, Herzog P, et al. Ultra-low-dose coro-
871881. nary artery calcium screening using multislice CT with retrospec-
40. Detrano R, Anderson M, Nelson J, et al. Effect of scanner type and tive ECG gating. Eur Radiol 2003;13:19231930.
calcium measure on the re-scan variability of calcium quantity by 50. Strong JP. Natural history and risk factors for early human athero-
computed tomography. Circulation 2002. genesis. Pathobiological Determinants of Atherosclerosis in Youth
41. Hong C, Bae KT, Pilgram TK. Coronary artery calcium: accuracy (PDAY) Research Group. Clin Chem 1995;41:134138.
and reproducibility of measurements with multi-detector row ct 51. Callister TQ, Cooil B, Raya SP, Lippolis NJ, Russo DJ, Raggi P.
assessment of effects of different thresholds and quantification Coronary artery disease: improved reproducibility of calcium scor-
methods. Radiology 2003;227:795801. ing with an electron-beam CT volumetric method. Radiology
42. Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means 1998;208:807814.
of electrocardiographically gated multisection spiral CT: initial 52. Ferencik M, Ferullo A, Achenbach S, et al. Coronary calcium quan-
experience. Radiology 2000;217:564571. tification using various calibration phantoms and scoring thresh-
43. Flohr T, Prokop M, Becker C, et al. A retrospectively ECG-gated olds. Invest Radiol 2003;38:559566.
multislice spiral CT scan and reconstruction technique with 53. Arad Y, Roth M, Newstein D, Guerci AD. Heart Scan May Be Better
suppression of heart pulsation artifacts for cardio-thoracic imag- Than Standard Risk Factors at Estimating Heart Disease Risk,
ing with extended volume coverage. Eur Radiol 2002;12:1497 American College of Cardiology 52nd Annual Scientific Session,
1503. Chicago, Ill, 2003.
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 111

12 Coronary Calcium Scoring

With Multidetector-Row CT
Rationale and Scoring Techniques

ROMAN FISCHBACH, MD AND DAVID MAINTZ, MD

INTRODUCTION ent scan protocols have been suggested and new approaches for
Coronary heart disease is the leading cause of death, illness, the quantification of coronary artery calcifications have been
and disability in populations worldwide. Strategies to reduce introduced. Today no accepted standards exist for performing
the risk of coronary heart disease include the early initiation of MDCT studies for quantification of coronary artery calcifications.
primary preventive measures including lifestyle changes and/ This chapter will give an overview of the basic concepts of
or medical therapy in patients with subclinical disease. The imaging of coronary artery calcification and will review cur-
reliable identification of presymptomatic patients, however, is rent techniques and indications for the calcium scoring
problematic by conventional risk assessment based on tradi- examination, with special regard to MDCT technique.
tional risk factors. Direct visualization and quantification of
CORONARY HEART DISEASE
the coronary atherosclerotic plaque burden would be desir-
able to more precisely determine a patients coronary heart Cardiovascular disease has been the leading cause of death
risk. In recent years, there has been an important increase of in the United States ever since 1900, with the exception of the
interest in the use of noninvasive measurement of coronary influenza epidemic in 1918. Although remarkable advances
arterial calcification as a screening test for coronary atheroscle- have been made in prevention and treatment of coronary heart
rosis. Since coronary artery calcification is conceived as a disease, it remains the major cause of mortality and morbidity
manifestation of atherosclerosis in the arterial wall, the detec- in the industrialized nations and accounts for 54% of all cardio-
tion of coronary calcifications may serve as a marker for the vascular deaths and 22% of all deaths in the United States (1).
presence of coronary artery disease. The most frequent appli- From 1990 to 2000, the death rate from coronary heart disease
cation of coronary calcium scoring has thus become the assess- declined 25%, but the absolute number of deaths decreased
ment of an individuals future risk for a myocardial event. This only 7.6%. Coronary heart disease typically manifests in
indication and the predictive value of coronary artery calcium middle-aged and older, predominantly male, individuals. The
measurement and its role in the assessment of myocardial event average age of a person having a first myocardial infarction is
risk has always been a matter of debate. 65 yr in males and 70 yr in women. In up to 50% of coronary
Until the introduction of subsecond mechanical computed heart disease victims, sudden coronary death or nonfatal myo-
tomography (CT) scanners, coronary calcium measurement cardial infarction is the first manifestation of disease, and
approx 50% of patients with acute myocardial infarction die
remained a domain of electron beam CT (EBCT). Owing to the
within the first month of the event (2,3). The identification of
limited number of EBCT scanners available in dedicated imag-
asymptomatic persons with subclinical disease who are at high
ing centers, access to coronary CT scanning has long been
risk of developing a future coronary event and who could
restricted. On the other hand, the scan protocol and the quanti-
potentially benefit from preventive efforts thus is of major eco-
fication method remained quite uniform for a decade. Many of
nomic and clinical importance. Coronary heart disease is the
the whole-body multidetector-row CT (MDCT) systems
product and manifestation (myocardial infarction, stable and
installed today are equipped with the necessary hard- and soft-
unstable angina, and myocardial dysfunction) of coronary
ware to perform examinations of the heart using either prospec-
artery disease (CAD), in the form of coronary atherosclerosis.
tive triggering of sequential scans or retrospective gating of
Traditionally, coronary heart disease risk stratification has
spiral CT scans. This development opens CT scanning of the
been based on well-known clinical and biochemical factors.
heart for a larger patient population. Parallel to the introduction
Moderately effective preventive treatment is available: lipid
of a growing range of MDCT systems capable of performing
lowering with HMG-CoA reductase inhibitors (statins) or
electrocardiogram (ECG)-synchronized cardiac scans, differ-
antiplatelet therapy (aspirin) have both resulted in decreased
From: Contemporary Cardiology: CT of the Heart:
incidence rates of coronary events (4,5). Coronary heart dis-
Principles and Applications ease thus seems to be a suitable target for screening efforts.
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ However, many myocardial events are not accounted for by
these risk factors alone. Therefore efforts have been made
111
112 FISCHBACH AND MAINTZ

to develop new diagnostic modalities that may provide an

improved approach to coronary heart disease risk assessment.
CORONARY ARTERY CALCIFICATION
Autopsy studies have shown that coronary artery calcifica-
tion is an excellent marker of CAD (6,7). Histomorphological
evaluation of coronary arteries at autopsy has produced good
correlation between coronary artery calcification and overall
plaque burden (810), but it is not known whether the amount
of calcium reflects the amount of total plaque over time or after
therapeutic intervention. Calcification seems to be a better
marker for overall coronary plaque burden than individual
stenosis based on residual lumen size (9). Calcifications are
frequently present long before clinical manifestation of athero-
sclerotic disease. Although coronary artery calcification is
found more frequently in advanced lesions, it may be histologi-
cally identified early in the disease process and has been
detected in lesions that are seen as early as the second decade
of life. On the other hand, atherosclerotic plaque can be present
when coronary calcium is either absent or not detectable by CT,
even though histopathologic studies confirmed an intimate
relation between coronary atherosclerotic plaque area and coro-
nary calcium area.
PATHOPHYSIOLOGY OF CORONARY
ARTERY CALCIFICATION
The development of coronary atherosclerosis can be
described as a sequence of changes in the arterial wall. An Fig. 1. Sequence of atherosclerotic lesions according to the classi-
fication system by the American Heart Association and the American
in-depth description of the pathophysiology would be beyond
College of Cardiology. The early lesion is characterized by isolated
the scope of this work, and the interested reader is referred to macrophage foam cells (type I lesion) or mainly intracellular lipid accu-
the literature (1113). Initial lesions seem to occur second- mulation (type II). Type II changes and early extracellular lipid pools
ary to an injury of the coronary artery endothelium. Circulating characterize the intermediate lesion. The more advanced lesions con-
histiocytes traverse the injured vascular endothelium and gather tain a core of extracellular lipids (atheroma, type IV), or multiple lipid
in the arterial wall. These histiocytes transform into macroph- cores and fibrotic layers or calcifications (type V). A lesion with a
surface defect and associated thrombus or hematoma is the compli-
ages and can accumulate lipid. Lipid accumulation can be iden- cated type VI lesion. Diagram modified according to (11).
tified as fatty streaks, the initial and intermediate lesions in
atherosclerotic plaque formation. Histologically, different
types of atherosclerotic lesions can be distinguished, which has form in calcified lesions. Atherosclerotic calcification is an
led to a classification system suggested by the American Heart organized, regulated process (similar to bone formation) that
Association. Lesions are designated by Roman numerals I occurs only when other aspects of atherosclerosis are also
through VI (see Fig. 1). present. Nonhepatic Gla-containing proteins like osteocalcin,
Initial and intermediate lesions are types I, II, and III. The which are actively involved in the transport of calcium out of
advanced atherosclerotic lesions are subdivided into lesions vessel walls, are suspected to have key roles in the pathogen-
type IV, V, and VI. The type IV lesion (atheroma) is character- esis of coronary calcification. Osteopontin and its mRNA,
ized by extracellular intimal lipid accumulation, the lipid core. known to be involved in bone mineralization, and mRNA for
The type V lesion contains fibrous connective tissue forma- bone morphogenetic protein-2a, a potent factor for osteoblas-
tions. If parts of the lesion are calcified, it becomes a Vb lesion. tic differentiation, have been identified in calcified atheroscle-
If the lipid core is absent, the type Va or Vb lesion is classified rotic lesions. Although the process of calcium accumulation in
as type Vc. Type IV and type V lesions may develop fissures, atherosclerotic lesions is not fully understood, it seems to be
hematoma, and/or thrombus (type VI lesion). Although this an active process, which involves mechanisms similar to bone
classification system suggests a linear sequence of events, formation and resorption (1618). Regardless of the mecha-
plaque development seems to be more complex (13). This is nisms involved, it is evident that calcium is very common in
suggested by angiographic studies, which have shown that rapid advanced atherosclerotic lesions.
progression of minor lesions to high-grade stenosis may occur Histologically, coronary artery plaque can be also classified
in only a few months (14,15). as plaque rupture, plaque erosion, vulnerable plaque,
PLAQUE MORPHOLOGY AND DEGREE fibroatheroma, fibrous plaque, plaque hemorrhage, and healed
OF CALCIFICATION plaque rupture. Rupture is characterized by an acute luminal
Apatite in the form of {Ca [Ca3 (PO4)2] 3}2+ 2 OH (hydro- thrombus with connection to a plaque with a lipid-rich core.
xyapatite) and carbonate apatite is the predominant mineral Plaque erosion represents a plaque with an intact fibrous cap
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 113

Table 1 than the prevalence in women well into the postmenopausal

Prevalence of Coronary Artery Calcification period, and it is obvious that the prevalence is age dependent.
Detected by CT in Asymptomatic Men Several investigators have since reported coronary calcium
and Women (adapted from ref. 22)
scores in large asymptomatic populations (see Fig. 2). The
Asymptomatic calcium score distributions published for these asymptomatic
mainly white American populations can be used to classify
Age Men (%) Women (%) patients compared with the expected norm. All investigators
29 11 6 (2325) report a steep increase in coronary calcium score with
3039 21 11 age: the 75th percentile ranges between 36 and 44 for males aged
4049 44 23 4550 yr, between 101 and 116 for males aged 5055 yr, and
5059 72 35 between 410 and 434 for males aged 6065 yr. Interestingly,
6069 85 67 the absolute calcium scores are quite similar in the different US
7079 94 89 populations studied.
8089 100 100
Even though there is a positive association of the amount of
coronary artery calcium and age, coronary artery calcification
is not an inevitable aspect of aging. In older adults with minimal
and acute luminal thrombus. The vulnerable plaque or clinical cardiovascular disease, 16% of the individuals exam-
thinned-cap fibroatheroma shows a plaque with a thin fibrous ined had no detectable coronary artery calcifications (26).
cap and infiltration of macrophages. The pattern of calcifica-
tion was correlated with the plaque morphology in a IMAGING OF CORONARY ARTERY CALCIFICATION
histomorphologic study of patients dying suddenly with severe Calcium strongly attenuates X-rays as a result of its rela-
coronary disease (19). The greatest amount of calcium was tively high atomic number; therefore, a variety of radiological
found in healed ruptures, followed by fibroatheroma. Acute imaging techniques, including chest X-ray, fluoroscopy,
plaque ruptures were most frequently seen in segments with EBCT, conventional, spiral, and MDCT are suitable methods
speckled or no calcification, but did also occur in areas with for detecting coronary calcifications.
diffuse calcifications. From these observations it can be CONVENTIONAL X-RAY TECHNIQUES
assumed that calcification reflects healing of plaque rupture The chest film has the least sensitivity, and only extensive
and intraplaque hemorrhage as well as a response to inflamma- calcifications can be identified (27). Fluoroscopy has a much
tion in the plaque. Calcification may thus be seen as an attempt higher sensitivity compared to plain chest radiography, but
of the arterial wall to stabilize itself, since calcified and fibrotic characterization, quantification, and documentation are prob-
plaques are much stiffer than lipid-rich lesion. The earlier con- lematic using a projection technique. Furthermore, fluoroscopy
cepts suggesting that calcification is associated with plaque requires a skilled and experienced examiner. Nevertheless,
instability are not supported by recent data. fluoroscopically detected calcifications were shown to carry
In a study of 40 patients with acute coronary syndromes and significant prognostic information. A prospective study using
no or moderate angiographic coronary disease, 36% of culprit fluoroscopy in high-risk asymptomatic populations found that
lesions were not calcified (20). Because most acute ruptures individuals with detectable coronary calcifications were three
seem to occur in areas with only little calcification, it is ques- times as likely to develop angina or to experience myocardial
tionable whether coronary calcification could be a marker for infarction or death than persons without coronary calcifica-
plaque instability. Furthermore, we will have to accept that tions (28). Margolis et al. assessed the significance of coronary
calcification pattern is not helpful in localizing potentially artery calcification in 800 symptomatic patients who under-
vulnerable lesions, as it has been shown that unstable plaque went cardiac fluoroscopy. Calcification was shown by fluoros-
can be present in vessels with a wide range of calcification copy in 250 patients, of whom 236 (94%) had greater than or
patterns as well as in vascular segments without detectable equal to 75% stenosis of one or more major coronary arteries at
calcium deposits. angiography. The 5-yr survival for patients with coronary artery
Even if the individual calcification or calcified plaque does calcifications was 58%, compared with 87% in patients without
not equal a possible future culprit lesion, the presence and extent coronary artery calcifications. The prognostic significance
of coronary artery calcifications reflect the total plaque bur- of coronary artery calcification was independent of sex, status
den, and as such the likelihood of the presence of potentially of the coronary vessels at angiography, left-ventricular func-
vulnerable lesions. Using this concept, the demonstration of tion, or results of exercise tests (29).
coronary artery calcium accumulation identifies the patient CT IMAGING TECHNIQUES
with coronary artery disease, and the degree of calcification Because of the high attenuation of calcium, CT is extremely
can be considered a potential risk factor (21). sensitive in depicting vascular calcifications (Fig. 3). Conven-
PREVALENCE OF CORONARY ARTERY CALCIFICATION tional single-section imaging with acquisition times in the range
The prevalence of coronary artery calcification in asymp- of 25 s allowed identification of coronary artery calcifica-
tomatic as well as symptomatic persons is well studied. Table 1 tions, but calcifications are blurred because of cardiac motion,
shows data from an early study investigating the prevalence of and small lesions may not be detected. Despite these draw-
coronary artery calcification in asymptomatic men and women backs, conventional CT seems to be more sensitive in detecting
of various ages (22). The prevalence in men is evidently higher calcifications than fluoroscopy (30). Still, motion artifacts and
114 FISCHBACH AND MAINTZ

Fig. 2. Coronary artery calcium score in asymptomatic populations. Age- and gender-adjusted 75th percentile of coronary artery calcium in
asymptomatic male patient populations reported by (24), (23), and (25). All investigators found a remarkable increase of calcium scores with
age. The reported 75th percentile calcium score is quite similar in the different groups.

Fig. 3. Maximum-intensity projection image of a multidetector-row CT coronary scan depicting severely calcified coronary arteries.
The right coronary artery (RCA), the left anterior descending artery (LAD), and the left circumflex (lCx) show diffuse calcifications.
Small calcium accumulations are seen in the wall of the ascending (arrow) and descending aorta (arrow head).
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 115

Table 2
Results of Reproducibility Measurements Using Electron Beam CT

Calcium score Volume score

Author (reference) Patients variability (%) variability (%)

Kajinami et al., 1993 (79) 25 34

Shields et al., 1995 (80) 50 38
Wang et al., 1996 (81) 72 29
Yoon et al., 1997 (82) 50 37 28.2
Callister et al., 1998 (39) 52 19 18
Devries et al., 1995 (83) 91 49
Achenbach et al., 2001 (47) 120 19.9 16.2
Mao et al., 2001 (43) 60 24
57 15a
a Triggering of the scan at 40% RR interval.

resulting volume averaging and breathing misregistration due scan reconstruction algorithms provided a significant improve-
to different inspiration depth precluded exact calcium quanti- ment in temporal resolution. Limited volume coverage and slow
fication. scan speed were major drawbacks of single-section CT. In
Electron Beam CT comparisons of EBCT and ECG-controlled sequential CT scan-
The introduction of EBCT in the mid-1980s made quantifi- ning (33,34), the variability between the two modalities was
cation of coronary artery calcifications feasible. The first rather poor (4284%). In combination with ECG-based syn-
reports on EBCT to quantify coronary artery calcifications were chronization of image reconstruction from subsecond single-
published in the late 1980s (31,32). EBCT was shown to be section spiral CT, agreement in the classification of subjects as
much more sensitive than fluoroscopy to detect small and less healthy or diseased based on calcium score categories was
densely calcified plaque. Only 52% of calcifications detected above 90% (35).
by EBCT had been detected by fluoroscopy in the study by The now widely available MDCT scanners offer the possi-
Agatston and coworkers. bility of cardiac imaging with similar or even better image
EBCT uses an electron beam to sweep stationary tungsten quality than EBCT. This is mainly a result of advantages of
target rings to generate images. A typical scan protocol for acquisition geometry, adjustable tube current, and the possibil-
coronary calcium detection consists of contiguous 3-mm sec- ity of continuous spiral scanning. Four-slice CT systems were
tions covering the heart with 100-ms exposure time per slice. introduced in 1998. These systems achieved rotation times as
The scans are prospectively triggered from the electrocardio- fast as 500 ms, resulting in a temporal resolution of a partial
gram (ECG) of the patient, usually at 80% of the RR interval. scan of 250 ms for a prospective triggering technique. Latest
The examination is performed during one or two breath-holds. generation systems offer rotation times of down to 370 ms and
Using EBCT with ECG triggering allowed investigators for the can acquire up to 16 slices with nearly isotropic resolution.
first time to obtain cross-sectional images of the heart with high Despite the fast technological development in MDCT systems,
spatial and temporal resolution. The current is currently limited these systems still have a potential disadvantage of inferior
to 630 mAs, which results in image noise, especially in obese temporal resolution compared to a 100-ms exposure time by
patients. Signal-to-noise ratio, however, is critical for distin- EBCT. Motion artifacts, especially in the rapidly moving right
guishing small calcifications from image noise. coronary artery, become increasingly problematic with higher
EBCT-based calcium quantification has been criticized for heart rates or cardiac contractility. However, the tube current
its rather high variability in measurement results. Interscan can be adjusted to improve the signal to noise ratio consider-
variability may hamper accurate serial measurements in fol- ably, which helps to distinguish small calcifications from im-
low-up studies. The reported mean percentage variability for age noise (36).
the calcium score in repeat examinations ranges between 15% Extensive validation studies to determine the equivalence of
and 49% (see Table 2). This variability is well within the EBCT and MDCT have not been performed. Initial experiences
reported annual calcium progression rate, rendering interpreta- show good correlation of EBCT measurement and prospec-
tion of yearly follow-up examinations difficult. tively triggered MDCT. A recent study compared different
Major reasons for this variability are artifacts resulting from quantification methods and found volume and mass indexes to
either cardiac motion in scan misregistration or patient motion be superior to the traditional calcium score for comparing the
occurring during the long breath-hold times. The most impor- results of EBCT and MDCT, and for determining significant
tant factor, however, seems to be the quantification method coronary artery disease (37).
(see below) and the rather thick section thickness, which ren-
ders the measurement susceptible to partial-volume effects. QUANTIFICATION OF CORONARY
Multidetector-Row CT ARTERY CALCIFICATION
The development of subsecond rotation speeds in mechani- CT allows for reliable measurements of lesion area and
cal scanners in conjunction with the development of partial lesion attenuation. On the basis of these measurements quanti-
116 FISCHBACH AND MAINTZ

Fig. 4. Calculation of the calcium score and effect of partial volumes. This figure illustrates the calcium quantification using the calcium score.
The grid represents the pixel matrix. A vessel is depicted containing three oval calcified lesions (white). The perivascular fat (dark grey) has
negative CT numbers. The CT number for pixels representing the calcified lesions is given. Every pixel with a CT number above 130 is
registered as calcified. The measured CT number in pixels that contain calcified plaque and perivascular fat or vascular structures are an
average of the underlying tissues and are scored as calcium containing plaque (speckled pattern) only if the CT number exceeds the threshold
of 130 HU. In (A) the left lesion contains 11 pixels with CT numbers greater than 130. The greatest CT number is 289, the lesion score is
11 2. The small lesion is scored as 2 2, and the plaque on the right contains 8 registered pixels with a peak density of 299 HU, resulting
in a lesion score of 8 2. The vessel score in this case would be 22 + 4 + 16 = 42. In (B) the same vessel is depicted with a slight change in
the position of vessel regarding the pixel matrix. This change in position results in important changes in registered CT numbers. The first lesion
now scores as 11 4, the small lesion has only one registered pixel, the right lesion is counted as 11 3. The vessel score in this case would
be 44 + 1 + 33 = 78.
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 117

fication of coronary artery calcifications can be performed the number of sections per volume will further increase.
using calcium area, calcium score based on area, and plaque Because the absolute CT number of a calcium deposit will dif-
density or volume, as well as mass measurement. fer with different CT systems and image protocols, not only the
TRADITIONAL CALCIUM SCORE measured area but also the maximum CT number and thus the
The initial quantification method introduced by Agatston weighting function for the Agatston score will be affected.
et al. used electron-beam tomography for the quantification of CALCIUM VOLUME SCORE
coronary calcification and introduced a method for coronary The rather poor reproducibility of calcium quantification
artery calcification quantification, which has since been referred results with the use of the Agatston score was realized when
to as the Agatston score or calcium score. The Agatston score calcium scoring was employed to follow the progression or
in its original form is assessed from 20 contiguous 3-mm sec- possible regression of coronary atherosclerotic plaque. The
tions of the heart. A threshold of 130 Hounsfield Units (HU) calcium volume score (CVS) has been introduced to improve
is empirically set to identify calcifications. This threshold was the reproducibility of calcium measurements (38,39). The CVS
introduced to exclude noise from the evaluation, since it was for each lesion is calculated as the number of voxels in the
more than two standard deviations of the average attenuation of volume data set that belong to the calcification, multiplied by
the aorta. In its original form, only plaques related to the coro- the volume of one voxel. The determination of calcified plaque
nary arteries with a minimum lesion size of 1 mm2 or two adja- volume is independent from the section thickness or image
cent pixels are included in the calculation, to exclude random overlap used. An increase in the number of sections scored
image noise. The analysis software will highlight any lesion that when overlapping image reconstruction is chosen or when
contains the minimum number of pixels with more than 130 HU. smaller slice thickness is generated, is automatically accounted
The observer has to identify scoreable coronary artery lesions, for. Therefore, the comparability between different modalities
and assigns each lesion to the left main coronary artery, the left is improved over the traditional calcium score. The lack of a
anterior descending artery, the left circumflex, or the right coro- density-dependent weighting factor further reduces the influ-
nary artery. A region of interest is placed around each lesion and ence of partial-volume effects.
area, and maximum CT number of each lesion is determined. Since the measured volume of a calcification depends on the
The calcium score is then calculated by multiplying the indi- attenuation of the lesion as well as the threshold used, the CVS
vidual lesion area with a weighting factor. The maximum CT does not represent the true volume of the calcification (40). If
number per lesion determines this weighting factor. The factor the traditional threshold 130 HU is employed, the CVS overes-
is 1 for a peak lesion attenuation of 130199 HU, 2 for 200299 timates the volume of very dense calcifications and underesti-
HU, 3 for 300399 HU, and 4 for a lesion with a density equal mates the volume of less dense calcifications. Since the tube
to or greater than 400 HU. The sum of all lesion scores is the current used also affects the attenuation of a calcified lesion,
vessel score, and all vessel scores are summed to calculate the and attenuation influences the lesion volume, a measurement
total calcium score. which is standardized, reproducible, and independent of scan-
The use of a density-dependent weighting factor makes the ner hardware as well as image acquisition parameters is desir-
area-based Agatston score nonlinear and can mean a dramatic able, especially in view of the rapidly increasing diversity of
change in the lesion and overall score based on just a single the MDCT systems available.
pixel. Only slight changes in the position of a lesion in the CALCIUM MASS
measured section, as in repeat examinations, can induce sig- It has been suggested that the most reliable calcium quanti-
nificant changes in the number of registered pixels and the fication method would be an absolute hydroxyapatite (Ca-HA)
calcium score of a given lesion. Fig. 4 uses an example to illus- mass measurement, which can be performed similar to bone
trate how differences in pixel attenuation of the same lesion mineral density assessment (40). Ca-HA mass is proportional
from partial-volume effect can influence score results. to the mean CT number of a calcification multiplied by the
Problems Using the Calcium Score in MDCT lesion volume. Because objects that are smaller than the section
When applying the Agatston method to coronary calcium thickness are displayed with decreased mean CT numbers,
scanning based on MDCT technology, it has to be kept in mind Ca-HA mass automatically corrects for linear partial-volume
that the Agatston score was initially designed for a special pro- effects. Furthermore, Ca-HA mass has the advantage of repre-
tocol and modality (EBCT). Any changes in image acquisition senting a real physical measurement, which can be rather reli-
parameters, such as scan volume, section thickness, image ably obtained regardless of differences in CT systems and
increment, image reconstruction kernels, and X-ray spectrum, scanning protocols when using appropriate calibration. To
will influence the quantification result. obtain absolute values for the calcium mass, a calibration
For example, calcium scans with a four-channel MDCT measurement of a sufficiently large calcification with known
using a 4 2.5-mm detector configuration and prospective trig- Ca-HA density has to be performed (41,40). The specific
gering will produce 2.5-mm sections instead of 3-mm sections, calibration factor needs to be determined for all scanners and
thus giving more images per volume. Consequently, the result- protocols. Several commercial analysis software packages
ing score would be falsely high, and mathematical operations allow for mass quantification. It can be expected that the need
need to be introduced to correct for this. If MDCT calcium for reproducible quantification results will result in the replace-
scanning is performed using spiral data acquisition with over- ment of the Agatston score by mass measurements. Large-size
lapping image reconstruction, as suggested by some groups, studies supporting this assumption are still lacking.
118 FISCHBACH AND MAINTZ

Fig. 5. Retrospective electrocardiogram gating and overlapping image reconstruction yields motion-free high-resolution images, which allow
multiplanar reformation showing the heavily calcified right coronary artery (RCA). High-attenuation calcification, intermediate-attenuation
vascular and cardiac structures, and the dark perivascular fat are well differentiated.

MDCT EXAMINATION TECHNIQUE stant, a prospective estimation of the duration of the next RR
Imaging of coronary calcifications is typically performed interval can be used to reliably position the scan in the mid to
using a low-dose technique without contrast enhancement. The late diastole. In patients with change in heart rate during the
scan time should be as short as possible to avoid patient motion necessary breath-hold or with arrhythmia, motion artifacts can
and breathing artifacts. Synchronization of the image acquisi- impose a major problem for reliable calcium quantification
tion or image calculation with the cardiac motion is mandatory (42). In EBCT scanning, most groups use an 80% delay of the
to scan the heart at reproducible positions to avoid gaps or RR interval for the scan. Recently, a 40% delay has been rec-
overlaps, which would result in image misregistration. There- ommended to increase the scan reproducibility (43). For
fore, continuous ECG recording is necessary to either trigger sequential MDCT scanning, no systematic evaluations have
image acquisition in sequential scanning or to synchronize the been performed to determine the optimal timing of scan trig-
image generation in spiral scans. The best possible temporal gering. As a result of the longer image exposure in MDCT, an
resolution is required for motion-free image acquisition. earlier trigger time in the RR interval than the 80% commonly
The coronary arteries are easily recognized in their used in EBCT seems to be better. In coronary MDCT angiog-
epimyocardial course surrounded by the low-attenuation peri- raphy, initial investigations based on spiral scanning and retro-
arterial fat. Owing to the high density of vascular calcifica- spectively gated image reconstruction have shown a wide
tions, even small calcium deposits are detected with high variation of optimal reconstruction windows. A reconstruction
sensitivity (Fig. 5). window of 5060% yielded the best results in terms of motion-
The patient is positioned as in any other examination of the free images (44,45). Since motion-free images are crucial for
thorax. The scan extends form the midlevel of the left pulmo- reliable calcium quantification, a similar timing can be expected
nary artery down to the diaphragm, and is planned on a scout to yield favorable results in prospectively triggered MDCT
image. All MDCT scanners are able to cover this 10- to 15-cm scanning. In our experience, individual test scans at the mid-
range in one breath-hold. Continuous table feed as in spiral cardiac level with 50%, 60%, or 70% RR interval can be used
technique allows for somewhat faster volume coverage than to check for the optimal timing (Fig. 6). The best result can then
sequential, stepwise scanning. In order to keep the scan proto- be used for the entire cardiac scan.
cols comparable with the increasing number of scanners and SPIRAL SCANNING
possible image-acquisition parameters, standard parameters for Spiral image acquisition has several advantages over
MDCT scanning should be developed. So far no official recom- sequential scanning. The continuous data acquisition speeds up
mendations exist. A recommendation of image and acquisition the scanning process, and images can be reconstructed at any
and evaluation parameters is given in Table 3. position in the scanned volume. A slow table motion during the
SEQUENTIAL SCANNING spiral data acquisition is necessary to allow for oversampling of
In prospectively triggered sequential scanning, each scan is spiral scan projections (46). This is necessary to allow consis-
started at a predefined position in the cardiac cycle. In this tent retrospective ECG-synchronized selection of data for
approach, the next RR interval is estimated from the previous image reconstruction in a predefined phase of the cardiac cycle.
RR intervals. As long as the heart rate and rhythm remain con- As with prospectively ECG-triggered scanning, a specific
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 119

Table 3
Recommendations for Image Acquisition and Evaluation Parameters
for Coronary Calcium Quantification With Multidetector-Row CT Scanners

Data acquisition
Scan mode Spiral
Tube voltage (kV) 120
Tube current 100
ECG gating Retrospective gating in mid to late diastole, usually 5060% RR interval;
individual selection of reconstruction window
with least motion artifacts is suggested
Breath hold Inspiration
Field of view (cm) 22
Matrix 512
Scan range Entire heart from 1 cm below tracheal bifurcation to diaphragm

Image reconstruction
Section thickness (mm) 3
Reconstruction increment 1.5
Filter kernel Medium smooth

Data evaluation
Threshold (HU) 130
Minimal lesion size 1 pixel
Motion artifacts Include in region of interest
Reporting Number of calcified vessels, calcium mass and calcium
volume score. The Agatston score cannot be recommended

Fig. 6. Scan in the mid-cardiac level at 50% RR interval (A) and 60% RR interval. Motion artifacts affect the depiction of the right coronary
artery (RCA) and the left circumflex (lCx). The LCx is not well delineated from calcifications of the mitral valve, which is better appreciated
on the motion-free image obtained at 60% RR interval.

position of the reconstruction window in relation to the R peaks assured. The user can perform additional reconstruction, if
of the recorded ECG trace is chosen. Since image reconstruc- motion artifacts should occur. Thus misregistration because of
tion is performed retrospectively after the data acquisition, an changes in heart rate is minimal, and spiral scanning can
optimal agreement with the desired phase of cardiac cycle is improve the robustness of image reconstruction in arrhythmia,
120 FISCHBACH AND MAINTZ

Fig. 7. Spiral scanning with tube output modulation. The tube current is reduced to 20% of the nominal value during the systole and is increased
to 100% during mid-diastole. The image reconstruction window is placed within the normal tube output.

assuring a direct relation of the positions of the reconstructed trolled tube output modulation can achieve a mean dose
images and heart rate. This renders spiral MDCT more stable reduction of 48% for males and 45% for females (50). Consid-
in terms of image quality and probably measurement repro- ering the advantages of MDCT spiral scanning regarding
ducibility than competing sequential scanning techniques. reproducibility of calcium measurements, this slightly higher
Another important factor is overlapping image reconstruc- effective dose for spiral data acquisition should be acceptable.
tion, which is possible with any desired image increment. The
CLINICAL APPLICATIONS OF CORONARY
use of overlapping image acquisition had already resulted in
improved reproducibility of sequential coronary scans (47). CALCIUM MEASUREMENT
The same effect was shown in an initial study using spiral Reports on the potential for CT imaging of coronary calci-
MDCT and overlapping image increment (48). Overlapping fications to identify patients with coronary heart disease began
image generation decreases the influence of partial-volume ef- early, with the papers of Tannenbaum (32) and Agatston et al.
fects and increases the sensitivity for small lesions. (31). In their study, Agatston et al. demonstrated that EBCT
The obvious disadvantage of spiral scan technique with was much more sensitive than fluoroscopy in detecting coro-
highly overlapping pitch is the continuous irradiation of the nary calcifications, and that patients with symptomatic coro-
patient even in cardiac phases that are not used for image recon- nary heart disease displayed higher mean calcium scores than
struction. Thus the radiation exposure is higher than in sequen- asymptomatic subjects (Fig. 8).
tial scanning. The use of a prospective tube-current modulation CORONARY ARTERY CALCIFICATION
can compensate for this disadvantage. In phases of the car- AND ANGIOGRAPHIC STENOSIS
diac cycle that are not important for image reconstruction, the In the following years, studies were conducted to confirm
tube current can be drastically reduced, and the nominal tube these initial results and to assess the ability of CT to identify
current is used only during the desired mid-diastolic phase patients with significant coronary artery obstruction. Accura-
(see Fig. 7). cies for predicting angiographic stenoses were reported by sev-
RADIATION EXPOSURE eral investigators. Table 4 lists the reported sensitivities and
Calcium measurement is mainly applied to apparently specificities of CT coronary calcium scanning for predicting
healthy individuals. Therefore, radiation exposure has been a angiographic stenoses. There is only a moderate correlation
concern all along. The radiation exposure should be kept as low between calcification score, as demonstrated by CT, and coro-
as reasonably achievable. Since image noise is a major concern nary narrowing, as shown by catheter angiography (51). Posi-
in differentiating small calcifications from noise, a lower level tive arterial remodeling may be an explanation for the rela-
is defined. If prospective triggering is used for MDCT scan- tive lack of correlation between the degree of calcification and
ning, the effective dose is approx 1 mSv, which is similar to luminal narrowing (52,9). In patients under age 50 yr, the sen-
EBCT (49). MDCT with constant tube output during spiral sitivity of calcification detected by CT is approx 85%, with a
scanning will almost double the radiation exposure. ECG-con- specificity of about 45% as compared to coronary angiography
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 121

Fig. 8. Comparison of mean absolute calcium scores in 475 symptomatic patients without and 109 patients with a diagnosis of coronary heart
disease. The calcium score increases with patient age, and there is a significant difference between patients with and without coronary heart
disease (31).

Table 4
Relationship of Coronary Calcium and Obstructive Coronary Artery Disease in Coronary Angiography

No. of Sensitivity Specificity Positive predictive Negative predictive

Author (reference) patients Age (yr) (%) (%) value (%) value (%)

Breen et al., 1992 (84) 100 47 100 47 62 100

Fallavollita et al., 1994 (85) 106 44 85 45 66 70
Devries et al., 1995 (83) 140 58 97 41 55 94
Rumberger et al., 1995 (86) 139 51 99 62 57 97
Budoff et al., 1996 (87) 710 56 95 44 72 84
Haberl et al., 2001 (88) 1764 56 99 39 57 97

for the detection of significant coronary artery stenosis. In older gested to apply calcium scanning to patients with a low to
individuals, the sensitivity of coronary calcification in predict- intermediate likelihood of coronary heart disease presenting
ing obstructive CAD is close to 100%, while the specificity with chest pain. Even though CT scanning does not allow for
remains low owing to the high prevalence of coronary calcium assessment of presence and severity of coronary stenoses, the
deposits in elderly individuals. knowledge of the presence or absence of coronary artery calci-
The pattern of calcification may be useful in predicting fication can be helpful in making further decisions regarding
luminal narrowing. The pattern of calcification can be classified diagnostic or therapeutic measures.
as absent, speckled, fragmented, or diffuse (53). Fragmented A negative predictive value of 98% has been reported for
calcifications represent single linear or wide calcification coronary chest pain or myocardial infarction in subjects with
>2 mm, whereas diffuse calcifications represent segments of acute symptoms and nonspecific ECG (56). A negative CT
continuous calcification >5 mm. Diffuse calcification is more calcium scan was furthermore shown to carry prognostic infor-
strongly associated with luminal narrowing than speckled mation. In 192 patients observed for a mean of 50 mo, after
calcification (54,55). undergoing an EBCT when presenting to the emergency room
CORONARY ARTERY CALCIFICATION because of chest pain, subjects without coronary calcium accu-
AND DIFFERENTIAL DIAGNOSIS OF CHEST PAIN mulations had a significantly lower coronary event rate than
A positive coronary calcium scan has only a low predictive subjects with a positive CT scan (57).
value for identifying stenoses but a concomitant high nega- It has to be kept in mind, however, that even though a nega-
tive predictive value for ruling out obstructive CAD in the tive CT scan result for calcium does imply a low likelihood of
absence of detectable calcifications. It has therefore been sug- significant luminal obstruction, the presence of atherosclerotic
122 FISCHBACH AND MAINTZ

Table 5
Risk Factors for Developing Coronary Heart Disease

Causal risk factors Predisposing risk factors Conditional risk factors

Hypercholesterolemia BMI > 30 kg/m3 Triglycerides

Arterial hypertension Sedentary life style Lipoprotein (a)
Cigarette smoking Family history of premature myocardial infarction Homocystein
Diabetes mellitus Male gender Fibrinogen
Metabolic syndrome Plasminogen acitvator inhibitor
Social factors C-reactive protein

plaque cannot be excluded. In this context, it has to be stressed events can potentially be prevented using lipid-lowering
that culprit plaque in sudden coronary death may contain only therapy as primary prevention in individuals with average cho-
a little calcium and consequently is not reliably identified by CT. lesterol levels (62). Therefore, risk-assessment tools or screen-
CORONARY ARTERY CALCIFICATION AND DISEASE ing tests are required that will reliably identify asymptomatic
PROGRESSION individuals at high risk of hard coronary events to efficiently
Coronary artery calcification constitutes a surrogate of total target therapy.
plaque burden, and plaque burden is closely related to the risk The risk of developing CAD depends on a wide range of
of future myocardial events. The possibility to observe changes environmental and biochemical factors, many of which have
in coronary atherosclerotic involvement noninvasively is an been identified in prospective epidemiological studies. Among
appealing concept for assessment of disease progression and these well-recognized risk factors are tobacco smoking, high
efficacy of primary or secondary preventive measures in CAD. low-density lipoprotein cholesterol levels, low high-density
Several studies have tracked changes of coronary calcium lipoprotein cholesterol, diabetes mellitus, arterial hypertension,
by EBCT. Score progression seems to be accelerated in patients and family history of premature myocardial infarction (Table 5).
with obstructive CAD compared with patients who have no These and other risk factors interact in a complex way, making
clinically manifest disease (27% vs 18%) (58). A mean annual risk assessment in the individual patient complicated. How-
rate of calcium score increase for untreated patients between ever, algorithms derived from large prospective epidemiologi-
24% and 36% has been found. Recent studies have shown the cal studies like the Framingham Study (63,64) in the United
ability of calcium quantification to monitor the progression of States and the Prospective Cardiovascular Mnster (PROCAM)
coronary calcification and to document the effect of risk-factor Study (65) in Germany can be used to calculate a persons risk
modification and medical treatment (38,59,60,61). In 66 of CAD. A person is said to be at increased risk if his or her
patients with coronary calcifications, the observed increase in absolute risk of suffering a future myocardial event within the
coronary calcium volume score was 25% without treatment next 10 yr exceeds 20%. Calculation of an individuals CAD
and decreased to 8.8% under treatment with statins (59). It has risk is possible either by using scoring systems and risk charts
not been shown whether a decrease in calcium progression also or computer-assisted algorithms.
represents a decrease in future event risks. Because multiple Subpopulations at a significantly increased global risk of
trials on lowering cholesterol levels have shown a decrease in CAD can be identified with a high level of exactness using
mortality, it is plausible that decrease in calcium score progres- scores or algorithms for risk assessment. Epidemiological stud-
sion may be a valuable tool in monitoring and comparing such ies have shown that in many individuals without clinically
therapies. apparent symptoms, the risk of developing a future myocardial
If progression is to be assessed in an individual person, a infarction may equal or even exceed that of persons with a
high reliability of the calcium quantification is mandatory. history of coronary heart disease (65). However, it remains
Reported variability from many EBCT studies ranges between questionable whether the presence of subclinical manifesta-
14% and 38%, however, precluding a meaningful interpreta- tions of atherosclerosis in these presymptomatic patients can
tion of calcium measurements in the individual patient. MDCT be diagnosed with sufficient certainty based on traditional risk
with spiral scanning, overlapping image reconstruction, and factors alone.
use of calcium mass measurement has resulted in a variability Even if standard or traditional risk-factor analysis is quite
of 5% in a first study (48). This study will need confirmation, accurate in identifying populations at risk, exercise stress test-
but a significant improvement for follow-up or therapy control ing, which is also commonly used to screen for CAD, is char-
studies may be expected. acterized by a rather low accuracy in predicting events. The
CORONARY ARTERY CALCIFICATION sensitivity of stress testing for identification of asymptomatic
AND SCREENING FOR CAD middle-aged males at risk for a future myocardial event is less
Conventional Identification of Subjects at Risk of CAD than 40% (66). This is not surprising, since more than two-
According to the Air Force/Texas Coronary Atherosclerosis thirds of acute coronary events in previously asymptomatic
Prevention Study (AFCAPS/TexCAPS) only 37% of acute individuals are due to rupture or erosion of nonobstructive
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 123

Table 6
Coronary Calcium Thresholds and Test Performance for Identifying Subjects
With Future Myocardial Infarction or Coronary Death (adapted from Guerci and coworkers [78])

Positive Negative Overall Odds ratio

Score Sensitivity Specificity predictive value predictive value accuracy (95% CI)

>80 0.89 0.74 0.05 0.99 0.74 22.3 (5.197.4)

>160 0.83 0.82 0.07 0.99 0.82 22.2 (6.477.4)
>600 0.55 0.94 0.13 0.98 0.94 20.3 (7.853.1)
CI, confidence interval.

coronary plaque (67,13). Nonobstructive CAD seldom causes scores greater than 160 vs event rates among subjects with
stress-induced myocardial ischemia, and thus perfusion imag- calcium scores less than 160 (see Table 6). The subjects were
ing does not improve the detection rate of subclinical CAD. followed for 3.6 yr, and 18 myocardial infarctions and coronary
Coronary Artery Calcification and Predicting Events deaths were registered. The positive predictive value increased
The use of noninvasive measurement of coronary artery as a function of the calcium score, indicating that calcium quan-
calcification as a screening test for coronary atherosclerosis tification may be a valid test for vulnerable atherosclerotic
has received remarkable interest in recent years and has gener- plaque.
ated much controversy. Coronary calcium detection by CT Raggi et al. gave a relative risk of 21 for myocardial infarc-
makes the assumption that direct demonstration of atheroscle- tion or cardiovascular death in subjects with a calcium score
rotic vessel-wall involvement in asymptomatic populations greater than the 75th percentile when compared to individuals
with increased coronary heart disease risk is helpful to identify with a calcium score lower than the 25th percentile (24). While
as well as to stratify individuals at risk. the absolute calcium score had a wide variation in the asymp-
To evaluate a possible benefit from CT calcium screening, tomatic individuals studied, a calcium score in the upper age-
we first should realize that the main purpose of CAD screening and gender-adjusted quartile seemed to identify subjects at
is classification of asymptomatic persons as likely or not likely risk better (see Table 7) than an increased absolute calcium
to have CAD. Early diagnosis of cases with subclinical CAD score. Twenty-seven events were registered in the 632 sub-
ought to reduce morbidity and mortality from the specific dis- jects during a mean observation of 32 mo. Approximately two-
ease among the population screened, because screening leads thirds of the observed events occurred in patients with mild to
to a course of action proven to save lives. If coronary CT screen- moderate amounts of coronary artery calcium (calcium scores
ing cannot achieve a significant reduction of morbidity and below 400). Subjects with massive calcifications (calcium
mortality, it cannot be deemed effective. Therefore, we need to scores >400) represented only 7% of the population scanned
prove that calcium screening of the entire adult population, or and accounted for 22% of all events. On the other hand, 70%
even a subpopulation, leads to effective preventive therapy with of the events observed (19 of 27) occurred in 181 subjects
a substantial benefit, before we can recommend this diagnostic with a calcium score above the 75th percentile adjusted for sex
test to healthy individuals and accept the involved expenditure and age.
of health care resources. Furthermore, calcium screening In a study of 926 subjects, who were either referred because
should identify high-risk individuals better then traditional risk- of risk factors or self-referred, the highest calcium score quartile
factor analysis. was associated with a relative risk of cardiovascular events of
In a comparison of persons dying of coronary heart disease 9 compared to subjects without coronary calcification (71). A
and age- and sex-matched individuals dying of other causes, limitation of this investigation is the high number of revascu-
three times as much calcium was found in the coronary heart larizations (23 of 28 reported events) during a mean follow-up
disease group (68). Coronary calcium was even nine times as of 3.3 yr after scanning.
abundant in young persons under 50 yr of age dying suddenly, In the South Bay Heart Watch Study, 29 myocardial
when compared to matched controls dying from accidents (10). infarctions and 17 coronary deaths were reported during 44 mo
Four prospective studies have addressed the prognostic of follow-up in 1196 asymptomatic high-risk subjects (70).
information of coronary calcium scanning in asymptomatic Calcium detection by CT failed to improve the identification of
individuals for predicting coronary events (69,70,24,71). patients at risk compared to traditional risk-factor assessment
Results of these studies are not conclusive, since most of the in this study, which was carried out in a population originally
studies report on a limited number of hard coronary events, the screened by fluoroscopy.
majority of endpoints being revascularizations (possibly trig- OMalley et al. performed a meta-analysis regarding the
gered by the scan results). Furthermore, some of the studies literature on using CT calcium to predict future events in
enrolled self-referred subjects or subjects referred for scanning asymptomatic adults (72). Their results show that the relative
because of cardiovascular risk factors. risk of calcification for a myocardial infarct or CAD-related
Arad et al. reported an odds ratio of 22 for the prediction of death varies from 1 to 22, with a weighted mean of 4.2. Thus,
myocardial infarction or death in 1173 self-referred men and there is a strong indication that coronary calcium scanning may
women, when comparing event rates in the group with calcium predict myocardial infarction in asymptomatic populations.
124 FISCHBACH AND MAINTZ

Table 7
Coronary Calcium Scores and Risk
of Myocardial Events (according to Raggi et al. [24])

In 632 asymptomatic persons, 8 fatal and 19 nonfatal myocardial infarcts were

observed. The use of adjusted calcium quartiles discriminates better between
risk statuses than the use of an absolute calcium score.

Absolute calcium scores

0 199 100400 >400

Patients 292 219 74 47
Annual event rate 0.11 2.1 4.1 4.8

Age- and gender-adjusted calcium score quartile

1st 2nd 3rd 4th

Patients 351 351 100 181
Annual event rate 0.2 0.2 1.4 4.5

The question remains, whether calcium scanning performs high risk for a future coronary event. While cardiovascular risk
better than conventional risk-factor analysis. factors have been shown to increase CAD risk 1.55 times
On the basis of the investigations in asymptomatic persons, (77,64), odds ratios for increased coronary calcium detected by
some would argue that the demonstration of coronary artery CT of 22 are reported. This indicates a strong association
calcium, which means the presence of subclinical CAD, should between coronary calcification and CAD risk, even though
shift an asymptomatic patient from a primary to a secondary formal proof in prospective population-based studies is still
prevention category (73). Even though this seems logical, it has lacking.
not been proven that a positive or increased calcium score is Because atherogenesis is a dynamic process, which repre-
equivalent to a prior myocardial infarction. Even if an increased sents the result of a life-long exposure of an individual to a
calcium score puts a person at increased risk, as several inves- variety of predisposing factors, manifestations of subclinical
tigations imply, one cannot be sure that any benefit would be coronary artery disease, as indicated by a positive CT calcium
reached by therapeutic intervention. It has been suggested that scan, may point to individuals with an increased future cardiac
the coronary plaque burden should provide a better indicator of risk. In this respect, coronary calcium may outperform clinical
the probability of developing an acute coronary syndrome than risk-factor analysis in asymptomatic subjects, since the simple
a persons age and could replace age as a risk factor in the presence of even a combination of risk factors does not mean
Framingham risk scores to improve risk assessment (74). presence of subclinical CAD. Coronary plaque burden has been
Ongoing Trials shown to be a good predictor for future coronary events in
Three prospective epidemiological trials under way will angiographic follow-up studies. Since coronary calcium is a
examine this issue. One part of the Prospective Army Coronary reliable marker for CAD and since the amount of coronary
Calcium (PACC) study (75) is a prospective cohort study of calcium reflects total coronary plaque burden, demonstration
2000 participants followed for at least 5 yr to establish the of coronary artery calcification may thus be of value in improv-
relation between coronary calcification and cardiovascular ing risk stratification of asymptomatic populations with mod-
events in an unselected, low-risk Army population. The erate to increased global coronary heart risk, or in assessing
Multi-Ethnic Study of Atherosclerosis (MESA) will use a disease progression.
cohort of 6500 American adults who will undergo CT scan- There is no formal proof that coronary heart disease risk can
ning and will be followed for coronary events for 7 yr. be reduced in patients with elevated coronary calcium scores.
Another population-based, prospective cohort study was Neither is there a consensus in the professional societies on
begun in Germany in 2000. The RECALL (Risk Factors, Evalu- how to use the results of calcium scoring for risk-factor man-
ation of Coronary Calcium and Lifestyle) study is designed to agement or therapeutic decisions, or who should undergo CT
define the relative risk associated with EBCT-derived coronary scanning at all. In groups with intermediate or increased risk,
calcium score for myocardial infarction and cardiac death in calcium scanning may be of value, but long-term prospective
4200 males and females aged 45 to 75 yr in an unselected urban studies will have to decide this issue.
population from the large, heavily industrialized Ruhr area (76). MDCT scanning, especially using spiral technique with
CONCLUSION overlapping slice reconstruction and calcium quantification
Risk-factor assessment and noninvasive imaging of athero- based on calibrated system and calcium mass measurement,
sclerosis by CT calcium scanning aim at selecting subjects at a holds promise to develop as the reference in the future.
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 125

REFERENCES 20. Schmermund A, Baumgart D, Adamzik M, et al. Comparison of

electron-beam computed tomography and intracoronary ultrasound
1. American Heart Association. Heart Disease and Stroke Statistics in detecting calcified and noncalcified plaques in patients with acute
2003 Update. 2002. Dallas, TX: American Heart Association. coronary syndromes and no or minimal to moderate angiographic
2. Chambless L, Keil U, Dobson A, et al. Population versus clinical coronary artery disease. Am J Cardiol 1998;81:141146.
view of case fatality from acute coronary heart disease: results from 21. Schmermund A, Erbel R. Unstable coronary plaque and its relation
the WHO MONICA Project 19851990. Multinational MONItoring
to coronary calcium. Circulation 2001;104:16821687.
of Trends and Determinants in CArdiovascular Disease. Circulation
22. Janowitz WR, Agatston AS, Kaplan G, et al. Differences in preva-
1997;96:38493859.
lence and extent of coronary artery calcium detected by ultrafast
3. Thaulow E, Erikssen J, Sandvik L, et al. Initial clinical presentation
computed tomography in asymptomatic men and women. Am J
of cardiac disease in asymptomatic men with silent myocardial ische-
Cardiol 1993;72:247254.
mia and angiographically documented coronary artery disease (the
23. Hoff JA, Chomka EV, Krainik AJ, et al. Age and gender distribu-
Oslo Ischemia Study). Am J Cardiol 1993;72:629633.
tions of coronary artery calcium detected by electron beam tomog-
4. Antiplatelet Trialists Collaboration. Collaborative overview of ran-
raphy in 35,246 adults. Am J Cardiol 2001;87:13351339.
domized trials of antiplatelet therapy. I: Prevention of death, myo-
24. Raggi P, Callister TQ, Cooil B, et al. Identification of patients at
cardial infarction, and stroke by prolonged antiplatelet therapy in
increased risk of first unheralded acute myocardial infarction by elec-
various categories of patients. Br Med J 1994;308:81106.
tron-beam computed tomography. Circulation 2000;101:850855.
5. Shepherd J, Cobbe S, Ford I, et al. West of Scotland Coronary Pre-
25. Wong ND, Budoff MJ, Pio J, et al. Coronary calcium and cardiovas-
vention Study Group. Prevention of coronary heart diesease with
cular event risk: evaluation by age- and sex-specific quartiles. Am
pravastatin in men with hypercholesterolemia. N Engl J Med 1995;
Heart J 2002;143:456459.
333:13011307.
26. Newman A, Naydeck B, Sutton-Tyrell K, et al. Coronary artery
6. McCarthy J, Palmer F. Incidence and significance of coronary artery
calcification. Br Heart J 1974;36:499506. calcification in older adults with minimal clinical or subclinical
7. Rifkin R, Parisi A, Folland E. Coronary calcification in the diagnosis cardiovascular disease. J Am Geriatr Soc 2000;48:256263.
of coronary artery disease. Am J Cardiol 1979;44:141147 27. Souza A, Bream P, Elliott L. Chest film detection of coronary artery
8. Rumberger JA, Simons DB, Fitzpatrick LA, et al. Coronary artery calcification: the value of CAC triangle. Radiology 1978;129:710.
calcium area by electron-beam computed tomography and coronary 28. Detrano R, Wong N, Tang W. Prognostic significance of cardiac
atherosclerotic plaque area. A histopathologic correlative study. cinefluoroscopy for coronary calcifc deposits in asymptomatic high
Circulation 1995;92:21572162 risk subjects. J Am Coll Cardiol 1994;24:354358.
9. Sangiori G, Rumberger J, Severson A, et al. Arterial calcification 29. Margolis JR, Chen JT, Kong Y, et al. The diagnostic and prognostic
and not lumen stenosis is correlated with atherosclerotic plaque significance of coronary artery calcification. A report of 800 cases.
burden in humans: a histologic study of 723 coronary artery seg- Radiology 1980;137:609616.
ments using nondecalcifying methodology. J Am Coll Cardiol 1998; 30. Rienmller R, Lipton M. Detection of coronary artery calcification
31:126133 by computed tomography. Dynam Cardiovasc Imag 1987;1:139145.
10. Schmermund A, Schwartz RS, Adamzik M, et al. Coronary athero- 31. Agatston A, Janowitz W, Hildner F, et al. Quantification of coronary
sclerosis in unheralded sudden coronary death under age 50: histo- artery calcium using ultrafast computed tomography. J Am Coll
pathologic comparison with healthy subjects dying out of hospital. Cardiol 1990;15:827832.
Atherosclerosis 2001;155:499508. 32. Tannenbaum S, Kondos G, Veselik K, et al. Detection of calcific
11. Stary HC, Chandler AB, Dinsmore RE, et al. A definition of deposits in coronary arteries by ultrafast computed tomography.
advanced types of atherosclerotic lesions and a histological classi- J Am Coll Cardiol 1989;15:827832.
fication of atherosclerosis. A report from the Committee on Vascu- 33. Becker CR, Knez A, Jakobs TF, et al. Detection and quantification
lar Lesions of the Council on Arteriosclerosis, American Heart of coronary artery calcification with electron-beam and conventional
Association. Circulation 1995;92:13551374. CT. Eur Radiol 1999;9:620624.
12. Stary HC, Chandler AB, Glagov S, et al. A definition of initial, fatty 34. Budoff MJ, Mao S, Zalace CP, et al. Comparison of spiral and elec-
streak, and intermediate lesions of atherosclerosis. A report from the tron beam tomography in the evaluation of coronary calcification in
Committee on Vascular Lesions of the Council on Arteriosclerosis, asymptomatic persons. Int J Cardiol 2001;77:181188.
American Heart Association. Arterioscler Thromb 1994;14:840856. 35. Carr JJ, Crouse JR, 3rd, Goff DC, Jr., et al. Evaluation of subsecond
13. Virmani R, Kolodgie FD, Burke AP, et al. Lessons from sudden gated helical CT for quantification of coronary artery calcium and
coronary death: a comprehensive morphological classification comparison with electron beam CT. AJR Am J Roentgenol 2000;
scheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol 174:915921.
2000;20:12621275. 36. Bielak LF, Kaufmann RB, Moll PP, et al. Small lesions in the heart
14. Ambrose JA, Tannenbaum MA, Alexopoulos D, et al. Angiographic identified at electron beam CT: calcification or noise? Radiology
progression of coronary artery disease and the development of myo- 1994;192:631636.
cardial infarction. J Am Coll Cardiol 1988;12:5662. 37. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium
15. Mancini GB, Pitt B. Coronary angiographic changes in patients with measurement: agreement of multirow detector and electron beam
cardiac events in the Prospective Randomized Evaluation of the CT. Am J Roentgenol 2001;176:12951298.
Vascular Effects of Norvasc Trial (PREVENT). Am J Cardiol 2002; 38. Callister T, Raggi P, Lippolis N, et al. Effect of HMG-CoA reduc-
90:776778. tase inhibitors on coronary disease as assessed by electron-beam
16. Doherty T, Detrano R. Coronary arterial calcification as an active computed tomography. N Engl J Med 1998;339:19721978.
process: a new perspective on an old problem. Calcif. Tissue. Int. 39. Callister TQ, Cooil B, Raya SP, et al. Coronary artery disease:
1994;54:224230. improved reproducibility of calcium scoring with an electron-beam
17. Jeziorska M, McCollum C, Wooley D. Observations on bone forma- CT volumetric method. Radiology 1998;208:807814.
tion and remodelling in advanced atherosclerotic lesions of human 40. Ulzheimer S, Kalender WA. Assessment of calcium scoring perfor-
carotid arteries. Virchows Arch 1998;433:559565. mance in cardiac computed tomography. Eur Radiol 2003;13:484497.
18. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: 41. Hong C, Becker CR, Schoepf UJ, et al. Coronary artery calcium:
pathophysiology, epidemiology, imaging methods, and clinical absolute quantification in nonenhanced and contrast-enhanced
implications. A statement for health professionals from the Ameri- multi-detector row CT studies. Radiology 2002;223:474480.
can Heart Association. Circulation 1996;94:11751192. 42. Mao SS, Oudiz RJ, Bakhsheshi H, et al. Variation of heart rate and
19. Burke AP, Weber D, Kolodgie F, et al. Pathophysiology of calcium electrocardiograph trigger interval during ultrafast computed tomog-
deposition in coronary arteries. Herz 2001;26:239244. raphy. Am J Card Imaging 1996;10:239243.
126 FISCHBACH AND MAINTZ

43. Mao S, Bakhsheshi H, Lu B, et al. Effect of electrocardiogram trig- 64. Wilson PW, DAgostino RB, Levy D, et al. Prediction of coronary
gering on reproducibility of coronary artery calcium scoring. Radi- heart disease using risk factor categories. Circulation 1998;97:
ology 2001;220:707711. 18371847.
44. Hong C, Becker CR, Huber A, et al. ECG-gated reconstructed multi- 65. Assmann G, Cullen P, Schulte H. Simple scoring scheme for cal-
detector row CT coronary angiography: effect of varying trigger culating the risk of acute coronary events based on the 10-year
delay on image quality. Radiology 2001;220:712717. follow-up of the prospective cardiovascular Munster (PROCAM)
45. Kopp AF, Schroeder S, Kuettner A, et al. Coronary arteries: retro- study. Circulation 2002;105:310315.
spectively ECG-gated multi-detector row CT angiography with 66. Detrano RC, Froelicher VF. Exercise testing: uses and limitations
selective optimization of the image reconstruction window. Radiol- considering recent studies. Prog Cardiovasc Dis 1988;31:837845.
ogy 2001;221:683688. 67. Petursson M, Jonmundsson E, Brekkan A, et al. Angiographic
46. Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of predictors of new coronary occlusions. Am Heart J 1995;129:
electrocardiographically gated multisection spiral CT: initial expe- 515520.
rience. Radiology 2000;217:564571 68. Eggen DA, Strong JP, McGill HC. Coronary calcification: Relation-
47. Achenbach S, Meissner F, Ropers D, et al. Overlapping cross-sec- ship to clinically significant coronary lesions and race, sex and
tions significantly improve the reproducibility of coronary calcium topographic distribution. Circulation 1965;32:948955.
measurements by electron beam tomography: a phantom study. 69. Arad Y, Spadaro LA, Goodman K, et al. Prediction of coronary
J Comp Assist Tomogr 2001;25:569573. events with electron beam computed tomography. J Am Coll Cardiol
48. Ohnesorge B, Flohr T, Fischbach R, et al. Reproducibility of coro- 2000;36:12531260.
nary calcium quantification in repeat examinations with retrospec- 70. Detrano RC, Wong ND, Doherty TM, et al. Coronary calcium does
tively ECG-gated multisection spiral CT. Eur Radiol 2002;12: not accurately predict near-term future coronary events in high-risk
15321540. adults. Circulation 1999;99:26332638.
49. McCollough CH. Patient dose in cardiac computed tomography. 71. Wong ND, Hsu JC, Detrano RC, et al. Coronary artery calcium
evaluation by electron beam computed tomography and its relation
Herz 2003;28:16.
to new cardiovascular events. Am J Cardiol 2000;86:495498.
50. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of
72. OMalley PG, Taylor AJ, Jackson JL, et al. Prognostic value of
the heart with retrospective ECG gating: reduction of radiation
coronary electron-beam computed tomography for coronary heart
exposure by ECG-controlled tube current modulation. Eur Radiol
disease events in asymptomatic populations. Am J Cardiol 2000;
2002;12:10811086.
85:945948.
51. Detrano R, Hsiai T, Wang S, et al. Prognostic value of coronary 73. Hecht HS. Impact of plaque imaging by electron beam tomography
calcification and angiographic stenoses in patients undergoing coro- on the treatment of dyslipidemias. Am J Cardiol 2001;88:406408.
nary angiography. J Am Coll Cardiol 1996;27:285290. 74. Grundy SM. Coronary plaque as a replacement for age as a risk
52. Glagov S, Weisenberg E, Zarins CK, et al. Compensatory enlarge- factor in global risk assessment. Am J Cardiol 2001;88:8E11E.
ment of human atherosclerotic coronary arteries. N Engl J Med 1987; 75. OMalley PG, Taylor AJ, Gibbons RV, et al. Rationale and design
316:13711375. of the Prospective Army Coronary Calcium (PACC) Study: utility of
53. Friedrich G, Moes N, Muhlberger V, et al. Detection of intralesional electron beam computed tomography as a screening test for coro-
calcium by intracoronary ultrasound depends on the histologic pat- nary artery disease and as an intervention for risk factor modifica-
tern. Am Heart J 1994;128:435441. tion among young, asymptomatic, active-duty United States Army
54. Kajinami K, Seki H, Takekoshi N, et al. Noninvasive prediction of personnel. Am Heart J 1999;137:932941.
coronary atherosclerosis by quantification of coronary artery calci- 76. Schmermund A, Mhlenkamp S, Stang A, et al. Assessment of clini-
fication using electron comuted tomography: comparison with elec- cally silent atherosclerotic disease and established and novel risk
trocardiographic and thallium exercise stress results. J Am Coll factors for predicting myocardial infarction and cardiac death in
Cardiol 1995;26:12091221. healthy middle-aged subjects: rationale and design of the Heinz
55. Tuzcu E, Berkalp B, De Franco A, et al. The dilemma of diagnosing Nixdorf RECALL Study. Risk Factors, Evaluation of Coronary
coronary calcification: angiography versus intravascular ultrasound. Calcium and Lifestyle. Am Heart J 2002;144:212218.
J Am Coll Cardiol 1996;27:832838. 77. Chambless LE, Heiss G, Folsom AR, et al. Association of coronary
56. McLaughlin VV, Balogh T, Rich S. Utility of electron beam com- heart disease incidence with carotid arterial wall thickness and major
puted tomography to stratify patients presenting to the emergency risk factors: the Atherosclerosis Risk in Communities (ARIC) Study,
room with chest pain. Am J Cardiol 1999;84:327328, A328. 19871993. Am J Epidemiol 1997;146:483494.
57. Georgiou D, Budoff MJ, Kaufer E, et al. Screening patients with 78. Guerci AD, Arad Y. Potential use of Ca++ scanning to determine the
chest pain in the emergency department using electron beam tomog- need for and intensity of lipid-lowering therapy in asymptomatic
raphy: a follow-up study. J Am Coll Cardiol 2001;38:105110. adults. Curr Cardiol Rep 2001;3:408415.
58. Janowitz WR, Agatston AS, Viamonte M, Jr. Comparison of serial 79. Kajinami K, Seki H, Takekoshi N, Mabuchi H. Quantification of
quantitative evaluation of calcified coronary artery plaque by coronary artery calcification using ultrafast computed tomography:
ultrafast computed tomography in persons with and without obstruc- reproducibility of measurements. Coron Artery Dis 1993;4(12):
tive coronary artery disease. Am J Cardiol 1991;68:16. 11031108.
80. Shields JP, Mielke CH, Jr, Rockwood TH, Short RA, Viren FK.
59. Achenbach S, Ropers D, Pohle K, et al. Influence of lipid-lowering
Reliability of electron beam computed tomography to detect coro-
therapy on the progression of coronary artery calcification: a pro-
nary artery calcification. Am J Card Imaging 1995;9(2):6266
spective evaluation. Circulation 2002;106:10771082.
81. Wang S, Detrano RC, Secci A, et al. Detection of coronary calcifi-
60. Budoff M, Lane K, Bakhsheshi H, et al. Rates of progression of cation with electron-beam computed tomography: evaluation of
coronary calcium by electron beam tomography. Am J Cardiol interexamination reproducibility and comparison of three image-
2000;86:811. acquisition protocols. Am Heart J 1996;132(3):550558.
61. Maher JE, Bielak LF, Raz JA, et al. Progression of coronary artery 82. Yoon HC, Greaser LE, 3rd, Mather R, et al. Coronary artery cal-
calcification: a pilot study. Mayo Clin Proc 1999;74:347355. cium: alternate methods for accurate and reproducible quantitation.
62. Downs J, Clearfield M, Weis S, et al. Primary prevention of acute Acad Radiol 1997;4(10):666673.
coronary events with lovastatin in men and women with average 83. Devries S, Wolfkiel C, Shah V, Chomka E, Rich S. Reproducibility
cholesterol levels: results of the AFCAPS/TexCAPS research. fo the measurement of coronary calcium with ultrafast computed
JAMA 1998;279:16151622. tomography. Am J Cardiol 1995;75(14):973975.
63. Anderson K, Wilson P, Odell P, et al. An updated coronary risk 84. Breen JB, Sheedy PF, Schwartz RS, et al. Coronary artery calcifica-
profilea statement for health professionals. Circulation 1991;83: tion detected with ultrafast CT as an indication of coronary artery
356362. disease. Radiology 1992;185:435439.
 CHAPTER 12 / MDCT CORONARY CALCIUM DETECTION 127

85. Fallavollita JA, Brody AS, Bunnell IL, Kumar K, Canty JM, Jr. 87. Budoff MJ, Georgiou D, Brody A, et al. Ultrafast computed
Fast computed tomography detection of coronary calcification in tomography as a diagnostic modality in the detection of coronary
the diagnosis of coronary artery disease. Comparison with angio- artery disease: a multicenter study. Circulation 1996;93(5):
graphy in patients <50 years old. Circulation 1994;89(1):285290. 898904.
86. Rumberger JA, Sheedy PF, III, Breen JF. Coronary calcium, as 88. Haberl R, Becker A, Leber A, et al. Correlation of coronary cal-
determined by electron beam computed tomography, and coronary cification and angiographically documented stenoses in patients
disease on arteriogram: Effect of patient's sex on diagnosis. Circu- with suspected coronary artery disease: results of 1,764 patients.
lation 1995;91:13631367. J Am Coll Cardiol 2001;37(2):451457.
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 129

13 Noninvasive Quantification
of Coronary Calcium
Quantification Methods, Scanner Types, Scan Protocols,
Accuracy, and Reproducibility

STEFAN ULZHEIMER, PhD, KAISS SHANNEIK, MSc,

AND WILLI A. KALENDER, PhD

INTRODUCTION Imaging of the heart has some very high demands to any
Calcium in the form of hydroxyapatite (HA) is regarded as imaging modality. For instance, it is of great importance to
a known marker for the presence of atherosclerotic lesions of have a very high reproducibility when detecting coronary cal-
the coronary arteries. Several studies have demonstrated that cifications in order to be able to track the progress of the disease
the risk for coronary events is associated and strongly corre- or the effectiveness of a certain medication. Despite its good
lated with the amount of coronary calcium (1,2). The absence temporal resolution, EBCT does suffer from a number of draw-
of coronary calcium does almost certainly imply the absence of backs, such as its restriction to cardiac applications as a result
coronary artery disease (CAD) (3), which, according to a World of limited performance in other fields compared to conven-
Health Report by the World Health Organization (WHO), is the tional CT scanners.
leading cause of mortality in the world, amounting to 13.7%. Therefore, conventional CT has tried to challenge EBCT in
Calcium strongly attenuates X-rays as a result of its rela- its domain of cardiac imaging since the first scanners became
tively high atomic number. Therefore, X-ray techniques are available. So, while conventional mechanical scanners without
quite suitable methods to detect and quantify coronary calcifi- any special arrangements to reduce motion artifacts have been
cations. In the past, plain chest radiographs and fluoroscopy used for the quantification of coronary calcium (611), much
were used to detect coronary calcifications, but they are both attention has been put to improve cardiac imaging by develop-
not sensitive enough and could not be used for the quantifica- ing new reconstruction approaches in image acquisition (12).
tion of calcium (4). However, owing to restrictions and limitations in image acqui-
Since its introduction in the late 1980s, electron beam com- sition speed and low volume coverage (single-slice scanners
puted tomography (EBCT) has been well established as a with 0.75-s gantry rotation time), this early generation of scan-
noninvasive imaging modality for the quantification of coro- ners has not completely achieved this ambitious goal. It was not
nary calcium (5) and the effective diagnosis of CAD. In addi- until the introduction of multislice computed tomography
tion, EBCT has offered the potential to image coronary arteries (MSCT) in 1998, with the acquisition of four slices simulta-
using CT angiography (CTA) for diagnostic and follow-up neously and an increase of the gantry rotation speed (0.5 ms),
studies (e.g., detecting and tracking any changes in the observed that several new clinical applications became much more fea-
parameters, such as the progress of the disease or the response sible, such as motion artifact-free imaging of the human heart.
to a certain medication). For the first time, these scanners The complex 3D motion of the heart can cause any single-slice
allowed the tomographic depiction of the entire heart volume CT scanner, and even the EBCT in a much higher probability,
(about 120 mm) to be imaged during one breath-hold and with to miss certain calcifications or even to count them more than
high temporal resolution (50 ms). For common protocols, once. The new mechanical CT scanners do not suffer from this
the total acquisition time was 100 ms for single axial sections problem, as they offer continuous spiral image data acquisi-
of 3 mm. tion. Electrocardiogram (ECG)-triggered and ECG-gated
reconstruction techniques have helped with new dedicated
From: Contemporary Cardiology: CT of the Heart: reconstruction algorithms to obtain a much better temporal
Principles and Applications resolution of typically 100250 ms. These scanners also
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ allowed for thinner slices (4 1-mm collimation) to be applied,

129
130 ULZHEIMER, SHANNEIK, AND KALENDER

Fig. 1. (A) Severe calcifications of the coronary arteries are clearly depicted even in contrast enhanced CT images for CTA studies. Obviously
contrast enhanced scan are not the method of choice to detect small amounts of calcium and for exact quantification of calcium. The scan was
performed on a Siemens SOMATOM Sensation 16 scanner with a standard cardiac scan protocol with a slice collimation of 12 0.75 mm and
a rotation time of 0.42 s. (B) shows a volume rendering of the showing the whole extent of the coronary arteries without any interruptions.

and offered excellent, almost isotropic resolution in the visual- to allow serial quantification of coronary calcium in individual
ization of the coronary anatomy, thereby making it more than patients over relatively short periods (<2 years) (24). In order
suitable in detecting CAD in an early stage of development to determine small changes with statistical significance, the
(Fig. 1). In 2001, a new generation of scanners was introduced standard deviation of the measurement has to be low.
by various manufacturers (13,14). These scanners offered fur- In the following sections we will review scoring methods and
ther improved spatial and temporal resolution, a reduced scan- concepts for quality assurance and standardization in this field.
ning time because of the faster gantry rotation (down to 0.42 s),
QUANTIFYING CORONARY CALCIUM
and an increased number of simultaneously acquired slices (up
to 16 submillimeter slices). Rapid developments in computer THE AGATSTON SCORE
and engineering technology have provided an opportunity to The first group that used EBCT images for the quantifica-
explore new clinical applications and to use cardiac CT imag- tion of coronary calcium were Agatston et al. in 1990. They
ing as a valuable quantification tool in clinical routine. introduced a pseudo-quantitative scoring methodlater called
With the advent of these new MSCT scanners, it becomes Agatston scorewhich became the traditional scoring method.
even more important to develop and establish standardized The Agatston score in its original form (5) is determined from
concepts for quality assurance in order to determine the 20 contiguous EBCT slices of the heart (3-mm slice thickness,
systems performance characteristics. It clearly becomes man- no interslice gaps, 100-ms data acquisition time per slice and
datory to know the errors of the methods applied and to develop breath-holding). The most cephalad slice is placed at the lower
tools which allow for calibration and direct comparison of margin of the main pulmonary artery, and image acquisition is
results obtained on arbitrary scanners. Unfortunately, in the triggered at 80% of the simultaneously recorded ECG signal.
case of the quantification of coronary calcium, this crucial task Calcified lesions are identified by applying a segmentation
has been neglected. Moreover, as the quantification of coro- threshold of 130 Hounsfield units (HU) on each of the 20 images
nary calcium with EBCT is still considered by many to be the and ignoring structures with sizes below 1 mm2 to exclude
gold standard, the new methods have to be compared to noise from evaluation. For each coronary artery i a region of
EBCT. Therefore, a phantom that provides an anthropomor- interest (ROI) is placed around each calcified lesion j in each
phic and reproducible environment has been designed and of the 20 images. In a succeeding step, the area Aij of the ROI
developed (15,16). in mm2 and the maximum CT number CT# max in HU are deter-
The issue of calcium score reproducibility has been mined. The Agatston score, Sij of lesion j in coronary artery i is
addressed by several clinical studies (1724). One paper deter- calculated as the product of the area of the lesion Aij with a
mined the reproducibility to be around 30% (25), while another weighting factor wij that depends on the maximum CT number
judged the method as being . . . not sufficiently reproducible CT#ijmax in the lesions for each image:
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 131

Fig. 2. Determination of the weighting factor for calculating the Agatston score.

Fig. 3. Evaluation of two measurements of the same small calcification with EBCT. Agatston score TCS, volume score V and calcium mass
m are calculated. The Agatston score shows a variability of 150%. For the volume score and the calcium mass the reproducibility errors are
much smaller.

Sij = wij Aij (1) SLAD = ijSLAD,j (4)

with
for the left anterior descending artery (LAD).
Figure 2 illustrates the determination of the weighting factor
1 if 150 HU = CT # ijmax
< 200 HU
w from the maximum CT number in the lesion, which is
2 if 200 HU = CT # ijmax < 300 HU described mathematically in equation 2. However, it should be
wij = noted that there is not a linear relationship when considering
max
3 if 300 HU = CT # ij < 400 0 HU (2) the Agatston score because of the definitive weighting factors
max
4 if 400 HU = CT # ij being used. Figure 3 demonstrates how the Agatston score is
calculated for two EBCT scans of one and the same small
The total calcium score (Stot or TCS) for all lesions in all artificial calcification with a diameter of 2 mm. By considering
coronary arteries is determined by summing up the scores of the this example it becomes evident what the drawbacks of this
lesions for all arteries in all images: score are (Table 1) and why the Agatston score can show a quite
high variability, which in this case is 150%. Noise and motion
Stot = ijSij = ijwij Aij (3)
are the cause for artifacts and result in small differences in
In later publications the score for each of the coronary arter- images. The strong dependence of the Agatston score on noise
ies alone was given separately, e.g., results from the fact that it is based on the maximum CT number
132 ULZHEIMER, SHANNEIK, AND KALENDER

Table 1
Pros and Cons of the Agatston Score

Pros Cons

Has been used for >10 yr, which Is unnecessarily complex

resulted in an accumulation of a large Was created for a special modality
amount of data and protocol
Has been shown to correlate with Uses absolute CT numbers that depend
coronary events upon reconstruction kernel, data
corrections, etc.
Does not correspond to any physical
measure
Is not a linear measureweighting
factors are used
Because its definition is based on
contiguous, nonoverlapping 3-mm
slices, an adaptation of the equations
must be undertaken for scanners other
than electron beam CT
Uses maximum CT numbers and is therefore
prone to noise and motion
Sensitive to noise and artifacts

that can lead to a difference in weighting factors by a factor of achieve an increase in reproducibility or sensitivity (29). There-
up to two. fore, even for EBCT the data acquired during the last 13 yr are
When scoring one and the same data set on different work- not necessarily validated and cannot be considered exactly
stations, the probability of getting different Agatston scores is comparable.
quite high (26,27). The quite complex description of the score Compared to the known high interscan variability of cal-
makes it difficult to obtain a decisive conclusion, as it leaves in cium scores of up to 46% for small calcifications determined
some points room for interpretation, while in others the pro- for EBCT and the Agatston score (21), even conventional
grammers knowingly have tried to make the score more robust. approaches without any dedicated improvements for motion-
However, these small differences obtained from the scoring artifact reduction yielded similar results. For example, the inter-
software seem to be negligible, as the definition of the score modality variability in a study which compared conventional
itself is quite arbitrary and is not based on any physical funda- 0.75-s sequential scans without ECG triggering to EBCT was
mentals. 42% (6).
Besides the already mentioned strong dependence on small It is possible to adapt the Agatston score to a certain degree
variations resulting from noise and motion, there are a number so that other scanners and other protocols yield similar absolute
of other factors that have a serious effect on this scoring method. scoring results as EBCT (Fig. 4) if the scan protocols are care-
These include its dependence on absolute CT numbers, the fully designed and corrections are applied when calculating the
number of slices per scan length, and its nonlinearity for score. Nevertheless, not all effects that influence the Agatston
increasing amounts of calcium, which is due to its reliance on score on different scanners or for different scan protocols can
nonlinear weighting factors. Furthermore, as the score does not be taken into account.
correspond to any physical measure, it cannot be easily com- VOLUME SCORES
pared to true values for exactly defined calcifications. Recent studies (30) based on estimating the volume of the
A large amount of data has accumulated over the years with calcification provide an alternative quantification method for
the use of the Agatston score, and the risk tables derived for coronary calcium. This scoring method, which can also be
coronary events (28) are the only, but of course very important, applied to EBCT, has become more and more popular as some
advantages of this score. studies indicate that it is more robust with respect to reproduc-
Even though the Agatston score shows a number of serious ibility (25). The determination is very similar to the Agatston
drawbacks and limitations, it has been used extensively in the score. The volume score represents the volume V of the voxels
past years and has established itself as a de facto standard for of the calcification that lie above a certain threshold. A ROI is
coronary calcium scoring. Some groups have suggested modi- drawn around each calcified lesion, and for each lesion the
fications for the Agatston score in order to improve its repro- number of voxels in the volume data set that lie above a certain
ducibility (29) or sensitivity (7), but all efforts were not rigorous threshold is multiplied by the volume of one voxel
enough and have not resulted in a better standard. Other groups
Vij = NVoxel VVoxel (5)
made modifications to their EBCT scan protocols (e.g., differ-
ent trigger positions than 80%, mean CT numbers instead of The easiest segmentation approach again is to take all voxels
maximum CT numbers, or different thresholds were used), to above a certain threshold, e.g., again 130 HU.
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 133

Fig. 4. Correlation of Agatston scores determined with two different scanners (EBCT and a conventional 4-slice scanner with prospective
triggering). When the scan protocols on scanners other than EBCT are carefully designed and corrections are applied when calculating the
Agatston score quite good agreement of scores on different machines can be achieved.

The total volume score is the sum of all individual lesions: ent scanners and protocols, and shows an increased reproduc-
ibility compared to the Agatston score (16,31). However, it also
Vtot = i,jVij
is dependent on the segmentation threshold.
Compared to the traditional scoring methodthe Agatston The mean CT number CT of the calcification in each slice
scorethis scoring method provides a higher reproducibility multiplied by the volume V of the calcification in that slice is
(30). However, there are also a few disadvantages when using directly proportional to the calcium mass m in that slice (Fig. 5).
the volume score as a quantification method. For example, it
should be noted that when using a relatively large slice thick- m i CT i Vi (6)
ness for small calcifications that have a high density, or when The density of the calcification, and with it the calcium
there is a fixed threshold, this score can lead to large deviations mass, is directly proportional to the mean CT number in the
from the true volume of the calcification, caused by the partial- calcification (Fig. 5). It automatically corrects for linear par-
volume effects. Thus, the volume score may strongly over- or tial-volume effects, as objects smaller than the slice thickness
underestimate the real volume of the calcification. Also, as the are displayed with accordingly decreased mean CT numbers.
volume score depends on the applied threshold, it is not a true To obtain the complete mass of the calcification, the single
physical representation of the volume of the calcification. masses for all slices are summed up. Although the value could
serve as a measure for the calcium mass, it is desirable and
CALCIUM MASS
much more descriptive to obtain absolute values for the cal-
Another more rigorously defined scoring method aims at cium mass m.
the determination of the absolute calcium mass, and thereby is To obtain absolute values for the calcium mass, a simple
the only scoring method that provides a truly quantitative mea- calibration measurement of a calcification with known HA
sure for the amount of calcium, e.g., given in milligrams hydro- density HA has to be carried out and a calibration factor c
xyapatite (HA) above a certain threshold. It possesses many determined so that
advantages compared to the other two scoring methods. For
instance, it can be easily determined, is comparable for differ- mij = c CTij Vij (7)
134 ULZHEIMER, SHANNEIK, AND KALENDER

The developed cardiac phantom has been manufactured in

cooperation with the company QRM GmbH, in Mhrendorf,
Germany. The exact dimensions of the phantom are shown in
Fig. 6. In order to achieve maximum flexibility in the assess-
ment of different image quality parameters, a modular concept
has been chosen. In Fig. 7, an example of the phantom body
with a calibration insert can be seen. The phantom consists of
a phantom body that provides an anthropomorphic and repro-
ducible environment, consisting of a cross-section of the human
thorax with artificial lung tissue and a spine insert surrounded
by tissue-equivalent material. At the position of the heart there
is an empty 10-cm circular space that is able to hold different
kinds of inserts. It allows for various quality assessment mea-
surements in coronary calcium scoring, such as the investiga-
tion of spatial and temporal resolution, contrast and noise,
Fig. 5. The density of a calcification and with it the calcium mass is patient dose and detection limits, reproducibility, and accuracy
proportional to the mean CT number in the calcification. The absolute in coronary calcium scoring. Additionally, extension rings for
calcium mass can be determined with a simple calibration measure-
ment carried out at a regular intervals for each scanner type and scan the phantom were developed in order to mimic different patient
protocol. sizes for the realistic assessment of image noise in larger
patients (Fig. 8).
The phantom inserts representing different sizes and con-
Using the relation mij = HAV, the calibration factor c is centrations of calcium hydroxyapatite and two calibration
calculated as inserts allow us to test both the calibration and accuracy of
arbitrary scanners and scan protocols, and the reproducibility
c = m/(CTV) = (HAV)/(CTV) = HA/CT. (8) of the measurements.
Equation 6 assumes that the mean CT number of water is 0 Small, exactly defined calcifications, consisting of different
by definition. If an exact CT value for water is available from HA densities immersed in water-equivalent plastic, can be
the same calibration measurement, and if it is not equal to 0, this inserted into the phantom body to simulate realistic calcifica-
can be taken into account by subtracting the CT value for water tions. One example is a solid insert that contains nine small
from the CT number of the calcification (baseline correction): calcifications (1-, 3-, and 5-mm diameter) with three different
HA densities (200, 400, and 800 mg HA/cm3) (Fig. 9). Addi-
(
c = HA CT CT water ) (9) tionally, the insert contains two large cylinders, one consisting
of water-equivalent material and one with an HA density of
In practical cases, the CT number of water always has to be 200 mg HA/cm3. These inserts can be used to determine the
checked and taken into account. The calibration factor c there- calibration factor c in equation 9 to calculate the calcium mass.
fore is given by the HA density HA of the known calcification Table 2 gives an overview of HA densities and the sizes of all
divided by the mean CT number difference in the calibration cylindrical calcifications with their respective values for the
measurement. Therefore, for each scanner and scan protocol, a areas of the circular cross-sections and the volumes of the cyl-
specific calibration factor is obtained. inders and the HA masses inside the calcifications. One cannot
The mass score of individual lesions is summed up to obtain expect to obtain exactly these values in measurements with the
the total mass score: phantom, as the measured values strongly depend on scan
parameters such as noise, slice thickness, and above all the
mtot = i,jmij.
segmentation threshold. They can be used in the calibration
process, however, in order to provide more accurate scoring
TOOLS FOR QUALITY ASSURANCE
results.
IN CORONARY CALCIUM SCORING An exemplary evaluation of all calcifications is given in
Assessing important performance parameters such as Table 3. It shows among other findings an extremely high sen-
reproducibility, accuracy, and comparability can be a difficult sitivity, and that even for small calcifications with a calcium
task, as there are many parameters that may vary. This is in mass of less than 1 mg and a relatively large slice thickness of
particular so in patient studies where it might be ethically prob- 2.5 mm, an acceptable quantification of the calcium can be
lematic to scan patients several times just to be able to obtain performed. Nevertheless, thinner slices would improve the
values for reproducibility or to compare calcium scores on dif- accuracy considerably.
ferent scanners. It is evident that it is necessary to design and In order to investigate the influence of cardiac motion, a
develop appropriate phantoms that allow the investigation of setup was built using the anthropomorphic cardiac phantom
such parameters. Furthermore, we will show that the anthropo- with a water tank is placed inside the space reserved for inserts
morphic phantoms later introduced can be used as calibration that also allows the movement of objects (e.g., calcifications)
standards for the assessment of absolute calcium mass values in an anthropomorphic manner. The setup is composed of a
(15,16,31). robot, seen at the back, which is able to mimic realistic heart
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 135

Fig. 6. Sketch of the anthropomorphic cardiac phantom for quality assurance and calibration purposes.

Realistic motion functions have been obtained from cine

angiography. Figure 11 shows a 4D plot of such a motion func-
tion. A set of linear translation tables was assembled, which
were mounted perpendicular to each other in order to reach all
points in 3D space. A system of metal rods that reaches into the
water tank was attached to one of the translation tables. The
water-equivalent rods containing the respective test objects can
be easily attached to this end of the metal rods. The setup also
produces a corresponding ECG signal that can be directly fed
into the ECG monitor used for triggering and gating recon-
struction techniques. As an application example, a plastic rod
with an immersed calcified specimen extracted from a cadaver
was moved with a realistic 3D motion function at a heart rate
of 60 beats per minute (bpm) and scanned on a spiral CT scan-
ner with retrospective gating technique (32). For the motion
Fig. 7. A photo of the anthropomorphic phantom body with a calibra- function that was used, the calcification is depicted nearly per-
tion insert that can be placed inside the circular space in the center of fectly in a slow-motion heart phase at around 60% of the RR
the body phantom. The space can hold different kinds of static inserts interval (Fig. 10C). In other phases the calcification is depicted
for various quality assessment measurements. with more or less severe motion artifacts, depending on the
motion speed and amplitude. This or similar phantom setups
are used in the validation of new approaches for cardiac imag-
movements in all three dimensions by simulating arbitrary 3D ing and in the comparison of the behavior of different scanners
trajectories inside the water tank that is inserted into the anthro- with respect to the depiction of the moving heart.
pomorphic phantom body (Fig. 10A,B). The robot is controlled The described phantoms were used to show that comparable
by means of a conventional PC or laptop, and its movements are Agatston scores can be obtained on different scanners and with
coupled to an ECG monitor to simulate the required heartbeats different scan protocols, but also that the calcium mass is an
later used for retrospective image reconstruction. easier and more reproducible measure (31). The calcium mass
136 ULZHEIMER, SHANNEIK, AND KALENDER

Fig. 8. Extension rings for the phantom body can mimic different patient sizes to investigate the influence of patient size on e.g., image noise
and quantification results. The phantoms are manufactured in co-operation with QRM GmbH, Mhrendorf, Germany (www.qrm.de).

Table 2
Properties of the Cylindrical Calcifications
Inside the Calibration Insert

HA density Length Diameter Area Volume HA mass

(mg/cm3) (mm) (mm) (mm2)a (mm3) (mg)

200 5.0 5.0 19.6 98.2 19.6

200 3.0 3.0 7.1 21.2 4.2
200 1.0 1.0 0.8 0.8 0.2

400 5.0 5.0 19.6 98.2 39.3

400 3.0 3.0 7.1 21.2 8.5
400 1.0 1.0 0.8 0.8 0.3

800 5.0 5.0 19.6 98.2 78.5

800 3.0 3.0 7.1 21.2 17.0
800 1.0 1.0 0.8 0.8 0.6
aArea of the circular cross-section of the cylindric calcifications.

Fig. 9. Static insert for the phantom body that contains various calci- the Agatston score (R2 = 0.9733) and that the mass can be
fications of different sizes and HA densities. There are three different
inserts for each of the three HA densities (200 mg HA/cm3, 400 mg accurately calculated from the Agatston score by multiplying
HA/cm3, 800 mg HA/cm3). it by a factor of, in this specific case, 0.1922. By achieving a
good correlation between the two scores we are able to assure
that old data collected over the past years would not be lost
completely when switching to new quantification methods.
is correlated to the Agatston score, so that it is possible to Reproducibility for a moving calcification with different scores
roughly calculate the calcium mass from the Agatston score. and scanning methods was investigated on three different scan-
The determination of the calcium mass and its accuracy with an ners (EBCT Evolution C-150XP, Imatron; Volume Zoom,
EBCT scanner is depicted in Fig. 12. From the graph it can be Siemens; SOMATOM Sensation 16, Siemens), in which the
easily shown that the calcium mass is quite well correlated to dependence of the Agatston score and HA mass on the heart
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 137

Table 3
Exemplary Determination of Agatston Score, Volume,
and Calcium Mass for All Calcifications in the Calibration Insert

Agatston score Volume (mm3) Calcium mass (mg)

measured measured true value deviation measured true value deviation

200 mgHA/cm3
5 mm 86.33 118.45 98.17 21% 19.41 19.63 1%
3 mm 14.65 18.79 98.17 21% 19.41 19.63 1%
1 mm 0.79 0.16

400 mgHA/cm3
5 mm 193.50 160.41 98.17 63% 42.11 39.27 7%
3 mm 43.18 42.25 21.21 99% 7.70 8.48 9%
1 mm (0.53)a 1.34 0.79 70% 0.16 0.31 49%

800 mgHA/cm3
5 mm 234.90 182.82 98.17 86% 92.79 78.54 18%
3 mm 83.47 52.17 21.21 146% 10.67 16.96 22%
1 mm 2.82 3.53 0.79 349% 0.53 0.63 15%
A sequence scan was carried out with 1.5-s rotation time and 2 10-mm slice collimation for the large calibration inserts
and 4 2.5-mm slice collimation and 0.5-s rotation time for the small calcifications. For both measurements, the tube
current was 150 mAs and the voltage 140 kVp. The field of view was chosen to be 100 mm with a 512 512 pixel matrix
for both measurements.
aThe area was below 1 mm2. According to the original version of the Agatston score, the score is 0.

Fig. 10. Setup of a motion phantom. A robot setup can move arbitrary objects on realistic 3D trajectories to simulate cardiac motion. (C) shows
an application example. A real specimen extracted from a cadaver scanned with a MSCT scanner and images were reconstructed at different
positions in the RR interval. The shaded surface displays show motion artifacts in all phases except at 60% of RR for this motion function.
138 ULZHEIMER, SHANNEIK, AND KALENDER

Fig. 11. A 4D plot of a realistic motion function.

Fig. 12. Correlation of the calcium mass compared to the Agatston score for EBCT and artificial and real calcifications. The close correlation
for calcifications that are not too dense allows a rough calculation of the mass from the Agatston score.

rate is depicted in Fig. 13 and Fig. 14. The results from the Zoom especially for retrospectively gated data. For the
16-slice CT scanner show an improvement over the 4-slice CT relationship of the calcium mass at various heart rates (Fig. 14)
scanner when comparing the data with EBCT. The Sensation a similar behavior can be observed as for the Agatston score
16 shows a better behavior at higher heart rates than the Volume (Fig. 13). However, the error for the calcium mass is lower than
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 139

Fig. 13. Diagram showing the dependency of the Agatston score on the heart rate for different protrocols for the EBCT, the 4-slice MSCT, and
a 16-slice MSCT scanner.

Fig. 14. Diagram showing the dependency of the calcium mass on the heart rate for different protocols on the EBCT, the 4-slice MSCT, and
a 16-slice MSCT scanner.

for the Agatston score, making it much more suitable as an articles (16,33), and a recent textbook on computed tomogra-
effective calcium quantification method. phy (32).
More details and more investigations on reproducibility, FUTURE PERSPECTIVES
detectability, and quality assurance in cardiac CT in general Several phantoms and concepts that have been developed
can be found in a recent PhD thesis (31), peer reviewed for the evaluation of calcium scoring performance during the
140 ULZHEIMER, SHANNEIK, AND KALENDER

last four years are readily available at the Institute of Medical 5. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte Mj,
Physics at the University of Erlangen, Germany (16,31,33,34). Detrano R. Quantification of coronary artery calcium using ultrafast
computed tomography. JACC 1990;15:827832.
These phantoms have been widely accepted among CT 6. Becker CR, Knez A, Jakobs TF, et al. Detection and quantification
manufacturers and within the scientific community during the of coronary artery calcification with electron-beam and conventional
last 4 yr and have been used in many studies that examined CT. Eur Radiol 1999;9:620624.
reproducibility and accuracy in the quantification of coronary 7. Broderick LS, Shemesh J, Wilensky RL, et al. Measurement of
calcium. The International Task Group on Standardization in coronary artery calcium with dual-slice helical CT compared with
coronary angiography: evaluation of CT scoring methods, inter-
Cardiac CT, consisting of representatives of all major CT manu- observer variations, and reproducibility. AJR Am J Roentgenol
facturers and experts from all over the world, was founded in 1996;167:439444.
2002 and is trying to push forward these efforts. 8. Shemesh J. Spiral methods quantify coronary calcification. CT
The most important outcome of all investigations carried 1994:3032.
9. Shemesh J, Apter S, Rozenman J, et al. Calcification of coronary
out with the phantoms up to now is that the calcium mass, as a arteries: detection and quantification with double-helix CT. Radiol-
true quantitative measure which corresponds to the amount of ogy 1995;197:779783.
calcium, should be used for future clinical studies. It can be 10. Shemesh J, Fisman EZ, Tenenbaum A, et al. Coronary artery calci-
easily calculated for all CT scanners and scan protocols in a fication in women with Syndrome X: usefulness of double-helical
CT for detection. Radiology 1997;205:697700.
comparable way, it corresponds to the true amount of calcium, 11. Shemesh J, Tenenbaum A, Fisman EZ, et al. Absence of coronary
is more reproducible than the Agatston score, and it can roughly calcification on double-helical CT scans: predictor of angiograph-
be calculated from the Agatston score for data acquired in the ically normal coronary arteries in elderly women? Radiology 1996;
past. As many physicians are used to the old scoring method, 199:665668.
12. Kachelriess M, Kalender WA. ECG-correlated image reconstruc-
the Agatston score can be given side by side during a transi- tion from subsecond spiral CT scans of the heart. Medical Physics
tional period. 1998;25:24172431.
However, it has to be stressed that the determination of cal- 13. Flohr T, Stierstorfer K, Bruder H, Simon J, Schaller S. New techni-
cium mass is not just another scoring method but appears to be cal developments in multislice CTPart 1: approaching isotropic
resolution with sub-millimeter 16-slice scanning. Rofo Fortschr Geb
the only reasonable physical quantification approach valid for Rontgenstr Neuen Bildgeb Verfahr 2002;174:839845.
arbitrary CT scanners. Adapting a scoring method like the 14. Flohr T, Bruder H, Stierstorfer K, Simon J, Schaller S, Ohnesorge B.
Agatston score (with many disadvantages, developed for a spe- New technical developments in multislice CT, part 2: sub-millime-
cial scanner) to new scanner types would introduce more errors ter 16-slice scanning and increased gantry rotation speed for cardiac
imaging. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr
and should be avoided. 2002;174:10221027.
Independently from the quantification method, it is manda- 15. Ulzheimer S, Kachelriess M, Kalender WA. New phantoms for
tory for a scientific approach that the error of a method be quality assurance in cardiac CT. RSNA 1999;213(P):402.
known. In the case of the quantification of coronary calcium, 16. Ulzheimer S, Kalender WA. Assessment of calcium scoring per-
formance in cardiac computed tomography. Eur Radiol 2003;13:
this has been neglected in the past, and the goal should be that 484497.
each manufacturer as well as each group that suggests a new 17. Miles KA. Measurement of tissue perfusion by dynamic computed
approach (scanner, scan protocol, evaluation method) also tomography. Brit J Radiol 1991;64:409412.
validate it by giving the respective error for the method. As the 18. Bielak LF, Sheedy PF, 2nd, Peyser PA. Coronary artery calcifica-
tion measured at electron-beam CT: agreement in dual scan runs and
error also depends on the characteristics of the calcifications change over time. Radiology 2001;218:224229.
(size, density, HA distribution) errors for typical calcifications 19. Achenbach S, Ropers D, Mohlenkamp S, et al. Variability of repeated
and the detection limits of the method should be given. coronary artery calcium measurements by electron beam tomogra-
For this purpose, the concepts and phantoms described in phy. Am J Cardiol 2001;87:210213, A8.
20. Becker CR, Knez A, Ohnesorge B, et al. Visualization and quanti-
this chapter can be used in the future to check and improve fication of coronary calcifications with electron beam and spiral
accuracy and reproducibility of the quantification methods, to computed tomography. Eur Radiol 2000;10:629635.
ensure uniform image quality for different approaches, and 21. Devries S, Wolfkiel C, Shah V, Chomka E, Rich S. Reproducibility
an important aspect that has not been addressed in this chap- of the measurement of coronary calcium with ultrafast computed
tomography. Am J Cardiol 1995;75:973975.
terto determine and limit patient dose (35). 22. Hernigou A, Challande P, Boudeville JC, Sn V, Grataloup C,
Plainfoss MC. Reproducibility of coronary calcification detection
REFERENCES with electron-beam computed tomography. European Radiology
1. Arad Y, Spadaro LA, Goodman K, Newstein D, Guerci AD. Predic- 1996;6:210216.
tion of coronary events with electron beam computed tomography. 23. Shields JP, Mielke CH, Rockwood TH, Short RA, Viren FK. Reli-
J Am Coll Cardiol 2000;36:12531260. ability of electron beam computed tomography to detect coronary
2. Raggi P, Callister TQ, Cooil B, et al. Identification of patients artery calcification. Am J Card Imaging 1995;9:6266.
at increased risk of first unheralded acute myocardial infarction 24. Wang S, Detrano RC, Secci A, et al. Detection of coronary calcifi-
by electron-beam computed tomography. Circulation 2000;101: cation with electron-beam computed tomography: evaluation of
850855. interexamination reproducibility and comparison of three image-
3. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification: acquisition protocols. Am Heart J 1996;132:550558.
pathophysiology, epidemiology, imaging methods, and clinical 25. Yoon HC, Greaser LE, Mather R, Sinha S, McNitt-Gray MF, Goldin
implications. A statement for health professionals from the American JG. Coronary artery calcium: alternate methods for accurate and
Heart Association. Writing Group. Circulation 1996;94:11751192. reproducible quantitation. Acad Radiol 1997;4:666673.
4. Stanford W, Thompson BH, Weiss RM. Coronary artery calcifica- 26. Knollmann FD, Hum mel M, Spiegelsberger S, Bocksch W, Hetzer
tion: clinical significance and current methods of detection. AJR R, Felix R. Comparison of algorithms for coronary artery calcium
Am J Roentgenol 1993;161:11391146. scoring. Radiology 2000;217(P):588.
 CHAPTER 13 / NONINVASIVE QUANTIFICATION METHODS 141

27. Adamzik M, Schmermund A, Reed JE, Adamzik S, Behrenbeck T, 31. Ulzheimer S. Cardiac Imaging with X-ray computed tomography:
Sheedy PF, 2nd. Comparison of two different software systems for new approaches to image acquisition and quality assurance. In:
electron-beam CT-derived quantification of coronary calcification. Kalender WA (ed), Reports from the Institute of Medical Physics,
Invest Radiol 1999;34:767773. University of Erlangen-Nrnberg. Shaker Verlag, Aachen: 2001:123.
28. ORourke RA, Brundage BH, Froelicher VF, et al. American Col- 32. Kalender WA. Computed Tomography. Wiley and Sons, New York,
lege of Cardiology/American Heart Association Expert Consensus NY: 2001.
Document on electron-beam computed tomography for the diag- 33. Kachelrie M, Ulzheimer S, Kalender WA. ECG-correlated imag-
nosis and prognosis of coronary artery disease. J Am Coll Cardiol ing of the heart with subsecond multislice spiral CT. IEEE Trans
2000;36:326340. Med Imaging 2000;19:888901.
29. Shemesh J, Tenenbaum A, Kopecky KK, et al. Coronary calcium 34. Ulzheimer S, Decker R, Kachelriess M, Kalender WA. ECG-corre-
measurements by double helical computed tomography. Using the lated image reconstruction from subsecond multislice spiral CT
average instead of peak density algorithm improves reproducibility. scans of the heart: methodological investigations, phantom mea-
Invest Radiol 1997;32:503506. surements and patient studies. 5th International SOMATOM CT
30. Callister TQ, Cooil B, Raya SP, Lippolis NJ, Russo DJ, Raggi P. Scientific User Conference, Zrich, 16./17.6.2000, 2000.
Coronary artery disease: improved reproducibility of calcium scor- 35. Schmidt B, Schmidt M, Kalender WA. A 3D voxel Monte Carlo
ing with an electron-beam CT volumetric method. Radiology 1998; program for scanner and patient specific dose calculations in CT.
208:807814. Eur Radiol 2000;10(Suppl 1):304.
 CHAPTER 14 / CT OF THE PERICARDIUM 143

CONTRAST-ENHANCED CT
OF THE HEART: IV
MORPHOLOGY AND FUNCTION
 CHAPTER 14 / CT OF THE PERICARDIUM 145

14 CT of the Pericardium

REINHARD GROELL, MD

INTRODUCTION ANATOMY AND FUNCTION

Because of its widespread availability and its cost-effective- The pericardium is a flask-like sac that surrounds the heart,
ness, echocardiography represents the primary method of and the neck of this sac is attached to the root of the great vessels
choice to image the pericardium. However, echocardiography at the base of the heart. The pericardium comprises an outer
is operator-dependent, and it often fails to detect the entire fibrous layer (the fibrous pericardium) and an inner serous sac
pericardium. Thus it is limited in the assessment of the severity (the serous pericardium) (2). The serous pericardium consists
of pericardial involvement in various diseases affecting the of an inner visceral layer (the epicardium), which is intimately
pericardium. Nataf et al. have demonstrated that echocardio- applied to the heart and the epicardial fat, and an outer parietal
graphy revealed thickened pericardium in only 62% of patients layer, which lines the fibrous pericardium. The visceral layer is
with constrictive pericarditis (1). Additionally, ultrasound can- reflected from the heart and the root of the great vessels onto the
not penetrate calcifications that occur frequently in patients inner surface of the fibrous pericardium.
with constrictive pericarditis. The pericardial cavity lies between the two layers of the
Magnetic resonance imaging (MRI) has a great potential in serous pericardium. Under physiological conditions, it con-
differentiating soft tissues of the pericardium and myocar- tains 2025 mL of serous fluid; however, the amount of fluid
dium. In MRI of the heart, image quality strongly depends on may vary considerably among individuals, particularly in chil-
the accuracy of cardiac gating. This can be problematic in the dren and infants (3). This fluid is an ultrafiltrate of plasma,
presence of arrhythmia, which frequently occurs in patients which is produced by the monolayer of mesothelial cells of the
with pericardial diseases. Flow artifacts may compromise the serous pericardium. It is drained into the right lymphatic duct
evaluation of pericardial fluid during the cardiac cycle. More- and the thoracic duct.
over, as a result of signal loss, the amount and distribution of On its outside, the pericardium is connected anteriorly to the
pericardial calcifications cannot be determined accurately sternum, inferiorly to the diaphragm, and posteriorly to the
with MRI. esophagus, the thoracic aorta, and the spine (2,4). The coronary
Fast computed tomography (CT) scanners that enable imag- vessels run in the subepicardial space between the epicardium
ing of the entire pericardium with high temporal and spatial and the myocardium, which is a connective tissue layer con-
resolution in virtually every patient can overcome some of these taining fat.
limitations with echocardiography and MRI. More than ever, The pericardium is considered to prevent the ventricles from
CT plays an important role in the diagnostic workup of pericar- extreme distension and to control the mechanics of ventricular
dial diseases. contractions (57). However, it is known that even with con-
The pericardium is frequently involved in myocardial genital absence of the entire pericardium, subjects usually do
pathologies, and vice versa. Additionally, clinical symptoms of not suffer from significant pathophysiological changes.
pericardial diseases may mimic those of myocardial patholo- On computed tomography studies of the thorax, the normal
gies. That is why CT imaging of the pericardium is generally pericardium appears as a thin band enveloping the heart
combined with morphologic imaging of the entire heart, and a (Fig. 1A,B). As the pericardium is surrounded by outer medi-
proper CT imaging protocol should follow these tasks. In our astinal and inner subepicardial fat, the visualization of the peri-
institution, we acquire subsecond, transverse CT sections of cardium on CT strongly correlates with the amount of fat. In
the heart during suspended respiration in the supine patient general, the more fat that is present, the better one can delineate
position, preferably using electrocardiogram (ECG) gating. the pericardium on CT; the presence of subepicardial fat is
Usually, a slice thickness of 1.53 mm is used, depending on especially important for the depiction of the pericardium. Typi-
the clinical question and the patients capability to hold the cally, subepicardial fat is well developed over the right ven-
breath. tricle, but it may be very thin or even invisible over the left
ventricle. That is why it is often problematic to delineate the
pericardium from the myocardium at the left lateral myocardial
From: Contemporary Cardiology: CT of the Heart: wall. The attachments to the sternum, diaphragm, or thoracic
Principles and Applications spine are rarely visualized, as they often do not represent dis-
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

145
146 GROELL

Fig. 1A,B. The pericardium (arrows) appears as a thin band enveloping the heart.

tinct anatomical ligaments but rather ill-defined fibrous strands behind and on the right lateral circumference of the superior
within the mediastinal fat. vena cava.
PERICARDIAL SINUSES AND RECESSES Transverse Sinus
At the serous reflections of the pericardium around the root The transverse sinus is located above the left atrium and
of the great vessels at the base of the heart, the pericardial cavity posterior to the ascending aorta and the pulmonary trunk.
forms the pericardial sinuses (8,9). They are not separate com- Between the ascending aorta and the superior vena cava, it is
partments but represent extensions of the pericardial cavity. connected with the pericardial cavity proper. It extends upwards
The Nomina Anatomica labels them the transverse sinus and along the ascending aorta, where it forms the superior aortic
the oblique sinus. Where the pericardium extends onto the great recess. This superior aortic recess is frequently visible, and it
vessels, the pericardial cavity proper as well as the sinuses form was one of the first recesses described in CT studies of the
recesses. Their differentiation is based on topographic land- mediastinum; it can be divided into an anterior, posterior, and
marks, since there is no histological difference between the right lateral portion (10). The anterior and right lateral portions
layers of the pericardial cavity, the sinuses, and the recesses. are directly related to the thymus, the posterior portion to
Recesses of the Pericardial Cavity Proper tracheobronchial lymph nodes. On unenhanced CT images and
The right and left pulmonic vein recesses are extensions of on magnetic resonance imaging studies, the superior aortic
the pericardial cavity proper located between the superior and recess may also simulate aortic dissection or intramural
inferior pulmonic veins on both sides. The postcaval recess lies hematoma (Fig. 2). The left pulmonic recess is located below
 CHAPTER 14 / CT OF THE PERICARDIUM 147

typical location and appearance helps the radiologist avoid a

misdiagnosis of lymphadenopathy and of other mediastinal
pathologies. Rarely, pericardial cysts or tumors that can mimic
cardiac tumors may develop in these sinuses and recesses (3).
As the pericardial sinuses and recesses represent extensions of
the pericardial cavity proper, it is very likely that pericardial
fluid might move from one area to another during the cardiac
cycle or during respiration (11).
DEVELOPMENTAL ANOMALIES
ABSENCE OF THE PERICARDIUM
Congenital absence of the pericardium may be partial or
complete, with a prevalence of 0.01% in postmortem studies
(12,13). The majority of cases consists of a partial defect of the
pericardium located over the left ventricle or left-lateral to the
left atrium and left auricle. Partial defects at the right side of
the pericardium or at the diaphragmatic parts are much less
common, and complete absence of the pericardium is extremely
rare. In 30% of the patients, congenital defects of the pericar-
dium may be associated with malformations of the heart such as
tetralogy of Fallot, atrial septal defects, or patent ductus arte-
riosus, or with congenital pulmonary malformations such as
bronchogenic cyst and lung sequestration. In most patients,
total absence of the left-sided pericardium does not result in
clinical symptoms; the same is true for very small defects.
However, medium-sized defects carry the risk of cardiac her-
niation and even strangulation, particularly of the left atrial
Fig. 2. Pericardial sinuses and recesses (arrows) are located at the appendage. In our institution we have examined a patient after
pericardial reflections around the root of the great vessels. left-sided pneumectomy and pericardiectomy, whose heart
herniated into the left thoracic space, resulting in compression
of the great vessels, which resulted in subsequent cardiac fail-
ure (Fig. 3). The symptoms resolved after subsequent re-opera-
tion with closure of the left-sided pericardial defect. It is also
the left pulmonary artery and posterolateral to the proximal reported that pericardial defects may predispose the heart to
portion of the right pulmonary artery. The right pulmonic recess pulmonary or mediastinal infections.
lies below the right pulmonary artery and above the left atrium. Frequently, the left ventricle is not surrounded by excessive
Its posterior circumference is directly related to inferior fatty tissue, which makes the delineation of the left-sided peri-
tracheobronchial lymph nodes. The inferior aortic recess is situ- cardium difficult even in normal subjects. Thus, defects of the
ated between the ascending aorta and the inferior portion of the left-sided pericardium are often not directly encountered on
vena cava superior or the right atrium, respectively. It may CT. More frequently, these defects are recognized indirectly by
extend down to the level of the aortic valve. a shift of the heart to the left orwhen the defect is smaller
Oblique Sinus by left-lateral prominence or herniation of the left atrium or left
The oblique sinus lies behind the left atrium. It is separated atrial appendage.
from the transverse sinus by a double reflection of the pericar- PERICARDIAL CYSTS AND DIVERTICULA
dium between the right and left superior pulmonic veins. On Pericardial cysts and diverticula are rare clinical entities
CT, the transverse sinus (including the right and left pulmonic (12,14,15). They represent well-demarcated fluid collections
recess) is always clearly separated from the oblique sinus that blend with the pericardium. Most often they occur in the
(including the posterior pericardial recess) by a fat plane. The right cardio-phrenic angle (Fig. 4); alternative locations in the
upper, right lateral extension of the oblique sinus is named the left cardio-phrenic angle or at the base of the heart are extremely
posterior pericardial recess. It is located behind the distal right uncommon. Usually they remain constant in size, but slow
pulmonary artery and medial to the bronchus intermedius. The progression has also been described. Rarely, calcifications may
esophagus runs posterior to the oblique sinus, and inferior tra- be present in the wall of a pericardial cyst. While pericardial
cheobronchial lymph nodes are in close proximity to these cysts represent encapsulated collections of fluid, diverticula
structures. are herniations of the pericardial cavity through a defect of the
In general, pericardial sinuses and recesses are frequently parietal pericardium (Fig. 5). Although pericardial cysts and
observed on CT studies of the heart. They may be problematic diverticula may reach a size of several centimeters, they rarely
in the differentiation from lymph nodes, esophageal or thymic cause clinical problems. Their radiological importance lies in
processes, or vascular abnormalities. The knowledge of their the differentiation against clinically relevant pathologies such
148 GROELL

Fig. 3. Dislocation of the heart to the left after left-sided pericardiectomy and pneumectomy, resulting in acute postoperative cardiac failure.
The symptoms resolved after surgical repositioning of the heart with fixation of the pericardial defect.

Fig. 4. Pericardial cyst (arrow) represents well-demarcated fluid collection that is typically located in the right cardio-phrenic angle.

Fig. 5. Pericardial diverticulum (long arrow) with open communication between the diverticulum and the pericardial cavity. Note contrast
material in the pericardial cavity proper (short arrows) after transthoracic puncture of the diverticulum and installation of diluted contrast
material.
 CHAPTER 14 / CT OF THE PERICARDIUM 149

as diaphragmatic hernias, abscesses, or bronchogenic cysts. PERICARDIAL CONSTRICTION

Uncommonly, pericardial cysts and diverticula may be Pericardial constriction is defined as fibrotic or calcific
acquired, such as after severe chest trauma or surgery. thickening and scarring of the pericardium, which then loses its
On CT, pericardial cysts and diverticula present as water- compliance and impairs diastolic filling of the cardiac cham-
equivalent, round or oval fluid collections surrounded by a thin bers (2224). It also leads to dissociation of intracardiac and
wall without significant wall enhancement at the above-men- intrathoracic pressures during respiration.
tioned typical locations. Pericardial constriction is a rare but potentially curable clini-
cal entity. Pericardiectomy is the only known treatment. Peri-
PERICARDIAL EFFUSION
cardial constriction is most often idiopathic, but it also may
The etiology of pericardial effusion comprises a variety of occur after surgery, radiation therapy, or inflammationpar-
clinical entities such as inflammation (e.g., viral, bacterial, or ticularly after tuberculosis, especially in developing countries.
fungal infections); collagenous and autoimmune disorders In most patients, pericardial constriction presents with insidi-
(e.g., lupus erythematosus); metabolic diseases (e.g., uremia); ous symptoms of venous congestion and is therefore difficult to
tumors; radiation, drug, or toxic reactions; and trauma (1618). diagnose and differentiate from other cardiac diseases. In par-
Within the first days following transmural myocardial infarc- ticular, pericardial constriction may mimic restrictive cardi-
tion, acute hemopericardium and tamponade may occur as a omyopathy, which is why their differentiation has to rely on
result of cardiac rupture, which is associated with a high mor- imaging studies that allow exact assessment of peri- and myo-
tality rate. Subacute cardiac tamponade may occur after even cardial structures.
nontransmural myocardial infarction. Postinfarct pericarditis According to the topographical distribution of pericardial
and Dresslers syndrome may appear from one week to several thickening and calcifications, global forms of constriction can
months after myocardial infarction, but it rarely leads to car- be differentiated from partial forms, in which the pericardium
diac tamponade. Chylopericardium is extremely uncommon is affected over the right or left side of the heart or in the atrio-
and mainly occurs after surgical or traumatic tears of the tho- ventricular grooves (annular constriction) (Fig. 7) (25). In the
racic duct or in association with neoplastic duct stenoses. majority of patients, calcifications are present in the thickened
While rapid accumulation of 150250 mL of fluid may lead pericardium, and often these calcifications extended into the
to cardiac tamponade, much higher volumes can be tolerated pericardial reflection zones at the root of the great vessels. It
without significant hemodynamic disturbances when they col- is particularly important to the surgeon to know the extent and
lect over a longer period of time. the topographic distribution of pericardial fibrosed and calci-
Usually, pericardial effusion is of low density in the range fied plaques, not only to establish the diagnosis but also to
from 0 to 20 Hounsfield Units (HU). When it contains higher determine the optimal approach and sites (sternotomy vs tho-
amounts of protein, such as in bacterial infections, or when it racotomy) for pericardiectomy. Frequently, the calcifications
is hemorrhagic, its density may rise to 50 HU and more. Fig. 6 may invade the right- or left-ventricular myocardium (Fig. 8).
shows two patients with high-density pericardial effusions CT is able to assess such possible myocardial invasion of peri-
resulting from hemopericardium. In inflammation, the pericar- cardial calcifications, which is a high risk factor during peri-
dial layers may show contrast enhancement (16). cardiectomy and which is not easily visible to the surgeon
PERICARDIAL THICKENING AND CONSTRICTION during surgery. Exact visualization of the subepicardial fat layer
on CT helps the surgeon determine the sites of potential resec-
PERICARDIAL THICKENING tability with relatively low risks. Occasionally pericardial con-
On CT studies, the thickness of the normal pericardium is striction is accompanied by myocardial fibrosis or atrophy.
12 mm when measured at the levels of the great vessels or the Knowledge of such areas of myocardial atrophy and fibrosis is
cardiac chambers (9,19,20). Inferiorly, near the diaphragmatic also crucial for the cardiac surgeon, owing to a high risk of
portion of the pericardium, it may appear slightly thicker, owing intraoperative ventricular aneurysms when the pericardium is
to partial-volume effects and the insertion of fibers for its dia- resected over segments of myocardial fibrosis and atrophy (25).
phragmatic attachment. Adequate assessment of pericardial calcifications is important
That is why measurements of pericardial thickness should in the diagnostic work-up of patients with constrictive peri-
be performed at more cranial, midventricular levels. Pericar- carditis, since calcifications are considered to be an indepen-
dial thickening is frequently observed after operation, radiation dent risk factor concerning pericardiectomy (26). Knowledge
therapy, or inflammation, and it may occur with or without of the topographic distribution of calcifications helps surgeons
accompanying effusion and calcifications. On autopsy, regional plan and calculate the risks of pericardiectomy. Secondary CT
thickening and calcifications of the pericardium are frequent signs of pericardial constriction include dilatation of the supe-
findings, and in the majority of cases they are not associated rior/inferior vena cava and of the right/left atrium as well as
with prior symptoms of constriction. Doppman et al. have dem- deformity of ventricular or septal contours.
onstrated that most thickened pericardia they observed on CT
studies of the chest were hemodynamically insignificant (21). TUMORS OF THE PERICARDIUM
It is important to consider that pericardial thickening alone Primary tumors of the pericardium are very rare. Among the
without clinical symptoms of cardiac constriction does not primary benign tumors of the pericardium are lymphangioma,
establish the diagnosis of constrictive pericarditis (22). hemangioma, lipoma, and teratoma (27). The most frequent
150 GROELL

Fig. 6. (A) Hemopericardium (arrows) with hemorrhagic distension of the superior aortic recess of the pericardium. (B) Hemopericardium (long
arrow) resulting from a ruptured ascending aortic aneurysm. Note compression of the right pulmonary artery (small arrows).

primary malignant pericardial tumor is pericardial malignant tastases occur frequently in patients with malignancies, and gen-
mesothelioma; other malignant primary tumors include heman- erally more often than usually recognized. The pericardium may
giosarcoma, fibrosarcoma, malignant teratoma, and liposar- also be directly infiltrated from adjacent tumors, such as carci-
coma. Apart from fat-containing tumors like lipoma and noma of the lung, breast, or esophagus. Apart from solid compo-
liposarcoma, CT usually cannot characterize these tumors; their nents, malignant tumors of the pericardium are often associated
definitive characterization has to rely on biopsy. with hemorrhagic pericardial effusion and possible tamponade.
By far the majority of pericardial tumors are secondary neo- Usually, the amount of effusion does not directly correlate with
plasms (28). The pericardium is frequently involved in tumor volumes. Finally, malignancies of the pericardium may
hematogeneous or lymphatic dissemination from various invade the myocardium and vice versa. Similar to constrictive
extrapericardial malignant tumors, such as malignant lymphoma pericarditis, infiltration of the myocardium is indicated by
or melanoma (Fig. 9). In autopsy examinations, pericardial me- nonvisualization of the subepicardial fat space.
 CHAPTER 14 / CT OF THE PERICARDIUM 151

Fig. 7. Global (A) and annular (B) forms of pericardial constriction (arrows). Dilatation of the atria (A,B) and deformity of ventricular and septal
contours (B) represent secondary CT signs of constriction. Persistent visualization of a subepicardial fat space (A) indicates potential surgical
respectability of calcified plaques.

CONCLUSION
With the advent of CT technologies during the last decade limitations of echocardiography can be overcome by CT,
providing high temporal and spatial resolution, CT gained such as operator dependence, restricted field of view not
more and more influence in imaging the pericardium. The covering the entire pericardium, or nonpenetration of calci-
latest spiral and electron beam CT scanners enable the imag- fied plaques.
ing of the normal pericardium and the most relevant diseases CT is especially important in the examination of patients
of the pericardium with excellent image quality and almost with constrictive pericarditis, since exact knowledge of the
free of motion artifacts. Although echocardiography still rep- extent and location of fibrosed and calcified plaques is cru-
resents the first-line imaging modality in assessing peri- cial to the cardiac surgeon to determine the optimal sites and
cardial morphology and pathology, some of the inherent the possible risks of pericardiectomy.
152 GROELL

Fig. 8. Myocardial infiltration of calcified plaques (arrow) is frequently encountered in constrictive pericarditis. Surgical attempts to remove
such infiltrating plaques are associated with high risks of developing intraoperative ventricular aneurysms.

Fig. 9. Pericardial metastasis (arrow) from malignant lymphoma.

REFERENCES
the presence of acute pulmonary hypertension. Ann Thorac
1. Nataf P, Cacoub P, Dorent R, et al. Results of subtotal pericardiec- Cardiovasc Surg 1998;4:332335.
tomy for constrictive pericarditis. Eur J Cardiothorac Surg 1993;7: 8. Vesely TM, Cahill DR. Cross-sectional anatomy of the pericardial
252255. sinuses, recesses and adjacent structures. Surg Radiol Anat 1986;
2. Shabetai R. Anatomy. In: The Pericardium. Grune and Stratton, New 8:221227.
York, NY: 1981;132. 9. Groell R, Schaffler GJ, Rienmueller R. The pericardial sinuses and
3. Moncada R, Baker M, Salinas M, et al. Diagnostic role of com- recesses: findings on ECG-triggered electron-beam CT studies.
puted tomography in pericardial heart disease: congenital defects, Radiology 1999;212:6973.
thickening, neoplasms, and effusions. Am Heart J 1982;103: 10. Kubota H, Sato C, Ohgushi M, Haku T, Sasaki K, Yamaguchi K.
263282. Fluid collection in the pericardial sinuses and recesses: thin-section
4. Hoit BD. Imaging the pericardium. Cardiol Clin 1990;8:587600. helical computed tomography observations and hypothesis. Invest
5. Shabetai R. Function of the pericardium. In: The Pericardium. Grune Radiol 1996;31:603610.
and Stratton, New York, NY: 1981;33107. 11. Suolen RL, Lapayowker MS, Cortes FM. Distribution of pericardial
6. Freeman GL. The effects of the pericardium on function of normal fluid: dynamic and static influences. AJR Am J Roentgenol 1968;
and enlarged hearts. Cardiol Clin 1990;8:579586. 103:583588.
7. Borrego JM, Ordonez A, Gutierrez E, et al. Integrity of the pericar- 12. Shabetai R. Congenital malformations of the pericardium. In: The
dium. Its beneficial effects on the protection of the right ventricle in Pericardium. Grune and Stratton, New York, NY: 1981;366373.
 CHAPTER 14 / CT OF THE PERICARDIUM 153

13. Jacob JL, Souza Jr. AS, Parro A, Jr. Absence of the left pericar- 21. Doppman JL, Rienmuller R, Lissner J. Computed tomography in
dium diagnosed by computed tomography. Int J Cardiol 1995;47: constrictive pericarditis. J Comput Assist Tomogr 1981;5:111.
293296. 22. Fowler NO. Constrictive pericarditis: its history and current status.
14. Olson MC, Posniak HV, McDonald V, Wisniewski R, Moncada R. Clin Cardiol 1995;18:341350.
Computed tomography and magnetic resonance imaging of the peri- 23. Myers RB, Spodick DH. Constrictive pericarditis: clinical and patho-
cardium. Radiographics 1989;9:633649. physiologic characteristics. Am Heart J 1999;138:219232.
15. Kutlay H, Yavuzer I, Han S, Cangir, AK. Atypically located pericar- 24. Ling LH, Oh JK, Schaff HV, et al. Constrictive pericarditis in the
dial cysts. Ann Thorac Surg 2001;72:21372139. modern era: evolving clinical spectrum and impact on outcome after
16. Hackney D, Slutsky RA, Mattrey R, et al. Experimental pericardial pericardiectomy. Circulation 1999;100:13801386.
inflammation evaluated by computed tomography. Radiology 25. Rienmller R, Grgan M, Erdmann E, Kemkes BM, Kreutzer E,
1984;151:145148. Weinhald C. CT and MR evaluation of pericardial constriction: a
17. Breen JF. Imaging of the pericardium. J Thorac Imaging 2001;16: new diagnostic and therapeutic concept. J Thoracic Imaging 1993;8:
4754. 108121.
18. Suchet IB, Horwitz TA. CT in tuberculous constrictive pericarditis. 26. Ling LH, Oh JK, Breen JF, et al. Calcific constrictive pericarditis:
J Comput Assist Tomogr 1992;16:391400. is it still with us? Ann Intern Med 2000;132:444450.
19. Delille JP, Hernigou A, Sene V, et al. Maximal thickness of the 27. McAllister HA, Jr. Tumors of the heart and pericardium. In: Silver
normal human pericardium assessed by electron-beam computed MD (ed.) Cardiovascular Pathology. Churchill Livingstone, New
tomography. Eur Radiol 1999;9:11831189. York, NY: 1983:909943.
20. Ling LH, Oh JK, Tei C, et al. Pericardial thickness measured with 28. Chiles C, Woodard PK, Gutierrez FR, Link KM. Metastatic involve-
transesophageal echocardiography: feasibility and potential clinical ment of the heart and pericardium: CT and MR imaging. Radio-
usefulness. J Am Coll Cardiol 1997;29:13171323. graphics 2001;21:439449.
 CHAPTER 15 / MDCT OF CARDIAC VALVES 155

15 Multidetector-Row CT
for Assessment of Cardiac Valves

JRGEN K. WILLMANN, MD AND DOMINIK WEISHAUPT, MD

INTRODUCTION cardiac valves. In patients with stenotic aortic and mitral dis-
Imaging of the heart with computed tomography (CT) is ease, precise quantification of mean and maximum trans-
challenging because the heart is continuously moving during valvular pressure gradients as well as determination of valve
data acquisition. As a result of the limited temporal resolution, area can be performed. Alternatively, transesophageal echo-
the use of single-detector helical CT for noninvasive cardiac cardiography can be performed, in particular in patients with
imaging was limited and resulted often in images with a high poor transthoracic acoustic windows, such as patients with large
content of artifacts. The introduction of multidetector-row CT and thick chest walls, small hearts, chest deformities, or in
(MDCT) scanners by the end of 1998 laid the foundation for elderly patients. Transthoracic as well as transesophageal
increased clinical use of CT for cardiac imaging. Partial view echocardiography allow assessment of the entire spectrum of
acquisition and retrospectively electrocardiogram (ECG)-gated valvular disease, in particular visualization of infective endo-
helical reconstruction offered by 4-channel MDCT scanners, carditis and its sequelae, including perivalvular abscesses.
allow for a temporal resolution of up to 125 ms, combined with However, both transthoracic and transesophageal echocardio-
both a high spatial resolution and a high signal-to-noise ratio graphy permit only a rough quantification of valvular calcifica-
(1). Last-generation 16-channel MDCT scanners permit scan- tion, as only indirect signs of calcification, including increased
ning of the whole heart with an even higher temporal resolution echogenicity and thickening of the valves, may be used for
of up to 105 ms within a convenient short breath-hold of about quantification (8).
18 s (2). Apart from visualization of coronary artery lumen and
MDCT
stenosis as well as detection and quantification of coronary
calcification (3,4), these technical developments in CT tech- RATIONALE
nology also improved visualization of morphological details of Aortic valve stenosis is the most common cardiac valve
the heart including the cardiac valves (5,6). lesion in the developed countries, with a prevalence of 2% to
In this chapter, we describe the potential clinical and research 7% in the population above 65 yr of age (9). Degeneration and
applications of MDCT, with special focus on the assessment of calcification of the aortic valve cusps and the aortic annulus are
the aortic and mitral valves. the most common cause of aortic valve stenosis (10). Several
studies have identified the presence and extent of aortic valve
CONVENTIONAL IMAGING TECHNIQUES calcification as a strong predictor both for the progression as
CONVENTIONAL ANGIOGRAPHY well as for the outcome of aortic valve stenosis (11,12). The
Conventional angiography allows assessment of functional positive therapeutic effect of a lipid-lowering pharmacologi-
data, including measuring the transvalvular pressure gradient cal therapy with HMG-CoA reductase inhibitors on the natural
as well as calculating the valve area using the Gorlin formula history of calcific aortic valvular disease has been demon-
(7). However, in patients with valvular regurgitation, conven- strated recently (13). Therefore, an imaging modality for quan-
tional angiography has been demonstrated to be inaccurate. In tification of aortic valve calcium in patients with aortic valve
addition, conventional angiography does not allow precise stenosis is desirable. From a surgical point of view, preopera-
quantification of valvular calcification. tive knowledge about anatomic details of the aortic valve is
TRANSTHORACIC AND TRANSESOPHAGEAL of particular interest, since the procedure of aortic valve sur-
ECHOCARDIOGRAPHY gery itself and the choice of the type of valve prosthesis to be
Transthoracic echocardiography is a widely available, non- implanted depend on various parameters, including the aortic
invasive, and safe imaging modality which allows an expedi- valve morphology (tricuspid vs bicuspid), its diameter, as well
tious assessment of the anatomic and functional status of the as the presence and extent of aortic valve and annulus calcifi-
cation (1416). Moreover, the presence of aortic valve calcifi-
From: Contemporary Cardiology: CT of the Heart: cation extending to the interventricular septum as a surgical
Principles and Applications finding at aortic valve replacement predicts the need for per-
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
manent pacing postoperatively (17). Similarly, two studies

155
156 WILLMANN AND WEISHAUPT

have demonstrated that mitral annulus calcification may be a region of interest (ROI) with an area of 1015 mm2 is placed
used as a marker for the severity of atherosclerosis and may in the lumen of the ascending aorta. This ROI serves as a ref-
be helpful to identify patients with a higher likelihood of dif- erence for the following dynamic measurements of contrast
fuse vascular atherosclerotic processes (18,19). enhancement after administration of 120 mL of nonionic iodi-
Since CT is a sensitive and objective method for the detection nated contrast medium. Repetitive low-dose monitoring scans
of calcifications, imaging of the aortic and mitral valve with (120 kV; 10 mAs; 0.5 s scanning time; 1 s interscan delay) are
MDCT combines both assessment of aortic and mitral valve then performed 10 s after the beginning of the contrast material
morphology and quantification of aortic valve and mitral valve injection. After reaching the preset contrast enhancement level
annulus calcifications (5,6). of 100120 HU, the MDCT scan is initiated automatically.
TECHNICAL CONSIDERATIONS VALVULAR MORPHOLOGY
MDCT for assessment of the cardiac valves may be per- MDCT allows reliable differentiation between bicuspid and
formed either with or without administration of intravenous tricuspid aortic valves. In a prospective study of 25 patients
contrast agents. Usually, for assessment of valvular morphol- with aortic valve stenosis, Willmann et al. (5) have shown
ogy, a contrast-enhanced CT scan is performed. For mere that contrast-enhanced MDCT is highly accurate for prediction
determination of valvular calcification, a nonenhanced CT scan of valve morphology when compared to surgery and echo-
is sufficient. cardiography. Using contrast-enhanced MDCT there was a
The scan is planned on a low-dose anteroposterior scout 100% agreement between MDCT and surgical or echocardio-
view (120 kV, 50 mAs) from the ascending aorta to the apex of graphic findings with regard to the morphology of the aortic
the heart. Four-channel MDCT acquisition is performed with valve (Fig. 1) (5). On nonenhanced MDCT data sets, the aortic
4 1-mm collimation (1.25-mm slice width for reconstruc- valve could be correctly classified in 87% of the cases (5).
tion), 500-ms rotation time, table feed of 2.54.5 mm per MDCT also allows for a precise measurement of the diameter
rotation, 300 mA at 120 kV, and a 0.6- to 0.8-mm image recon- of the aortic valve annulus (Fig. 2). In the same study by
struction increment. The acquisition parameters for a 16-chan- Willmann et al. (5), the diameter of the aortic valve annulus
nel MDCT acquisition include a 12 0.75-mm collimation as measured on contrast-enhanced 4-channel MDCT images
(0.75-mm slice width for reconstruction), a table feed of correlated highly with the intraoperative measurement with
6.7 mm/s, a gantry rotation time of 420 ms, and a reconstruction only an overestimation by 0.7 mm on contrast-enhanced MDCT
increment of 0.4 mm. For retrospective reconstruction of the data sets (5).
MDCT data set, a digital ECG file from the patient is simulta- The morphology of the mitral valve and its apparatus,
neously recorded during MDCT scanning. Image reconstruc- including the mitral valve annulus, mitral valve leaflets, tendi-
tion is performed with a medium-sharp body convolution kernel nous cords, and papillary muscles can also be visualized on
in the mid- to end-diastolic phase of the cardiac cycle using MDCT (Fig. 3). Recently, our group reported on the prelimi-
between 50% and 70% relative delay to the R waves of the ECG nary experience with mitral valve imaging in 20 patients (6).
signal. With 4-channel MDCT scanning, this results in a voxel Good-to-excellent image quality of the mitral valve annulus
size of 0.35 0.35 1.25 mm3 (field of view [FOV], 18 cm; and its leaflets were obtained in 15 of 20 consecutive patients
matrix size, 512 512 pixels). For 16-channel MDCT, the (75%). In 19 of 20 patients (95%) papillary muscles could also
resulting voxel size is 0.35 0.35 0.75 mm3. For data analy- be visualized to good or excellent advantage. However, visibil-
sis, the reconstructed MDCT data are best transferred to an ity of tendinous cords was inferior. In 14 of 20 patients (70%)
independent workstation, which allows multiplanar reconstruc- tendinous cords were not or only moderately visible (6).
tions in sagittal, coronal, and oblique planes. For assessment of VALVULAR CALCIFICATION
the valvular morphology and quantification of valvular cal- MDCT is in particular helpful for the assessment of aortic
cium, routine reconstruction of volume renderings or multiplanar valve calcification. According to the echocardiographic grad-
reformations is not necessary. ing of aortic valve calcifications (11), we use a four-point
If the MDCT scanning is performed with intravenous grading scale for assessment of the degree of the calcification
administration of an iodinated contrast agent, we first deter- of the aortic valve. This grading scale is as follows: grade 1,
mine the optimal scan delay using the test bolus method. For no calcification; grade 2, mild calcification (small, isolated
this purpose, a test bolus of 20 mL of iodinated contrast agent spots of calcification); grade 3, moderate calcification (mul-
followed by a 50-mL chaser of saline is injected by a power tiple larger spots of calcification); and grade 4, heavy calci-
injector. Delay times are determined by visually evaluating the fication (extensive calcification of all aortic valve leaflets)
contrast material at the level of the aortic valve by using ten (Figs. 4 and 5). By using this grading scale there was an 84%
consecutive transverse images obtained every 2 s without table agreement between contrast-enhanced MDCT findings and
feed. The MDCT scan is then performed after administering the true calcification status of the aortic valves as assessed
120 mL of nonionic contrast material through a 20-gage needle during surgery (5).
placed in an antecubital vein at 35 mL/s followed by a 50-mL As with the aortic valve, the calcifications of the mitral valve
saline chaser bolus. Alternatively, the delay time can be may be located either on the mitral valve annulus or on the
obtained by using a bolus-tracking technique (e.g., CARE- mitral valve leaflets. MDCT yielded a 95100% agreement
Bolus, Volume Zoom Navigator, Siemens). For this purpose, a compared to echocardiography and intraoperative findings with
single nonenhanced low-dose scan (10 mAs) at the level of the regard to assessment of calcifications of the mitral valve annu-
ascending aorta is performed. Based on this transverse image, lus and mitral valve leaflets (6).
 CHAPTER 15 / MDCT OF CARDIAC VALVES 157

Fig. 2. Precise assessment of aortic valve diameter is possible with

contrast-enhanced, retrospectively electrocardiogram-gated
multidetector-row CT (MDCT). Preoperative assessment was per-
formed on a sagittal oblique reconstruction of the heart in this 63-yr-
old female with aortic valve stenosis. There was only a minimal
difference between the diameter as assessed by MDCT and the diam-
eter as measured during surgery.

Fig. 1. (A) 57-yr-old male with severe aortic valve stenosis who
underwent contrast-enhanced retrospectively electrocardiogram-
gated multidetector-row CT (MDCT) before surgery. Double oblique
reconstruction shows heavily calcified bicuspid aortic valve with large
calcific deposit (arrows). (B) Image obtained during surgery shows
excellent agreement between both MDCT and intraoperative status.
Note presence of bicuspid aortic valve with large deposits of calcium
at the free edges of the aortic valve (arrows).

All these findings imply the potential of MDCT for a longi-

tudinal monitoring of patients with aortic and mitral valve cal-
cifications by measuring the change of valvular calcification Fig. 3. 73-yr-old female with mitral regurgitation. Long-axis view
over time. Apart from a visual semi-quantitative assessment of reconstruction contrast-enhanced, retrospectively electrocardiogram-
gated multidetector-row CT data set demonstrates anterior mitral
valvular calcification, future studies will demonstrate whether valve annulus (large arrow), mitral valve leaflets (small arrows), ten-
MDCT imaging may also allow an absolute quantification of dinous cord (arrowhead), and the anterior papillary muscle (P). LV,
valvular calcium. In vitro studies with absolute determination left ventricle; LA, left atrium; AA, ascending aorta.
158 WILLMANN AND WEISHAUPT

Fig. 4. Diagrams of different grades of aortic valve calcification (modified from ref. 5). (A) Grade 1, no calcification; (B), grade 2, mild
calcification (small isolated spots of calcification); (C), grade 3, moderate calcification (multiple larger spots of calcification); and (D), grade
4, heavy calcification (extensive calcification of all aortic valve leaflets).

of aortic valve calcium by ashing or enzymatic digesting of A second limitation includes the radiation dose, which is
noncalcified areas of the aortic valve and subsequent measure- inherent to CT scanning. Since the data are acquired with an
ment of the mass of the remaining aortic valve calcium are overlapping helical pitch and continuous radiation exposure,
warranted to correlate the real amount of calcium and the cal- there is a considerable applied radiation dose. Mean effective
cium obtained on MDCT. Furthermore, the reliability of differ- dose values of up to 13 mSv have been calculated for imaging
ent calcium scores used for quantification of coronary artery the entire heart using a 4-channel MDCT scanner (23). Com-
calcium need to be evaluated when quantifying aortic valve pared to the mean effective dose values calculated for conven-
calcium, including the Agatson score (20), the volume score tional coronary angiography (usually between 2.1 and 2.5 mSv),
(21), the mass equivalent score, as well as the calibrated min- the mean effective dose caused by MDCT angiography is higher
eral mass score (22). by a factor of about 5 (23). However, by reducing the tube
LIMITATIONS OF MDCT IMAGING output during heart phases that are not likely to be targeted
So far, MDCT imaging has demonstrated its capability only by the ECG-gated reconstruction (i.e., reconstruction intervals
for evaluation of the morphology of mitral and aortic valves, excepting 5070% of the cardiac cycle), a dose reduction of up
including the semi-quantitative assessment of calcifications. to 48% is possible (24). Furthermore, improvements of dose
As a result of the limited temporal resolution of 4-channel utilizations of recent generation 16-channel MDCT scanners
MDCT scanners, functional assessment of cardiac valves as may also help in reducing radiation dose (25).
well as quantification of valvular area, in particular of the aortic CONCLUSION AND FUTURE DIRECTIONS
valve, has not been possible yet. With the introduction of MDCT is an emerging tool for noninvasive imaging of the
16-channel scanners, as well as further reduction of the rotation heart. Apart from evaluation of the coronary arteries, a compre-
time of the CT tube, additional improvement of spatial and hensive work-up of the heart also includes assessment of mor-
temporal resolution may also allow accurate assessment of phological details of the heart chambers, including cardiac
valvular area. valves. Preliminary studies have shown that the aortic and
 CHAPTER 15 / MDCT OF CARDIAC VALVES 159

mitral valves can be visualized with high quality. At this point,

MDCT of the aortic valve seems to be of special interest for the
assessment of aortic valvular calcification. Future studies are
warranted to exploit the potential of MDCT for evaluation of
the pulmonary and tricuspidal valves as well as to establish
tools for quantitative analysis of valvular function and quanti-
fication of the calcium deposits.
REFERENCES
1. Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of
electrocardiographically gated multisection spiral CT: initial expe-
rience. Radiology 2000;217:564571.
2. Flohr T, Bruder H, Stierstorfer K, Simon J, Schaller S, Ohnesorge B.
New technical developments in multislice CT, part 2: sub-millime-
ter 16-slice scanning and increased gantry rotation speed for car-
diac imaging. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr
2002;174:10221027.
3. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary
artery stenoses by contrast-enhanced, retrospectively electrocardio-
graphically-gated, multislice spiral computed tomography. Circula-
tion 2001;103:25352538.
4. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium
measurement: agreement of multirow detector and electron beam
CT. AJR Am J Roentgenol 2001;176:12951298.
5. Willmann JK, Weishaupt D, Lachat M, et al. Electrocardiographi-
cally gated multi-detector row CT for assessment of valvular mor-
phology and calcification in aortic stenosis. Radiology 2002;225:
120128.
6. Willmann JK, Kobza R, Roos JE, et al. ECG-gated multi-detector
row CT for assessment of mitral valve disease: initial experience.
Eur Radiol 2002;12:26622669.
7. Gorlin R, Gorlin SG. Hydraulic formula for calculation of the area
of the stenotic mitral valve, other valves and central circulatory
shunts. Am Heart J 1951;41:129.
8. Otto CM, Lind BK, Kitzman DW, Gersh BJ, Siscovick DS. Associa-
tion of aortic-valve sclerosis with cardiovascular mortality and
morbidity in the elderly. N Engl J Med 1999;341:142147.
9. Stewart BF, Siscovick D, Lind BK, et al. Clinical factors associated
with calcific aortic valve disease. Cardiovascular Health Study.
J Am Coll Cardiol 1997;29:630634.
10. Otto CM, Burwash IG, Legget ME, et al. Prospective study
of asymptomatic valvular aortic stenosis. Clinical, echocardio-
graphic, and exercise predictors of outcome. Circulation 1997;95:
22622270.
11. Rosenhek R, Binder T, Porenta G, et al. Predictors of outcome
in severe, asymptomatic aortic stenosis. N Engl J Med 2000;343:
611617.
12. Otto CM. Timing of aortic valve surgery. Heart 2000;84:211218.
13. Shavelle DM, Takasu J, Budoff MJ, Mao S, Zhao XQ, OBrien KD.
HMG CoA reductase inhibitor (statin) and aortic valve calcium.
Lancet 2002;359:11251126.
14. Adams DH, Chen RH, Kadner A, Aranki SF, Allred EN, Cohn LH.
Impact of small prosthetic valve size on operative mortality in eld-
erly patients after aortic valve replacement for aortic stenosis: does
gender matter? J Thorac Cardiovasc Surg 1999;118:815822.
15. Delius RE, Samyn MM, Behrendt DM. Should a bicuspid aortic
valve be replaced in the presence of subvalvar or supravalvar aortic
stenosis? Ann Thorac Surg 1998;66:13371342.
16. Luciani GB, Casali G, Mazzucco A. Risk factors for coronary com-
plications after stentless aortic root replacement. Semin Thorac Car-
diovasc Surg 1999;11:126132.
17. Boughaleb D, Mansourati J, Genet L, Barra J, Mondine P, Blanc JJ.
[Permanent cardiac stimulation after aortic valve replacement: inci-
Fig. 5. Two examples of two different grades of aortic valve calcifi- dence, predictive factors and long-term prognosis]. Arch Mal Coeur
cations. Double-oblique reconstruction of multidetector-row CT Vaiss 1994;87:925930.
angiographic data set demonstrates small isolated spots of calcifica- 18. Adler Y, Fink N, Spector D, Wiser I, Sagie A. Mitral annulus calci-
tion (grade 2, [A]) as well as multiple larger spots of calcification ficationa window to diffuse atherosclerosis of the vascular sys-
(grade 3, [B]). tem. Atherosclerosis 2001;155:18.
160 WILLMANN AND WEISHAUPT

19. Tenenbaum A, Shemesh J, Fisman EZ, Motro M. Advanced mitral and contrast-enhanced multi-detector row CT studies. Radiology
annular calcification is associated with severe coronary calcifica- 2002;223:474480.
tion on fast dual spiral computed tomography. Invest Radiol 2000; 23. Hunold P, Vogt FM, Schmermund A, et al. Radiation exposure
35:193198. during cardiac CT: effective doses at multi-detector row CT and
20. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, Jr., electron-beam CT. Radiology 2003;226:145152.
Detrano R. Quantification of coronary artery calcium using ultrafast 24. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of
computed tomography. J Am Coll Cardiol 1990;15:827832. the heart with retrospective ECG gating: reduction of radiation
21. Callister TQ, Cooil B, Raya SP, Lippolis NJ, Russo DJ, Raggi P. exposure by ECG-controlled tube current modulation. Eur Radiol
Coronary artery disease: improved reproducibility of calcium scor- 2002;12:10811086.
ing with an electron-beam CT volumetric method. Radiology 1998; 25. Flohr T, Stierstorfer K, Bruder H, Simon J, Schaller S. New techni-
208:807814. cal developments in multislice CTPart 1: approaching isotropic
22. Hong C, Becker CR, Schoepf UJ, Ohnesorge B, Bruening R, Reiser resolution with sub-millimeter 16-slice scanning. Rofo Fortschr Geb
MF. Coronary artery calcium: absolute quantification in nonenhanced Rontgenstr Neuen Bildgeb Verfahr 2002;174:839845.
 CHAPTER 16 / MSCT IN CHD PATIENTS 161

16 Multislice CT of the Heart and Great Vessels

in Congenital Heart Disease Patients

JEAN-FRANOIS PAUL, MD

INTRODUCTION Magnetic resonance imaging (MRI) and computed tomogra-

Congenital heart disease (CHD) is defined as abnormal phy (CT) have been proposed for this purpose: MRI has been
development of heart structures during intrafetal life, resulting shown to be helpful, and this modern technique is recommended
in more or less complex malformations of the heart and great as a first-line technique by the task force report for many CHD
vessels. Many patients benefit by surgical intervention or per- entities (1). At first look, MRI seems to be an ideal technique
cutaneous interventional radiological procedures to correct the because of its absence of radiation burden, which is a very
anatomical anomalies. These interventions have become more important issue especially in neonates or young children. Some
and more successful, allowing for longer life spans for patients limitations are important, however: the major limitation is the
with CHD, and thus enabling most of them to reach adulthood. long sedation and close monitoring required, especially in
For the clinical management of patients with complex CHD, infants with cyanotic heart disease with often unstable condi-
3D accurate evaluation of their morphologic conditions is criti- tions. Pediatric intensive care physicians must be present dur-
cal. Three-dimensional imaging should demonstrate the shape ing MRI examinations. Spatial resolution obtained with MRI is
and spatial relation of great arteries, proximal branch pulmo- lower than resolution of CT images, which may be a critical
nary arteries, and anomalous pulmonary venous or systemic issue for visualization of small anatomical structuresfor
connections. The 3D information of extra-cardiac morphologic example, coronary arteries.
characteristics may determine the choice of surgical interven- More recently, helical CT has been proposed as the tool for
tion technique. Usually, echocardiography and angiography 3D anatomical visualization in patients with CHD (2). Helical
are used for assessment of diagnosis and as a guide for treat- technology brings volume acquisition in a short period of time,
ment. Echocardiography is the major imaging technique for the making good 3D vascular image quality possible, even in neo-
exploration of the intracardiac structures. High-resolution, real- nates or infants. Multislice CT technology now available brings
time images depict the intracardiac anomalies, and Doppler much faster acquisition time, which drastically reduces respi-
ultrasounds may demonstrate, for example, atrial or septal ratory artifacts, as well as a potential reduction of heart motion
defects by direct visualization of the defect and associated with the new possibility of image synchronization with the
abnormal blood flows. Valve positions and dysfunctions are cardiac rhythm.
also clearly detectable, and anomalies of the great-vessel ori- In our surgical center specialized in CHD, multislice CT has
gins are recognizable. However, extracardiac anomalies are rapidly become an important complementary imaging tech-
very frequently associated with intracardiac anomalies. nique for pre- or postoperative management of patients. The
Echocardiography may be insufficient for visualisation of aim of this chapter is to present our initial experience over a
extracardiac anatomy, especially for assessment of abnormal 2-yr period, using first a 4-slice and then a 16-slice CT.
pulmonary or systemic venous connections, pulmonary steno-
sis diagnosis, or aortic arch anomalies. TECHNICAL ASPECTS: GENERALITIES
Percutaneous angiography is the usual complementary Compared to previous helical CT, multislice CT allows
imaging method for assessment of pulmonary artery status, faster acquisition time, thinner slice thickness, and electrocar-
aorta and branches morphology including coronary arteries, diogram (ECG)-gated acquisition.
and venous connections. In addition, blood pressure in the The first issue is whether one should use ECG-gated acqui-
different cardiac structures is directly measurable with percu- sition in CHD patients; the second is which protocol is best
taneous catheters. However, angiography is a plane imaging suited: in particular, the radiation dose delivered should be
technique, and the major drawback is the absence of direct 3D estimated for each protocol in order to minimize radiation
visualization of the abnormal vascular structures in CHD. exposure.
In our center, the first criterion we consider for the choice of
From: Contemporary Cardiology: CT of the Heart: the protocol is whether an apnea is possible or not during the
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
acquisition. Generally, it is not possible until 6 or 7 yr of age.

161
162 PAUL

In children between ages 6 and 12, we usually test the breath- diagnostic examinations, although representing 4% of those
hold two or three times before the acquisition, in order to ensure studies (5). Multislice CT offers even more diagnosis capa-
thoracic immobility during the scanning process. bilities but tends to increase the radiation dose as a result of
NEONATES AND YOUNG INFANTS routine use of thinner slice thicknesses, extension of the vol-
In neonates or young infants, breath-hold is not possible, ume of acquisition, or multiple phase acquisitions. Following
and we do not use ECG-gated acquisition, for four main reasons: the As Low As Reasonably Achievable (ALARA ) principle,
1. Respiratory artifacts are responsible for substantial degra- dose reduction is necessary but examination quality must be
dation of images. These respiratory artifacts cause more preserved without losing diagnostic information. The thorax is
important artifacts than heart motion does. ECG-gated a low-attenuation region, though substantial dose reduction dur-
acquisition is much slower than nongated acquisitions, ing chest CT is feasible because of the high inherent contrast.
causing more respiratory artifacts. In August 2001, the ALARA conference of the Society for
2. In babies, normal heart rate is over 100 beats per minute Pediatric Radiology raised the question of dose reduction by
(bpm), which is too fast to prevent any heart motion arti- reduction of kilovoltage (6).
facts. First reports on cardiac CT showed that artifacts are In our center, we decided to apply the ALARA principle as
frequent in adults when cardiac rhythm is over 64 (4-slice far as possible in neonates and babies with CHD, and then apply
technology) or 70 bpm (16-slice technology) (3). In babies
some systematic rules:
with cyanotic CHD, cardiac rhythm is much higher, gen-
erally between 140 and 180 bpm. No preview scan (responsible for unnecessary additional
3. Generally, the clinical question concerns extracardiac radiation dose). Scan length is calculated with laser.
anatomy. Images of extracardiac structures are less sensi- Systematic use of 80-kV settings.
tive to heart motion than heart images are. Adaptation of the mAs to childs weight.
4. ECG-gated cardiac retrospective acquisitions require Only one phase acquisition when possible.
much higher radiation doses than nongated thoracic CT, Systematic protection of nonscanned organs.
because only a part of the dose (i.e., dose delivered during
diastole) is used for creating images. If a standard cardiac 80-kV protocols have not been developed yet, although in
adult protocol were applied for a 3-kg baby, dose delivered our experience, it is possible to scan the thorax in adults of less
would be about 50 times more than the dose we currently than 75 kg without substantial loss of image quality. Reducing
use in these conditions. In babies, organ sensitivity to the kilovoltage from 120kVp to 80 kVp decreases the radiation
radiation is much higher than in adults. In addition, better dose by 65% at constant current setting, as radiation dose varies
imaging quality is not guaranteed, because of the two first with the square of kV. This setting is sufficient for good quality
points mentioned above, and a risk of developing cancer in images, as long as the mAs are adjusted according to the childs
the future could not be totally ruled out (4).
weight.
In our experience, the principle go as fast as possible was We always use 80 kV as the standard kilovoltage setting in
adequate to get good image quality in neonates with CHD. children. Current exposure is adapted to the body weight in
Short acquisition time is the best way to minimize respiratory neonates and infants, using the following rule:
artifacts. With a 4-slice CT (Volume Zoom, Siemens), we cur-
3 kg: 30 mAs
rently scan the thoraxes of babies in 3 or 4 s, using 2.5-mm
4 kg: 35 mAs
collimation and a table speed of 20 mm/s (pitch 8). The images
5 kg: 40 mAs
are of superior quality to those using 1-mm slice thickness, 6 kg: 45 mAs
which also require longer acquisition time and cause more res-
piratory artifacts. Very short acquisition times (5 s or less) Using this protocol, radiation exposure is estimated under
make apnea possible in intubated babies. In these cases, images 1 mSv for a neonate. 1 mSv in equivalent to less than 6 mo
were found free from respiratory artifacts. With the 16-slice natural radiation. Even lower dose may be used in premature
CT, we may choose now either to scan the thorax of a baby in babies or when high quality images are not mandatory for
about 4 s using 0.75-mm slices or in 2 s using 1.5-mm collimation. diagnosis. 80 kV and 17 mAs is the minimum exposure setting
CHILDREN OVER SEVEN YEARS OF AGE, AND ADULTS allowed with the 16-slice CT from Siemens (Sensation 16). A
In this group of older children and young adults, two options previous study compared radiation dose with electron beam CT
are possible: (1) breath-hold angio-CT acquisition; or (2) ECG- (EBCT) and conventional angiography, which were favourable
gated acquisition. In this group, the protocol must be chosen from far to EBCT (7) (25 to 50 times less skin exposure than
depending on the clinical questions. If, for example, coronary with conventional angiography as estimated by the authors).
visualization is required to look for a possible anatomical vari- Anatomical data acquired from CT may be judiciously used to
ant, ECG-gated protocols are recommended. In the other cases, limit the number of views acquired with angiography, and
ECG gating is often not mandatory. In any case, one should sometimes replace conventional angiography. CT may be then
always keep in mind that the radiation dose delivered is much advantageous to reduce global radiation exposure in CHD
higher when using ECG-gated acquisition. patients.
DOSE CONSIDERATIONS The other advantage of 80 kV settings is the possibility for
Radiation exposure is a major public health issue. CT con- reduction of amount of contrast medium injection, because low
tributes greatly to the population dose arising from medical kilovoltage is more sensitive to contrast (iodine has a high
exposure, accounting for 35% of the dose delivered during atomic number) than standard 120 or 140 kV settings.
 CHAPTER 16 / MSCT IN CHD PATIENTS 163

INJECTION PROTOCOL
Dose injection must be adapted to the babys weight. We
currently use 2 cc per kg in our institution, which was sufficient
in all cases encountered. Sometimes CT is performed just after
conventional angiography, in order to benefit from central
venous access. The limit of the total contrast amount of the two
procedures may be up to of 8 cc/kg without complication in our
experience.
Basic protocol of injection for pulmonary arteries or sys-
temic vascular enhancement:
Using 80 kV, rate of injection can be as low as 0.6 cc per
second in neonates with catheter placed in the vein of the hand.
Higher rate may be used in case of central catheter (femoral or
jugular). Power injector is currently used to ensure a continous
and regular flow rate. Our current rules used for rate of injection
are as follows:
3 kg: 30 mAs
4 kg: 35 mAs
5 kg: 40 mAs
6 kg: 45 mAs
The start delay in neonates and infants is 15 s for peripheral Fig. 1. Pulmonary atresia with ventricular septal defect: Multislice
injection, 10 s for central venous injection. To be sure of CT showed small but confluent pulmonary arteries (seagull sign) in
a 8-yr-old patient. These pulmonary arteries were not seen with con-
having vascular contrast during the acquisition, we sometimes ventional angiography. This finding has been of fundamental impor-
slightly increase the amount of contrast medium in order to tance for surgical decision (staged repair). Complete repair was finally
follow this rule: obtained with normalization of blood oxygen saturation.
Time of injection = start delay + time of acquisition
Using this rule, acquisition is never too late for good vas-
cular enhancement, because acquisition ends with the end of edge of baby management and a quiet attitude is of primary
injection, so contrast medium is still in the peripheral veins importance.
when acquisition ends. Our sedation protocol in infants includes intrarectal: mid-
Precautions for Venous Access azomal at 0.3 mg/kg, given 15 min before examination. Addi-
Peripheral venous access is performed in the pediatric unit. tional sedative drugs may be useful (hydroxyzine at 1 mg/kg,
Right arm injection is preferable (but not mandatory) to avoid per os, 1 h before examination). With experienced technolo-
striking artifacts on the innominate left brachio-cephalic vein. gists, mean total examination time in the CT room is 20 min.
In some cases, venous connections are congenitally different or Qualified medical monitoring may sometimes be necessary
surgically modified. This information, when available, is during the examination, depending on the clinical conditions of
important before the scan procedure, as it may change the scan the babies. In all cases, oxygen saturation should be closely
injection protocol. Venous visualization may be realized at first monitored.
pass, with a high concentration of contrast medium, or some-
times later, at the time of venous return. The optimal injection ANATOMICAL ASSESSMENT
protocol depends on each particular venous anatomy. Although CHDs may be a difficult topic for radiologists
Catheter permeability is checked before injection. It is who are not familiar with the exploration of these pathologies,
essential to avoid any air injection during the scan procedure. the main clinical question for which the CT is performed is
All bubbles should be removed when connecting catheter to often simple, making interpretation of the image accessible to
power injector. Because many patients with CHD have right to any radiologist. Examples of questions include: What is the
left shunt, air injection through venous access could cause air status of the pulmonary artery (present, absent, or stenotic)?
systemic embolism, with possible fatal consequences. What is the size of the aorta? Where are the collateral arteries
Sedation originating? Do these collaterals cross behind or beyond the
General anesthesia is never necessary in our experience. In central airways? Where are the pulmonary venous connec-
neonates we do not use any sedative drugs. In infants, we rec- tions situated? We identified several main follow-up ques-
ommend oral or intrarectal sedation (or both) before CT proce- tions to be raised, which include: pulmonary artery status and
dure, to prevent baby agitation during the acquisition, which morphology, coronary artery origins, aorta and collaterals,
may be responsible for poorer image quality and occasionally suspicion of airway compression of vascular origin, suspicion
require re-examination. Sedation is not always mandatory if of anomalous venous return, and postoperative evaluation.
the baby is quiet. Experienced technologists are necessary in Questions may sometimes be multiple, so protocols should be
the CT room for good management of the babies: precise knowl- adapted accordingly.
164 PAUL

Fig. 2. Tetralogy of Fallot: 3D visualization of severe left pulmonary Fig. 4. Left anterior descending (LAD) coronary artery arising from
artery (LPA) stenosis in an infant (10 mo, 8 kg). PA, pulmonary the right coronary artery (RCA) in an infant (5 mo, 4.5 kg) with
artery; Ao, aorta. pulmonary atresia with ventricular septal defect.

pulmonary sling. For pulmonary artery visualization, we usu-

ally do not use ECG-gated acquisition. With a 4-slice CT, we
use 2.5-mm collimation (3-mm slice width) with an increment
of 1.5 mm in order to get fast acquisitions (of less than 5 s in
babies), for reduction of motion and respiratory artifacts on
pulmonary arteries. With a 16-slice CT, we usually use 0.75-
mm collimation and get 1-mm slice width with an increment of
0.5 mm. Compared to 4-slice acquisition, images of higher
resolution are acquired in a similar acquisition time. High reso-
lution is advantageous for pulmonary artery stenosis evalua-
tion. For neonates or infants before 7 yr of age, start delay is
either 15 s (peripheral venous access) or 10 s (central venous
access). In patients over 6 yr of age, we sometimes use the bolus
tracking technique for optimization of start delay. For recon-
structions, 3D images are currently performed using maximum
intensity projection, multiplanar reformations, and volume-
rendering technique.
CORONARY ARTERIES
Anomalous coronary arteries are frequently associated with
CHD. The most frequent anomalous finding is a left coronary
artery originating from right coronary sinus (Fig. 4), but many
Fig. 3. Right and left pulmonary artery (PA) stenosis after complete variants are possible, even coronary artery originating from
repair of pulmonary atresia with ventricular septal defect: control
following endovascular kissing stents (arrowheads) implantation in a pulmonary arteries. The normal position of coronary origins
14-yr-old patient. may be different from usual in PAVSD or in tetralogy of Fallot,
because of rotation of the aorta: the origin of the left main is
typically at 6 oclock, and the origin of the right coronary artery
PULMONARY ARTERIES at 1 oclock. Detection of an anomalous origin of coronaries is
Pulmonary artery evaluation is a priority in pulmonary atre- especially important before surgery when a ventriculotomy is
sia with ventricular septal defect (PAVSD) (Fig. 1), tetralogy planned, as accidental lesion of the coronary artery crossing the
of Fallot (Figs. 2 and 3), truncus arteriosus, or suspicion of right ventricle during intervention can be fatal.
 CHAPTER 16 / MSCT IN CHD PATIENTS 165

Fig. 5. Multislice CT shows a large fistula connecting the left main

coronary artery and the superior vena cava (SVC) in an adult patient. Fig. 7. Severe aortic coarctation in a newborn (4 d, 3 kg). Multislice
Cx, left circumflex artery;. LAD, left anterior descending coronary CT showed absence of opacification of aortic arch after the origin of
artery. supra-aortic trunks (arrowheads). A severe stenosis of left carotid
artery was also depicted (arrow). DA, descending aorta.

In neonates we usually do not use ECG-gated acquisition

because the heart rate is too high and also to prevent excessive
radiation dose. Using the same protocol of injection as for
pulmonary arteries, coronary origins can be clearly seen in
about 70% of the cases, but visualization remains difficult in
babies of less than 4 kg. The value of multislice CT for detec-
tion of anomalous coronaries has not yet been assessed.
In older CHD patients, if the patient can hold his or her
breath for a sufficient time, ECG-gated acquisition may be the
technique of choice. Visualization is then possible free of
motion artifact, making possible the precise evaluation of the
coronary artery tree. To avoid heart motion artifacts, heart
rate must be regular and below 70 per minute; excellent results
are generally obtained at 5560 bpm. We usually use ECG-
pulsing technique, in order to reduce radiation dose by about
35% (authors data), all other factors being constant. If axial
and maximum-intensity projection images are sufficient for
diagnosis, volume-rendering technique may provide compre-
hensive imaging of the coronary artery anomaly (Figs. 5 and 6).
AORTA AND COLLATERALS
Aortic anatomy evaluation is essential in case of aortic
coarctation (Fig. 7) or for suspicion of aortic arch anomalies;
complete or incomplete double aortic arch (Fig. 8), right aortic
arch, or cervical arch are very clearly seen with 3D CT images.
In case of PAVSD, major aorto-pulmonary collateral arteries
Fig. 6. Anomalous origin of coronary arteries in a patient with Tetral-
ogy of Fallot: left anterior descending (LAD) coronary artery is origi- (MAPCA) often originate from the descending aorta (Fig. 9);
nating from right coronary sinus. RCA, right coronary artery. AA, evaluation of size and spatial relationship of these arteries is of
ascending aorta; PA, pulmonary artery. primary importance for planning surgical intervention.
166 PAUL

Fig. 8. Severe respiratory distress in a newborn: multislice CT showed Fig. 10. Severe respiratory distress in a newborn (20 d, 3.8 kg) with
clearly a complete double aortic arch responsible for tracheal com- pulmonary valves agenesis with ventricular septal defect. Multislice
pression (2-mo-old patient, 4 kg). RAA, right aortic arch; LAA, left CT showed a compression of the right bronchus between the dilated
aortic arch; AA, ascending aorta; PA, pulmonary artery. right pulmonary artery (RPA) and the right sided aorta. DA, descend-
ing aorta; AA, ascending aorta.

used to get high-resolution images. High-resolution images

are useful for aorta stenosis evaluation, especially in case of
aortic coarctation, for a better evaluation of vessel narrowing.
For neonates or children before 7 yr of age, start delay is either
15 s (peripheral venous access) or 10 s (central venous access).
In patients over 6 yr of age, we usually use the bolus tracking
technique for optimization of start delay. The ROI is placed at
the center of the ascending aorta. For reconstructions, 3D images
are currently performed using maximum-intensity projection,
mutliplanar reformations, and volume-rendering technique.
UPPER AIRWAYS EVALUATION
Compression of central airways from vascular origin may
result from various situations; the most frequently observed are
aortic arch anomalies, pulmonary artery sling, dilated pulmo-
nary arteries, posteriorly displaced aorta (switch intervention),
and so on. In 20 consecutive cases of vascular compression, we
observed six cases of aortic arch anomalies (four double aortic
arch, one incomplete vascular ring, one circling aorta), five
cases of dilatation of pulmonary arteries (one case of pulmo-
nary valve agenesia [Fig. 10], and four cases of pulmonary
Fig. 9. Multislice CT shows four major aorto-pulmonary collateral hypertension associated with ventricular septal defect), three
arteries (arrowheads) originating from descending aorta (DA), in a
patient with pulmonary atresia with ventricular septal defect. Three- cases of posteriorly displaced aorta after intervention for trans-
dimensional evaluation of the course of each collateral was of primary position of great vessels, three cases of left main bronchus
importance for surgical planning. compression by descending aorta, one pulmonary artery sling,
and two compressions by a brachio-cephalic artery.
Nonenhanced CT is sufficient to look for a stenosis of cen-
For aorta visualization, we usually not use ECG-gated tral airways, but contrast enhancement is mandatory to search
acquisition for babies, infants, or adult patients. Thin collima- for a vascular origin. In babies, we currently use the same pro-
tion (1 mm with a 4-slice CT, 0.75 mm with a 16-slice CT), is tocol as for pulmonary artery or aorta visualization. In addition
 CHAPTER 16 / MSCT IN CHD PATIENTS 167

Fig. 11. Multislice CT using 3D rendering technique showed total Fig. 13. Postoperative control of a Blalock (B) anastomosis between
abnormal pulmonary venous return in the superior vena cava (SVC) the innominate artery (IA) and the right pulmonary artery (RPA) in a
and left brachio-cephalic vein (LBVC). LPV, left pulmonary vein; newborn with tetralogy of Fallot (15 d, 3.5 kg). LCA, left carotid
RPV, right pulmonary vein. artery.

compression easy. However, airway compression can be more

distal: in this case, diagnosis of vascular compression may be
difficult. Evidence of air trapping on parenchymal windowing
may be of interest for distal compressions.
ANOMALOUS VENOUS RETURN
CT and MRI are very efficient imaging techniques to detect
pulmonary or systemic anomalous venous return. Both tech-
niques allow 3D visualization of these anomalies. Three-dimen-
sional images can be used as a primary method for diagnosis
(Figs. 11 and 12). ECG-gated acquisitions are usually not nec-
essary because venous structures are not very sensitive to car-
diac motion. Injection site and timing for acquisition must be
chosen carefully, because time for best opacification depends on
the venous drainage, and any anomalous venous drainage may
interfere with the optimal timing. An additional delayed acqui-
sition may be necessary to opacify the whole venous system. To
avoid striking artifacts from concentrated contrast medium in
the veins, it is preferable to reduce the rate of injection.
POSTINTERVENTIONAL EVALUATION
Postoperative evaluation includes various clinical situations,
such as evaluation of bypass patency (Fig. 13), suspicion of
Fig. 12. Demonstration of anomalous systemic venous return in a mediastinitis, suspicion of airway compression (Fig. 14), and
newborn patient with tetralogy of Fallot: multislice CT showed
absence of right superior vena cava and anomalous left azygos vein pulmonary embolus (Fig. 15). The CT protocol should be
ending in the coronary sinus (15-d-old patient, 1.9 kg). adapted to the question. Generally speaking, radiation expo-
sure can often be reduced, because there is generally no need
for detailed anatomic information. For example, for a medias-
to maximum intensity projection (MIP) reconstructions, vol- tinitis search, only one delayed (35 min) acquisition may be
ume-rendering technique (VRT) is very effective to show cen- enough to evidence a mediastinal collection. Conversely, con-
tral airway narrowing (Fig. 10). Using VRT, it is easy to colorize trol of conduit patency can be carried out with an acquisition at
low-density airway structures and high-density vascular struc- the arterial phase alone. CT evaluation of new vascular anatomy
tures to get high-contrast images, making diagnosis of vascular may be useful after complex surgical repair (Fig. 16).
168 PAUL

Fig. 14. External tracheal compression by dense material (found to be

biological glue) following complete repair of aortic coarctation and
ventricular septal defect. Note that the right pulmonary artery (RPA)
is also narrowed due to the presence of the glue. DA, descending
aorta; AA, ascending aorta; LPA, left pulmerary artery; MPA, main
pulmonary artery.

Fig. 16. Complete unifocalization in an infant (3 mo, 4 kg) with pul-

monary atresia with ventricular septal defect and no confluent pulmo-
nary arteries: the major aortopulmonary collaterals were unifocalized
and connected to the right ventricle with a tube. (A) Maximum inten-
sity projection view in axial plane. (B) Volume-rendering technique
view allowing evaluation of the surgical repair from any angle.

EVALUATION OF MAIN CHD WITH MULTISLICE CT

PULMONARY ATRESIA WITH VENTRICULAR
SEPTAL DEFECT
Status of Pulmonary Arteries
Presence or absence pulmonary arteries
Fig. 15. Persistent hypoxia following surgery in a 13-mo-old baby:
multislice CT using axial MIP view showed an obstructive thrombus Confluent pulmonary arteries (seagull sign, Fig. 1).
at the origin of the right pulmonary artery (arrows). DA, descending Size of the pulmonary arteries
aorta; AA, ascending aorta; LPA, left pulmonary artery; MPA, main Presence of pulmonary artery stenosis
pulmonary artery. Patent ductus arteriosus
 CHAPTER 16 / MSCT IN CHD PATIENTS 169

Status of MAPCA Aortic Coarctation

Number and location (Fig. 9) Type of coarctation: preductal, juxtaductal, or postductal
Stenosis (usually proximal) (Fig. 7)
Position in front of or behind the central airways (essential Degree of narrowing
for surgical planning) Presence of thrombus
Coronary Arteries Visualisation of collateral arteries
Left anterior descending artery originating from right Associated anomalies
coronary artery (Fig. 4) Anomalous Pulmonary Venous Return
Double LAD (anterior and posterior) 3D visualization of anomalous venous return (Figs. 11, 12)
Others Distance from left atrium for evaluation of surgical repair
Venous connections possibilities
Bronchial anomalies
Follow-up CONCLUSIONS
Increase of PA Precise 3D visualization of anomalous extracardiac anat-
Patency of conduits omy can be routinely carried out using multislice CT in CHD
New pulmonary anatomy (unifocalization) (Fig. 16) patients. Multislice CT angiography represents an important
TETRALOGY OF FALLOT additional low-invasive diagnostic tool for evaluation of CHD,
Status of pulmonary arteries and may be used occasionally as an alternative to angio-
Size of pulmonary arteries graphy. Dose consideration is the only relative limitation of
Presence of pulmonary artery stenosis this technique, especially in neonates and infants, but excel-
Pulmonary artery stenosis (evaluation of stenosis and plan- lent image quality may be achieved using very low exposure
ning before stent placement [Fig. 2]) parameters.
Coronary arteries
Left anterior descending artery originating from right coro- REFERENCES
nary artery (contraindication for ventriculotomy) 1. The clinical role of magnetic resonance in cardiovascular disease.
Double LAD (anterior and posterior) Task Force of the European Society of Cardiology, in collaboration
Follow-up with the Association of European Paediatric Cardiologists. Eur Heart
Size of PA J 1998;19:1939.
Patency of conduits or stent (Fig. 3 and 13) 2. Kawano T, Ishii M, Takagi J, et al. Three-dimensional helical com-
puted tomographic angiography in neonates and infants with com-
Others plex congenital heart disease. Am Heart J 2000;139:654660.
Venous connections 3. Schroeder S, Kopp AF, Kuettner A, et al. Influence of heart rate on
Bronchial anomalies vessel visibility in noninvasive coronary angiography using new
Transposition of Great Arteries multislice computed tomography: experience in 94 patients. Clin
Long-term follow-up Imaging 2002;26:106111.
Position of aorta and pulmonary artery 4. Brenner D, Elliston C, Hall E, Berdon W. Estimated risks of radia-
tion-induced fatal cancer from pediatric CT. AJR Am J Roentgenol
Pulmonary artery stenosis 2001;176:289296.
Re-implanted coronary arteries 5. Nagel HD et al. In: Radiation exposure in computed tomography:
Atrial Septal Defect Fundamentals, influencing parameters, dose assessment, optimiza-
Generally the defect is not well visible tion, scanner data, terminology. 2nd ed. Offizin Paul Hartung Druck,
CT may be useful to detect associated anomalous venous Hamburg, Germany: 2002;13.
return 6. Slovis TL. ALARA conference proceedings. The ALARA concept
in paediatric CT: intelligent dose reduction (Editorial). Pediatr
Double Aortic Arch Radiol 2002; 32:217231.
Size of both arches (Fig. 8) 7. Westra SJ, Hurteau J, Galindo A, McNitt-Gray MF, Boechat MI,
Right or left sided anterior arch Laks H. Cardiac electron-beam CT in children undergoing surgical
Right or left sided descending aorta repair for pulmonary atresia. Radiology 1999;213:502512.
 CHAPTER 17 / MASS AND PSEUDOTUMOR IMAGING 171

17 Imaging of Cardiac and Paracardiac

Masses and Pseudotumors

BERND J. WINTERSPERGER, MD

INTRODUCTION studies are not the primary focus of cardiac CT imaging, CT

Although the use of computed tomography (CT) in cardiac is a valuable tool in the work-up of patients with suspected
imaging mainly focuses on coronary artery disease and its cardiac tumors.
sequelae, there are many more indications in which CT can As noninvasive cross-sectional imaging modalities, EBCT
provide excellent and valuable information. This has already and MDCT offer a large field of view without major limita-
been reported based on studies with electron beam CT (EBCT). tions based on the patients constitution. In addition, with the
The rapidly emerging and vastly growing interest and knowl- use of retrospectively electrocardiogram (ECG)-gated data
edge in cardiac CT today is mainly caused by the development acquisition techniques, MDCT provides a 3D data set that
and widespread use of multidetector-row CT (MDCT) systems. enables secondary reconstruction of any desired image plane.
Initially introduced in 1992 (Elscint CT-Twin), MDCT led to These features overcome the limitation of providing only
a more widespread use of CT in cardiac imaging, starting with transaxial images in CT or the approximation of individual
4-row MDCT. Within recent years, further technical develop- cardiac axes in EBCT. With the recent advent of 16-row
ments have led to the use of 8-, 16-, and 64-row MDCT sys- MDCT, even submillimeter collimated images are possible
tems. Cardiac imaging has been one of the major and most within a short breath-hold of 1720 s. However, temporal reso-
exciting focuses of MDCT since then. Although the focus of lution of EBCT is superior to that of MDCT and also allows
interest in cardiac MDCT is again coronary artery disease and visualizing very tiny valve-attached tumors, although this still
its prevention, most of the information provided by EBCT can remains a challenge (11).
also be revealed by MDCT (14).
IMAGING TECHNIQUES
Compared to other referrals, imaging of cardiac or para-
cardiac masses is a rather rare request for a CT study. The CT imaging of cardiac masses is basically not different from
examination of and screening for cardiac tumors is among the other cardiac applications like coronary artery imaging or car-
rare causes for performing cross-sectional cardiac imaging diac morphology in congenital heart disease (CHD). Detailed
studies. The modality of choice in screening for cardiac masses information and background about cardiac CT imaging algo-
is still 2D echocardiography (57). Because of its real-time rithms using EBCT and MDCT can be found in Part II, Tech-
approach, echocardiography can detect cardiac and even small nical Background.
cardiac valve-attached masses and their impact on valve and However, depending on the individual scanner and data
global cardiac function within the same examination (8). How- acquisition settings, a few basics have to be considered for
ever, there are certain restrictions to echocardiography affect- imaging of cardiac masses.
ing its clinical use. Although echocardiography can be easily DATA ACQUISITION
performed in any patient with bedside capabilities, image qual-
ity is mainly dependent on the patients habitus and the Especially in MDCT cardiac imaging, different algorithms
examiners skills. Besides restrictions caused by limited acous- and data acquisition techniques can be used. However, to allow
tic windows due to obesity, right-ventricle assessment may be a 3D assessment of the heart with adaptation to individual car-
hampered by pulmonary emphysema. In addition, tissue char- diac axis cardiac chambers, retrospective ECG-gated algo-
acterization is often not possible even though tumors may rithms are strongly recommended, although they basically lead
present with typical echocardiographic signs. EBCT and to a higher radiation exposure. Recently developed techniques
MDCT are cross-sectional techniques for acquiring images allow again for a subsequent reduction of redundant radiation
with high spatial and temporal resolution to freeze cardiac without losing the benefits of retrospective gating. In the
motion and to follow the cardiac cycle (9,10). Although tumor assessment of coronary artery calcifications, Jakobs et al.
showed a significant reduction of 4550% without change in
From: Contemporary Cardiology: CT of the Heart: image quality (12). As there are major differences in basic tech-
Principles and Applications niques, only general recommendations can be given. Technical
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ backgrounds and imaging algorithms of EBCT and MDCT are

171
172 WINTERSPERGER

Fig. 1. Axial multidetector-row CT (MDCT) image of a 4-row MDCT scanner system showing artifacts within the right atrium and ventricle
due to contrast influx (A). By optimization of contrast application using 16-row MDCT these artifacts can be avoided, leading to a better
depiction of the right coronary artery (arrow). However, for mass depiction, an enhancement of all cardiac chambers would be necessary.

substantially different. This implies differences in the exact tion protocols, opacfication of the right atrium and ventricle is
scanning protocol used. However, there are basic demands and not guaranteed and mass diagnoses may fail.
requirements for assessment of cardiac masses in CT that Therefore, modification of injection protocols is necessary
determine the details of the imaging protocol. in order to ensure opacification of all cardiac chambers. Espe-
MDCT allows for very thin collimated slices in the range of cially where right atrial tumors are suspected, additional
1 mm or even less. This is favorable for the assessment of delayed scanning during the second or third pass of contrast
coronary arteries but is usually unnecessary for the assessment media will improve the homogeneity of atrial opacification and
of cardiac masses. This can also satisfactorily be done with a help to avoid misinterpretation of influx of nonopacified blood
slice thickness in the range of 23 mm. This allows for a cutoff from the inferior vena cava.
in data acquisition time that makes it even possible to be per-
EPIDEMIOLOGY OF CARDIAC MASSES
formed easily within a short breath-hold on 4-row MDCT. With
overlapping slice reconstruction, multiplanar reformations Basically, when talking about cardiac masses, primary and
(MPR) can be used to determine the exact location and extent secondary tumors have to be differentiated. Primary tumors
of tumors. typically originate within the heart, whereas secondary tumors
Although infrequently used at present, dynamic ECG-trig- most commonly result from metastasis of malignancies prima-
gered or -gated scanning can provide dynamic tumor enhance- rily located outside the heart or from direct tumor spread and
ment and may provide further information concerning tumor invasion of masses located adjacent to the heart (e.g.,
type. However, to date there has been no study published focus- bronchgenic carcinoma). Primary tumors of the heart are by far
ing on such an imaging strategy in CT. less common that secondary tumors. Autopsy studies report a
prevalence of primary masses in the range of 0.0017%0.33%
CONTRAST APPLICATION (13). However, although there is a wide range of variation
Although some features of cardiac tumors (e.g., fatty tissue, between different studies mainly based on the year of publica-
calcifications) may even be visible on plain scans without tion and inclusion criteria, primary cardiac masses can defi-
application of contrast agents, allocation of these features nitely be considered as rare tumors when compared to other
and mass diagnosis necessitates the use of iodinated contrast types. Secondary cardiac tumors on the other hand are about
media. There is extensive literature on the optimization of 2040 times more frequent than primary ones (13). Modern
contrast injection regimens, primarily focusing on the assess- noninvasive imaging modalities such as echocardiography in
ment of coronary arteries. With the use of 16-row MDCT, particular, but also CT and magnetic resonance imaging (MRI),
contrast agent volume could be reduced, therefore avoiding now allow mass diagnosis prior to death and allow for therapy
streak artifacts from high contrast influx within the right atrium planning. In addition to real neoplasms, no matter whether
and ventricle affecting the assessability of the proximal right benign or malignant, primary, or secondary, CT imaging is also
coronary artery (10) (Fig. 1A,B). However, using these injec- increasingly performed when cardiac thrombi are suspected. In
 CHAPTER 17 / MASS AND PSEUDOTUMOR IMAGING 173

Table 1 has been associated with a variety of hyperpigmented skin

Frequency of Benign Cardiac Tumors (adapted from ref. 13) lesions (14). Also, the development of other, extracardiac
Benign tumors Approximate frequency (%) tumors and endocrine abnormalities may occur (15).
Although the diagnosis can be easily made using different
Myxoma 29 imaging modalities, patients with cardiac myxoma can be
Papillary fibroelastoma 8 symptomatic for a long time before myxomas are considered as
Rhabdomyoma 5
a potential differential diagnosis (16). Patients may present
Fibroma 5
Hemangioma 4 with a triad of symptoms including constitutional symptoms,
Lipoma 4 signs of valvular obstruction (based on myxoma prolapse), and
Others 8 embolic events. The target vasculature of embolic events is
Total 63 strongly dependent on the location of the tumor. Embolic phe-
nomena of left-sided myxomas, considered the most severe
complication, may lead to ischemia of either the extremities or
the viscera. In case of supra-aortic emboli, symptoms of acute
stroke, transient ischemic attacks, or seizures may occur. In
right atrial myxomas, symptoms are less frequent.
Atrial myxoma often shows a pedunculated appearance, with
its origin in the area of the oval fossa, although the stalk may not
be identified with CT (Fig. 3A,B). With increasing size, those
mobile masses can even prolapse into the mitral annulus, lead-
ing to changes in cardiac hemodynamics. Right-sided myxo-
mas are less frequently attached to the oval fossa. Myomas can
even present as a mass within the area of the oval fossa extend-
ing to both atria.
Intracavitary tumor masses usually appear as filling defects
within the opacified blood pool (Figs. 2 and 3). This necessi-
tates homogeneous contrast distribution and cavitary opacifi-
cation. However, especially within the right atrium, inflow
artifacts need to be differentiated. Mass appearance in CT is
strongly dependent on gross pathologic features. Based on their
gelatinous nature, myxomas usually have heterogeneous low
attenuation at CT (Figs. 2 and 3). Regressive changes may lead
to recurrent hemorrhage and calcification, which is frequently
Fig. 2. Patient with a hypodense round mass within the left atrium seen in CT (17) (Fig. 3A). In addition, adherent thrombi can
attached to the interatrial septum (arrow heads). Mass was proven as
a myxoma at surgery. Reconstruction was performed during mid- occur. The mobility of myxomas can be demonstrated by either
diastole with open mitral valve leaflets (arrows). the cine mode in EBCT or multiple data reconstructions at dif-
ferent time points of the cardiac cycle in MDCT with retrospec-
tive ECG gating.
addition to imaging features and appearance, cardiac masses Although papillary fibroelastomas represent the second
can often be differentiated based on location and morphology. most common benign primary cardiac neoplasm, they are rarely
demonstrated with CT modalities. They usually appear at
BENIGN CARDIAC TUMORS endocardial surfaces and are most commonly located (90%) on
Approximately 6070% of primary cardiac tumors are valve surfaces. The vast majority of these tumors are less than
benign (13). Cardiac myxoma accounts for about half of all 1 cm in diameter (18). Their appearance on the rapidly moving
benign tumors, followed by papillary fibroelastoma, fibroma, valves and their small size makes it rather difficult to identify
and lipoma (Table 1) (13). However, there is also a difference those lesions using CT, although they are easily detected at
in frequency of benign tumors based on a patients age. Whereas echocardiogaphy. Whereas MDCT techniques currently can-
myxoma represents the most common tumor type in adults, in not easily follow valve movement because of the rather bad
childhood, rhabodomyoma accounts for the majority of cases. temporal resolution (approx 100250 ms), EBCT may allow
Myxomas can be found in all cardiac chambers; however, in the depiction of those tumors (11) (Fig. 4). Recently published
approx 75% of the cases, the mass is located within the left MDCT data report the depiction of even small valvular pathol-
atrium (Fig. 2). Another 20% of myxomas are located within ogy (19,20). However, at present, especially in diagnosis of
the right atrium, whereas only 5% arise within either the left or fibroelastomas, echocardiography is usually far superior to CT.
the right ventricle. In rare instances, myxomas can be found at Although symptoms are less frequent than in cases of myxoma,
multiple locations within the heart. Although most cases of papillary fibroelastomas may also lead to embolic events. This
cardiac myxoma are sporadic, possibly part of the Carney com- is mostly based on adherent thrombo-embolic material.
plex, an autosomal dominant syndrome of cardiac myxomas Although the suspicion of a fibroelastoma cannot be consid-
174 WINTERSPERGER

Fig. 3. (A) Round-shaped myxoma of the right atrium with small calcifications (arrows). RV, right ventricle; LV, left ventricle. (B) Correspond-
ing magnetic resonance imaging in long-axis orientation show the stalk of the myxoma (arrows) attached to the oval fossa.

Contrary to other tumors, lipomas are usually found on the

epicardial surface, and may present at any site of the atria or
ventricles. However, lipomas may also extent intracavitary. In
combination with other tissues, e.g., in angiolipoma, typical
image features of fatty tissue may disappear and therefore ham-
per diagnosis (Figs. 5AC and 6A,B).
Lipomatous hypertrophy of the interatrial septum consists
of an accumulation of mature fat as well as of adipose cells
resembling brown fat cells. Deposits at the level of the oval
fossa with a diameter larger than 2 cm are considered lipoma-
tous hypertrophy (Fig. 7). These changes may be accompanied
by arrhythmias. Similar to lipomas, lipomatous hypertrophy
usually shows typical homogeneous low attenuation at CT,
which can be found even in nonenhanced scans (21,22).
Other benign tumors such as rhabdomyoma and fibroma
usually appear within the ventricular wall or the interventricu-
lar septum. They usually show a density comparable to normal
myocardium, which precludes easy diagnosis and delineation.
Rhabdomyoma is the most frequent tumor in infancy and child-
hood, and is often associated with tuberous sclerosis (23). Up
to 50% of patients with rhabdomyoma suffer from tuberous
sclerosis syndrome, and 60% of children with tuberous sclero-
Fig. 4. Axial image of an electron beam CT data set at the level of
the aortic valve. A small papillary fibroelastoma can be depicted at
the tip of the aortic valve cusps (arrows). RVOT, right-ventricular
outflow tract.
Fig. 5. (A) (right and page 176) Axial slices of an extensive mass
arising from the right atrial roof, narrowing the right pulmonary
ered as a reason for cardiac CT scanning, CT may be helpful to artery and the superior vena cava. The mass shows heterogeneous
exclude thrombi within the cardiac cavities. enhancement with central necrosis and calcifications. The mass
Based on the typical appearance of fatty tissue, lipomatous was proven to be a benign angiomyolipoma based on several tissue
probes using CT-guided needle biopsy and open thoracotomy.
tumors are usually easily diagnosed using CT. Based on pathol- (B) The sagittal reconstruction also shows central necrosis and cal-
ogy findings, two different types have to be differentiated cification (arrow). VC, vena cava. Fatty tissue is found only at the
lipomas and lipomatous hypertrophy of the interatrial septum. very bottom of the mass (arrow heads).
CHAPTER 17 / MASS AND PSEUDOTUMOR IMAGING 175
176 WINTERSPERGER

Fig. 5. (Continued from page 174) (C) Sequential electrocardiogram-triggered axial multidetector-row CT images at a constant level of the
tumor shows a marked enhancement of most parts of the mass. This is based on the high degree of vascularization in this angiomyolipoma.

Fig. 6. Corresponding T1 weighted axial magnetic resonance imaging sections before (A) and after (B) contrast administration show massive
enhancement of most tumor parts except a central necrosis (N).

sis have detectable cardiac masses (24). Fibroma also prima- Beside these solid tumors, pericardial cysts represent a
rily affects children and is often detected in infants or in utero benign lesion that needs to be differentiated from other
by ultrasound (25). tumors.
 CHAPTER 17 / MASS AND PSEUDOTUMOR IMAGING 177

Fig. 7. Axial multidetector-row CT images of a patient with lipomateous hypertrophy of the atrial septum. Note the thickened septum and the
low-density mass between both atria, consistent with fatty tissue (arrow heads). RA, right atrium; LA, left atrium; RV, right ventricle; LV, left
ventricle.

MALIGNANT CARDIAC TUMORS invade cardiac valves, and tend to recur after resection
PRIMARY CARDIAC TUMORS (Fig. 9A,B). There are numerous other types of malignant
Primary cardiac malignancies are rather rare. Only 25% of primary cardiac tumors, including malignant fibrous histiocy-
all primary cardiac tumors are malignant (13). In addition, they tomas, osteosarcomas, hemangiopericytomas, and lymphomas
are much more uncommon than metastatic lesions to the heart. (Fig. 10). The approximate frequency of different types of
The distribution and frequency of different types varies within malignant tumors is shown in Table 2.
published data (13). Primary cardiac malignancies represent a Both CT and MRI are used in follow-up of malignant tumors
clinical dilemma. They are often asymptomatic until they after chemotherapy, resection, or even cardiac transplantation.
become large, and even then they produce nonspecific symp- When focusing on the primary lesion itself, cardiac gating is
toms (26). Before the advent of cross-sectional imaging, pri- recommended, whereas in examinations for assessment of
mary cardiac malignancies were rarely diagnosed before death. metastatic lesions from primary cardiac tumors, nongated tech-
Nowadays they are being diagnosed within living patients, niques are adequate. Compared to MRI, CT allows for a com-
allowing for conservative or even surgical treatment, including prehensive appreciation of the primary lesion and possible
heart transplantation (27). However, based on the usual delayed tumor spread with a single injection of contrast agent. Beside
diagnosis of then extended disease including metastasis, these time-saving considerations, work-flow and patient manage-
tumors show a rather bad outcome. CT can be used to accu- ment have to be taken into account
rately image the heart and the surrounding mediastinum, and SECONDARY CARDIAC TUMORS
therefore to evaluate the extent of disease. Secondary malignancies are 2040 times more frequent than
Angiosarcoma represents the most common (approx 35 primary ones. As already mentioned, not only metastatic disease
40%) cardiac sarcoma. Patients usually present with right-sided but also direct tumor involvement and invasion of the heart from
heart failure or tamponade based on its tendency to occur within primary lesions adjacent to the heart account for these cases. In
the right atrium and to invade the pericardium (Fig. 8). addition to direct invasion, tumor spread to the heart can also
Approximately 75% of angiosarcomas are located within the arise via predefined venous structures, from hepatocellular car-
right atrium (28). Based on their composition, they usually cinomas or renal cell carcinomas, for example. Based on their
show a major contrast uptake, which may be combined with proximity to the heart, the most common tumors affecting the
areas of necrosis. In contrast to angiosarcoma, rhabdomyosa- heart by direct invasion are bronchogenic carcinomas and breast
rcomas do not show a predominant location within the heart. tumors. An overview of the frequency of cardiac or pericardial
On nonenhanced scans, they may show a density identical to involvement of different tumor types is given in Table 3. Imag-
normal myocardium, whereas they can usually be well differ- ing features of secondary cardiac malignancies are often but not
entiated from myocardium in enhanced scans. They may even mandatory, similar to those of the primary lesion.
178 WINTERSPERGER

Fig. 8. Multidetector-row CT images of a patient with recurrent angiosarcoma involving the right atrium and parts of the right ventricle (arrow
heads). The mass almost occludes the superior vena cava (arrows).

Fig. 9. (A) Recurrence of a rhabdomyosarcoma within the left atrium (arrow heads). (B) In addition to the tumor within the left atrium, the mass
infiltrates the mitral valve and extends to the left ventricle (arrows).

IMAGING OF CARDIAC THROMBI Therefore, early identification with subsequent therapy is of

Thrombotic deposits within the heart are often referred to as paramount importance.
pseudotumors or pseudomasses. Intracardiac thrombi may be Occurring in virtually any part of the vascular system, the
caused by a variety of different pathologies and may occur in development of thrombotic clots may be based either on
any cardiac chamber. They are responsible for approx 15% of changes of surface properties, flow dynamics, or changes in
all ischemic strokes and put patients at a major risk of stroke. blood coagulation. Patients with artificial devices (e.g., pros-
 CHAPTER 17 / MASS AND PSEUDOTUMOR IMAGING 179

Fig. 10. A set of axial multidetector-row CT images of a malignant hemangiopericytoma involving the right coronary artery and the chest wall
(arrows). The tumor encases the right atrium and ventricle. Within the left lower lobe of the lungs, metastatic disease in shown.

Table 2 Table 3
Distribution and Frequency of Primary Frequency of Cardiac/Pericardial Involvement in Cases
Cardiac Malignancies (adapted from ref. 13) of Extracardiac Malignancies (adapted from ref. 13)

Approximate Malignant tumors Frequency (%)

Malignant tumors frequency (%)
Melanoma 49
Angiosarcoma 9 Germinoma 42
Malignant fibrous histiocytoma 4 Leukemia 34
Osteosarcoma 3 Bronchogenic carcinoma 28
Leimyosarcoma 3 Sarcomas 22
Rhabdomyosarcoma 2 Lymphomas 21
Lymphoma 2 Breast Cancer 20
Others 14 Esophageal carcinoma 17
Renal cell carcinoma 15
Total 37 Thyroid carcinoma 12

thetic valves, pacemakers), atrial fibrillation, or wall motion depends on location. In regions of wall motion, abnormalities,
abnormalities (including ventricular aneurysms) are at special or ventricular aneurysms after myocardial infarction, they usu-
risk. As already stated, thrombotic layers may also occur at the ally lie adjacent to the areas of infracted myocardium. They can
surface of myxomas or papillary fibroelastomas. usually be differentiated from normal myocardium based on
Thrombi may be solitary or multiple, and present as filling their rather low attenuation. However, in close contact to scar
defects within opacified cardiac chambers. Their appearance tissue in chronic infarction, the exact extent may be overesti-
180 WINTERSPERGER

Fig. 11. Four-chamber view of a multidetector-row CT data set. Ventricular aneurysm (A) within the apex after myocardial infarction and a
thrombus can be delineated (arrow heads). The thrombus show tiny calcifications (arrows).

Fig. 12. (A) Multiplanar reformation (MPR) of the left upper extremity with an acute embolic occlusion of the axillary artery (arrow heads).
The subclavian artery shows normal diameter and patency (arrow). (B) Reconstruction of the same data set focused in the chest shows residual
thrombus (arrow heads) within the left atrial appendage (arrow). AA, ascending aorta; DA, descending aorta; PA, pulmonary artery.
 CHAPTER 17 / MASS AND PSEUDOTUMOR IMAGING 181

mated. Long-standing thrombi tend to get organized and may 9. MacMillan RM. Magnetic resonance imaging vs. ultrafast computed
calcify (Fig. 11). Clots within the atria are commonly related to tomography for cardiac diagnosis. Int J Card Imaging 1992;8:
217227.
contraction abnormalities (e.g., atrial fibrillation). They may 10. Wintersperger BJ, Nikolaou K, Jakobs TF, Reiser MF, Becker CR.
often be found within the atrial appendages (Fig. 12A,B). Cardiac multidetector-row computed tomography: initial experience
Echocardiography represents the basic modality for throm- using 16 detector-row systems. Crit Rev Comput Tomogr 2003;44:
bus screening. However, based on the large field of view, car- 2745.
11. Wintersperger BJ, Becker CR, Gulbins H, et al. Tumors of the cardiac
diac CT allows for reliable depiction or occlusion of thrombi. valves: imaging findings in magnetic resonance imaging, electron
Although not primarily used for cardiac imaging, routine CT of beam computed tomography, and echocardiography. Eur Radiol
the chest may reveal unknown cardiac thrombi. 2000;10:443449.
12. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of
CONCLUSION the heart with retrospective ECG gating: reduction of radiation
exposure by ECG-controlled tube current modulation. Eur Radiol
Cardiac CT shows promise in the assessment of cardiac 2002;12:10811086.
masses. Especially with the widespread use of MDCT, this 13. Burke A, Virmani R. Tumors of the heart and great vessels. Armed
modality is rapidly growing in cardiac imaging. Its use in Forces Institute of Pathology; Washington D.C.: 1995.
14. Carney JA, Gordon H, Carpenter PC, Shenoy PV, Go VL. The com-
imaging of cardiac masses and thrombi represents only a niche plex of myxomas, spotty pigmentation and endocrine overactivity.
indication. However, the requirements are less demanding than Medicine 1985;64:270283.
those of coronary CT angiography and can also be performed 15. Casey M, Mah C, Merliss AD, et al. Identification of a novel genetic
with less sophisticated MDCT scanner generations. However, locus for familial cardiac myxomas and Carney complex. Circula-
tion 1998;98:25602566.
as a prerequisite for cardiac CT imaging, ECG-based data
16. Burke AP, Virmani R. Cardiac myxomas: a clinicopathologic study.
acquisition strategies and algorithms are necessary. With the Am J Clin Pathol 1993;100:671680.
use of MDCT, a new modality of cardiac imaging competes 17. Tsuchiya F, Kohno A, Saitoh R, Shigeta A. CT findings of atrial
with MRI. The acquisition of 3D data sets even allows for myxoma. Radiology 1984;151:139143.
multiplanar imaging of cardiac tumors. There are specific fea- 18. Abu Nassar SG, Parker JC, Jr. Incidental papillary endocardial
tumor. Its potential significance. Arch Pathol 1971;92:370376.
tures of tumors (e.g., calcifications) that may be visualized and 19. Willmann JK, Kobza R, Roos JE, et al. ECG-gated multi-detector
depicted only in CT. On the other hand, CT may be inferior to row CT for assessment of mitral valve disease: initial experience.
MRI in exact evaluation of the tumor type in most cases, based Eur Radiol 2002;12:26622669.
on its lower soft-tissue contrast properties. 20. Willmann JK, Weishaupt D, Lachat M, et al. Electrocardiographically
gated multi-detector row CT for assessment of valvular morphology
and calcification in aortic stenosis. Radiology 2002;225:120128.
REFERENCES 21. Kamiya H, Ohno M, Iwata H, et al. Cardiac lipoma in the interven-
1. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral tricular septum: evaluation by computed tomography and magnetic
computed tomography angiography for determination of coronary resonance imaging. Am Heart J 1990;119:12151217.
artery stenoses. Am J Cardiol 2001;88:11911194. 22. Hayashi H, Wakabayashi H, Kumazaki T. Ultrafast computed
2. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography tomography diagnosis of an epicardial lipoma in the pericardial
with multi-slice computed tomography. Lancet 2001;357:599603. sac: the split pericardium appearance. J Thorac Imaging 1996;11:
3. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama 161162.
PM, de Feyter PJ. Reliable noninvasive coronary angiography with 23. Smythe JF, Dyck JD, Smallhorn JF, Freedom RM. Natural history
fast submillimeter multislice spiral computed tomography. Circula- of cardiac rhabdomyoma in infancy and childhood. Am J Cardiol
tion 2002;106:20512054. 1990;66:12471249.
4. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery 24. Smith HC, Watson GH, Patel RG, Super M. Cardiac rhabdomyo-
stenoses with thin-slice multi-detector row spiral computed tomo- mata in tuberous sclerosis: their course and diagnostic value. Arch
graphy and multiplanar reconstruction. Circulation 2003;107: Dis Child 1989;64:196200.
664666. 25. Beghetti M, Gow RM, Haney I, Mawson J, Williams WG, Freedom
5. Felner JM, Knopf WD. Echocardiographic recognition of intracar- RM. Pediatric primary benign cardiac tumors: a 15-year review. Am
diac and extracardiac masses. Echocardiography 1985;2:3. Heart J 1997;134:11071114.
6. Salcedo EE, Cohen GI, White RD, Davison MB. Cardiac tumors: 26. Araoz PA, Eklund HE, Welch TJ, Breen JF. CT and MR imaging of
diagnosis and management. Curr Probl Cardiol 1992;17:73137. primary cardiac malignancies. Radiographics 1999;19:14211434.
7. Link KM, Lesko NM. MR evaluation of cardiac/juxtacardiac masses. 27. Uberfuhr P, Meiser B, Fuchs A, et al. Heart transplantation: an
Top Magn Reson Imaging 1995;7:232245. approach to treating primary cardiac sarcoma? J Heart Lung Trans-
8. Olson LJ, Tajik AJ. Valvular heart disease. In: Skorton DJ, Schelbert plant 2002;21:11351139.
HR, Wolf GL, Brundage BH. (eds.) Cardiac Imaging. 2nd ed. 28. Janigan DT, Husain A, Robinson NA. Cardiac angiosarcomas: a
Saunders Company, Philadelphia: 1996;365394. review and a case report. Cancer 1986;57:852859.
 CHAPTER 18/ MDCT ASSESSMENT OF LEFT-VENTRICULAR FUNCTION 183

18 Multidetector-Row CT Assessment
of Left-Ventricular Function

KAI UWE JUERGENS, MD AND ROMAN FISCHBACH, MD

INTRODUCTION Transthoracic echocardiography is a widely available imag-

For the diagnosis, disease stratification, treatment planning, ing modality, relatively cheap and mobile; however, image
and prognosis of different cardiac diseases clinically present- acquisition is acoustic-window and operator dependent. Owing
ing with myocardial dysfunction and regional or global to geometric assumptions of the left-ventricular shape, accu-
ischemia, respectively, the accurate and reproducible determin- racy in regards to quantitative assessment of left-ventricular
ation of left-ventricular myocardial function is fundamental. function is hampered in remodeled hearts with complex irregu-
Heart failure afflicts about 4.5 to 5 million people in the United lar shape changes. Good visualization of the entire endocardial
States, with more than 500,000 new cases developing each year. border cannot always be provided using conventional 2D
Clinical assessment with regard to left-ventricular function is echocardiography, resulting in a high amount of subjectivity
generally difficult; although an entirely normal electrocardio- and dependency on the operators clinical experience. Thus,
gram (ECG) provides a close correlation to normal systolic reliable measurement of myocardial thickness and thickening
function, electrographical findings are nonspecific. The ideal is frequently elusive. Three-dimensional echocardiography has
imaging modality to determine left-ventricular function in a improved its clinical impact; however, dependency of acoustic
clinical setting would be noninvasive, accurate, reproducible, windows and clinical practicability have to be evaluated in
easily available, as well as cost and time effective (13). further studies. Analysis of left-ventricular volume-time curves
The analysis of left-ventricular myocardial function includes by real-time 3D echocardiography provides quantitative data
the determination of regional and global function characteris- on global left-ventricular function, such as filling rates (4).
tics. Regional function parameters are myocardial wall thick- CVG while performing coronary catheter angiography is
ness and wall thickening, calculated as end-diastolic and currently a clinically accepted standard for the assessment of
end-systolic wall thickness, systolic wall thickening, and per- cardiac volumes and function. However, this method is inva-
cent systolic wall thickening. Clinical information on global sive and is limited as a result of geometric assumptions made
left-ventricular function are determined from measurements of from projection images.
diastolic and systolic left-ventricular volumes and consecu- Radionuclide ventriculography is commonly used to mea-
tively calculated ejection fraction, stroke volume, and cardiac sure cardiac function in terms of left-ventricular ejection frac-
output. With regard to reproducibility and accuracy of global tion; however, it is rarely performed clinically, because it is
functional parameters, an imaging modality is only as good as hampered by limited temporal and spatial resolution, and prepa-
its measurements of left-ventricular volumes. ration and scanning times are relatively prolonged (5). Gated
perfusion SPECT is not used to measure ventricular function
ESTABLISHED TECHNIQUES FOR ASSESSING alone, although it enables 3D assessment of cardiac function
LEFT-VENTRICULAR FUNCTION and is especially useful when perfusion needs to be assessed
Various noninvasive imaging modalities (such as echo- with a high reproducibility. Its diagnostic accuracy might be
cardiography, radionuclide ventriculography, gated perfusion limited both in small and large ventricles due to limited spatial
single photon emission computed tomography [SPECT], elec- resolution. Owing to very low counts, it is difficult to define
tron-beam computed tomography [EBCT], and cardiac mag- ventricular borders in left-ventricular areas with circumscript
netic resonance imaging [MRI]) and invasive techniques (such thinning as a result of transmural infarction. In both nuclear
as cine ventriculography [CVG]) are in use for the determina- techniques, the need for repeated radionuclide doses in sequen-
tion of regional and global left-ventricular function. tial studies is problematic due to radiation exposure. Using
99mTc-tetrofosmin, the dose equivalent for myocardium and a

blood pool marker is about 8 101 mSv per 100 MBq. In a

clinical setting, about 750 to 900 MBq 99mTc-tetrofosmin are
From: Contemporary Cardiology: CT of the Heart: commonly used to perform a first-pass radionuclide ventricu-
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
lography.

183
184 JUERGENS AND FISCHBACH

Table 1
Left-Ventricular (LV) Parameters (mean 1 SD) With 95% Confidence Interval (1.96 SD) in Parenthesesa

ALL Males Females

(N = 75) (N = 47) (N = 28)

LV end-diastolic volumes [mL] 121 34 (55187) 136 30 (77195) 96 23 (52141)

LV end-diastolic volumes BSA [mL/m2] 66 12 (4489) 69 11 (4792) 61 10 (4181)
LV end-systolic volumes [mL] 40 14 (1367) 45 14 (1972) 32 9 (1351)
LV ejection fraction [%] 67 5 (5778) 67 5 (5678) 67 5 (5678)
LV stroke volume [mL] 82 23 (36127) 92 21 (5183) 65 16 (3397)
a Derived from reference method cardiovascular magnetic resonance imaging. BSA, body surface area; SD, standard
deviation. N = 75. Modified from ref. 11.

EBCT scanners providing a temporal resolution down to patients ECG. A major drawback of prospectively ECG-trig-
50100 ms have been successfully used for noninvasive coro- gered image acquisition, however, is its vulnerability to cardiac
nary calcium measurements, coronary arteriography, and arrhythmia or changes in heart rate.
determination of left-ventricular (LV) mass and function data. Continuous imaging of the heart volume with spiral tech-
However, these systems are costly, and the limited number of nique and ECG-synchronized image reconstruction was
scanners available restricts access to this modality (6,7). Pro- severely hampered because of limited volume coverage in the
spectively ECG-triggered sequential single-slice scanning is patient-longitudinal (z) axis when only one detector line was
performed with a selected delay from the preceding R peak of available. In contrast, multidetector spiral CT technique enables
the ECG. This prospective triggering technique renders the scanning of larger anatomical volumes at a given scan time or
images vulnerable to sudden changes in heart rate or cardiac smaller volumes at a narrower collimation, resulting in a higher
rhythm. The fixed setup of an EBCT scanner impairs assessing axial resolution. Thus, high-quality data sets for 3D postpro-
LV functional parameters in the anatomically true short-axis cessing are provided. The entire volume of the heart can be
orientation, as it is applied in echocardiography and cardiac MRI. scanned within a single breath-hold. Motion of the heart can
In comparison to other noninvasive imaging techniques, be virtually frozen using retrospective ECG-gating technique,
cardiac MRI is now widely accepted as the reference standard when the ECG signal of the patient is recorded simultaneously
of noninvasive assessment of cardiac function, being accurate during spiral CT data acquisition. The method of retrospective
and reproducible in normal and also in abnormal ventricles ECG gating enables greater robustness with regard to arrhyth-
using conventional breath-hold turbo-gradient echo sequences mia, because prospective estimation of the heart rate is not
(TGrE) as well as newly developed steady-state free precession needed for the triggers placement (14). In the clinical setting,
(SSFP) cine sequences (810). Cardiac MRI provides an excel- left-ventricular function information can be derived from
lent temporal and spatial resolution, with image acquisition in multidetector-row CT (MDCT) coronary angiography data
any desired plane (e.g., in vertical and horizontal long-axis and sets. From axial thin-section CT spiral data sets acquired dur-
short-axis views). Image acquisition can be performed rapidly ing the entire cardiac cycle, images at diastolic and systolic
to allow visual and qualitative assessment of global and regional window can be subsequently reconstructed using retrospec-
left-ventricular function. A synopsis of global left-ventricular tive ECG-gating technique (see Data Acquisition and Image
functional parameters in healthy volunteers (11) as determined Reconstruction below).
by cardiac MRI is given in Table 1. DETERMINATION OF LEFT-VENTRICULAR
VOLUMETRIC DATA
ASSESSMENT OF LEFT-VENTRICULAR For clinical and research purposes, left-ventricular volum-
FUNCTION USING CT etric data are determined using different anatomical and geo-
Application of conventional CT to cardiac imaging has long metric models. The standard approaches for reliable and
been limited by insufficient temporal resolution as a result of reproducible volumetrics are the area-length method based on
slow gantry rotation, and long total acquisition time resulting the long-axis view as well as the Simpsons method applied to
from slow volume coverage with single-slice imaging. Analy- contiguous short-axis reformations (Fig. 1). For reproducible
sis of cardiac function has been possible only in experimental assessment of volumetric data, the imaging of the left ven-
setups and was restricted to a transaxial slice orientation tricle along to the cardiac imaging axes is advantageous,
(12,13). Cardiac imaging using conventional single-slice CT because the body axesi.e., transverse, sagittal, and coronal
scanners had become available with the introduction of sub- planes (Figs. 24)are not perpendicular to the left-ventricu-
second rotation in 1994. Initial promising results were achieved lar cavity or the myocardial wall, resulting in an overestimation
using prospective ECG-triggering technique combined with a of true left-ventricular dimensions owing to partial-volume
possible temporal resolution of 500 ms. As in EBCT, prospec- effects and obliqueness of those axes. High z-axis resolution
tive ECG-triggering techniques enable determination of volu- nearly approaching isotropic voxels, provided by MDCT tech-
metric data of the heart from sequential scans, which are started nique, enables multiplanar reformations (MPRs) from primary
following a predefined delay to the onset or the R waves of the axial image reconstructionsin particular, the left-ventricular
 CHAPTER 18/ MDCT ASSESSMENT OF LEFT-VENTRICULAR FUNCTION 185

Fig. 1. Diagram showing two different geometric models for determination of left-ventricular volumetric data: the standard approaches for
reliable and reproducible volumetric are the Simpsons method applied to contiguous short-axis reformations using 3D reconstruction (A) as
well as the area-length method based on the long-axis view of the heart using the biplane ellipsoid model (B). A = cross-sectional area within
endocardial borders determined from one section (modified from 9).

Fig. 2. Multidetector CT postprocessing and generation of multiplanar

reformations in an anatomically optimized vertical long-axis view.
C.t., chordae terdinae; MV, mitral valve; PPM, posterior papillary
muscle.

long axis and the short-axis plane: starting with axial recon-
structions, a vertical long-axis plane is obtained connecting the
left-ventricular apex and the middle of the atrioventricular junc-
tion. A further plane connecting the left-ventricular apex and Fig. 3. Multidetector CT postprocessing and generation of multiplanar
the middle of the mitral valve ring is chosen, resulting in a reformations in an anatomically optimized horizontal long-axis view.
IVS, intraventricular septum; LA, left atrium; MV, mitral valve; RCA,
horizontal long-axis plane. Perpendicular to the horizontal long right coronary artery; T.s., Trabecular septomarginalia.
axis, the stack of double-oblique short-axis images are orien-
tated by an inclination parallel to the mitral-valve ring. Second-
ary MPRs in horizontal and vertical long-axis and true Area-length method. From a long-axis view, the area A
short-axis orientation, respectively, can be performed from within traced endocardial contours and the length L from the
diastolic and systolic reconstructions, enabling determination left-ventricular apex to the level of mitral valve ring are used to
of left-ventricular end-diastolic and end-systolic volumes calculate the LV volume (VLA) according to equation 1:
(Figs. 5 and 6). Consecutively, the calculation of left-ventricu-
lar ejection fraction and stroke volume is feasible as follows: VLA = (8/3) A2/L (1)
186 JUERGENS AND FISCHBACH

Fig. 4. (A) The standard sagittal, coronal, and axial orientation of multiplanar reformations (MPR) as provided by standard postprocessing
software. Based on those standard angulations, MPR in the long-axis view and four-chamber view can be generated, resulting in a true short-
axis orientation (B). With regard to functional analysis, a stack of double contiguous short-axis-orientated sections (thickness 8 mm without
any intersection gap) is acquired.
 CHAPTER 18/ MDCT ASSESSMENT OF LEFT-VENTRICULAR FUNCTION 187

Fig. 5. Ten to eighteen sections are used to encompass the entire left ventricle from base to apex.

Fig. 6. Endocardial contours were manually traced using standard planimetric software implemented in the workstation. Midventricular short-
axis multidetector-row CT (MDCT) images in end diastole (A) and end systole (B) showing the manually traced endocardial borders and the
respective cross-sectional areas (A: 17.61 cm2, B: 8.55 cm2) in a patient with arrhythmogenic right-ventricular cardiomyopathy. Coronary
artery disease was ruled out using MDCT coronary angiography; global left-ventricular (LV) function was normal (LV-ejection fraction
72.9%). Calculation of diastolic and systolic left-ventricular volumes using Simpsons method based on the short-axis reformations.

Simpsons method. Endocardial contours of all short-axis (LV-EDV) and end-systolic (LV-ESV) volumes according to
images showing left-ventricular cavity are traced with papil- equation 3
lary muscles being included into the left-ventricular cavity to
calculate the cross-sectional area A of each section. Left-ven- LV-EF = [(LVEDV LVESV)/LVEDV] 100% (3)
tricular volumes (VSA) are calculated by adding all measured
cross-sectional areas AN multiplied by the intersection thick- PRACTICAL GUIDE TO FUNCTIONAL MDCT
ness S: SCAN PROTOCOL, IMAGE ACQUISITION,
AND DATA ANALYSIS
VSA = AN S (2)
DATA ACQUISITION AND IMAGE RECONSTRUCTION
Left-ventricular ejection fraction (LV-EF) is calculated for In initial studies, left-ventricular function data have been
both, long-axis and short-axis reformations, from end-diastolic determined from MDCT coronary angiography studies using
188 JUERGENS AND FISCHBACH

standard technique. The preparation and carrying-out of the DATA ANALYSIS USING DEDICATED
MDCT studies are performed according to the manufacturers ANALYSIS SOFTWARE
recommendations. For optimal contrast bolus timing, the use of Alternatively, left-ventricular volumes can be determined
a bolus tracking technique is not necessary; however, a good by analysis software adapted from cardiac MRI, which has
contrast between endocardium and left-ventricular cavity is been validated in research and clinical studies for the past
mandatory, enabling optimized delineation of endocardial con- decade. Following data acquisition and postprocessing, MPRs
tours. Following CT data acquisition, retrospectively ECG- are performed in short-axis orientation. For diastolic and sys-
gated image reconstruction in axial orientation is performed tolic MPRs, basal and apical sections are visually identified
using standard reconstruction parameters in overlapping slices and manually marked. To ensure reproducible data analysis,
(increment < slice thickness) in set increments. the first short-axis orientated section should be placed at the
In order to calculate volumetric data, the maximal systolic hearts base, covering the most basal portion of the left ven-
constriction and diastolic relaxation phase of the left ventricle tricle just forward of the atrioventricular ring. Endocardial
are identified by performing an axial image test series at borders are traced semiautomatically in both image series, start-
midventricular level in 5% steps through the entire RR interval. ing with a software-generated ellipsoid or circular figure placed
It is recommended that axial images show the anterior left- in the middle of the left-ventricular cavity. Contours are visu-
ventricular papillary muscles as well as the anterior and poste- ally reviewed for correctness and, if necessary, are manually
rior leaflet of the mitral valve (Fig. 7). End-diastolic and adjusted to endocardial contours.
end-systolic phase are identified electrocardiographically and
controlled visually as the images showing the largest and small- ACCURACY AND REPRODUCIBILITY OF MDCT
est left-ventricular cavity area, respectively. The reconstruc- ASSESSMENT OF LEFT-VENTRICULAR FUNCTION
tion windows determined in this way finally are used for So far, only initial results have been reported from studies
diastolic and systolic image calculation. According to a current with small numbers of patients evaluating left-ventricular vol-
study, a mean reconstruction window for diastolic image series ume and function assessment from MDCT in comparison to
was 83.1 4.1% (range 75 to 90) of the RR interval, whereas such established techniques as CVG (1619), echocardio-
systolic image reconstructions from MDCT data sets were per- graphy (20), and cardiac MRI (15,2123).
formed at 23.1 3.8% (range 20 to 35) of the RR interval (15). As has been demonstrated for cardiac MRI, an early study
From axial image reconstructions, short-axis orientated comparing MDCT and biplane CVG determination of left-
MPRs are created for diastolic and systolic image series. To ventricular ejection fraction found a better correlation if
encompass the entire left ventricle from the base of the heart to Simpsons method on short-axis reformations was used in
its apex, usually ten to eighteen sections of short-axis-orien- comparison to the area-length method. Regions with previous
tated MPRs are needed. In principle, any section thickness is myocardial infarction could clearly be delineated, showing a
possible. A section thickness of 8 mm using no intersection gap thinned LV wall and reduced or missing systolic myocardial
seems to be sufficient, according to experiences from cardiac wall thickening (16). Those results were confirmed by others
MRI (Fig. 5 and 6). (1719), showing that 3D data assessed from retrospectively
DATA ANALYSIS OF LEFT-VENTRICULAR ECG-gated MDCT studies enable calculation of left-ventricu-
VOLUMETRY AND FUNCTION lar volumes to estimate systolic function. All authors consis-
tently found an overestimation of left-ventricular end-systolic
Left-ventricular volumetric analysis can be performed
volumes, resulting in underestimation of left-ventricular ejec-
from MPRs orientated to left-ventricular long axis and short
tion fraction in MDCT. In comparison to transthoracic
axis, respectively, using standard 3D postprocessing software
echocardiography, no significant differences in left-ventricu-
as provided by several manufacturers. Recently, dedicated
lar volumes and ejection fraction were detected from MDCT
analysis software enabling semiautomated contour detection
in patients with both ischemic and nonischemic cardiomyopa-
might has been developed as an alternative for clinical and
thies (20). Measurements of end-diastolic and end-systolic
research purposes.
volumes as determined by MDCT revealed a close correlation
DATA ANALYSIS USING STANDARD 3D SOFTWARE to respective cardiac MRI measurements (Figs. 811, Table 2).
Diastolic and systolic left-ventricular volumes can be calcu- Acceptable limits of agreement between the two modalities
lated using standard planimetric software provided by the dif- were found, demonstrating a mean difference below 1% and
ferent manufacturers according to the area-length method based 1.5 mL regarding determination of left-ventricular ejection
on the long-axis view as well as to Simpsons method applied fraction and stroke volume (Fig. 12), respectively (16,22,23).
to contiguous short-axis reformations (see Assessment of Left The use of MDCT for the analysis of global left-ventricular
Ventricular Function Using CT above). The tracing of left- function has been limited by the lack of standardized analysis
ventricular endocardial contours as well as the determination software in clinical practice. Semi-automated or automated
of left-ventricular area and length in the long-axis view are methods of border detection have been developed in such car-
done using standard software tools with manual interaction. As diac imaging techniques as echocardiography or monoplane
in cardiac MRI, papillary muscles and endocardial trabeculae and biplane CVG (24). An initial study demonstrated that evalu-
should be included in the left-ventricular cavity and excluded ation of MDCT data sets lasted 1015 min using analysis soft-
from the left-ventricular volume. ware to perform semi-automated contour detection. Although
189
Fig. 7. Screenshots acquired from the scanners main workstation (Navigator) illustrating an axial-image-orientated test series in 5% steps (AJ) through the entire RR interval at
midventricular level. Axial sections show the anterior left ventricular papillary muscles as well as the anterior and posterior leaflet of the mitral valve. End-diastolic and end-systolic phase
CHAPTER 18/ MDCT ASSESSMENT OF LEFT-VENTRICULAR FUNCTION

were identified electrocardiographically and controlled visually as the images showing the largest and smallest left-ventricular cavity area, respectively. The corresponding reconstruction
window was used for image calculation.
189
190 JUERGENS AND FISCHBACH

Fig. 8. Short-axis reformations comparing four-slice multidetector-row CT (MDCT) technology (A and B) to cardiac magnetic resonance
imaging (MRI) (C and D) using conventional turbo gradient-echo sequences. (A) MDCT diastole; (B) MDCT systole (minor stair step artifacts
resulting from a mean heart rate of 69 beats per minute); (C) MRI diastole; (D) MRI systole. Left-ventricular ejection fraction was 51.8% as
determined from MDCT study in comparison to 57.9% as calculated from cine MRI.

Fig. 9. Short-axis reformations comparing four-slice multidetector-row CT (MDCT) technology and cardiac magnetic resonance imaging
(MRI) using steady-state free precession (SSFP) cine sequences. Diastolic (A,C) and systolic (B,D) short-axis reformations from MDCT
(A,B) and SSFP cine magnetic resonance imaging (MRI: C,D) studies from a 58-yr-old female patient with one-vessel coronary artery disease.
Normal dimensions of both ventricles and myocardial wall thickness, normal global left-ventricular function (left-ventricular ejection fraction
= 67.5% MDCT vs 67.6% MRI).
 CHAPTER 18/ MDCT ASSESSMENT OF LEFT-VENTRICULAR FUNCTION 191

Fig. 10. Diastolic (A,C) and systolic (B,D) short-axis reformations from multidetector-row CT (MDCT) (A,B) and steady-state free precession
cardiac magnetic resonance imaging MRI (C,D) studies obtained in a 72-yr-old male patient with three-vessel coronary artery disease who had
experienced previous inferior myocardial infarction. Hypokinetic infero-septal and akinetic inferior wall of left ventricle can be clearly
identified (white arrow; left-ventricular ejection fraction = 64.1% MDCT vs 62.9% MRI).

endocardial left-ventricular contours had to be manually cor-

rected in some cases, there was not a relevant difference in
comparison to cardiac MRI data sets, with a mean analysis time
of approx 10 min. Thus, in selected patients evaluated for coro-
nary artery disease, semiautomated analysis software enabled
reliable left-ventricular volumetric measurements and func-
tion analysis from MDCT data sets (Table 3) in comparison to
SSFP cardiac MRI (15).
With regard to left-ventricular end-diastolic and end-sys-
tolic volumes as well as ejection fraction as determined by
MDCT studies, an inter-observer variability between 6 and 8%
was reported (16,22).
It has been generally accepted that a temporal resolution of
50 msachievable by established techniques as CVG, nuclear
techniques, EBCT, and cardiac MRIis mandatory for a repro-
ducible evaluation of global left-ventricular function. Currently
available MDCT techniques achieve a temporal resolution
Fig. 11. Systolic short-axis reformation from sixteen-slice
multidetector CT at midventricular level from a 72-yr-old man with down to 105 ms, depending on the patients heart rate. How-
a history of three-vessel coronary artery disease having undergone ever, initial studies published by different investigators suggest
bypass surgery (white arrow indicates a tiny vascular clip from a that left-ventricular volumetric and functional evaluation by
bypass to the left descending coronary artery). As a result of recurrent MDCT gives reliable results in comparison to reference modali-
myocardial infarction the multiplanar reformation in the short-axis ties. A current study compared data of patients with a mean
demonstrates an absent regional contraction of left-ventricular myo-
cardium in the lateral and inferior wall of left-ventricular myocar- heart rate below and above 65/min, respectively, revealing no
dium including the posterior papillary muscles (dark arrows). APM, significant differences with regards to reliability of volumetric
anterior papillary muscle; PPM, posterior papillary muscle. measurements (15) by MDCT (Table 4).
192 JUERGENS AND FISCHBACH

Table 2
Comparison of Left-Ventricular Volumetric and Functional Parameters As Determined
From MDCT of the Heart in Comparison to CVG and MRI Using TGrE and SSFP Cine Sequences

Modality LV-EF:
compared MDCT vs
Reference N to MDCT LV-EDV LV-ESV LV-EF other modality

Juergens et al. (22) 28 TGrE-MRI 0.92 0.90 0.90 7.9 5.6 %

Wintersperger et al. (17) 25 CVG 0.59 0.88 0.82 17 9 %
Hundt et al. (19) 30 CVG 0.72 0.88 0.76 13.7 11 %
Erhard et al. (27) 7 SSFP-MRI 0.93 0.95 0.83 3.8 9.4 %
Mahnken et al. (23) 16 SSFP-MRI 0.99 0.99 0.98 0.9 3.6 %
Juergens et al. (15) 30 SSFP-MRI 0.93 0.94 0.89 0.25 4.9 %
CVG, cine ventriculography; MDCT, multidetector-row computed tomography; MRI, magnetic resonance imaging; SSFP, steady-state free
precession; TGrE, Turbo gradient-echo.

Fig. 12. Bland-Altman plots of left-ventricular end-diastolic volume

(LVEDV) (A), left-ventricular end-systolic volume (LVESV) (B),
and left-ventricular ejection fraction (LVEF) (C), showing the rela-
tion between differences and means between multidetector-row CT
(MDCT) and cine steady-state free precession magnetic resonance
imaging (MRI) for each parameter. The difference (y axis) between
each pair ([mean MDCT] [mean MRI]) is plotted against the average
value (x axis) of the same pair ([(mean MDCT) + (mean MRI)]/2). (A
C: solid lines = mean value of differences; short dotted lines = mean
value of differences 2 standard deviation.)

limitations, however, a further significant increase in tempo-

ral resolution may be realized only by separating physiological
and physical acquisition time. Multisegment-reconstruction
algorithms have become available for MDCT that process data
from several gantry rotations and can reduce the temporal reso-
lution down to 7090 ms (25,26). Initial results, however, sug-
gest that the resulting decrease in spatial resolution limits
left-ventricular function analysis by multisegment-algorithms
FURTHER DEVELOPMENTS in comparison to cardiac MRI.
Progress concerning a more accurate determination of end-
systolic frames, and possibly analysis of regional myocardial MDCT VS CARDIAC MRI
function, can be expected from MDCT systems with reduced In comparison to MDCT, the advantages of cardiac MRI are
gantry rotation time (below 500 ms) and a concomitant increase the lack of radiation exposure, avoidance of iodinated contrast
in temporal resolution (down to 105 ms). As a result of physical media, and better temporal resolution. Furthermore, short-axis
 CHAPTER 18/ MDCT ASSESSMENT OF LEFT-VENTRICULAR FUNCTION 193

Table 3
Volumetric and Functional Left-Ventricular Parameter
As Determined by Means of MDCT and SSFP Cine MRI, N = 30

MDCT SSFP Cine MRI MDCT vs MRI

Mean SD Range Mean SD Range r= p<

LV-EDV [mL] 13831 87228 14232 88211 0.93 0.001

LV-ESV [mL] 5321 20108 5422 27123 0.94 0.001
LV-EF [%] 6110 4077 6210 4175 0.89 0.001
LV-SV [mL] 8420 48140 8621 51142 0.88 0.001

CI, confidence interval; LV-EDV, left-ventricular end-diastolic volume; LV-ESV, left-ventricular end-
systolic volume; LV-EF, left-ventricular ejection-fraction; LV-SV, left-ventricular stroke volume; MDCT,
multidetector-row computed tomography; MRI, magnetic resonance imaging; SD, standard deviation; SSFP,
steady-state free precession. From ref. 15.

Table 4
Volumetric and Functional Left-Ventricular Parameter As Determined by MDCT and SSFP Cine MRIa

HR < 65/min HR 65/min

N = 20 N = 10
Mean SD Mean SD
HRMean 58.1 5.5 71.2 6.2 < 0.001
HRRange 4764 6588 < 0.001
MDR-CT Cine MRI p MDR-CT Cine MRI p

LV-EDV (mL) 142.7 32.8 143.7 34.3 n.s. 129.3 22.3 134.6 25.8 n.s.
LV-ESV (mL) 57.3 20.9 56.5 23.8 n.s. 45.9 17.3 48.1 17.7 n.s.
LV-EF (%) 58.9 10.6 61.2 10.4 n.s. 65.3 9.8 64.7 8.9 n.s.
HR, heart rate; LV-EDV, left-ventricular end-diastolic volume; LV-ESV, left-ventricular end-systolic volume; LV-EF,
left-ventricular ejection-fraction; MDCT, multidetector-row computed tomography; MRI, magnetic resonance imaging; SD,
standard deviation; SSFP, steady-state free precession; n.s., not statistically significant.
aComparison of MDCT data sets reconstructed by biphasic adaptive cardiac volume algorithm with regard to mean HR
of below 65/min vs 65/min. From ref. 15.

images are readily available, and time-consuming secondary SUMMARY

reformations as required in MDCT are not needed. Therefore, The assessment of global functional and volumetric left-
the software-assisted automatic or even the interactive creation ventricular parameters has become feasible using MDCT stud-
of short-axis orientated reformations during data acquisition, ies of the heart. Although this application is still in an
as provided by state-of-the art cardiac MR scanners, would lead experimental stage, initial studies from different study groups
to an improved applicability of global left-ventricular volumet- have revealed promising results in comparison to established
ric and functional measurements in a clinical setting. techniques as CVG, echocardiography, and cardiac MRI. Recent
Considering contrast media application, radiation exposure, developments in postprocessing and analysis software have
and limited temporal resolution, the use of MDCT solely for improved clinical applicability.
analysis of cardiac function parameters seems not unreason-
able at present time. Regarding radiation exposure, a substan- REFERENCES
tial reduction may be achieved by the application of an 1. White HD, Norris RM, Brown MA, Brandt PWT, Whitlock RML,
ECG-triggered tube current modulation. However, the combi- Wild CJ. Left ventricular end-systolic volume as the major determi-
nation of survival after recovery from myocardial infarction. Circu-
nation of noninvasive coronary artery imaging and determina-
lation 1987;76:4451.
tion of global left-ventricular function within one single 2. Schocken DD, Arrieta MI, Leaverton PE, Ross EA. Prevalence and
breath-hold MDCT study, however, might be an interesting mortality rate of congestive heart failure in the United States. J Am
approach to a fast and conclusive cardiac workup in different Coll Cardiol 1992;20:301305.
cardiac diseases, e.g., patients with suspected coronary artery 3. Gerber TC, Behrenbeck T, Allison T, Mullan BP, Rumberger JA,
Gibbons RJ. Comparison of measurement of left ventricular ejection
disease. This application of MDCT might be of clinical rel- fraction by Tc-99 m sestamibi first-pass angiography with electron
evance for patients who are not suitable for cardiac MRI be- beam computed tomography in patients with anterior wall acute
cause of specific contraindications. myocardial infarction. Am J Cardiol 1999;83:10221026.
194 JUERGENS AND FISCHBACH

4. Zeidan Z, Erbel R, Barkhausen J, Hunold P, Bartel, Buck T. Analysis 17. Wintersperger BJ, Hundt W, Knez A, et al. Left ventricular systolic
of global systolic and diastolic left ventricular performance using function assessed by ECG gated multirow-detector spiral computed
volume-time curves by real-time three-dimensional echocardio- tomography (MDCT): comparison to ventriculography. Eur Radiol
graphy. J Am Soc Echocardiogr 2003;16:2937. 2002;12:S192.
5. Lethimonnier F, Furber A, Balzer P, et al. Global left ventricular 18. Herzog C, Mehtap A, Abolmaali N, Balzer JO, Schaller S, Vogl TJ.
cardiac function. Comparison between magnetic resonance imag- Value and reproducibility of multidetector-row CT in the assess-
ing, radionuclide angiography, and contrast angiography. Invest ment of cardiac function. Eur Radiol 2002;12:S192.
Radiol 1999;34:199203. 19. Hundt W, Siebert K, Becker C, Knez A, Rubin GD, Reiser MF.
6. Lipton MJ, Higgins CB, Boyd DP. Computed tomography of the Assessment of global left ventricular function: comparison of multi-
heart: evaluation of anatomy and function. J Am Coll Cardiol 1985;5: detector computed tomography with left ventriculography. Radiol-
55S69S. ogy 2002;252:388(P).
7. Woo P, Mao S, Wang S, Detrano RC. Left ventricular size deter- 20. Sakuma T, Nishioka M, Fukuda K. Noninvasive determination of
mined by electron beam computed tomography predicts significant left ventricular volumes and ejection fraction assessed by multi-
coronary artery disease and events. Am J Cardiol 1997;79:12361238. detector-row CT. Radiology 2001;221:452(P).
8. Bloomgarden DC, Fayad ZA, Ferrari VA, Chin B, St. John Sutton 21. Halliburton S, Petersilka M, Schvartzman P, Obuchowski N, White
M, Axel L. Global cardiac function using fast breath-hold MRI: R. Validation of left ventricular volume and ejection fraction mea-
validation of new acquisition and analysis techniques. Magn Reson surement with multi-slice computed tomography: comparison to cine
Med 1997;37:683692. magnetic resonance imaging. Radiology 2001;221:452(P).
9. Thiele H, Paetsch I, Schnackenburg B, et al. Improved accuracy of 22. Juergens KU, Fischbach R, Grude M, et al. Evaluation of left ven-
quantitative assessment of left ventricular volume and ejection tricular myocardial function by retrospectively ECG-gated multislice
fraction by geometric models with steady-state-free precession. spiral CT in comparison to cine magnetic resonance imaging. Eur
J Cardiovasc Magn Reson 2002;4:327339. Radiol 2002;12:S191.
10. Miller S, Simonetti OP, Carr J, Kramer U, Finn JP. MR imaging of 23. Mahnken AH, Spntrup E, Wildberger JE, et al. Quantification of
the heart with cine true fast imaging with steady-state precession: cardiac function with multislice spiral CT using retrospective ECG-
influence of spatial and temporal resolutions on left ventricular func- gating: comparison with MRI. Fortschr Rntgenstr 2003;175:
tional parameters. Radiology 2002;223:263269. 8388.
11. Lorenz CH, Walker ES, Morgan VL, Klein SS, Graham TP Jr. 24. Waiter GD, McKiddie FI, Redpath TW, Semple SIK, Trent RJ.
Normal human right and left ventricular mass, systolic function, and Determination of normal regional left ventricular function from cine-
gender differences by cine magnetic resonance imaging. J Car- MR images using a semi-automated edge detection method. Magn
diovasc Magn Reson 1999;1:721. Reson Imag 1999;17:99107.
12. Harell GS, Guthaner DF, Breiman RS, et al. Stop-action cardiac 25. Boese JM, Bahner ML, Albers J, van Kaick G. Optimization of
computed tomography. Radiology 1977;123:515517. temporal resolution in CT using retrospective ECG-gating. Radio-
13. Lackner K, Thurn P. Computed tomography of the heart: ECG-gated loge 2000;40:123129.
and continuous scans. Radiology 1981;140:413420. 26. Halliburton SS, Boese JM, Flohr TG, Lieber ML, Kuzmiak SA,
14. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial White RD. Improved volumetric analysis of left ventricle using
and temporal resolution for electrocardiogram-gated multislice cardiac multi-slice computed tomography (MSCT) with high
spiral CT of the heart. J Comput Assist Tomogr 2001;25:907923. temporal resolution image reconstruction. Radiology 2002;252:
15. Juergens KU, Grude M, Maintz D, et al. Multi-detector row CT of 389(P).
left ventricular function with dedicated analysis software versus MR 27. Ehrhard K, Oberholzer K, Gast K, Mildenberger P, Kreitner KF,
imaging: initial experience. Radiology 2004;230(2):403410. Thelen M. [Multi-slice CT (MSCT) in cardiac funciton imaging:
16. Juergens KU, Grude M, Fallenberg EM, et al. Using ECG-gated threshold-value-supported 3D volume reconstructions to determine
multidetector CT to evaluate global left ventricular myocardial func- the left ventricular ejection fraction in comparison to MRI] [Article
tion in patients with coronary artery disease. Am J Roentgenol 2002; in German]. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr.
179:15451550. 2002;174(12):15661569.
 CHAPTER 19 / IMAGING INTRAMYOCARDIAL MICROCIRCULATORY FUNCTION 195

19 Imaging Intramyocardial Microcirculatory

Function Using Fast Computed Tomography

STEFAN MHLENKAMP, MD, AXEL SCHMERMUND, MD, BIRGIT KANTOR, MD,

RAIMUND ERBEL, MD, AND ERIK L. RITMAN

WHY IMAGE CORONARY cardial coronary microvessels (1115). Accordingly, risk fac-
MICROVASCULAR FUNCTION? tor-induced epicardial atherosclerosis is frequently, but not
Coronary arterial disease is currently diagnosed and treated always, preceded by impaired vasomotion of coronary
primarily on the basis of its impact on the large-diameter epi- microvessels (1618). Different risk factors, coronary athero-
cardial arteries. A structural change, usually a localized nar- sclerosis, and also other diseases such as aortic valve stenosis
rowing (stenosis) of a coronary artery lumen, is generally or cardiomyopathies may all ultimately result in reduced coro-
detected and quantitated by selective coronary angiography. nary flow reserve, but their functional impact on the coronary
However, by the time the epicardial artery stenosis results in microvasculature, which predominantly regulates regional
reduced epicardial flow and the patient becomes symptomatic, myocardial blood flow (4,1922), is markedly different.
it is generally too late to arrest (much less reverse) the disease We have explored fast X-ray computed tomography (CT) as
process in that artery. Therefore, a noninvasive test that iden- a minimally invasive method to quantitate intramyocardial
tifies presymptomatic, subclinical disease should result in ini- microcirculatory function in different clinically relevant set-
tiation of therapy at a time when the disease process is still tings, and we propose to use the blood volume-to-flow relation-
reversible. Causal risk factorsi.e., dyslipidemia, arterial ship to identify characteristic changes in the functional behavior
hypertension, diabetes mellitus, and smokingare responsible of the coronary microcirculation.
for the majority of coronary artery disease cases, and risk-fac-
FAST CT-BASED INDICES OF INTRAMYOCARDIAL
tor modification in high-risk asymptomatic individuals has
MICROVASCULAR FUNCTION
been shown to improve outcome (1). Early risk stratification,
aggressive preventive counseling, and therapy in high-risk The intramyocardial microcirculation consists of arteries
subjects is therefore recommended (2,3). However, limited that are up to 0.5 mm in lumen diameter and progressively
economic resourses warrant careful patient selection and branch to the 5-m-diameter capillaries. Unfortunately, these
appropriate therapeutic aggressiveness. Hence, any imaging small microvessels cannot be individually visualized by clini-
tool for this purpose is required (1) to allow identification and cally applicable imaging methods. Consequently, some indi-
quantification of early disease, and (2) to be sensitive enough rect estimates of microvascular vasomotion must be made. We
to ascertain therapeutic efficacy over time to justify continua- do this using the intramyocardial intravascular blood volume-
tion or modification of the initiated therapy. to-flow relationship. Minimally invasive estimates of micro-
Coronary atherosclerosis affects the intramyocardial arte- vascular blood volume and perfusion in absolute numbers can
rial microcirculation early in the disease process, well before be obtained from the analysis of whole-body CT images using
the epicardial vessels are hemodynamically compromised intravascular contrast-medium dilution curves. The principles
(46). If symptoms are present, they frequently cannot suffi- of indicator dilution techniques using X-ray CT and contrast
ciently be explained by the degree of epicardial lumen narrow- agent as the indicator have been described in detail elsewhere
ing. Even a normal coronary angiogram may be associated (2325). Its basis is the StewartHamilton Equation (2631):
with typical angina pectoris and inducible ischemia in Blood volume (BV) = Flow (F) Mean transit time (MTT)
noninvasive stress tests (7,8), which is usually attributed to
With the development of fast-CT technology such as the
microvascular functional impairment and termed Syndrome X
dynamic spatial reconstructor (DSR) (32) and electron beam
(9) or microvascular angina (10). Risk factors can frequently
CT (EBCT) scanners (GE-Imatron, South San Francisco, CA)
be identified, that can induce endothelial dysfunction and hence
(33), quantitative estimates of intramyocardial fractional blood
be responsible for a reduced vasomotor capacity of intramyo-
volume and perfusion in vivo have been obtained in the beating
heart (24,25,3443) and other organs like the kidney (4446),
From: Contemporary Cardiology: CT of the Heart: the brain (23,47), the liver (48), and the lung (4950).
Principles and Applications To estimate intramyocardial (fractional) blood volume and
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
flow in the beating heart in vivo, we used the following equa-

195
196 MHLENKAMP ET AL.

Fig. 1. (A) Cross-sectional CT image through a pig thorax at a mid-

ventrucular level showing the left-ventricular cavity (LV) and the
surrounding LV wall (left). The left anterior descending artery
(LAD) region of interest was previously highlighted using selective
injection of contrast agent into the LAD. The image sequence pro-
vides time-intensity curves (right), where each data point represents
the average attenuation (in Hounsfield units [HU]) at enddiastole
within the region of interest. These curves are used to obtain indexes
of intramyocardial blood volume and perfusion (see 1B and text;
modified from ref. 41). (B) Schematic illustration of time-density
curves obtained from left-ventricular and intramyocardial regions
of interest. The areas under the curves and the first moment of these
functions can be computed and used to obtain indices of
intramyocardial blood volume and perfusion (see text for details).
AT, arrival time, MTT, mean transit time.

tions, the components of which can be derived in absolute (24), against direct concurrent volumetric measurements of
numbers from indicator dilution curves (Fig. 1A,B). myocardial blood volume (35), and against intracoronary
Doppler ultrasound (40,41), with a scan-to-scan variability for
1) BV[mL/cm3] = Cmyo(t) dt / CLV(t) dt both blood volume and perfusion below 10% (41).
0 0
For the interested reader we add the derivation of equation 3:

Myocardial perfusion (mL/min/g) relates myocardial blood
2) MTT[s] = [ C(t) t dt / C(t) dt] AT
0 0 flow (mL/min) to the mass of exanguinated myocardial muscle
(g). In order to compute perfusion from fractional blood vol-
3) F[mL/min/g] = [BV/(1 BV)] (60/MTT) (1/1.05), ume (mL/cm3) and mean transit time (s), which we obtain from
with BV = intramyocardial fractional blood volume = the frac- our image sequences, the right side of the equation F = BV /
tion of myocardium that is blood, MTT = mean transit time, MTT (1) must be divided by (1 BV) and the specific gravity
F = myocardial perfusion, of myocardial tissue, i.e., 1.05 g/mL, for the following reason:
We first do BV = Vb/(Vb + VM) (2), which is the fraction of

CLV/myo(t) dt = the areas under the left-ventricular (input) myocardium that is blood, where Vb/M are the intravascular
0 volumes of blood/myocardium within the region of interest.
curve and intramyo-cardial (intravascular) time-density curves, Next, VM = gM/dM (3), where gM is the mass and dM is the
and AT = arrival time = the time interval between the start of specific density of myocardial muscle. From (2) follows (4): Vb
injection of the bolus and the initial rise (appearance) of the = [BV/(1 BV] VM, and from substituting (3) in (4) follows
myocardial tissue curve above baseline (Fig. 1). Computation (5): Vb/gM = [BV/(1 BV] 1/dM = the intravascular volume of
of myocardial perfusion and blood volume has been validated blood per gram of myocardial muscle. Substituting (5) into (1)
against radiolabeled microsphere-based estimates of perfusion immediately gives equation 3).
 CHAPTER 19 / IMAGING INTRAMYOCARDIAL MICROCIRCULATORY FUNCTION 197

(Note that for historic reasons we continue to use F as diameter so much as to change in the number of vessels per-
the symbol and synonym for perfusion, given in [mL/min/g], fused. The vascular resistance in these vessels is thus inversely
but this is distinctly different from flow, given in [mL/min].) related to the number of recruited vessels. The intraluminal
When nonionic iodinated contrast material is used, first- blood volume of the recruitable vessels (BVr) should be pro-
pass myocardial image intensities do not right away return to portional to the number of recruited vessels; hence it should
baseline after reaching their peak value, but arrive at a plateau follow that F BVr. Because the recruitable vessels and
before slowly returning to baseline. This phenomenon is pri- nonrecruitable vessels are in series, the following relationship
marily and immediately attributable to extravascular accumu- should hold:
lation of contrast medium in small part, and after a delay caused
BV= BVr + BVnr = aF + b F
by recirculation of contrast material. While recirculation can
where a and b are coefficients to be established experimentally.
be observed near the end of the scan sequence, a measurable
Obviously, such a model is simplistic and limited in its
amount of contrast material leaves the intravascular space
ability to reflect complicated microvascular vasomotor control
almost immediately during its first pass through the myocar-
mechanisms, and should be considered a preliminary and ongo-
dium (5153). The extravascular volume of distribution of
ing attempt to obtain functional information about the coronary
contrast material is about 30% of the total intramyocardial
microcirculation in vivo (42).
volume (54) and therefore has a significant impact on
quantitation of intramyocardial blood volume and perfusion EXPERIMENTAL APPLICATIONS OF CARDIAC CT
(52) even if only a small percentage leaves the vascular com- TO IMAGE MICROVASCULAR (DYS)FUNCTION
partment. Thus, extravascular accumulation of dye violates an
CHANGES IN CORONARY MICROVASCULAR
assumption of indicator dilution principles (28) but can be FUNCTION WITH GROWTH
corrected for as described below. Extravascular accumulation The coronary microcirculation undergoes functional and
of contrast material can also be used to estimate microvascular morphologic changes not only in a number of frequent diseases
permeability: as stated by the CroneRenkin equation, extrac- but also under various physiological conditions, such as growth,
tion rate depends on the flow (F) and the permeability-surface maturation, and exercise (6164). We assessed the ability of
product (PS): E = 1 e(PS/F) (5557). Accordingly, an index of EBCT to quantitatively image such changes over a period of
PS can be computed from CT images if the extraction fraction 3 mo using a porcine model (41). In 13 male pigs we measured
(E) and F are known: PS = F ln(1 E) (53). The degree of myocardial blood volume and perfusion at baseline (approx
extravascular accumulation of contrast material can be esti- 3-mo-old animals) and again 3 mo later.
mated from the residual image intensities (24,46,53,58) and be At rest, fractional blood volume and perfusion values
used to quantitatively estimate the microvascular extraction remained unchanged from baseline to follow-up. However, the
fraction (53). Lerman et al. proposed to separate the intravas- percent increase in blood volume and perfusion in response to
cular from the extravascular component of the myocardial time- selective intracoronary infusion of adenosine was significantly
intensity curve by virtue of the components of an extended higher at follow-up as compared with baseline values. These
gamma variate (45). This approach was also used to obtain increases in intramyocardial microvascular response to adenos-
indices for extraction rate as a basis to estimate microvascular ine from baseline to follow-up were consistent with changes in
permeability (59,60). Doppler ultrasound-based intracoronary blood flow reserve
(Fig. 2). Similar findings were previously described in children
THE BLOOD VOLUME-TO-FLOW during growth (65) and in rabbit hearts (66). Especially the data
(BV-TO-F) RELATIONSHIP from the latter study suggested an immature vasoreactivity as
The intramyocardial microcirculation (i.e., vessels with a the main reason for the attenuated response to adenosine in
lumen diameter of less than 500 m that are not individually young animals (66).
resolvable on whole-body CT scans) is modeled as two sets of Both microvascular componentsrecruitable and
vessels arranged in series. One set consists of functionally non- nonrecruitable microvesselscontributed to the long-term
recruitable conducting vessels, which are assumed to be always increase in microvascular functional reserve, with no discern-
patent and perfused but can change their diameter in response ible change in blood volume distribution among the microvas-
to various stimuli. The vascular resistance in these vessels is cular components despite changes in microvascular reactivity.
assumed to follow Poiseuilles law, which relates the flow to In the BV-to-F relationship, all resting and adenosine values at
the fourth power of the vessel radiusthat is, the square of the baseline and 3 mo later followed one curvilinear line (Fig. 3),
vessel cross sectional area (CSA); thus, F CSA2. Because which, in agreement with previous findings (65,66), supported
the length of these vessels does not change much relative to the growth-, maturation-, or age-related change in response to
the change in cross-sectional area, the intraluminal volume adenosine with no change in the relation of recruitable and non-
of these nonrecruitable vessels (BVnr) should be proportional recruitable microvessels in the healthy porcine heart.
to the cross-sectional area. Hence, it should follow that Our data demonstrated that X-ray CT permits quantitation
of longitudinal changes in microvascular function over a period
F BVnr2 or F BVnr . of at least 3 mo. This is not only important for minimally inva-
The other microvascular component consists of function- sive experimental and clinical research but, together with find-
ally recruitable vessels (presumably mostly capillaries, small ings demonstrated below, it indicates the ability to image
arterioles, and venules), which are assumed not to change in long-term changes in response to therapy.
198 MHLENKAMP ET AL.

functional components of the coronary microcirculation in

normal compared to hypercholesterolemic pigs (40,59,60).
Intramyocardial microvascular functional reserve in response
to selective intracoronary infusion of adenosine was impaired in
hypercholesterolemic vs normal control pigs, consistent with
our intracoronary Doppler ultrasound measurements (Fig. 4)
and experimental and clinical observations by others (59,67).
In contrast to our findings in young presumably immature nor-
mal control pigs (above), the BV-to-F relationship demon-
strated a significantly attenuated increase in blood volume per
increase in perfusion in hypercholesterolemic versus normal
mature control pigs (Fig. 5), which can be explained by HC-
induced endothelial dysfunction and its consequences for
Fig. 2. Quantitation of long-term changes in microvascular func-
tional response to adenosine infusion at baseline (3-mo-old pigs) and microvascular vasomotion: in hypercholesterolemia it is pri-
at follow-up (6-mo-old pigs). Within the baseline and follow-up stud- marily the oxidized low-density lipoprotein- and free oxygen
ies separately, Doppler-ultrasound-based intracoronary blood flow radicals-induced inhibition of nitric oxide (NO)-dependent
(CBF), EBCT-based blood volume (BV), and perfusion significantly vasorelaxation (68,69), in combination with reduced NO
increased in response to adenosine (p < 0.001, respectively). The bioavailability (70,71), that plays a key role for the observed
increases in BV and perfusion at follow-up in response to adenosine
were significantly greater compared to baseline, consistent with CBF endothelial microvascular dysfunction both in animal experi-
measurements (p < 0.001, respectively) (see text for details; modified ments and in humans (72,73). Stepp et al. demonstrated that
from ref. 41). flow-induced vasodilation in arteries with diameters >160 m
is also regulated to some degree by endothelial NO release (74).
This explains the observed effects of adenosine, a substance
that results in vasodilation of resistance and conducting
microvessels (21,75) and capillary recruitment (76), predomi-
nantly via endothelium-independent mechanisms.
At rest, blood volume in the recruitable component was
decreased, but it was increased in the nonrecruitable compo-
nent in hypercholesterolemic pigs (40). Lerman et al. previ-
ously reported a similar change in microvascular blood volume
distribution after administration of NG-monomethyl-L-argin-
ine (L-NMMA), a competitive inhibitor of NO synthase (39).
Since experimental hypercholesterolemia is associated with a
decrease in endogenous NO bioavailability (70,71), it probably
induces physiological effects similar to those with L-NMMA.
In response to 5-min continuous infusion of nitroglyceride
(NTG), we observed an increase in both the recruitable and
Fig. 3. Fast CT-based blood volume (BV) to flow (F) relationship in nonrecruitable microvascular component in control pigs,
normal hearts imaged twice before and after 3 mo (same data as in although NTG primarily acts on microvessels >200 m in
Fig. 2). Values are given at baseline (resting conditions = open red diameter. These findings can be explained by a modulating
circles, adenosine = closed red circles) and at follow-up (resting con-
ditions = open blue squares, adenosine = closed blue squares). Despite
effect of upstream resistance vessel function on capillary
a different microvascular functional response to adenosine at baseline recruitment (40). In hypercholesterolemic pigs however, blood
and follow-up, all values follow the same curvilinear line (see text for volume remained unchanged in both the recruitable and the
details; modified from ref. 41). nonrecruitable component, likely a result of impaired resis-
tance vessel vasomotion.
The characteristic behavior of different functional micro-
vascular components in hypercholesterolemic animals as
MICROVASCULAR ENDOTHELIAL DYSFUNCTION imaged by fast CT is consistent with experimental observa-
IN HYPERCHOLESTEROLEMIA tion by others using different methodology, which supports the
Imaging Diseased Coronary Microvascular Function ability of fast CT to obtain meaningful information on
in Chronic Hypercholesterolemia intracoronary microvessels in chronic diseases that alter
Chronic hypercholesterolemia (HC) leads to endothelial microvascular function.
damage and dysfunction, as well as impaired coronary micro- Imaging Therapeutic Reversal of Coronary Microvascular
vascular function. Using a porcine model of chronic hyper- Dysfunction in Chronic HC
cholesterolemia, EBCT technology was used to evaluate the Using the same porcine model of hypercholesterolemia and
degree to which it is able to detect, discriminate, and quantify similar methodology, the effect of HC with and without high-
microvascular permeability and blood volume in the different dose antioxidants (vitamins E and C) and statins on fast-CT-
 CHAPTER 19 / IMAGING INTRAMYOCARDIAL MICROCIRCULATORY FUNCTION 199

Fig. 4. Quantitation of changes in microvascular functional response

to adenosine infusion in normal control pigs (C) and chronic hyper-
cholesterolemic (HC) pigs (HC). Doppler-ultrasound-based intra- Fig. 6. Relative change (percent compared with resting conditions) of
coronary blood flow (CBF) and electron beam CT-based blood microvascular permeability in response to intravenous adenosine in
volume (BV) significantly increased in response to adenosine (p < three groups: normal controls, hypercholesterolemic (HC) pigs, and
0.001, respectively). However, the increase in microvascular response HC pigs treated with statins (S) or antioxidants (AO, vitamins E and
to adenosine was significantly blunted in HC-pigs compared to nor- C). Microvascular permeability is significantly increased in HC pigs,
mal controls (see text for details; modified from ref. 40). but long-term therapy preserves permeability (see text for details;
modified from refs. 59 [shaded bars] and 60 [full colors]).

the established role of lipid oxidation in endothelial dysfunction,

it was suggested that the beneficial effects are mediated via
antioxidative, vasoprotective properties of statins.
The above findings are in accordance with our own observa-
tions in hypercholesterolemic pigs and also support the ability
of fast-CT technology to quantitatively assess coronary
microvascular dysfunction in chronic diseases such as HC.
Most importantly, these studies demonstrate that reversal of
HC-induced endothelial dysfunction using statins and vitamins
can be quantitated using minimally invasive fast-CT image
analysis.
CORONARY MICROVASCULAR DYSFUNCTION
IN SUBCLINICAL CORONARY ARTERY STENOSIS
Fig. 5. Blood volume-to-flow relationship in normal controls and Coronary angiography remains the gold standard to
hypercholesterolemic (HC) pigs. A 3-mo high-cholesterol diet led to assess advanced stages of epicardial coronary artery disease.
endothelial dysfunction with significant impairment in mobilization While there is no doubt about the hemodynamic relevance of
of microvascular blood volume for any given perfusion value (see text
high-grade lesions (84), functional assessment of apparently
for details; modified from ref. 40)
mild or intermediate lumen diameter reductions (5075%)
remains difficult (85,86). Lumen narrowing less than 50% is
frequently associated with a normal epicardial flow reserve
based myocardial perfusion and microvascular permeability and therefore considered clinically nonsignificant. However,
has been assessed at rest and during intravenous infusion of such nonsignificant lesions may already be associated with
adenosine (59,60). These studies found a significant increase in functional microvascular alterations (17). In a canine study,
myocardial perfusion and unaltered microvascular permeabil- Wu et al. demonstrated that intramyocardial blood volume
ity in response to adenosine in normal control pigs, whereas in increases with increasing pressure gradients across a nonsig-
hypercholesterolemic animals, myocardial perfusion remained nificant stenosis (P 40 mmHg Stenosis 55%), while
unaltered after adenosine infusion and microvascular perme- perfusion initially remains unaltered (58,87). When the pres-
ability significantly increased (Fig. 6). However, animals sure gradient across a more severe epicardial stenosis rises to
treated with antioxidants and statins showed a response to ad- values much above 50 mmHg, compensatory mechanisms
enosine comparable to normal control pigs (59,60), which is will exhaust and regional blood flow drops in proportion to a
consistent with clinical observations (7779). Since statins do decrease in blood volume (58). Similar results have been found
not alter plasma cholesterol levels to the same degree as they do using echocardiography in a canine open-chest model (88).
in humans (80), preservation of microvascular (endothelial) These observations are consistent with intramyocardial release
function is likely independent of their lipid-lowering proper- of adenosine in response to mild local ischemia, leading to
ties (81). In accordance with previous studies (82,83) and with compensatory vasodilation of resistance vessels and preserva-
200 MHLENKAMP ET AL.

tion of regional perfusion (89). The increase in microvascular

blood volume in nonsignificant epicardial lumen reduction in
order to preserve microvascular perfusion is also reflected in
the BV-to-F relationship (Fig. 7) (87), which may be a more
sensitive index for microvascular impairment than the coro-
nary flow reserve in early CAD (87).
CORONARY MICROVASCULAR DYSFUNCTION
IN CORONARY MICROEMBOLIZATION
Data from large clinical trials suggest that embolization of
plaque material or microthrombi from dysfunctional epicardial
arterial endothelium into the distal coronary microcirculation
has long been an underappreciated, possibly common, event
(90-92), particularly during acute coronary syndromes (93,94)
and intravascular coronary interventions (95,96). Interestingly,
resting epicardial arterial blood flow may even be elevated
despite being associated with myocardial ischemia, necrosis, Fig. 7. Effect of experimental canine coronary artery stenosis on
and contractile dysfunctiona phenomenon termed perfusion- myocardial blood volume and perfusion. Increasing nonsignificant
stenosis, which is not associated with decreased epicardial flow
contraction mismatch (92,97). The functionally frustrate
reserve, leads to a progressive increase in blood volume relative to
increase in coronary flow was attributed to vasodilation of perfusion values. This is characteristically different from changes
adjacent nonembolized vessels in response to adenosine release observed in microvascular disease (see text for details; modified
from the surface of embolized myocardium (98100). It may be from ref. 87).
accompanied by a reduction in maximum achievable epicardial
coronary blood flow despite normal or restored coronary artery
lumen diameter (101), suggesting involvement of the coronary explained by the increase in blood volume and perfusion within
microvasculature for these observations. neighboring terminal arterioles and capillaries that is greater
In another porcine study, we assessed the degree to which than the loss of blood volume and perfusion resulting from the
different sized coronary microvessels contribute to the patho- embolized vessels. Such an excess mobilization of microvas-
physiological changes observed after microembolization (43). cular functional reserve may be required to compensate for the
Specifically, we studied the consequences of embolization of greater intercapillary distance between the recruited effective
10 m and 100 m intramyocardial arteries and arterioles on capillary exchange surface area and the tissue previously per-
microvascular blood volume, perfusion, and transit time in fused by the embolized microvessel. When microspheres are
vivo. We found appreciable differences in microvascular blood injected during adenosine infusion, the already mobilized
volume and perfusion dynamics depending on the size of microvascular functional reserve cannot counteract the loss of
embolized microvessels and the initial perfusion state (Fig. 8). microvessels so that blood volume and perfusion decrease even
At rest, injection of small amounts of 10 m microspheres ini- after injection of small amounts of small microspheres. When
tially result in a very brief decrease in coronary blood flow, large diameter microspheres embolize, then a large contiguous
which is soon followed by increases in blood volume and per- volume of myocardium is deprived of flow. As the functional
fusion. This hyperemic response usually lasts for several min- capillary reserve can now be mobilized only at the surface of
utes (92). As previously observed in a canine model, the the embolized perfusion territory, this limited increase in cap-
hyperemic response progressively diminishes with increasing illary blood volume and perfusion cannot overcome the loss of
doses of small microspheres (Fig. 8) (98,102). In canines and capillaries at the center of the embolized perfusion territory
in pigs, the local hyperemic response can be attributed to (43,100).
adenosine release, because it was prevented by the adenosine When microvascular blood volume progressively decreased
receptor antagonist theophylline (98,99) and the 1-adreno- from maximum vasodilation to resting state by infusing
ceptor antagonist prazosin (103). Unlike the response to 10 m decreasing adenosine concentrations, transit time significantly
microspheres, resting blood volume and perfusion was not increased in nonembolized myocardium. In contrast, when
preserved with injection of 100 m microspheres. When the intramyocardial blood volume decreased in response to
microvasculature was maximally dilated before injection of microvascular embolization, transit time remained unchanged
microspheres, blood volume and perfusion also decreased (Fig. 9). Because transit time is inversely related to perfusion,
immediately and progressively with repetitive injections of this leads to perfusion values higher than expected for the given
microspheres, irrespective of microsphere size, both in porcine volume of perfused microvessels and hence to pseudo-preser-
and canine myocardium (43,98). vation of microvascular perfusion. This may contribute to a
Our EBCT-based findings support the concept of mobiliza- perfusion-contraction mismatch (97) by reducing the time
tion of a functional arteriolar and capillary blood volume and available for nutrient exchange at the capillary level. Our find-
perfusion reserve in response to microspheres injection. The ings are consistent with the previously reported microspheres-
increase in blood volume and perfusion after injection of low induced decrease in arterio-venous O2 difference and lactate
amounts of 10 m-microspheres at resting conditions may be extraction (98), and may, in part, explain microembolization-
 CHAPTER 19 / IMAGING INTRAMYOCARDIAL MICROCIRCULATORY FUNCTION 201

Fig. 8. Myocardial perfusion in response to selective injections of 10-m microspheres into the left anterior descending coronary artery.
Injection of small amounts of small microspheres at rest leads to an initial increase in myocardial perfusion, which then decreases with
further injections (left panel). When microspheres are injected at maximum vasodilation, the microvasculature cannot compensate for the
loss of embolized microvessels, and perfusion decreases even after small amounts of microspheres are injected (right panel). Injection of
larger microspheres leads to immediate decrease in perfusion and blood volume also at resting conditions (not shown) (see text for details;
modified from ref. 43).

related regional contractile dysfunction despite (pseudo-) nor- Under resting conditions, and using 4 to 16 contiguous
mal regional flow. The degree to which myocardial perfusion myocardial ROIs, the range of perfusion values has been shown
is preserved is described by the BV-to-F relationship (Fig. to be between 2.0 and 5 mL/g/min, and the blood volume has
10A,B) (104). a corresponding range in values (114). Importantly, the BV-to-
Our study demonstrated the ability of fast CT to image acute F relationship generated from the BV and F values obtained
changes in microvascular function, which in this experiment from those multiple ROIs follows the same BV = aF + bF
was induced by selective injection of microspheres. The tech- relationship compared to the one that can be generated with a
nique provided new and unique insight into microvascular func- single large ROI sampled in scans repeated under resting and
tional behavior in vivo. The alterations in response to different vasodilated states (Fig. 11).
sized microspheres of blood volume, perfusion, and transit Consequently, the BV-to-F relationship that characterizes
time, and the way in which these parameters are interrelated, the myocardial microcirculatory function can be generated from
were highly specific for coronary microembolization. a single scan sequence, which reduces both radiation and con-
trast material exposure.
QUANTITATION AND ASSESSMENT
OF MYOCARDIAL MICROVASCULAR CONCLUSION
FUNCTIONAL HETEROGENEITY Using fast X-ray CT and indicator dilution principles, indi-
Under normal conditions, intramyocardial blood volume ces of intramyocardial fractional blood volume, myocardial
and perfusion have been shown to be heterogeneously distrib- perfusion, and microvascular transit time can be quantitated
uted in adjacent regions of interest (= spatial heterogeneity) with minimal invasiveness in vivo in absolute numbers. These
and over time (= temporal heterogeneity) (105107). It was data provide important information on coronary microvascular
recently suggested that increased heterogeneity may be present function in healthy and diseased myocardium, and contribute
in early atherosclerosis (108). Assuming a bifurcation at every to a better understanding of the role of the coronary microvascu-
coronary artery branch (109), myocardial heterogeneity can lature in different physiological and pathophysiological settings.
be well represented by a fractal process (19,110) and can be Our data provide evidence that the BV-to-F relationship is
measured by relating the relative dispersion (= standard devia- sensitive enough to detect characteristic differences in the
tion divided by the mean = the coefficient of variation) to the behavior of healthy compared to diseased coronary microvas-
size of the region of interest (ROI) (111,112). It has been shown cular function, and specific enough to identify distinct differ-
that the logarithm of the coefficient of variation between con- ences in microvascular function between different diseases. In
tiguous ROIs within the myocardium varies linearly with the healthy myocardium, blood volume relative to flow always
logarithm of the size of those ROIs: log [SD/mean] = (1-D) log follows a single curvilinear line. A reduced response to adenos-
[area ROI], with 1-D being the slope and D being the fractal ine results in reduced increases in blood volume and perfusion,
dimension (111,113). The value of D is expected to increase but shows no deviation of values from the relationship seen in
with increasing heterogeneity both in normal and diseased a presumably healthy microvasculature. In contrast, epicardial
myocardium. and microvascular alterations lead to characteristic deviations
202 MHLENKAMP ET AL.

Fig. 9. Blood volume (BV)-to-transit time relationship in normal,

nonembolized myocardium and in myocardium embolized with
10-m and 100-m microspheres. In normal, nonembolized myo-
cardium, transit time increases when maximal functionally avail-
able blood volume decreases. However, when functionally available
volume of myocardium is progressively reduced by embolization
with increasing amounts of 10-m and 100-m microspheres, tran-
sit times remain unaltered, leading to pseudo-preservation of myo-
cardial perfusion (from ref. 43).

from this normal relationship. In mild epicardial coronary

stenosis, coronary microvessels are recruited to increase the
transstenotic pressure gradient and hence to preserve flow,
which leads to an increase in blood volume for any given
amount of microvascular blood flow.
Our approach to the evaluation of microvascular function is
not limited to fast X-ray CT. Other imaging techniques, espe-
cially cardiac ultrasound and magnetic resonance imaging , are
under extensive investigation and continuous refinement in
Fig. 10. Blood volume-to-flow (BV-to-F) relationship normal con-
their ability to obtain absolute values of microvascular blood trols compared to repetitive selective injection of microspheres into
volume and perfusion (115118). Once blood volume and per- the left anterior descending coronary artery (LAD). Microembol-
fusion are both accurately and concurrently quantifiable, then izationi.e., plugging of intramyocardial microvesselsleads to a
the subsequent use of the BV-to-F relationship easily follows, reduced number of patent microvessels and hence to a decrease in
fractional blood volume. Despite the decrease in fractional blood
irrespective of the imaging modality.
volume, perfusion values are preserved within limits, which are
Evaluating myocardial microvascular function should be described by the BV-to-F relationship (see also Fig. 9 and text for
useful in other clinically relevant settings. In nondiseased details; from ref. 104).
myocardium, the approach may be used to quantify the long-
term effect of physiological stimuli such as aging and exercise
on normal coronary microvascular functional reserve (64,119).
It is likely that other diseases that affect intramyocardial
microvessels (e.g., diabetes mellitus, arterial hypertension, or
REFERENCES
cardiomyopathies) (20) also show characteristic patterns in the 1. Grundy SM. Primary prevention of coronary heart disease. Inte-
grating risk assessment with intervention. Circulation 1999;100:
BV-to-F relationship. Fast CT may further be used to monitor 988998.
the efficacy of therapeutic strategies that target the coronary 2. Smith SC, Greenland P, Grundy SM. Prevention Conference V
microcirculation such as medication (120,121), or poten- Beyond secondary prevention: Identifying the high-risk patient for
tially transmyocardial laser revascularization (122) and gene primary preventionexecutive summary. Circulation 2000;101:
therapy to induce myocardial angiogenesis in ischemic myo- 111116.
3. Greenland P, Smith SC, Grundy SM. Improving coronary heart
cardium (123). Preliminary reports on the assessment of coro- disease risk assessment in asymptomatic people. Role of traditional
nary microvascular function in patients (38,124) suggest a risk factors and non-invasive cardiovascular tests. Circulation 2001;
potential of fast-CT technology to evaluate early microvascu- 104:18631867.
lar disease also in clinical practice, and may become a useful 4. Marcus, ML, Chilian WM, Kanatsuka H, et al. Understanding the
coronary circulation through studies at the microvascular level.
adjunct to noninvasive assessment of epicardial coronary Circulation 1990;82:17.
calcium quantitation (125) and noninvasive coronary angio- 5. Naseri A, Crea KF, Crainflone D. Myocardial ischemia caused by
graphy (126). distal coronary vasoconstriction. Am J Cardiol 1992;7:16021605.
 CHAPTER 19 / IMAGING INTRAMYOCARDIAL MICROCIRCULATORY FUNCTION 203

14. Nitenberg A, Valensi P, Sachs R, et al. Impairment of coronary

vascular reserve and ACh-induced coronary vasodilation in dia-
betic patients, angiographically normal arteries and normal left ven-
tricular systolic function. Diabetes 1993;42:10171025.
15. Bartel T, Yang Y, Mller S, Wenzel RR, Baumgart D, Philipp T,
Erbel R. Non-invasive assessment of microvascular function in
arterial hypertension by transthroacic Doppler harmonic
echocardiography. J Am Coll Cardiol 2002;39(12):20122018.
16. Baumgart D, Haude M, Liu F, et al. Current concepts of coronary
flow reserve for clinical decision making during cardiac catheter-
ization. Am Heart J 1998;136:136149.
17. Ge J, Qian J, Bhate R, et al. Prevalence of microvascular dysfunc-
tion in patients with severe coronary artery disease. Eur Heart J
1999;20(Suppl):476482.
18. Kuo L, Davies MJ, Cannon MS, Chilian WM. Pathophysiological
consequences of atherosclerosis extend to the microcirculation. Circ
Res 1992;70:465476.
19. Chilian WM, Eastham CL, Layne SM, Marcus ML. Small vessel
phenomena in the coronary microcirculation: phasic intramyocar-
dial perfusion and coronary microvascular dynamics. Prog Car-
diovasc Dis 1988;31:1738.
20. Chilian WM. Coronary microcirculation in health and disease. Sum-
mary of an NHLBI workshop. Circulation 1997;95:522528.
21. Kuo L, Davies MJ, Chilian WM. Longitudinal gradients for endo-
thelium-dependent and -independent vascular responses in the
coronary microcirculation. Circulation 1995;92:518525.
22. Kanatsuka H, Lamping KG, Eastham CL, Marcus ML. Heterog-
enous changes in epimyocardial microvascular size during graded
coronary stenosis. Circ Res 1990;60:389396.
23. Axel L. Cerebral blood flow determination by rapid-sequence com-
puted tomography. A theoretical analysis. Radiology 1980;137:
679686.
24. Wang T, Wu S, Chung N, Ritman EL. Myocardial blood flow esti-
Fig. 11. Multiple-ROI analysis of blood volume and perfusion het- mates by synchronous, multislice high-speed CT. IEEE Transact
erogeneity: to measure blood volume (BV) and flow (F) heterogene- Med Imag 1989;8(1):7077.
ity, one large myocardial region of interest (ROI) is consecutively 25. Rumberger JA, Bell MR. Measurement of myocardial perfusion
halved (top). BV and F are determined within each ROI. Because of using electron beam (ultrafast) computed tomography. In: Marcus
the fractal nature of BV and F, variation of measurements (or relative ML (ed), Cardiac Imaging. A Companion to Braunwalds Heart
dispersion) increases around the mean value within progressively Disease. WB Saunders, Philadelphia: 1996;835852.
smaller ROIs (middle left). These values have been shown to follow 26. Stewart GN. Researches on the circulation time and on the influ-
the BV-to-F relationship, which can therefore be determined indi- ences that affect it. IV. The output of the heart. J Physiol 1897;22:
vidually based on just one image sequence (middle right and bottom, 159183.
AC) (see text and refs. 112 and 114 for details). 27. Hamilton WF, Moore JW, Kinsman JM, Spurling RG. Studies on
circulation IV. Further analysis of the injection method and of
changes in hemodynamics under physiological and pathophysi-
ological conditions. Am J Physiol 1932;99:534551.
6. Hasdai D, Sangiorgi G, Spagnoli LG, et al. Coronary artery apopto- 28. Meier P, Zierler KL. On the theory of the indicator-dilution method
sis in experimental hypercholesterolemia. Atherosclerosis 1999; for measurement of blood flow and volume. J Appl Physiol 1954;
142:317325. 6(12):731744.
7. Kemp HG, Kronmal RA, Vliestra RE, Frey RL. Seven year survival 29. Zierler KL. A simplified explanation of the theory of indicator
of patients with normal or near normal coronary arteriograms: a dilution for measurement of fluid flow and volume and other dis-
CASS registry study. J Am Coll Cardiol 1986;7:479483. tributive phenomena. Bull Johns Hopkins Hosp 1958;103:199217.
8. Hasdai D, Gibbons RJ, Holmes DR Jr., Higgano ST, Lerman A. 30. Zierler KL. Theoretical basis of indicator-dilution methods for
Coronary endothelial dysfunction in humans is associated with measuring flow and volume. Circ Res 1962;10:393407.
myocardial perfusion defects. Circulation 1997;96:33903395. 31. Zierler KL. Equations for measuring blood flow by external moni-
9. Kemp HG. Left ventricular function in patients with angina and toring of radioisotopes. Circ Res 1965:16(4):309321.
normal coronary angiograms. Am J Cardiol 1973;32:375376. 32. Ritman EL, Kinsey JH, Robb RA, Harris LD, Gilbert BK. Physics
10. Cannon R, Epstein S. Microvascular angina as a cause of chest and technical considerations in the design of the DSR: a high tem-
pain with angiographically normal coronary arteries. Am J Cardiol poral resolution volume scanner. AJR 1980;135:369374.
1988;43:13381343. 33. Boyd DP, Lipton MJ. Cardiac computed tomography. Proc IEEE
11. Schachinger V, Britten MB, Elsner M, et al. A positive family his- 1983;71(3):298307.
tory of premature coronary artery disease is associated with im- 34. Wolfkiel CJ, Ferguson JL, Chomka EV, et al. Measurement of
paired endothelium-dependent coronary blood flow regulation. myocardial blood flow by ultrafast CT. Circulation 1987;76(6):
Circulation 1999;100:15021508. 12621273.
12. Nahser PJ, Brown RE, Oskarsson H, et al. Maximal coronary flow 35. Liu YH, Bahn RC, Ritman EL. Dynamic intramyocardial blood
reserve and metabolic coronary vasodilation in patients with diabe- volume: evaluation with a radiological opaque marker method. Am
tes mellitus. Circulation 1995;91:635640. J Physiol 1992;263:H963H967.
13. Kaufmann PA, Gnecci-Ruscone T, DiTerlizzi M, et al. Coronary 36. Liu YH, Bahn RC, Ritman EL. Microvascular blood volume to flow
heart disease in smokersvitamin C restores coronary microcircu- relationship in porcine heart wall: Whole body CT evaluation in
latory function. Circulation 2000;102:12331238. vivo. Am J Physiol 1995;269(38):H18201826.
204 MHLENKAMP ET AL.

37. Weiss RM, Otoadese EA, Noel MP, DeJong SC, Heery SD. 59. Rodriguez-Porcel M, Lerman A, Best PJM, et al. Hypercholester-
Quantitation of absolute regional myocardial perfusion using cine olemia impairs myocardial perfusion and permeability: role of oxi-
CT. J Am Coll Cardiol 1994;23:11861193. dative stress and endogenous scavenging activity. J Am Coll Cardiol
38. Bell MR, Lerman LO, Rumberger JA. Validation of minimally 2001;37:608615.
invasive measurement of myocardial perfusion using electron beam 60. Bonetti PO, Wilson SH, Rodriguez-Porcel M, et al. Simvastatin
computed tomography and application in human volunteers. Heart preserves myocardial perfusion and coronary microvascular per-
1999;81:628635. meability in experimental hypercholesterolemia independent of
39. Lerman LO, Siripornpitak S, Maffei NL, et al. Measurement of in lipid lowering. J Am Coll Cardiol 2002;40:546554.
vivo myocardial microcirculatory function with EBCT. J Comput 61. Senneff MJ, Geltman EM, Bergmann SR, Hartman J. Noninvasive
Assist Tomogr 1999;23(3):390398. delineation of the effects of moderate aging on myocardial perfu-
40. Mhlenkamp S, Lerman LO, Lerman A, et al. Minimally invasive sion. J Nucl Med 1991;32:20372042.
evaluation of coronary microvascular function by EBCT. Circula- 62. Czernin J, Muller P, Chan S, et al. Coronary and peripheral blood
tion 2000; 102(19):24112416. flow: influence of age and hemodynamics on myocardial blood
41. Mhlenkamp S, Behrenbeck TR, Lerman LO, Lerman A, Sheedy flow and flow reserve. Circulation 1993;88(1):6269.
PF II, Ritman EL. Quantitation of long-term changes in coronary 63. Rakusan K, Cicutti N, Flanagan MF. Changes in the microvascular
microvascular functional reserve using EBT. Radiology 2001;221: network during cardiac growth, development and aging. Cell Mol
229236. Biol Res 1994;40:117122.
42. Mhlenkamp S, Lerman LO, Bajzer Z, Lund PE, Ritman EL. 64. White FC, Bloor CM, McKirnan D, Carroll SM. Exercise training
Quantitation of myocardial microcirculatory function with X-ray in swine promotes growth of arteriolar bed and capillary angiogen-
CT. Ann NY Acad Sci 2002;972:307316. esis in heart. J Appl Physiol 1998;85:11601168.
43. Mhlenkamp S, Bailey PE, Pfeifer EA, Behrenbeck TR, Sheedy PF 65. Hamaoka K, Onouchi Z, Ohmochi Y, Sakata K. Coronary arterial
II, Ritman EL. Intramyocardial blood volume, perfusion and transit flow-velocity dynamics in children with angiographically normal
time in response to embolization of different sized microvessels. coronary arteries. Circulation 1995;92:24572462.
Cardiovasc Res 2003;57(3):843852. 66. Buss DD, Hennemann III WW, Posner P. Maturation of coronary
44. Jaschke W, Gould RG, Cogan MG, Sievers RG, Lipton MJ. Cine- responsiveness to exogenous adenosine in the rabbit. Basic Res
CT measurement of cortical renal blood flow. J Comput Assist Cardiol 1987;82:290296.
Tomogr 1987;11:779784. 67. Schachinger V, Zeiher A. Alterations of coronary blood flow and
45. Lerman LO, Bell MR, Lahera V, et al. Quantification of global and myocardial perfusion in hypercholesterolemia. Heart 1996;76(4):
regional renal blood flow with EBCT. Am J Hypertens 1994;7: 295298.
829837. 68. Kelm M, Schrader J. Control of coronary vascular tone by nitric
46. Bentley MD, Lerman LO, Hoffman EA, et al. Measurement of renal oxide. Circ Res 1990;66:15611575.
perfusion and blood flow with fast computed tomography. Circ Res 69. Komaru T, Lamping K, Eastham CL, et al. Effect of an arginine
1994;74:945951. analogue on acetylcholine-induced coronary microvascular dilata-
47. Gobbel TG, Cann CE, Fike JR. Measurement of regional cerebral tion in dogs. Am J Physiol.1991;261:H2001H2007.
blood flow using ultrafast CT. Theoretical aspects. Stroke 70. Lerman A, Burnett JC Jr. Intact and altered endothelium in regula-
1991;22:768771. tion of vasomotion. Circulation 1992;86[Suppl III]:III12III19.
48. Jaschke W, Lipton MJ, Boyd D, et al. Dynamic CT scanning of the 71. Mathew V, Cannan CR, Miller VM, et al. Enhanced endothelin-
normal canine liver: interpretation of time density curves resulting mediated coronary vasoconstriction and attenuated basal nitric
from an intravenous bolus injection of contrast material. Eur J Radiol oxide activity in experimental hypercholesterolemia. Circulation
1985;5:25660. 1997;96(6):193036.
49. Clough AV, Al-Tinami A, Linehan JH, Dawson CA. Regional tran- 72. Hein TW, Kuo L. LDLs impair vasomotor function of the coronary
sit time estimation from image residue curves. Ann Biomed Eng microcirculation. Role of superoxide anions. Circ Res 1998;83:
1994; 22:128143. 404414.
50. Wu X, Latson LA, Wang T, et al. Regional pulmonary perfusion 73. Cooper A, Heagerty AM. Endothelial dysfunction in human
estimated by high-speed volume scanning CT. Am J Physiol Imag intramyocardial small arteries in atherosclerosis and hypercholes-
1988;3:7380. terolemia. Am J Physiol 1998;275:H14821488.
51. Olsson B, Aulie A, Svenn K, Andrew E. Human pharmakokinetics 74. Stepp DW, Nishikawa Y, Chilian WM. Regulation of shear stress
of iohexol, a new non-ionic contrast medium. Invest Radiol 1983; in the canine coronary microcirculation. Circulation 1999;100:
18:177182. 15551561.
52. Canty JM, Judd RM, Brody AS, Klocke FJ. First-pass entry of non- 75. Kanatsuka H, Lamping KG, Eastham CL, et al. Comparison of the
ionic contrast agent into the myocardial extravascular space. Effects effects of increased myocardial oxygen consumption and adenos-
on radiographic estimates of transit time and blood volume. Circu- ine on the coronary microvascular resistance. Circ Res 1989;65:
lation 1991;84:20712078. 12961305.
53. Ritman EL. Myocardial capillary permeability to iohexol. Invest 76. Grover GJ, Rosolowski M, Kedem JK, et al. Effect of hypoxia and
Radiol 1994;29:612617. adenosine on the microvascular reserve in the rabbit heart. Microcirc
54. Judd RM, Resar JR, Yin FCP. Rapid measurements of diastolic intra- Endothel Lymph 1986;3:359382.
myocardial vascular volume. Am J Physiol 1993;265:H1038H1047. 77. DellOmo G, Bandinelli S, Penno G, et al. Simvastatin, capillary per-
55. Levick JR. An Introduction to Cardiovascular Physiology, 2nd Ed. meability, and acetylcholine-mediated vasomotion in atherosclerotic,
Butterworth-Heinemann Ltd., Oxford: 1992. hypercholesterolemic men. Clin Pharmacol Ther 2000;68:427434.
56. Crone C, Levitt DG. Capillary permeability to small solutes. In: 78. Egashira K, Hirooka Y, Kai H, et al. Reduction in serum cholesterol
Renkin EM, Michel CC, Geiger SR (eds), Handbook of Physiology, with pravastatin improves endothelium-dependent coronary vaso-
Vol. IV: The Cardiovascular System, Part I, Microcirculation, Sec- motion in patients with hypercholesterolemia. Circulation
tion 2, The Cardiovascular System. American Physiological Soci- 1994;89:25192524.
ety, Bethesda, MD: 1984:411466. 79. Gould KL, Martucci JP, Goldberg DI, et al. Short-term cholesterol
57. Renkin EM. Capillary transport of macromolecules: pores and other lowering decreases size and severity of perfusion abnormalities by
endothelial pathways. J Appl Physiol 1985;58(2):315325. positron emission tomography after dipyridamol in patients with
58. Wu XS, Ewert DL, Liu YH, Ritman EL. In vivo relation of intra- coronary artery disease. A potential noninvasive marker of healing
myocardial blood volume to myocardial perfusion. Evidence sup- endothelium. Circulation 1994;89:15301538.
porting microvascular site for autoregulation. Circulation 1992;85: 80. Hasler-Rapacz J, Kempen HJ, Princen HMG, et al. Effects of
730737. simvastatin on plasma lipids and apolipoproteins in familial hyper-
 CHAPTER 19 / IMAGING INTRAMYOCARDIAL MICROCIRCULATORY FUNCTION 205

cholesterolemic swine. Arterioscler Thromb Vasc Biol 1996;16: 101. Herrmann J, Haude M, Lerman A, et al. Abnormal coronary flow
13743. velocity reserve after coronary intervention is associated with car-
81. Lefer AM, Scalia R, Lefer DJ. Vascular effects of HMG CoA- diac marker elevation. Circulation 2001;103:23392345.
reductase inhibitors (statins) unrelated to cholesterol lowering: new 102. Herzberg RM, Rubio R, Berne RM. Coronary occlusion and embo-
concepts for cardiovascular disease. Cardiovasc Res 2001;49: lization: effect on blood flow in adjacent arteries. Am J Physiol
281287. 1966;210:169175.
82. Wu CC, Chang SW, Chen MS, Lee YT. Early change of vascular 103. Hori M, Tamai J, Kitakaze M, et al. Adenosine-induced hyperemia
permeability in hypercholesterolemic rabbits. Arterioscler Thromb attenuates myocardial ischemia in coronary microembolization in
Vasc Biol 1995;15:529533. dogs. Am J Physiol 1989;257:H244H251.
83. Wilson SH, Simari RD, Best PJM, et al. Simvastatin preserves 104. Mhlenkamp S, Lerman LO, Behrenbeck TR, et al. EBCT-based
coronary endothelial function in hypercholesterolemia in the ab- evaluation of coronary microembolization: evidence for a faster
sence of lipid lowering. Arterioscler Thromb Vasc Biol 2001;21: transit of blood through non-embolized microvessels. Circulation
122128. 2000;102(18):II 112.
84. Gould KL, Lipscomb K, Calvert C. Physiologic basis for assessing 105. Balaban RS, Arai A. Function, metabolic and flow heterogeneity
critical coronary stenosis. Instantaneous flow response and regional of the heart. The view is getting better. Circ Res 2001;88: 265
distribution during coronary hyperemia as measures of coronary 267.
flow reserve. Am J Cardiol 1974;33:8794. 106. Bassingthwaighte JB, King RB, Roger SA. Fractal nature of
85. Pijls NHJ, de Bruyne B, Peels KH, et al. Measurement of fractional regional myocardial blood flow heterogeneity. Circ Res 1989;65:
flow reserve to assess functional severity of coronary-artery 578590.
stenoses. N Engl J Med 1996;334:17031708. 107. Deussen A. Blood flow heterogeneity in the heart. Basic Res Cardiol
86. Erbel R, Ge J, Grge G, et al. Neue bildgebende Verfahren zur 1998;93:430438.
Darstellung der Koronararterien. Z Kardiol 1998;87(Suppl 2): 108. Wieneke H, Schmermund A, Ge J, Altmann C, et al. Increased
6173. heterogeneity of coronary perfusion in patients with early coronary
87. Wu XS, Bahn RC, Ritman EL. Myocardial microcirculation as atherosclerosis. Am Heart J 2001;142(4):691697.
evaluated with CT. In: Sideman S, Beyar R (eds), Molecular and 109. Van Bavel E, Spaan JAE. Branching patterns in the porcine coro-
Subcellular Cardiology: Effects of Structure and Function. Plenum, nary arterial tree. Estimation of flow heterogeneity. Circ Res 1992;
New York: 1995;261268. 71:12001212.
88. Lindner JR, Skyba DM, Goodman NC, et al. Changes in myocardial 110. Bassingthwaighte JB, VanBeek JH, King RB. Fractal branchings:
blood volume with graded coronary stenosis: an experimental evalu- the basis of myocardial flow heterogeneities? Ann NY Acad Sci
ation using myocardial contrast echocardiography. Am J Physiol 1995;591:392401.
1997;272:H567H575. 111. King RB, Bassingthwaighte JB. Temporal fluctuations in regional
89. Berne RM. The role of adenosine in the regulation of coronary myocardial flows. Pflugers Arch 1989;413:336342.
blood flow. Circ Res 1980;47:808813. 112. Ritman EL. Temporospatial heterogeneity of myocardial perfusion
90. Topol EJ, Yadav JS. Recognition of the importance of emboliza- and blood volume in the porcine heart wall. Ann Biomed Eng 1998;
tion in atherosclerotic vascular disease. Circulation 2000;101: 26:519525.
570580. 113. King RB, Bassingthwaighte JB, Shales JR, Rowell LB: Stability of
91. Erbel R, Heusch G. Coronary microembolization. J Am Coll Cardiol heterogeneity of myocardial blood flow in normal awake baboons.
2000;36:2224. Circ Res 1985;57:285295.
92. Heusch G, Schulz R, Baumgart D, Haude M, Erbel R. Coronary 114. Shu N-H, Ritman EL. Heterogeneity of intramyocardial blood flow,
microembolization. Prog Cardiovasc Dis 2001;44:217230. blood volume and transit timefractal analysis of fast CT images.
93. Falk E. Unstable angina with fatal outcome: dynamic coronary The Physiologist 1989;32:188 (abstract).
thrombosis leading to infarction and/or sudden death. Autopsy 115. Passariello R, De Santis M. Magnetic resonance imaging evalua-
evidence of recurrent mural thrombosis with peripheral emboliza- tion of myocardial perfusion. Am J Cardiol 1998;81:68G73G .
tion culminating in total vascular occlusion. Circulation 1985; 116. Wei K, Jayaweera AR, Firoozan S, Linka A, Skyba DM, Kaul S.
71:699708. Quantification of myocardial blood flow with ultrasound-induced
94. Willerson JT, Cohen LS, Maseri A. Pathophysiology and clinical destruction of microbubbles administered as a constant venous
recognition. In: Willerson JT, Cohn JN (eds), Cardiovascular Medi- infusion. Circulation 1998;97(5):473483.
cine. Churchill Livingston, New York: 1995;333365. 117. Wilke N, Kroll K, Merkle H, et al. Regional myocardial blood vol-
95. Van Liebergen RAM, Piek JJ, Koch KT, de Winter RJ, Li KI. ume and flow: first-pass MR imaging with polylysine-Gd-DTPA. J
Immediate and long-term effect of balloon angioplasty or stent Magn Res Imag 1995;2:227237 .
implantation on the absolute and relative coronary blood flow 118. Wu CC, Feldman MD, Mills JD, et al. Myocardial contrast
velocity reserve. Circulation 1998;98:21332140. echocardiography can be used to quantify intramyocardial blood
96. Kern MJ, Puri S, Bach RG, et al. Abnormal coronary flow velocity volume: new insights into structural mechanisms of coronary auto-
reserve after coronary artery stenting in patients. Role of relative regulation. Circulation 1997;96:10041011.
coronary reserve to assess potential mechanisms. Circulation 119. Rakusan K, Flanagan MF, Geva T, Southern J, Van Praagh R. Mor-
1999;100:24912498. phometry of human coronary capillaries during normal growth and
97. Drge H, Neumann T, Behrends M, Schulz R, Erbel R, Heusch G. the effect of age in left ventricular pressure-overload hypertrophy.
Perfusion-contraction mismatch with coronary microvascular Circulation 1992;86:3846.
obstruction: role of inflammation. Am J Physiol 2000;279: 120. Lscher TF, Barton M. Endothelins and endothelin receptor antago-
H25872592. niststherapeutic considerations for a novel class of cardiovascu-
98. Hori M, Inoue M, Kitakaze M, et al. Role of adenosine in hyperemic lar drugs. Circulation 2000;102:24342440.
response of coronary blood flow in microcirculation. Am J Physiol 121. Schmermund A, Lerman LO, Rumberger JA, et al. Effects of acute
1986;250:H509H518. and chronic angiotensin receptor blockade on myocardial vascular
99. Grund F, Sommerschild HT, Lyberg T, Kirkeboen KA, Ilebekk A. blood volume and perfusion in a pig model of coronary micro-
Microembolization in pigs: effects on coronary blood flow and embolization. Am J Hypertens 2000;13:838845.
myocardial ischemic tolerance. Am J Physiol 1999;277: H533 122. Frazier OH, March RJ, Horvath KA. Transmyocardial revasculari-
H542. zation with a carbon dioxide laser in patients with end-stage coro-
100. Malyar NM, Lerman LO, Goessl M, Mhlenkamp S, Ritman EL. nary artery disease. New Engl J Med 1999;341:10211028.
Myocardial microvascular permeability after microembolization. 123. Losordo DW, Vale PR, Symes JF, et al. Gene therapy for myocar-
The FASEB J (Part I) 2003;17(4):136. dial angiogenesis: initial clinical results with direct myocardial
206 MHLENKAMP ET AL.

injection of phVEGF165 as sole therapy for myocardial ischemia. method should we choose? Electron-beam computed tomography.
Circulation 1998;98:28002804. Eur Heart J 2002;4:F47F53.
124. Mhlenkamp S. Myocardial blood flow measured by EBT. Eur 126. Mhlenkamp S, Schmermund A, Gerber T, et al. Nicht-invasive
Radiol 2002;12(11):H6. CT Koronarangiographie als Ergnzung zur Koronarkalk-
125. Erbel R, Budde T, Kerkhoff G, Mhlenkamp S, Schmermund A. quantifizierung bei symptomatischen Patienten. Herz 2003;28:
Understanding the pathophysiology of the arterial wall: which 106118.
 CHAPTER 20 / ASSESSING MYOCARDIAL VIABILITY WITH MDCT 207

20 Approaches for Assessing Myocardial

Viability With Multidetector-Row CT

YASUSHI KOYAMA, MD AND TERUHITO MOCHIZUKI, MD

INTRODUCTION The protocol for the contrast-enhanced CT was as follows:

Because the myocardial reperfusion status is pivotal to the One hundred mL of an iodine contrast medium was intrave-
prognosis of patients who have suffered an acute myocardial nously injected at a rate of 1.5 mL/s, with the early image being
infarction (AMI), the assessment of microvascular flow after taken 45 s after the start of the injection.
reperfusion therapy is of great importance. Left ventriculography (LVG) was performed immediately
Gibson et al. (1) have recently reported that patients with after the acute-phase intervention and again at the chronic phase
both normal epicardial flow (thrombolysis in myocardial (1 mo) to evaluate regional wall motion. The wall motion was
infarction [TIMIsee Note 1] grade 3 flow) and normal tissue- scored into six grades (5: normal; 4: mild hypokinesia; 3: mod-
level perfusion (TIMI myocardial perfusion [see Note 2] grade erate hypokinesia; 2: severe hypokinesia; 1: akinesia; 0: dyski-
3) are at an extremely low risk for mortality. nesia) and used as the regional wall motion score (RWM). The
Conventionally, myocardial perfusion has been evaluated corresponding LVG segment to the CT site is shown in Fig. 2.
mainly using nuclear imaging techniques (2,3). Contrast- Myocardial enhancement patterns were classified into
enhanced dynamic magnetic resonance imaging (MRI) (46) three groups: Group N (normal), showing no ED, was consid-
and myocardial contrast echocardiography (710) are newcom- ered as the normal group; Group SE (region retained in the
ers that can assess myocardial perfusion. Electron beam com- subendocardium), the region in which ED accounted for
puted tomography (EBCT) also provides reliable information less than 50%; and Group TM (region existing transmurally),
on myocardial perfusion (1113). the region in which ED accounted for more than 50%. Addi-
In this chapter, we describe the meaning of X-ray CT tionally, the mean myocardial wall thickness of the seven
enhancement and the classification of enhancement patterns as regions was calculated and compared in both acute and 1-mo
clinical predictors in patients with AMI after successful percu- phases.
taneous coronary interventions (PCI). ED VS WALL THICKNESS
A case of Group SE is shown in Fig. 3A, and a case in the TM
ENHANCEMENT IN ACUTE MYOCARDIAL
group is shown in Fig. 3B.
INFARCTION (EARLY DEFECT, LATE
As shown in Fig. 4A, wall thickness in Group N showed no
ENHANCEMENT, RESIDUAL DEFECT)
significant difference between the acute and chronic phases.
Early defect (ED) is observed as a myocardial perfusion In Group SE, the wall thickness decreased slightly in the
defect (dark zone) in the early image (3060 s). Residual defect chronic phase (p < 0.05). In Group TM, the wall thickness
(RD) is observed as smaller dark regions observed in the significantly decreased in the chronic phase (p < 0.001), whose
subendocardium, surrounded by a partially hyperenhanced zone rate of decrease was larger than that of the SE group. As the
of late enhancement (LE) in the late image (510 min) (Fig. 1). depth of the ED increased, wall thickness in the chronic phase
The density of the ED (30.2 11 Hounsfield units [HU]) was decreased.
significantly lower than that of normal myocardium (102.1 ED VS WALL MOTION (RWM SCORE)
9.0 HU, p < 0.0001). The LE (112.9 18.5 HU) presented with The RWM in Group N exhibited no difference between the
higher density than RD (59.3 11 HU) (p < 0.0001). acute and the chronic phases. In Group SE, the RWM improved
ED IN RELATION TO WALL THICKNESS AND in the chronic phase, and in Group TM, the RWM did not
WALL MOTION IN CHRONIC PHASE (1 MO) (14) improve (Fig. 4B). The RWM in Group SE improved, while in
In AMI patients after successful PCI, the depth (subendocar- Group TM it remained worse.
dial or transmural) of the ED can predict wall thickness and Thus ED by contrast-enhanced CT is useful as a predictor of
wall motion in the chronic phase (1 mo). wall thickness and regional wall motion at 1 mo after successful
reperfusion therapy in AMI.
From: Contemporary Cardiology: CT of the Heart:
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

207
208 KOYAMA AND MOCHIZUKI

Fig. 1. Early defect in the early phase (left), and residual defect and delayed enhancement in the delayed phase (right).

Fig. 2. Corresponding segments of CT (A) and left ventriculography (B).

ED VS TL-201 MYOCARDIAL PERFUSION SINGLE mural AMI after primary PCI. Figure 6 shows an antero-septal
PHOTON EMISSION CT (15,16) transmural AMI after primary PCI. Location and extent of the
perfusion defects were well concordant.
When the early defect from the contrast-enhanced CT was
compared with the defects on Tl-201 myocardial perfusion TWO-PHASE CONTRAST-ENHANCED CT (17)
single photon emission CT (SPECT), the location and extent METHOD OF TWO-PHASE CONTRAST-ENHANCED CT
were well concordant, indicating that myocardial perfusion can Information regarding myocardial perfusion in the late phase
be evaluated with CT. Figure 5 shows an infero-lateral trans- (510 min), along with information in the early phase, seems to
209
Fig. 3. (A) A case of Group-SE. The early defect (ED) was observed in the endocardium of the inferior wall by the contrast-enhanced helical CT in the acute phase. At the site, there was
no decrease in the wall thickness at 1 mo. (B) A case of Group-TM. By the contrast-enhanced CT in the acute phase, ED was observed in the whole layer of the antero-septal wall. At the
CHAPTER 20 / ASSESSING MYOCARDIAL VIABILITY WITH MDCT

site, the wall thickness decreased at 1 mo. As shown by the arrow, there was no improvement of wall motion in the left ventriculography.
209
210 KOYAMA AND MOCHIZUKI

Fig. 4. (A) Wall motion (regional wall motion score, RWM) in the chronic phase (1 mo) in Group-N, Group-SE, and Group-TM). RWM in
group-SE improved, while in Group-TM it became worse. (B) Wall thickness in the chronic phase (1 mo) in Group-N, Group-SE, and Group-
TM). As the depth of early defect increased, wall thickness decreased in the chronic phase.

Fig. 5. A case of infero-lateral transmural acute myocardial infarction after successful primary percutaneous coronary intervention. Location
and extent of the perfusion defects were well concordant. The hypoenhanced (dark) region between arrows indicates the perfusion defect by
contrast-enhanced CT. Note that the perfusion defect by CT is similar in extent and location to the perfusion defect on Tl-201 myocardial
perfusion single photon emission CT. Tx = transaxial image; Sx = short axis; and Be = Bulls eye map.
 CHAPTER 20 / ASSESSING MYOCARDIAL VIABILITY WITH MDCT 211

Fig. 6. A case of antero-septal transmural acute myocardial infarction after successful primary percutaneous coronary intervention. Location
and extent of the perfusion defects were well concordant. The hypoenhanced (dark) region between arrows indicates the perfusion defect by
contrast-enhanced CT. Note that the perfusion defect by CT is similar in extent and location to the perfusion defect on Tl-201 myocardial
perfusion single photon emission CT. Tx, transaxial image; Sx, short axis; and Be, Bulls eye map.

Fig. 7. Three enhancement patterns. Illustration (A) and typical cases (B). Group 1: the absence of early defect (ED) in the early phase, and
the presence of late enhancement (LE) without residual defect (RD) in the late phase. Group 2: the presence of ED in the early phase, and the
presence of LE without RD in the late phase. Group 3: the presence of ED in the early phase, and the presence of both LE and RD.

yield a more detailed prediction of functional recovery in the infused at a rate of 0.1 mL/s for the second scan to acquire the
chronic phase. late image, which was taken 7 min after the start of the admin-
The protocol for the contrast-enhanced CT was as follows. istration. A total of 150 mL of the contrast medium was used.
A nonion iodine contrast medium (300 mg iodine/mL) was CLASSIFICATION OF ENHANCEMENT PATTERNS
intravenously administrated at a rate of 1.5 mL/s for the first Enhancement patterns can be classified into three groups
(early) scan, with the early image being taken 45 s after the start (Fig. 7): Group 1, the absence of ED in the early phase, and the
of the administration. Then the same contrast medium was presence of LE without RD in the late phase; Group 2, the
212 KOYAMA AND MOCHIZUKI

Fig. 8. Left-ventricular volume (end-diastolic volume [EDV], end-systolic volume [ESV] and functional parameters (left-ventricular ejection
fraction) in Group 1 to 3 are summarized. In Group 1, the ejection fraction (EF) significantly improved from 61 14% to 78 10% (p < 0.05).
In Group 2, the EF was not significantly different. In Group 3, the EDV significantly increased from 117 38 mL to 147 41 mL and the EF
significantly decreased from 63 13% to 51 15%.

presence of ED in the early phase, and the presence of LE between Group 1 and Group 3. Thus, two-phase contrast-
without RD in the late phase; and Group 3, the presence of ED enhanced CT predicts left-ventricular functional recovery in
in the early phase, and the presence of both LE and RD in the patients with AMI after successful reperfusion therapy.
late phase. Figure 7B shows typical cases of the enhancement ENHANCEMENT PATTERNS ON TWO-PHASE
patterns. CONTRAST-ENHANCED CT VS 201TL/99MTC-PYROPHOSPHATE
ENHANCEMENT PATTERNS ON TWO-PHASE DUAL-ISOTOPE SPECT
CONTRAST-ENHANCED CT VS LEFT-VENTRICULAR Patients with RD > LE zones showed a perfusion defect of
FUNCTION AND VOLUME IN THE CHRONIC PHASE (17) 201Tl and strong uptake of 99mTc-pyrophosphate (see Note 3).
Thirty AMI patients, after successful PCI (4.3 1.2 h after Patients with RD < LE zones showed an overlap of 201Tl and
onset), underwent a two-phase contrast-enhanced CT (37 4 h 99mTc- pyrophosphate.

after direct angioplasty). Conventional LVG was performed A few patients with subendocardial-early defects, having only
immediately after the PCI as the acute study. They had both wide LE without RD, showed normal perfusion without 99mTc-
coronary angiography (CAG) and LVG as the chronic study pyrophsphate accumulations. Thus, a large RD compared with
(28 3 d). LE indicates the least viability. In other words, RD in the late
COMPARISON OF LEFT-VENTRICULAR VOLUME image suggests an abundance of necrotic tissue injured by AMI.
AND FUNCTION IN EACH GROUP BY CONVENTIONAL LVG Fig. 9 shows a comparison of two-phase contrast-enhanced CT
The left-ventricular volume and functional parameters (end- with 201Tl/99mTc-pyrophosphate dual isotope SPECT.
diastolic volume [EDV], end-systolic volume [ESV], and ejec-
tion fraction [EF]) in each group are summarized in Fig. 8. THE MEANING OF EACH ENHANCEMENT PATTERN
In Group 1, the left-ventricular ejection fraction (LV-EF) SIGNIFICANCE OF AN ED
improved. In Group 2, the LVEF was not significantly differ- The contrast medium is thought to reach the microvascular
ent. In Group 3, the EDV increased and the LV-EF decreased. bed in the early phase after an intravenous administration. An
The lack of evidence of the ED, as seen in Group 1, indicates EBCT study (13) reported that myocardial enhancement in the
a functional improvement in the chronic phase. Residual defect, early phase reflected the volume of the vascular bed.
as seen in Group 3, indicates the worst functional outcome, i.e., Recently, reduced signal intensity on first-pass MRI con-
deterioration in the chronic phase. Diminishment of the ED, as trast images has been shown to indicate reduced blood flow
seen in Group 2, exhibited an intermediate functional recovery (18). Therefore, the ED observed on CT would also reflect a
 CHAPTER 20 / ASSESSING MYOCARDIAL VIABILITY WITH MDCT 213

Fig. 9. Comparison of two-phase contrast-enhanced CT with 201Tl/99mTc-pyrophosphate dual isotope single photon emission CT. Case 1 was
a patient with acute myocardial infarction (AMI) after left ascending coronary artery reperfusion (white arrows), such as Group 1 (an almost
normal enhancement pattern of CT), who had a normal uptake of 201Tl and an absence of 99mTc-pyrophosphate (99mTc-PYP) accumulations.
Case 3 was a patient with AMI after left ascending coronary artery reperfusion, and Case 4 was a patient with AMI after left ascending coronary
artery reperfusion, such as Group 3 (early defect [ED] (+), late enhancement [LE] (+), residual defect [RD] (+)). The ED is similar to the uptake
of 201Tl. However, subendo-ED of Case 2, who was a patient with AMI after circumflex coronary artery reperfusion, such as Group 2 (ED [+],
LE [+], RD [-]), had a mismatch of uptake of 201Tl and ED. Especially the infracted area of LE in Case 2 had an absence of 99mTc-PYP, on the
other hand, the infarct area of LE with small RD of Case 3 had a small accumulation of 99mTc-PYP. That is, LE was not the necrotic area of
AMI. RDs in Case 3 and Case 4 appear to be the necrotic areas, because these RD areas correlated well with the accumulation of 99mTc-PYP.
These findings are important in evaluating the myocardial viability of using CT.

decrease in the volume of the vascular bed, i.e., a decrease of some extentthat is, tissue showed mild to severe micro-
the myocardial blood flow. vascular damage and myocardial necrosis, owing to the wall
Using an experimental infarct-reperfusion model in dogs, thickness in Group 2 and Group 3, showing that the ED was
Braunwald et al. (19) classified the condition of myocardial significantly reduced.
tissue into four layers from the endocardial side. The first layer SIGNIFICANCE OF RD AND LE
corresponded to a viable and very thin myocardium, receiving After the contrast medium reaches the microvascular bed,
oxygen directly from the left ventricle; the second layer to it gradually flows into the interstitium (extracellular space),
myocardial necrosis with extensive capillary (microcirculation) remains for some time, and then washed out slowly. Therefore,
disorder; the third layer to myocardial necrosis, in which blood myocardial enhancement in the late phase mainly reflects the
supply was preserved to some extent; and the fourth layer to characteristics of the interstitiumthat is, the volume of the
stunned myocardium that had escaped from necrosis. interstitial space (12,13).
The ED in our study may correspond to myocardial necro- When the RD was detected, as in Group 3, functional recov-
sis with extensive capillary (microcirculation) disorder, or to ery was not observed. However, when ED turned into LE, as in
myocardial necrosis in which blood supply was preserved to Group 2, deterioration of left-ventricular function was mini-
214 KOYAMA AND MOCHIZUKI

mal, or less than that observed in Group 3. We speculate that the sisted after successful reperfusion. Additionally, in patients
area of the RD might correspond to myocardial necrosis with showing RD in our study, this no-reflow phenomenon exists,
extensive capillary disorderi.e., microvascular no-reflow regardless of its various degrees and causes, in the coronary
and that LE might correspond to the layer where blood supply microcirculation system after successful reperfusion in coro-
was preserved to some degree, indicating the possibility of nary arteriography. We believe that the no-reflow at the level
residual myocardial viability, although microvascular flow was of regional microcirculation is involved in decreases in cardiac
disturbed by edema within 48 h after PCI. function and wall motion in the chronic phase.
Considering the findings of SPECT studies and our results, We conclude that contrast-enhanced CT is useful in evalu-
we conclude that the RD indicated a necrotic area as a result of ating myocardial enhancement, which may serve as a predictor
severe microvascular obstruction microvascular no-reflow of changes in wall motion and thickness, left-ventricular func-
by red blood cells and necrotic debris (20), as seen in the wave tion, and myocardial viability after PCI in patients with AMI.
front phenomenon of ischemic necrosis (21,22). Since the PITFALLS OF ENHANCEMENT PATTERNS
presence of 99mTc-pyrophosphate in Group 3 was well corre-
lated with RD, the myocardial enhancement pattern in Group 3 In the assessment of perfusion, there are some pitfalls.
indicates less antegrade microvascular flow beyond the point In patients with old myocardial infarction, or re-AMI, we
of microvascular obstruction than seen in Group-2. As a result may also detect lipid degeneration in the left-ventricular myo-
of the incomplete perfusion in the microvascular level, the cardium (Fig. 10A). This is always clearly detectable in plain
necrotic area of Group 3 increased among three groups. images without the administration of a contrast medium. This
Basically, the pharmacokinetic properties of X-ray con- should not be interpreted as ED or RD on two-phase contrast-
trast agents are similar to those of the well-known gadolinium enhanced CT. In patients with old myocardial infarction, or
complexes (2325) for MRI. Judd et al. reported that the re-AMI, we may detect a thrombus in the left ventriculum
hyperintense regions observed on delayed MRI images were (Fig. 10B), which is always clearly detectable in both early and
generally smaller than the risk region and larger than regions late images. This should not be misinterpreted as ED or RD. In
of necrosis as defined by triphenyltetrazolium chloride stain- patients with old myocardial infarction or re-AMI, we may
ing (26). This would indicate that at least part of the delayed even detect calcium in the left-ventricular myocardium (Fig.
hyperintense region was viable myocardium. Yokota et al. 10C). This is clearly detectable in plain CT, and should not be
investigated Gd-enhanced MRI images 5 to 10 min after an misinterpreted as LE on two-phase contrast-enhanced CT.
intravenous gadopentate dimeglumine (Gd-DTPA) injection STUDY LIMITATIONS AND FUTURE OF MDCT
in patients with nonreperfused myocardial infarction (27). ADVANTAGES OF CARDIAC CT FOR PATIENTS WITH AMI
They qualitatively analyzed Gd-enhanced MRI findings in The advantages of cardiac CT are generally as follows: (1) no
relation to peak creatine phosphokinase levels, wall motion, blind area; (2) shorter acquisition time (less than 30 s); (3) metal
and coronary angiography, and concluded that subendo- devices such as an infusion pump, pacemaker, and intra-aortic
cardial or transmural hyper-enhancement could reflect the balloon pump (IABP) are acceptable, whereas they are com-
existence of viable myocardium, while subendocardial hypo- monly contraindicated for MRI, and therefore the CT is able to
enhancement was associated with necrotic myocardium. utilize them for the acute phase of AMI; (4) coronary arteries
In general, there is a consensus that delayed hyper-enhance- can be evaluated with the same data.
ment on MRI reflects nonviable myocardium (28). But in our DISADVANTAGES OF CT FOR PATIENTS WITH AMI
study, the LE presented both in Group 1 and in Group 2, indi- The disadvantages of CT over other noninvasive modalities
cating that the LE included viable myocardium, at least in con- are the use of an iodine contrast medium (however, there were
trast-enhanced CT studies within 48 h after reperfusion therapy, neither major nor minor complications in this study) and X-ray
as seen in our results in this chapter. exposure of the two-phase contrast CT. Although the two-phase
The myocardium and the microvasculature changes dynami- study requires double dose compared to the single-phase study,
cally during the healing stage in the acute phase after successful overlapping reconstruction does not increase the radiation
reperfusion; therefore, timing of the CT study after reperfusion dosei.e., it is exactly the same as nonoverlapping reconstruc-
therapy is important in order to assess the myocardial enhance- tion. The exposed range for cardiac CT (12 cm) is smaller than
ment pattern with contrast-enhanced CT. that of whole-lung or abdominal CT scans. Therefore, the
RELATIONSHIP BETWEEN ED, RD, radiation dose is not a limitation.
AND THE NO-REFLOW PHENOMENON CLINICAL IMPLICATIONS OF CARDIAC CT
The causes of the no-reflow phenomenon include experi- Although cardiac CT is a retrospective analysis, and not a
mental obstruction of the lumen of vascular vessels by neutro- real-time analysisas is echocardiography, it provides a lot of
phils or platelets, compression towing to edema of myocardial information from data obtained during a 30-s breath-hold
cells out of the vascular vessels, and a change in the viscosity acquisition (3035). The assessment of the myocardial
of blood (29). enhance-ment patterns was one of clinical implications of car-
In our study, the ED and RD existed even after successful diac CT.
reperfusion at the epicardial coronary artery level and improve-
ment of coronary blood flow, indicating that the circulation NOTES
disorder at the level of coronary microcirculation level, that is, 1. TIMI grade: grade 0 perfusion is no antegrade flow beyond
the no-reflow at the regional microcirculation level, still per- the point of occlusion; grade 1 is minimal incomplete per-
 CHAPTER 20 / ASSESSING MYOCARDIAL VIABILITY WITH MDCT 215

Fig. 10. Pitfalls of enhancement patterns. Fatty degeneration in old myocardial infarction (A). Left-ventricular thrombus (B). Left-ventricular
calcium deposit (C).

fusion of the contrast medium around the clot; grade 2 in the Kitaishikai Hospital; and Shigru Nakata, RT, Katsuji Inoue,
(partial perfusion) is complete but delayed perfusion of the MD, Tsuyoshi Matsunaka, MD, and Jitsuo Higaki, MD, the
distal coronary bed with contrast material; and grade 3 Chairman of the 2nd Department of Internal Medicine in Ehime
(complete perfusion) is antegrade flow to the entire distal University School of Medicine, for their excellent assistance in
bed at a normal rate.
publication. We are very grateful to Masaya Doi, RT, Masato
2. TIMI myocardial perfusion grade (TMPG): in TMPG 0,
there is minimal or no myocardial blush; in TMPG 1, dye Imai, RT, Yasuyuki Takahashi, RT, Hideyuki Chiba, RT,
stains the myocardium and this stain persists until the next Takashi Okamoto, RT, Isao Ouchi, RT, and Hiroshi Miguchi,
injection; in TMPG 2, dye enters the myocardium but RT, for their excellent technical assistance.
washes out slowly so that the dye is strongly persistent at
REFERENCES
the end of the injection; and in TMPG 3, there is normal
entrance and exit of dye in the myocardium so that dye is 1. Gibson CM, Cannon CP, Murphy SA, et al. Relationship of TIMI
mildly persistent at the end of the injection. myocardial perfusion grade to mortality after administration of
thrombolytic drugs. Circulation 2000;101:125130.
3. 99mTc-pyrophosphate localizes to denatured proteins
2. Abe M, Kazatani Y, Fukuda H, Tatsuno H, Habara H, Shinbata H.
within the mitochondria in areas of myocardial necrosis, Left ventricular volumes, ejection fraction, and regional wall motion
highlighting AMI region during the 7 to 10 d after the calculated with gated technetium-99m tetrofosmin SPECT in
onset. reperfused acute myocardial infarction at super-acute phase: com-
parison with left ventriculography. J Nucl Cardiol 2000;7:569574.
ACKNOWLEDGMENTS 3. Watanabe K, Sekiya M, Ikeda S, Miyagawa M, Kinoshita M,
We are grateful to Taketoshi Ito, MD, Hiroshi Matsuoka, MD, Kumano S. Comparison of adenosine triphosphate and dipyridamole
Hiroshi Higashino, MD, Hideo Kawakami., MD, Jun Aono, MD, in diagnosis by thallium-201 myocardial scintigraphy. J Nucl Med
1997;38:577581.
Kana Sakamoto, MD, and Junko Kato, MD, in the Department 4. Al-Saadi N, Nagel E, Gross M, et al. Noninvasive detection of
of Cardiology and Radiology, Imabari Hospital; Kazuhisa myocardial ischemia from perfusion reserve based on cardiovascu-
Nishimura, MD, Hideki Okayama, MD, Takumi Sumimoto, MD, lar magnetic resonance. Circulation 2000;101:13791383.
216 KOYAMA AND MOCHIZUKI

5. Rogers WJ Jr, Kramer CM, Geskin G, et al. Early contrast-enhanced 18. Rochitte CE, Lima JA, Bluemke DA, et al. Magnitude and time
MRI predicts late functional recovery after reperfused myocardial course of microvascular obstruction and tissue injury after acute
infarction. Circulation 1999;99:744750. myocardial infarction. Circulation 1998;98:10061014.
6. Pislaru SV, Ni Y, Pislaru C, et al. Noninvasive measurements of 19. Braunwald E, Kloner RA. Myocardial reperfusion: a double-edged
infarct size after thrombolysis with a necrosis-avid MRI contrast sword? J Clin Invest 1985;76:17131719.
agent. Circulation 1999;99:690696. 20. Engler RL, Schmid-Schonbein GW, Pavelec RS. Leukocyte capil-
7. Bogaert J, Maes A, Van de Werf F, et al. Functional recovery of lary plugging in myocardial ischemia and reperfusion in the dog.
subepicardial myocardial tissue in transmural myocardial infarction Am J Pathol 1983;111:98111.
after successful reperfusion: an important contribution to the 21. Reimer KA, Lowe JE, Rasmussen MM, Jennings RB. The wavefront
improvement of regional and global left ventricular function. Circu- phenomenon of ischemic cell death. 1. Myocardial infarct size vs
lation 1999;99:3643. duration of coronary occlusion in dogs. Circulation 1977;56:786794.
8. Ito H, Tomooka T, Sakai N, et al. Lack of myocardial perfusion 22. Reimer KA, Jennings RB. The wavefront phenomenon of myo-
immediately after successful thrombolysis. A predictor of poor cardial ischemic cell death. II. Transmural progression of necrosis
recovery of left ventricular function in anterior myocardial infarc- within the framework of ischemic bed size (myocardium at risk) and
tion. Circulation 1992;85:16991705. collateral flow. Lab Invest 1979;40:633644.
9. Scherrer-Crosbie M, Liel-Cohen N, Otsuji Y, et al. Myocardial per- 23. Mutzel W, Speck U, Weinmann HJ. Pharmacokinetics of iopromide
fusion and wall motion in infarction border zone: assessment by in rat and dog. Fortschr Geb Rontgenstrahlen Nuklearmed
myocardial contrast echocardiography. J Am Soc Echocardiogr Erganzungsbd 1983;118:8590.
2000;13:353357. 24. Allard M, Doucet D, Kien P, Bonnemain B, Caille JM. Experimental
10. Lepper W, Hoffmann R, Kamp O, et al. Assessment of myocardial study of DOTA-gadolinium. Pharmacokinetics and pharmacologic
reperfusion by intravenous myocardial contrast echocardiography properties. Invest Radiol 1988;23 Suppl 1:S271274.
and coronary flow reserve after primary percutaneous transluminal 25. Tauber U, Mutzel W, Schulze PE. Whole body autoradiographic
coronary angioplasty [correction of angiography] in patients with distribution studies on nonionic x-ray contrast agents in pregnant
acute myocardial infarction. Circulation 2000;101:23682374. rats. Fortschr Geb Rontgenstrahlen Nuklearmed Erganzungsbd
11. Hamada S, Naito H, Takamiya M. Evaluation of myocardium in 1989;128:215219
ischemic heart disease by ultrafast computed tomography. Jpn Circ 26. Judd RM, Lugo-Olivieri CH, Arai M, et al. Physiological basis of
J 1992;56:627631. myocardial contrast enhancement in fast magnetic resonance images
12. Naito H, Saito H, Takamiya M, et al. Quantitative assessment of of 2-day-old reperfused canine infarcts. Circulation 1995;92:
myocardial enhancement with iodinated contrast medium in patients 19021910.
with ischemic heart disease by using ultrafast x-ray computed 27. Yokota C, Nonogi H, Miyazaki S, et al. Gadolinium-enhanced mag-
tomography. Invest Radiol 1992;27:436442. netic resonance imaging in acute myocardial infarction. Am J Cardiol
13. Naito H, Saito H, Ohta M, Takamiya M. Significance of ultrafast 1995;75:577581.
computed tomography in cardiac imaging: usefulness in assessment 28. Wu K, Zerhouni E, Judd R, et al. Prognostic significance of
of myocardial characteristics and cardiac function. Jpn Circ J 1990; microvascular obstruction by magnetic resonance imaging in patients
54:322327. with acute myocardial infarction. Circulation 1998;97:765772.
14. Koyama Y, Matsuoka H, Higashino H, et al. Wall motion and wall 29. Engler RL, Schmid-Schnbein GW, Pavelec RS. Leukocyte capil-
thickness analysis of after receiving successful reperfusion therapy lary plugging in myocardial ischemia and reperfusion in the dog.
in acute myocardial infarction: an assessment of early perfusion Am J Pathol 1983;111:98111.
defect by enhanced helical CT. Jpn J Interv Cardiol 2001;16: 30. Becker CR, Ohnesorge BM, Schoepf UJ, Reiser MF. Current devel-
233242. opment of cardiac imaging with multidetector-row CT. Eur J Radiol
15. Koyama Y, Matsuoka H, Higasino H, et al. Assessment of Myocar- 2000;36:97103.
dial Infarct Volume by Contrast Enhancement CT. The Radiological 31. Mochizuki T, Murase K, Higashino H, et al. Two- and three-dimen-
Society of North America 87th Scientific Assembly and Annual sional CT ventriculography: a new application of helical CT. AJR
Meeting November 2530, 2001 Supplement to Radiology 2001; 2000;174:203208.
221(P):196. 32. Mochizuki T, Murase K, Koyama Y, Higashino H, Ikezoe J. LAD
16. Koyama Y, Matsuoka H, Higasino H, et al. Assessment of Myocar- stenosis detected by subsecond spiral CT. Circulation 1999;99:1523.
dial Infarct Size by Contrast Enhanced CT. American Heart Associa- 33. Mochizuki T, Koyama Y, Tanaka H, Ikezoe J, Shen Y, Azemoto S.
tion Scientific Sessions 2001, November 1114, 2001 Supplement Left ventricular thrombus detected by two- and three-dimensional
to Circulation 2001;104(17):II-477. computed tomographic ventriculography: a new application of
17. Koyama Y, Matsuoka H, Higasino H, et al. Early Myocardial Perfu- helical CT. Circulation 1998;98:933934.
sion Defect and Late Enhancement on Enhancement CT Predict 34. Koyama Y, Matsuoka H, Higasino H, et al. Four-dimensional cardiac
Clinical Outcome in Patients with Acute Myocardial Infarction after image by helical computed tomography. Circulation 1999;100:e61e62.
Reperfusion Therapy. The Radiological Society of North America 35. Koyama Y, Matsuoka H, Higashino H, et al. Left ventricular hyper-
87th Scientific Assembly and Annual Meeting November 2530, trophy demonstrated by four-dimensional myocardiography by
2001. Supplement to Radiology 2001;221(P):196. helical computed tomography. Circulation 2001;103:e15e17.
 CHAPTER 21 / CORONARY ARTERIES AND VEINS IN CT 217

CONTRAST-ENHANCED CT
OF THE HEART: V
CORONARY ARTERIES
 CHAPTER 21 / CORONARY ARTERIES AND VEINS IN CT 219

21 Anatomy of the Coronary Arteries

and Veins in CT Imaging

ROBERT J. M. VAN GEUNS, MD AND FILIPPO CADEMARTIRI, MD

INTRODUCTION able, but usually 12 cm (1). In a small proportion of cases, the

The cardiologist and radiologist interpreting coronary com- LCA is very short and bifurcates almost immediately. In 0.41%
puted tomography angiography (CTA) should be familiar with of the cases, the LCA is not developed and there are two orifices
coronary artery anatomy. It has a standard logical structure in the left coronary sinus (2). In two-thirds of the subjects, the
with some common variations and only a few rare abnormali- main LCA divides beneath the left atrial appendix, into the left
ties. In a conventional selective coronary angiography, blood anterior descending (LAD) and the circumflex arteries. When
in the chambers and coronary veins does not interfere with the using VRTs for off-line evaluation, the atrial appendix is usu-
visualization of the coronary arteries. In addition, myocar- ally excluded from the data set (Fig. 2) (3). The LAD artery
dium and other soft tissues are hardly seen because of their passes to the left of the pulmonary trunk and turns forwards to
low absorption of X-rays. Invasive selective coronary run downwards in the anterior interventricular groove towards
angiograms use projections performed in various orientations the apex. When MPR is used for evaluation of CTA data sets,
so that the cardiologist can perceive the 3D anatomy of the a series of parallel slices or curved MPR along the intraven-
coronary arteries. This is quite different for imaging tech- tricular groove have to be created for optimal visualization of
niques such as CTA. In CTA the contrast agent is intrave- the vessel (Fig. 3). The LAD artery provides two main groups
nously injected, which results in enhancement of the of branches. First, the septal branches, which supply the ante-
myocardium and blood in the cavities, and projection tech- rior two-thirds of the septum, and second, the diagonal
niques such as maximum intensity projection (MIP) are there- branches, which lie on the lateral aspect of the left ventricle.
fore of limited use. Overlap of structures that obscure coronary The septal branches arise form the LAD at approximately
imaging can be avoided by multiplanar reformation (MPR) 90-degree angles. They vary in size, number, and distribution.
using thin slices in any desired orientation. However, in that The first large septal branch may divide into a fork where both
case much of the 3D information is not used. With modern branches run parallel into the septum. In other cases, a septal
postprocessing tools, such as MIP or the volume-rendering branch may run parallel to the LAD through the myocardium of
technique (VRT), 3D impressions on a 2D surface can be cre- the septum. By convention, the first septal branch separates the
ated. These images look much like the gross anatomy of the proximal LAD from the middle part of the LAD. The diagonals
heart, but they do not resemble the images known from inva- also vary in number and course (Fig. 4). Usually at least one
sive selective coronary angiography. In this chapter we will diagonal is present, and if none is visualized, a total occlusion
therefore review the normal coronary artery and venous may be expected. A normal variation of the large diagonal is a
anatomy as it may be seen on MPR and VRT images of CTA. parallel course to the LAD. In the majority of the patients, the
We will also review the anomalies that may be encountered LAD itself courses around the apex to reach the inferior wall
during investigations for coronary artery disease. and septum. In the other cases, the distal right coronary artery
(RCA) is larger and supplies the blood flow for the apex. This
CORONARY ARTERY ANATOMY is one of the potential collateral routes if either the RCA or LAD
LEFT CORONARY ARTERY is occluded. The left circumflex artery (LCx) turns backwards
The left main coronary artery (LCA) arises from the left shortly beyond its origin to run downwards in the left
posterior aortic sinus. In tomographic imaging, the preferred arterioventricular groove. It too gives rise to a variable number
orientation displays the left of the patient in the right side of the of branches, which lie on the lateral aspect of the left ventricle
images, as when the patient is viewed from the feet. In such (the marginal branches, Fig. 4). In one-third of the subjects, the
images the LCA starts at the right side of the aorta just posterior left main coronary artery trifurcates into the aforementioned
to the right-ventricular outflow tract (Fig. 1). Its length is vari- branches and an intermediate artery, which follows a course
between the circumflex and LAD arteries over the anterolateral
From: Contemporary Cardiology: CT of the Heart: wall of the left ventricle (Fig. 5) (3). Additional branches of the
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
LCx are small atrial branches that supply the lateral and poste-
rior regions of the left atrium.

219
220 VAN GEUNS AND CADEMARTIRI

Fig. 2. Volume rendering of the left coronary artery after removal of

Fig. 1. Transverse slice through the aortic root. The left main coronary the left atrial appendage and cranial part of the left atrium (LA) in the
artery (LM) can clearly be seen, the proximal left anterior descend- original topographic slices over 15 levels. Without this processing
ing artery (LAD) turns around the pulmonary artery (PA) anteriorly. tool, the left atrial appendage normally overlaps the left main (LM)
AAo, ascending aorta; SCV, superior caval vein; LA, left atrium; and proximal circumflex coronary artery (LCX). The right atrial
dAo, descending aorta; A, anterior thoracic wall; P, posterior (spine); appendage (RAA) may sometimes override the right coronary artery
L, left; R, right. (RCA), but in this case manual removal was not necessary. LAD, left
anterior descending coronary artery; RVOT, right-ventricular out-
flow tract; MO, margo obtusus of the circumflex coronary artery.

Fig. 3. Left anterior descending artery (LAD) and proximal right coronary artery (RCA) in a single plane. (A) Starting from a transverse plane
through the middle of LAD, a curved reconstruction plane is selected through the proximal RCA, ascending aorta, left main and proximal LAD
(dashed line). (B) Curved multiplanar reformation along LAD and RCA. The LAD follows a course over the anterior wall of the left ventricle
to the apex of the heart. RVOT, right-ventricular outflow track; Ao, aorta; LV, left ventricle; RV, right ventricle.
 CHAPTER 21 / CORONARY ARTERIES AND VEINS IN CT 221

Fig. 4. (A) Anatomical view of the left coronary artery (reproduced with permission from [13]). The auricle of the left atrium (L.A.A.)
overlapping the circumflex coronary artery is removed. The left main (LM) artery divides beneath the L.A.A. in the left anterior descending
(LAD) and circumflex (LCX) coronary arteries. From the LAD artery diagonal branches (D) arise. The margo obtusus (MO) arises from the
LCX artery. (B) A comparable noninvasive coronary angiogram with computed tomography. Ao, aorta; PT, pulmonary trunk; LA, left atrium,
after removal of the auricle; LV, left ventricle; RVOT, right-ventricular outflow track.

RIGHT CORONARY ARTERY

The RCA arises from the anterior aortic sinus, somewhat
inferior to the origin of the LAD. It passes forward and then
downward in the right atrioventricular groove (Fig. 6) and
continues around the margin of the heart towards the crux, a
point below where the atrioventricular groove and the poste-
rior interventricular groove meet. Sometimes a single MPR
image displays a long segment of the RCA if an imaging plane
through the right atrioventricular groove is selected (Fig. 6). A
longer part may be visible if this plane is tilted with the caudal
side to the back of the patient. The first branch of the RCA is
generally the conus artery that runs over the anterior surface of
the right- ventricular outflow tract (Fig. 7). The second branch
is usually the sinoatrial node artery; alternatively, the sinus
node is supply by a proximal branch of the LCx, and in some
cases both routes are available. In the majority (80%) of indi-
viduals, the RCA continues forwards from the crux along
the posterior interventricular groove to become the posterior
descending artery (PDA), running to the apex of the heart
(Fig. 8). This is by convention called RCA dominance (4).
Septal branches supplying the posterior third of the septum
arise from the PDA and can connect with the septal branches
Fig. 5. Coronary CT angiography, trifurcation of the left main artery from the LAD and form a collateral circulation. The postero-
into left anterior descending (LAD), circumflex (LCX), and interme-
diate (MI) arteries. The LAD is occluded after the first diagonal branch lateral (PL) branch supplying the postero-inferior aspect of the
(D) but shows some contrast filling through collateral vessels (arrow- left ventricle also arises from the RCA close to the crux. Shortly
heads). the PL is a continuation of the RCA in the left atrioventricular
222 VAN GEUNS AND CADEMARTIRI

Fig. 6. Localization of the right coronary artery (RCA). (A) A transverse plane between the right ventricle outflow track (RVOT) and right
atrium (RA) through the proximal RCA is selected (dashed line). (B) Image along the proximal and middle segment of the RCA. RV, right
ventricle; LV, left ventricle.

Fig. 7. (A) A pressure-fixed anatomical specimen showing the proximal and middle right coronary artery (RCA) with its side branches (conus
branch [CB] and right-ventricular branch [RVB]). Reproduced from McAlpine (13) with permission of Springer-Verlag. (B) A 3D rendering
of the right coronary artery. Because of the small size, only the proximal part of the conus branch can be seen. The RCA shows atherosclerotic
disease over its full length. At the right side of the picture, the left anterior descending (LAD) coronary artery can be clearly seen. Ao, ascending
aorta; PT, pulmonary trunk; LV, left ventricle; RV, right ventricle; RA, right atrium.

grove but within 1 or 2 cm the crux follows an epical course CORONARY VENOUS ANATOMY
over the myocardium of the left ventricle parallel to the PDA. With the growing possibilities in electrophysiology where
Here the RCA can serve as a collateral for an occluded LCx. ablation catheters or pacemaker leads are positioned in the
Also close to the crux a small artery arises that passes upwards coronary veins, there has been a renewed interest in the venous
to the atrioventricular node of the conduction system. Left anatomy. There are two major systems of epicardial cardiac
coronary dominance exists when the PDA arises from the cir- veins: tributaries of the coronary sinus and the anterior cardiac
cumflex artery (Fig. 9). veins (Fig. 10). In principle, the veins run parallel to the arter-
 CHAPTER 21 / CORONARY ARTERIES AND VEINS IN CT 223

Fig. 8. Distal right coronary artery (RCA), diaphragmatic view. (A) Coronary CT angiography. At the crux the RCA divides into the postero-
descending artery (PDA) and postero-lateral (PL) branch over the inferior wall of the left ventricle (LV). RV, right ventricle; RA, right atrium.
(B) Anatomical view. (Reproduced with permission from Sobotta Atlas der Anatomie des Menschen, Elsevier GmbH, Munich.)

parallel to the circumflex artery, where it drains into the coro-

nary sinus. The anatomical transition of the GCV into the coro-
nary sinus is at the site of entrance of the oblique vein of the
left atrium (5). The coronary sinus continues parallel to the
circumflex artery and drains into the right atrium. The ostium
of the coronary sinus in the right atrium is most frequently
covered by a thick valve (the valve of the coronary sinus or
Thebesian valve) (5).
The GCV and coronary sinus encircle most of the left atrio-
ventricular connection and therefore are frequently used for
location of assessorial electrical pathways that are present in
the WolffParkinsonWhite syndrome.
The middle cardiac vein (MCV), receiving blood from the
posterior third of the septum, runs parallel to the PDA and
enters the coronary sinus in 87% of the cases (6). In only 36%
of the cases is there a small cardiac vein, draining the blood of
the right ventricle into the coronary sinus (6). Additional coro-
nary veins drain the lateral wall of the left ventricle and enter
the coronary sinus between the GCV and MCV. The largest
of these are used for implantation of the second lead of biven-
tricular pacemakers. These lateral veins cover the area of the
Fig. 9. Left coronary artery dominance. Posterior descending artery heart that is depolarized the latest in the presence of a left-
(PDA) and postero-lateral (PL) branch originate from the circumflex ventricular bundle branch block, which makes it the most effec-
(CX) coronary artery. LV, left ventricle; RV, right ventricle.
tive side for additional left-ventricular pacing.
The other epicardial venous system, that of the anterior
ies. The great cardiac vein (GCV), receiving blood from the veins, drains the blood from the right-ventricular wall into the
anterior two-thirds of the septum, runs parallel to the LAD in right atrium via atrial sinuses (7). Sometimes this so-called
the anterior interventricular groove. At the origin of the LAD, sinus coronarius atrii dextri is quite large (6) and can be con-
the GCV turns into the left atrioventricular groove, running fused with the RCA.
224 VAN GEUNS AND CADEMARTIRI

Fig. 10. Coronary veins. (A) Anatomical view. Two cardiac venous systems: anterior cardiac veins (ACV) and tributaries of the coronary sinus
(great cardiac vein [GCV], middle cardiac vein [MCV], and small cardiac vein [SCV]). (B) Conventional coronary angiography, venous phase.

Fig. 11. The great cardiac vein (GCV) turns from the anterior interventricular groove into the atrioventricular groove, crossing all the
branches of the left coronary artery and forming the triangle of Brocq and Mouchet together with the left anterior descending (LAD) and
circumflex (LCX) coronary arteries. (A) CTA: the view at the LCX artery is obstructed by the GCV. (B) Comparative anatomical view. LM, left
main artery; RCA, right coronary artery; ACV, anterior cardiac veins.

RELATION BETWEEN ARTERIES AND VEINS lower and the middle parts of the interventricular groove, the
LCA AND GCV GCV runs most often to the right of its related artery (8). The
The GCV is the longest venous vessel of the heart. The vein GCV crosses over the LAD artery and all of it branches in
originates at the anterior interventricular groove, near the apex 49% of the cases (Fig. 11). On reaching the atrioventricular
of the heart, and it empties into the coronary sinus. In the groove, the GCV crosses the LAD and circumflex arteries,
 CHAPTER 21 / CORONARY ARTERIES AND VEINS IN CT 225

Fig. 12. Renderings of the left coronary arterial and venous systems. (A) Great cardiac vein (GCV) running parallel to the left anterior
descending coronary artery (LAD), crossing under the circumflex artery (LCX) and entering the coronary sinus (CS). Triangle of Brocq and
Mouchet formed by the proximal left anterior descending coronary artery, the proximal circumflex coronary artery, and the great cardiac vein
crossing from the anterior interventricular groove to the atrioventricular groove. (B) 3D rendering of the same dataset, offering a more lateral
and posterior view of the heart, arteries, and veins. This clearly shows the possibility of the 3D rendering technique to view the object from
any angle. Reproduced with permission from (14).

forming the base of the triangle of Brocq and Mouchet (8).

The distance from the GCV of the left main coronary artery is
variable (07 mm [8]), and sometimes the GCV touches the
left main coronary artery and turns with a very sharp angle to
the left atrioventricular groove, crossing under the branches
of the left main coronary artery (Fig. 12). The circumflex
artery is covered by the GCV in 60% of the cases so that the
underlying anatomy of circumflex artery is obscured or inad-
equately visualized.
RCA AND CORONARY SINUS
At the crux of the heart the RCA is, with very rare excep-
tions, inferior to the coronary sinus. The middle cardiac vein
crosses over the postero-lateral branch of the RCA and stays
left of the PDA when running in the posterior interventricular
groove (Fig. 13). In cases of left circumflex artery dominance,
veins draining blood from the inferior wall of the left ventricle
cross over the artery before entering the coronary sinus.
CORONARY ARTERY ANOMALIES
A large number of coronary artery anomalies have been
described in the literature (2,9). Most extreme is the origin of
one of the coronary arteries from the pulmonary artery; most
frequently, this is the LAD, probably owing to its normal
course underneath and around the pulmonary artery (Fig. 1).
Because of the large amount of ischemia in such a case, symp-
toms are normally noticed the first 4 mo of life (9,10). The
Fig. 13. Right coronary artery (RCA) in the inferior atrioventricular
most frequent anomalies are the different fistulae, which nor- groove partly covered by the middle cardiac vein (MCV) and the
mally arise from the RCA; less frequently they arise from the coronary sinus. PCV, posterior cardiac vein; RA, right atrium; LV,
LAD or LCx. Drainage usually occurs into the right ventricle, left ventricle.
226 VAN GEUNS AND CADEMARTIRI

Fig. 14. Different pathway of aberrant origin of the coronary artery arteries. (A) Normal anatomy. Caudal view. (B) Right coronary artery
(RCA) from left coronary cusp, inter-arterial course. (C) Left main (LM) from right coronary cusp, anterior course. (D) LM from right coronary
cusp, inter-arterial course. (E) LM from right coronary cusp, posterior course.

Fig. 15. Left anterior descending (LAD) coronary artery from right coronary cusp. (A) Curved axial MIP demonstrating the inter-arterial course
of the LAD between the pulmonary artery (PA) and the aorta (Ao). (B) VRT image. The proximal and distal LAD are clearly visible; the inter-
arterial course is covered by the PA. LCX, left circumflex coronary artery; RCA, right coronary artery; LV, left ventricle.

right atrium, or pulmonary artery. Occasionally, they drain in Origin of a coronary artery from the contralateral sinus
the left ventricle or superior caval vein. Symptoms of coro- (Fig. 14) is normally detected only during selective coronary
nary artery fistulae are related to congestive heart failure as angiography for suspected coronary artery disease (11,12).
a result of left-to-right shunting, infective endocarditis, or They are categorized into anomalies expected to cause myocar-
myocardial ischemia. dial ischemia or those unlikely to cause myocardial ischemia.
 CHAPTER 21 / CORONARY ARTERIES AND VEINS IN CT 227

CONCLUSION
Three-dimensional data sets from noninvasive 3D coronary
imaging techniques such as CTA are displayed with MPR or
VRT. This provides images of the coronary arteries and veins
much like their real anatomy (13), which are not always famil-
iar to the practicing cardiologist. Knowledge of the course of
the epicardial coronary arteries and veins is required for accu-
rate analysis.

REFERENCES
1. James T. Anatomy of the coronary arteries in health and disease.
Circulation 1965;32:10201033.
2. Yamanaka O, Hobbs R. Coronary artery anomalies in 126,595
patients undergoing coronary arteriography. Cathet Cardiovasc
Diagn 1990;21:2840.
3. Levin DC, Harrington DP, Bettmann MA, Garnic JD, Davidoff
A, Lois J. Anatomic variations of the coronary arteries supplying
the anterolateral aspect of the left ventricle: possible explanation
for the Unexplained anterior aneurysm. Invest Radiol 1982;
17:458462.
4. Braunwald E. Heart Disease. A Textbook of Cardiovascular
Medicine. 4th ed. W.D. Saunders Company, Philadelphia: 1992.
5. Maric I, Bobinac D, Ostojic L, Petkovic M, Dujmovic M. Tribu-
taries of the human and canine coronary sinus. Acta Anat (Basel)
Fig. 16. Right coronary artery (RCA) from left coronary cusp with in 1996;156:6169.
inter-arterial course. Ao, aorta; PA, pulmonary artery; RV, right ven- 6. von Ludinghausen M. Clinical anatomy of cardiac veins, Vv.
tricle; LAD, left anterior descending artery; LCX, left circumflex cardiacae. Surg Radiol Anat 1987;9:159168.
coronary artery. 7. Pina JA. Morphological study on the human anterior cardiac veins,
venae cordis anteriores. Acta Anat (Basel) 1975;92:145159.
8. Pejkovic B, Bogdanovic D. The great cardiac vein. Surg Radiol
Anat 1992;14:2328.
The latter have been associated with sudden death of young 9. Levin DC, Fellows KE, Abrams HL. Hemodynamically signifi-
athletes during exercise. This category includes the interarte- cant primary anomalies of the coronary arteries. Angiographic
rial course or septal course of the left main or left anterior aspects. Circulation 1978;58:2534.
descending coronary artery originating from the right coronary 10. Wilson CL, Dlabal PW, Holeyfield RW, Akins CW, Knauf DG.
Anomalous origin of left coronary artery from pulmonary artery.
cusp and the interarterial course of a right coronary artery origi- Case report and review of literature concerning teen-agers and
nating from the left coronary artery (Fig. 14B,D). The danger adults. J Thorac Cardiovasc Surg 1977;73:887893.
is explained by the slit-like ostium with a sharp angle between 11. Chaitman BR, Lesperance J, Saltiel J, Bourassa MG. Clinical,
the coronary artery and aorta (Fig. 15). Another explanation is angiographic, and hemodynamic findings in patients with
a possible compression of the vessel between the aorta and anomalous origin of the coronary arteries. Circulation 1976;53:
122 131.
pulmonary artery or compression within the myocardium of the 12. Kimbiris D, Iskandrian AS, Segal BL, Bemis CE. Anomalous
interventricular septum in such a course (Fig. 16). These abnor- aortic origin of coronary arteries. Circulation 1978;58:606615.
mal vessels may be prone to earlier atherosclerosis and suc- 13. McAlpine. Heart and Coronary Arteries. Springer-Verlag, Ber-
cessive myocardial ischemia. lin: 1975.
14. Rensing BJ, Bongaerts AHH, van Geuns RJ, et al. In vivo assess-
Coronary artery anomalies not related to ischemia follow a ment of three dimensional coronary anatomy using electron beam
course anterior (Fig. 14C) to the pulmonary artery or a long tra- computed tomography after intravenous contrast administration.
jectory posterior around the aorta (retroaortic course) (Fig. 14E). Heart 1999;82(4):523525.
 CHAPTER 22 / ACQUISITION AND DISPLAY OF CTA 229

22 Techniques and Protocols for Acquisition and

Display of Contrast-Enhanced CT Angiography

CHRISTOPH R. BECKER, MD

INTRODUCTION the human heart may change rapidly, in particular under special
Contrast-enhanced computed tomography (CT) studies of conditions such as breath-holding and Valsalva maneuver.
the coronary arteries were performed first to visualize the ves- When multisector reconstruction is performed, the temporal
sel lumen and to achieve an angiographic-like presentation of resolution then depends on the actual heart rate and may vary
the coronary arteries in combination with 3D postprocessing during the acquisition. For this reason, this technique is of lim-
methods (1). Because short exposure times are essential for ited practical use under clinical conditions.
coronary CT angiography (CTA), investigations initially were In general the redundant radiation exposure in the systole
performed with electron beam CT (EBCT) scanners. These can substantially be reduced by a technique called prospective
dedicated cardiac CT scanners were originally designed to ECG tube current modulation. On the basis of the ECG signal,
measure myocardial perfusion (2). For morphological assess- the X-ray tube current is switched to its nominal value during
ment of cardiac structures, a restrictive scan protocol with the diastole phase and is reduced significantly during the sys-
EBCT allows for 100-ms exposure time, 3-mm slice thickness, tole phase of the heart, respectively. This technique is most
and 130-kVp, 630-mA electron-gun power. The acquisition of effective in patients with low heart rates. If the heart rate is
every single slice is triggered prospectively by the electrocar- lower that 60 beats per minute (bpm), the radiation exposure
diogram (ECG) signal at the end systolic phase of the cardiac will be reduced by approx 50% (3).
cycle. In addition, retrospective ECG gating allows for reconstruc-
Multidetector-row CT (MDCT) scanners operate with a dif- tion of images at any time within the cardiac cycle. However,
ferent acquisition mode called retrospective ECG gating. The with the currently available temporal resolution in MDCT, the
combination of fast gantry rotation, slow table movement, and image quality is poor only if the images are reconstructed in any
multislice helical acquisition allows for acquisition of a high other than the mid-diastolic phase. The accuracy of the func-
number of X-ray projection data. The ECG trace is recorded tional assessment by MDCT may be influenced if motion arti-
simultaneously during the helical scan acquisition. The X-ray facts are present or if a -blocker have been administered.
projections of the mid-diastolic phase are selected to recon-
PATIENT PREPARATION
struct images from the slow-motion diastole phase of the heart.
In MDCT the temporal resolution (approx 200 ms) is longer For a 16-row-detector CT scanner with 420 ms gantry
and the radiation exposure (approx 10 mSv) is higher than in rotation, an optimized partial scan view lasts about 200 ms.
EBCT. On the other hand, image quality is superior in MDCT Reasonably good image quality with this temporal resolution
compared to EBCT because of higher spatial resolution, lower can be achieved only in patients with low heart ratese.g.,
image noise, and continuous volume acquisition. <70 bpm (4). Therefore, caffeine or any drug like atropine or
In MDCT an attempt has been made to improve temporal nitroglycerin that increases the heart rate should be avoided
resolution by multisector reconstruction. For this technique, prior to a cardiac CTA investigation. Instead the use of a
X-ray projections of more than one heartbeat are used to recon- -blocker may become necessary for patient preparation aim-
struct an image. This technique requires absolute consistent ing at a heart rate of 60 bpm or even less.
data from at least two consecutive heartbeats for successful To consider -blocker for patient preparation, contraindica-
image reconstruction. For multisector reconstruction, the table tions (bronchial asthma, AV block, severe congestive heart
feed needs to be reduced to collect enough projections. This failure, aortic stenosis, and so on [5]) have to be ruled out and
leads to longer scan times and higher radiation exposure com- informed consent must be obtained from the patient. In case the
pared to a partial scan reconstruction. However, the rhythm of heart rate of a patient is significantly above 60 bpm, 50 to 200 mg
of Metoprololtartrat can be administered orally 3090 min prior
to the investigation. Alternatively, 520 mg of Metopro-
From: Contemporary Cardiology: CT of the Heart: loltartrat divided into 4 doses can be administered intravenously
Principles and Applications (5) immediately prior to scanning. Monitoring of vital func-
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
tions (heart rate and blood pressure) is essential during this

229
230 BECKER

Table 1
Scan Parameters for 4- and 16-Detector-Row CT
4-Detector CT 16-Detector CT
Slice thickness 1 mm 0.75
Tube current 400 mAs 500 mAs
Tube voltage 120 kVp 120 kVp
Gantry rotation 500 ms 420 ms
Table feed 3 mm/s 6 mm/s
Breath hold 40 s 20 s
Spatial resolution 8 linepairs (lp)/cm 8 lp/cm
Temporal resolution 250 ms 200 ms
Number of slices 200 240

approach. There are four positive effects of -blocker on MDCT The point of time for the least coronary motion may be different
scanning: better patient compliance, less radiation exposure for every coronary artery. Minimal motion artifacts may result
and cardiac motion artifacts, and higher vascular enhancement. for reconstructing the right coronary artery (RCA), left anterior
descending (LAD) artery, and left circumflex (LCx) artery at
MDCT SCANNING
50%, 55%, and 60% of the RR interval, respectively. Indi-
For retrospective ECG gating with MDCT, the pitch factor vidual adaptation of the point of time for reconstruction seems
(detector collimation divided by table feed per gantry rotation) to further improve image quality (4). However, the lower the
must not exceed 0.3 to allow for scanning the heart at any heart heart rate, the easier it is to find the best interval for all three
rate higher than 40 beats per minute. In a 4-detector-row CT major branches of the coronary artery tree. Images are recon-
with 1-mm slices and 500-ms gantry rotation, a typical scan structed with spatial in-plane resolution of 8 linepairs (lp)/cm
range of 12 cm lasts approx 40 s. In a 16-detector-row CT with (0.6 mm). The slice thickness and reconstruction increment is
0.75-mm slices and a gantry rotation time of 420 ms, the entire 1.3 with 0.7 mm and 1 with 0.5 mm in 4- and 16-detector-row
scan lasts about 20 s. CT, respectively. Therefore, near-isotropic spatial resolution
In 4- and 16-detector-row CT, tube current is set to 400 and may be achieved only with the 16-detector-row CT.
500 effective mAs, respectively. The higher tube current is Depending on the preparation, patient investigation time
needed to compensate for higher image noise caused by both may last approx 1015 min. Image reconstruction and post
thinner slices and the higher temporal resolution used with a processing can be performed within approx 10 min. Because
16- compared to 4-detector-row CT scanner. Higher temporal coronary CTA is performed with thin slices and low image
resolution equates to shorter exposure times and therefore does noise, the radiation dose with tube current modulation is sig-
not necessarily translate into higher radiation exposure. The nificantly higher (approx 5 mSv) that for coronary calcium
patient may either be exposed for a longer time with a lower screening (approx 1 mSv). However, the radiation of a CTA
tube current or for a shorter exposure time with higher tube investigation is comparable to what is applied during a diag-
current; however, the final amount of radiation for the same nostic coronary catheter procedure. Scan parameters are sum-
image noise remains the same. marized in Table 1.
The patients should be instructed not to press when taking a
deep breath in, to avoid the Valsalva maneuver. During the IMAGE POSTPROCESSING
Valsalva maneuver, the intraabdominal pressure increases with The detection of coronary artery pathology in axial CT
blood from the inferior vena cava entering the right atrium. images is difficult since every slice displays only a small frag-
With an increased pressure in the right atrium, blood mixed ment of the entire coronary artery. Two- or three-dimensional
with contrast medium from the superior vena cava is prevented postprocessing display such as multiplanar reformations and
from entering the right atrium. Furthermore, as a reflex the maximum intensity projections (MIP) or shaded surface dis-
increased blood volume in the right atrium leads to a decrease play, virtual coronary endoscopy, and volume-rendering tech-
in heart rate. nique (VRT), respectively, may be helpful for a first glance at
The Valsalva maneuver may last several seconds under the course of the coronary arteries or for angiographic-like
breath-hold conditions. After it releases, high-density contrast projections. It should be considered that any 2D or 3D recon-
medium may enter the right atrium and the heart rate recovers structions may come along with a loss of spatial resolution.
to its original or even slightly higher frequency (6). As a result, Smaller side branches, for instance, which are visible in the
the CT may not be homogeneously enhanced and image quality axial source, may not be seen in volume-rendered images, and
may suffer from rapid changes of the heart rate. the presence of coronary artery stenosis may be overestimated.
After scan acquisition has been completed, reconstruction In particular, motion and ECG trigger artifacts in 2D and 3D
of the axial slices always begins with a careful analysis of the images may appear as pathologies or even stenoses that do not
ECG trace recorded with the helical scan. The image recon- exist. In addition, the primary axial slices are superior to any
struction interval is best been placed in between the T and the postprocessing method to rule out coronary artery disease
P wave of the ECG, corresponding to the mid-diastole interval. (CAD) and atherosclerosis. Therefore, none of the available
 CHAPTER 22 / ACQUISITION AND DISPLAY OF CTA 231

Fig. 1. Volume-rendering technique after segmentation of the bones and auricle performed along the long axis of the heart (A). Selective images
display the course of the left anterior descending and circumflex coronary artery (B) as well as the proximal and middle right coronary artery
(C). At the base of the heart, the distal right and posterior descending coronary artery can be seen (D).

postprocessing tools is better than the axial slices to detect This projection plane spreads the branches of the LAD and is
coronary artery stenosis (7). therefore called the spider view. The spider view is well suited
VRT requires the segmentation of the sternum, the rips, and to demonstrate the proximal part of all three major coronary
the vertebral column. The auricle of the left atrium may also be arteries (Fig. 4).
segmented to gain access to the left main coronary artery. Stan-
dardized projection views around the long axis of the heart ANALYSIS OF AXIAL CT IMAGES
allow for display of the superficial course of all three major All findings from postprocessed images have to be con-
coronary arteries (Fig. 1). firmed in the original axial CT slices. The primary axial slices
Standardized thin (3 mm) MIP slices with 1.5-mm incre- are superior to any postprocessing method to rule out CAD.
ment between the slices may be reconstructed in three different Image analysis begins with identification of the coronary artery
planes, similar to standard cardiac catheter projections (8). MIP segments in the axial CT slices. Coronary segments can be
along the interventricular groove creates images in the right numbered (Fig. 5) according to the model suggested by the
anterior oblique plane that best displays the course of the LAD American Heart Association (9).
(Fig. 2). MIP along the atrioventricular groove in the left ante- In patients with acute or chronic myocardial infarctions, a
rior oblique plane displays best the course of the right and lack of contrast uptake in the myocardium or thinning of the
circumflex coronary artery (Fig. 3). In addition, a left anterior myocardial wall may be detected only in the axial slices (10).
oblique projection can be reconstructed following the course of Depending on the coronary vessel affected, different anatomic
the left anterior descending coronary artery in the lateral view. regions may be involved (Fig. 6). In case of an involvement of
232 BECKER

Fig. 2. The right anterior oblique (RAO) view reconstructed along the
septum (A) is designed to display the course of the left anterior
descending coronary artery (B).

the left anterior descending, a myocardial infarction may be

identified in the anterior left-ventricular wall, the roof of the
left ventricle, the apex, the superior part of the septum, or in the
anterior papillary muscle. The posterior left-ventricular wall
and the posterior papillary muscle are supplied by the circum-
flex coronary artery. The inferior left-ventricular wall and the
inferior part of the septum finally are supplied by the right
coronary artery.
With later development, a subendocardial or transmural
myocardial infarction may result in an aneurysm. As a result of
hypokinesia in the aneurysm, thrombus formation is likely to
develop in the cardiac chamber and can be detected by CTA
even better than by transthoracic ultrasound (11).

Fig. 3. (right) The left anterior oblique (LAO) view (A) reconstructed
along the atrio-venticular groove is best suited to display the right (B)
and circumflex coronary artery (C).
 CHAPTER 22 / ACQUISITION AND DISPLAY OF CTA 233

Fig. 4. Left anterior oblique (Spider) views are reconstructed along

the plane of the left anterior descending coronary artery (A) for dis-
playing this vessel and its major diagonal branches (B).

A late uptake of contrast media after first pass in the myo-

cardium of patients after infarction was observed in CT about
17 yr ago (11). It is rather likely that this kind of myocardial
enhancement may correspond to nonviable myocardium due to
interstitial uptake of contrast media within necrotic myocytes,
6 wk to 3 mo after onset. To allow for superior detection of the
late myocardial enhancement with the new generation MDCT
scanners, images should be acquired with thick slices, high
signal-to-noise ratio, and reconstructed with a very soft tissue
kernel. The optimal point of time for scanning may be between
10 and 40 min after first pass (12).
with a larger field of view, or a more dedicated (CT) inves-
tigation should then be recommended.

Fig. 5AG. (right and page 224) In the axial slices, the coronary
artery segments are identified and numbered according the American
Heart Association model.
234 BECKER

Fig. 5AG. (continued from page 223)

Fig. 6. The myocardium can be divided into three different territories:

the anterior, inferior, and posterior wall, supplied by the left anterior,
right, and circumflex coronary artery. In this particular image, an
infarction scan can be seen in the posterior wall.

CONCLUSION
With 16-detector-row scanners, investigation of the coro-
nary arteries and the myocardium becomes a feasible and prac-
tical CT application. Patient preparation and instruction is
mandatory to acquire diagnostic image quality for the coronary
arteries. Under these conditions, postprocessing in a standard-
ized fashion may help to identify the pathology. However, any
finding needs to be confirmed by the source images to identify
artifacts that may arise from motion. Pathologies arising from
the peri- and myocardium may be assessable only in the axial
slices.
 CHAPTER 22 / ACQUISITION AND DISPLAY OF CTA 235

REFERENCES 7. Vogl TJ, Abolmaali ND, Diebold T, et al. Techniques for the detec-
tion of coronary atherosclerosis: multi-detector row CT coronary
1. Moshage WE, Achenbach S, Seese B, Bachmann K, Kirchgeorg M. angiography. Radiology 2002;223(1):212220.
Coronary artery stenoses: three-dimensional imaging with electro- 8. Johnson M. Principles and practice of coronary angiography. In:
cardiographically triggered, contrast agent-enhanced, electron-beam Skorton D, Schelbert H, Wolf G, Brundage B (eds), Marcus Cardiac
CT. Radiology 1995;196(3):707714. Imaging: A Companion to Braunwalds Heart Disease. 2nd ed. WB
2. Boyd D. Computerized transmission tomography of the heart using Sanders Company, Philadelphia: 1996;220250.
scanning electron beams. In: Higgins C (ed), CT of the Heart and the 9. Austen WG, Edwards JE, Frye RL, et al. A reporting system on
Great Vessels: Experimental Evaluation and Clinical Application. patients evaluated for coronary artery disease. Report of the Ad Hoc
Futura, Mount Kisco, New York: 1983. Committee for Grading of Coronary Artery Disease, Council on
3. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT of Cardiovascular Surgery, American Heart Association. Circulation
the heart with retrospective ECG gating: reduction of radiation 1975;51(4 Suppl):540.
exposure by ECG-controlled tube current modulation. Eur Radiol 10. Paul JF, Dambrin G, Caussin C, Lancelin B, Angel C. Sixteen-slice
2002;12(5):10811086. computed tomography after acute myocardial infarction: from
4. Hong C, Becker CR, Huber A, et al. ECG-gated reconstructed multi- perfusion defect to the culprit lesion. Circulation 2003;108(3):
detector row CT coronary angiography: effect of varying trigger 373374.
delay on image quality. Radiology 2001;220(3):712717. 11. Masuda Y, Yoshida H, Morooka N, Watanabe S, Inagaki Y. The
5. Ryan T, Anderson J, Antman E, et al. ACC/AHA guidelines for the usefulness of x-ray computed tomography for the diagnosis of myo-
management of patients with acute myocardial infarction. A report cardial infarction. Circulation 1984;70:217225.
of the American College of Cardiology/American Heart Associa- 12. Huber D, Lapray J, Hessel S. In vivo evaluation of experimental
tion Task Force on Practice Guidelines (Committee on Manage- myocardial infarcts by ungated computed tomography. AJR 1981;
ment of Acute Myocardial Infarction). J Am Coll Cardiol 1996;28: 136:469473.
13281428. 13. Horton KM, Post WS, Blumenthal RS, Fishman EK. Prevalence of
6. Mao SS, Oudiz RJ, Bakhsheshi H, Wang SJ, Brundage BH. Varia- significant noncardiac findings on electron-beam computed tomog-
tion of heart rate and electrocardiograph trigger interval during raphy coronary artery calcium screening examinations. Circulation
ultrafast computed tomography. Am J Card Imaging 1996;10:239243. 2002;106(5):532534.
 CHAPTER 23 / CONTRAST MATERIAL INJECTION TECHNIQUES 237

23 Contrast Material Injection Techniques

for CT Angiography of the Coronary Arteries

FILIPPO CADEMARTIRI, MD AND KOEN NIEMAN, MD

INTRODUCTION jection (MIP) algorithms, curved planar reformations (CPR),

Conventional catheter angiography is based on capturing and 3D reconstructions are more important. Those techniques
X-ray images of the iodinated contrast material (CM) while it need a high vascular contrast in order to provide diagnostic
flows into the vessels. Computed tomography angiography images.
(CTA) is based on the same concept, scanning the patient and For all of these reasons, CM volume, concentration, rate,
his/her heart while a high concentration of iodinated CM flows and ultimately the synchronization between CM material pas-
through the coronary arteries (1). The CM inside the vessels sage and data acquisition, need to be optimized in order to
increases the density of the vessel lumen compared to the sur- exploit the potential of MSCT technique.
rounding tissues, and allows us to distinguish between lumen The bolus timing can be based on the knowledge of bolus
and soft tissues. geometry (e.g., demographics-based delay or fixed delay), but
Because the CM is administered intravenously in CTA, the can better rely on synchronization techniques. An optimal tim-
optimal phase to scan coronary arteries will be the arterial one. ing is achieved by predicting the veno-arterial transit time (e.g.,
In fact, during the delayed phase, several tissues enhance due test bolus), or synchronizing data acquisition with CM passage
to the CM perfusion, and the CM itself dilutes into the extracel- (e.g., bolus tracking).
lular fluid, reducing the intravascular attenuation. Then vessels
BASICS OF BOLUS GEOMETRY
are no longer easily visualized as in the arterial phase (1).
Spiral computed tomography technology was limited by low Bolus geometry is the pattern of enhancement plotted on a
speed and spatial resolution. Multislice computed tomography time(s)/attenuation (Hounsfield unitsHU) curve, after intra-
(MSCT) technology provides better image quality because of vascular injection of contrast material, and measured in a region
thinner slice thickness, and faster acquisition time because of of interest (ROI) (1). The value of enhancement is extracted by
reduced scan rotation time and multiple detector rows. The subtracting the attenuation value in an unenhanced baseline
advantages of MSCT technology related to CM use are: (1) use scan from the attenuation values in the enhanced scans (1).
of less CM (3050%); (2) the injection rate can be increased Optimal bolus geometry, for the purpose of CTA, corresponds
with a concomitant better enhancement of the vessels; and (3) to an immediate rise to the maximum value of enhancement
most of the data can be acquired during a defined phase (e.g., (high attenuationHU) just before the start of the acquisition
arterial phase for CTA) (1). of CT data, and a steady state in which the enhancement does
The latest generation of spiral computed tomography (CT) not alter during data acquisition (Fig. 1A). Actual bolus geom-
scanner, featuring up to 16-row MSCT technology, provides etry is different. After intravascular injection of CM, there is a
fast data acquisition, increasing the need for an accurate opti- steady increase in enhancement; the peak of the curve will be
mization of CM administration and synchronization tech- reached after the end of contrast injection, followed by a steady
niques. decline in the enhancement. Normally, CTA will be performed
Scanning during the arterial phase in CTA, with low or no during the upslope and downslope of the enhancement curve,
venous enhancement, enables optimal analysis of the acquired and the peak of maximum enhancement (PME) will be inside
images and is critical for postprocessing techniques. Although the scan period (Fig. 1B).
axial images are commonly used for other conventional CT The actual bolus geometry can be defined by the peak of
applications such as thoracic and abdominal imaging, for visualiza- maximum enhancement in HU (PME) and the time to reach that
tion and diagnosis of cardiac CT data sets, 2D multiplanar peak in seconds (tPME) (1).
reformations (MPR) with standard and multiple intensity pro- PARAMETERS INFLUENCING BOLUS GEOMETRY
Demographics
The time to peak (tPME) is not affected by age (24), weight
From: Contemporary Cardiology: CT of the Heart: (25), height (3,4), body surface (3,4), blood pressure (3), heart
Principles and Applications rate (3,5), and gender (2,3), and PME is not affected by age or
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ gender (2). A heavier body weight is associated with a higher

237
238 CADEMARTIRI AND NIEMAN

Fig. 1. Optimal and actual bolus geometry in CT angiography (CTA). (A) Optimal bolus geometry. The abrupt rise and the steady plateau of
attenuation characterize the optimal pattern of bolus geometry. Ideally, the scan time tightly overlaps with the length of the plateau of
enhancement in order to use the entire volume of contrast material (CM) administered for the acquisition. (B) Actual bolus geometry. The
pattern of actual bolus geometry is quite different from the ideal one. Before and after the peak of maximum enhancement there are slopes of
enhancement. Generally the up-slope is steeper than the down-slope, especially with CM administration protocols for CTA characterized by
a high injection rate.

extracellular intravascular fluid volume. This results in a dilu- ship is independent of iodine concentration and CM injection
tion of CM with a lower iodine concentration in blood, and volume (2,68,10,11).
therefore a reduced PME. Iodine Concentration
Diseases Higher iodine concentration produces a higher PME (Fig.
A reduced cardiac output produces a proportionally higher 2C) (18,12). Different iodine concentrations with constant rate
PME and longer tPME in the aortic bolus geometry (6). This is and volume, do not affect the injection duration, and this means
a result of the increase in circulation time and CM pooling, that the tPME remains unchanged.
which occur during reduced cardiac output conditions (6).
Injection Volume
A higher volume of CM shifts the time/attenuation curve Fig. 2. (right and page 240) Parameters affecting bolus geometry. (A)
upwards and rightwards (Fig. 2A) (1). This determines a higher The influence of contrast material (CM) volume. Increasing the
PME and a longer tPME. The relation is independent of injec- volume of injected CM produces an increase in peak of maxiumum
tion rate and iodine concentration (79). enhancement (PME), and a delayed time to peak (tPME). (B) The
Injection Rate influence of CM injection rate. Increasing the rate of injected CM
produces an increase in PME, and an earlier tPME. (C) The influence
Increasing the injection rate produces a proportionally of CM iodine concentration. Increasing the iodine concentration of
higher PME and earlier tPME with a shift of the time/attenua- the injected CM produces an increase in PME, without any influence
tion curve upwards and leftwards (Fig. 2B) (1). The relation- on tPME.
CHAPTER 23 / CONTRAST MATERIAL INJECTION TECHNIQUES 239
240 CADEMARTIRI AND NIEMAN

Fig. 2. (continued from page 238) (D) The influence of saline chaser. The saline chaser pushes the injected contrast medium through the veins
of the forearm, providing a result similar to the injection of a larger contrast volume. The example shows the effect of a 50-mL saline chaser
(thicker curve), using a bolus with the same volume, rate, and iodine concentration. Moreover, the saline chaser prevents the decrease of the
CM in the arm veins, which may normally cause an increase in the CM concentration after the end of the contrast injection. (E) The influence
of multiphasic protocols. Multiphasic protocols allow producing a longer plateau of enhancement. If the scan time is long (approx 3035 s),
then advantages are consistent, but with a shorter scan time, as with 16-row multislice CT, the importance of a long plateau of enhancement
is reduced if compared to the impact of a very high PME.

Bolus Chaser a bolus chaser is added to the CM injection (16), while no

A bolus chaser is a saline solution pushed through the injec- increase in PME and tPME are expected if a bolus chaser is
tion line immediately after the injection of the main bolus (1). added to the injected volume with a concomitant decrease in the
Advantages are reported from this technique applied in CT contrast volume, resulting in an unchanged total injection vol-
(1315). Bolus chaser is generally administered by a double- ume (Fig. 2D).
head power injector system (13). With bolus chaser, less con- Multiphasic Protocols
trast medium volume (up to 40% less) can be used for a CTA Multiphasic protocols are characterized by a decreasing of
scan, without affecting arterial enhancement and diagnostic the injection rate during CM administration (6,1720). The aim
accuracy (15). The PME is higher and the tPME is longer when of multiphasic protocols is to create a steady plateau of enhance-
 CHAPTER 23 / CONTRAST MATERIAL INJECTION TECHNIQUES 241

ment during the scan by means of a higher injection rate at an artery close to the region which has to be studied, and a
the beginning of the injection and a lower rate in the second trigger attenuation value (threshold) that is arbitrarily chosen
part (Fig. 2E) (17). The introduction of the last generation of before starting the CTA data acquisition (1). A single level
MSCT scanners with shorter scan times reduced the impor- dynamic scan (e.g., monitoring scan) is performed at short
tance of multiphasic protocols. A higher PME can be far more intervals (12 s) during the injection of CM. When the CM
important for the qualitative outcome of the CTA scan. arrives at the level of the ROI, the change in attenuation is
Injection Site detected and a CT scan is started after reaching the triggering
Contrast material can be administered through an antecu- threshold (Fig. 3C) (1).
bital vein that drains directly into the deep venous circulation Bolus tracking provides a better timing and allows the use of
of the arm (basilic vein), or a forearm vein that drains into the less CM with a higher injection rate. A pitfall of this technique
deep venous circulation of the forearm and subsequently in the occurs when the threshold is not reached. Even though it is a
deep venous circulation at the elbow joint or into the subclavian very rare event when the protocol is optimized, it is always
vein through the cephalic vein. Larger veins allow higher rates possible to start the scan manually. In this case it is difficult to
and more safety. obtain a high-quality CTA because of the later start of the scan,
PREDICTION AND SYNCHRONIZATION OF BOLUS the decreased amount of contrast medium inside the vessels,
GEOMETRY and the prominent venous enhancement.
Patient Demographics
No or poor correlation was reported between a calculated PATTERNS OF ENHANCEMENT
tPME and the actual tPME (3,21). Nevertheless, it is not pos- OF CORONARY ARTERIES
sible to exclude that in the future a relationship between demo- Coronary arteries originate at the root of the ascending
graphics and bolus timing parameters will be found. aorta, and the intravenous CM arrives at that level already
Fixed Delay diluted, after pulmonary circulation and left-ventricle con-
Fixed delay is a routinely applied technique for CTA. The traction. Therefore, it can be assumed that the characteristics
increased scan speed of MSCT scanners needs a more careful and the dynamics of bolus geometry inside coronary arteries
CM administration. The risks of a fixed-delay technique are are the same as in the ascending aorta. Differences in attenu-
related to the fact that, in a percentage of patients, circulation ation between ascending aorta and coronary arteries can be
time is quite different from the protocol applied. In those cases determined by stenosis or occlusions. In this case, in fact,
the scan could be successful if the circulation time is shorter there can be various degrees of attenuation inside coronary
than that fixed delay, even if a prominent venous enhancement arteries and their branches depending on the flow through the
can be observed; but when the circulation time is longer than stenotic vessel or on the backflow provided by collateral cir-
fixed delay, the scan fails because of the lack of proper attenu- culation. In a patent coronary artery, the flow speed guaran-
ation inside the coronary arteries (Fig. 3A). tees an optimal enhancement in a few seconds. With stenosis
Test Bolus and occlusions, the flow speed can be reduced. The MSCT
Test bolus technique entails that a ROI is plotted inside the scanners with 16 slices and approx 0.4-s rotation time provide
lumen of an artery close to the region that needs to be studied. a scan time of less than 20 s. With a good synchronization
A small amount of CM (1015 mL, or around 10% of the main technique, the heart has at least 46 s before the smaller
bolus) is injected at the same rate as the main bolus while a branches (diagonal and marginal) of the coronary arteries are
single level dynamic scan (e.g., monitoring scan) is performed acquired with a scan performed in the cranio-caudal direc-
at short intervals (approx 12 s). When CM arrives in the lumen tion. This time span is generally enough to allow an arterial
of the artery at the level of the ROI, test bolus geometry is perfusion and even the collateral circulation to fill in the case
assessed and the time between the start of the test bolus injec- of stenosis or occlusion of one or more vessels.
tion and a determined point of the time/attenuation curve of the BOLUS TIMING TECHNIQUES WITH 16-ROW MSCT
test bolus is used as delay time for the injection of the main For cardiac and coronary MSCT imaging, fixed delay, test
bolus (Fig. 3B) (1). bolus, and bolus tracking techniques have been applied
The use of test bolus is based on a relationship between the (2328). There are no published data comparing bolus timing
geometry of the test bolus and that of the main bolus. There is techniques applied to coronary imaging in the 4-slice era. In
no or poor correlation between test bolus tPME and main bolus fact, the modalities applied for bolus timing have been severely
tPME (2,5,22), while there is a strong correlation between test influenced by the speed of acquisition and by the hyperventi-
bolus tPME and time to reach determined attenuation thresh- lation performed just before the scan to allow a longer apnea.
olds like T50, T100, T150, and T200 (2,5) (i.e., the time from In some cases, the use of oxygen to increase the apnea has been
the beginning of the injection to reach 50/100/150/200 HU in reported. For these reasons, bolus tracking was not possible
the ROI) in main bolus (Fig. 3B). The result of this is that with with 4-row MSCT scanners.
conventional test bolus technique the scan is safe but can be too With scanners that have 6 or more rows, the scan time can
early, especially if the vessels of interest are located at the be reduced significantly, especially if there is a parallel reduc-
beginning of the scan range. tion in gantry rotation time (<500 ms). With those features,
Bolus Tracking scan time can be reduced below 30 s. Apnea becomes more
Bolus tracking technique is real-time bolus triggering tech- affordable for a larger number of patients without prescan
nique. It is based on an ROI that is plotted inside the lumen of hyperventilation or oxygen administration.
242 CADEMARTIRI AND NIEMAN
 CHAPTER 23 / CONTRAST MATERIAL INJECTION TECHNIQUES 243

Fig. 3. (left) Timing and synchronization techniques. (A) Fixed delay technique. The image shows the main pitfall of fixed delay technique.
The continuous curve shows the ideal situation when the fixed delay and bolus geometry correspond. In some cases the bolus geometry of the
patient can be different from the expected one. If actual bolus geometry is faster than fixed delay (dotted curve), the enhancement of vascular
structures will be still present but reduced, and prominent venous enhancement can be evident. If actual bolus geometry is slower than fixed
delay (tittled curve), the risk is to scan the passage of contrast material (CM) only in the distal portion of the scan range with poor contrast
enhancement of arteries. (B) Test bolus technique. The correlation between test bolus time to peak of maximum enhancement (tPME) and main
bolus tPME is displayed. In this case, the test bolus tPME correlates with the time to reach 150 HU in main bolus. A main scan based on this
information will be successful, even though not optimal as a CT angiography (CTA). In fact the vascular attenuation of 150 HU at the beginning
of the scan is too low for optimal CTA. Worse results are obtained when the correlation is with the time to reach 50 HU or 100 HU. Test bolus
actually has a different geometry than main bolus. The lack of injection power after the end of the injection of test bolus determines a pooling
of the test bolus in the venous system without any vis a tergo. In other words, test bolus is left alone in the venous system of the arm, without
the help of saline solution (bolus chaser) or adjunctive contrast media (main bolus) that pushes it forward. (C) Bolus tracking technique. The
sequence for real-time bolus tracking is displayed. After the topogram is acquired, the monitoring scan is set at the level of aortic root and the
region of interest (ROI) inside the lumen of ascending aorta. The trigger threshold is set at 100 HU. Then, CM administration and monitoring
sequence are started at the same time, and when the attenuation in the ROI reaches a value greater than 100 HU at the triggering point (TP),
a transition delay (TDgenerally 4 s are enough for this procedure) starts while the table reaches its starting position and the patient receives
breath-holding instructions.

Fig. 4. Examples of different contrast material administration protocols in 16-row multislice coronary CT angiography. Axial (A, B, and C)
and sagittal (D, E, and F) multiplanar reformations of three different protocols of CM administration are displayed: 100 mL of 320 mgI/mL
CM at 4 mL/s (A and D); 140 mL of 320 mgI/mL CM at 4 mL/s (B and E); 100 mL of 320 mgI/mL CM + 40 mL of saline chase at 4 mL/s (C
and F). The images from the protocol with 100 mL and the images with 100 mL + 40 mL are similar, as expected, and show a decreasing cranio-
caudal gradient of attenuation inside the pulmonary artery (D and F; black asterisks), that is not present in the protocol with 140 mL (E). The
enhancement in the descending aorta is preserved in every protocol (D, E, and F; black +), as well as the enhancement in the left ventricle (A,
B, and C; white +).

Bolus tracking is an optimal method for CM synchroniza- prevents suboptimal synchronization (determined by inter-
tion in noninvasive CTA (Fig. 4). Bolus tracking, in fact, is individual veno-arterial circulation time with fixed delay, and
less time consuming than test bolus techniques (because of the by the lack of reliable relationship between the tPME of test
calculation needed with test bolus), allows the use of less CM bolus and main bolus). No significant differences are observed
(the larger bolus needed to have a wider enhancement plateau regarding the attenuation reached at the level of aortic root,
with fixed delay, or the 20 mL needed for test bolus), and whether an injection volume of 100 mL or 140 mL is used.
244 CADEMARTIRI AND NIEMAN

Therefore, a volume of 100 mL, especially if followed by a 2. Platt JF, Reige KA, Ellis JH. Aortic enhancement during abdominal
saline chaser, provides optimal enhancement with reduced CT angiography: correlation with test injections, flow rates, and
patient demographics. AJR Am J Roentgenol 1999;172:5356.
volume and renal toxicity for the patient. A rate of injection of 3. Puskas Z, Schuierer G. [Determination of blood circulation time
4 mL/s is optimal because it is feasible in almost every patient for optimizing contrast medium administration in CT angiogra-
through a 18-gauge venflon (green), with a very low risk of phy] Kreislaufzeitbestimmung zur Optimierung der Kontrast-
CM extravasation. Iodine concentration is proportional to the mittelapplikation bei der CT-Angiographie. Radiologe 1996;36:
750757.
attenuation value obtained, and concentrations of 320350 4. Kirchner J, Kickuth R, Laufer U, Noack M, Liermann D. Optimized
mgI/mL are optimal for coronary angiography purposes. enhancement in helical CT: experiences with a real-time bolus track-
Bolus chaser is always recommended because it allows using ing system in 628 patients. Clin Radiol 2000;55:368373.
a reduced volume of CM more effectively. 5. van Hoe L, Marchal G, Baert AL, Gryspeerdt S, Mertens L. Deter-
mination of scan delay time in spiral CT-angiography: utility of a
GUIDELINES FOR CONTRAST test bolus injection. J Comput Assist Tomogr 1995;19:216220.
MATERIAL ADMINISTRATION 6. Bae KT, Heiken JP, Brink JA. Aortic and hepatic contrast medium
enhancement at CT. Part II. Effect of reduced cardiac output in a
CM administration technique for the purpose of nonin- porcine model. Radiology 1998;207:657662.
vasive CTA with multislice scanners can be summarized as 7. Garcia P, Genin G, Bret PM, Bonaldi VM, Reinhold C, Atri M.
Hepatic CT enhancement: effect of the rate and volume of contrast
follows: medium injection in an animal model. Abdom Imaging 1999;24:
Choice of CM: use nonionic iodinated CM. Higher iodine 597603.
concentrations will provide higher attenuation, and different 8. Han JK, Kim AY, Lee KY, et al. Factors influencing vascular and
molecules will not affect the final result significantly. The hepatic enhancement at CT: experimental study on injection pro-
iodine concentration ranges between 300 mgI/mL and 400 tocol using a canine model. J Comput Assist Tomogr 2000;24:
400406.
mgI/mL. 9. Nakayama M, Yamashita Y, Oyama Y, Ando M, Kadota M,
Site of administration: start with the right antecubital access Takahashi M. Hand exercise during contrast medium delivery at
and then use the left one if there are difficulties. Use forearm or thoracic helical CT: a simple method to minimize perivenous arti-
hand only if the antecubital accesses are not available. When fact. J Comput Assist Tomogr 2000;24:432436.
10. Coche EE, Muller NL, Kim KI, Wiggs BR, Mayo JR. Acute pulmo-
using central IV lines, reduce the rate of infusion slightly to nary embolism: ancillary findings at spiral CT. Radiology 1998;
2.53.0 mL/s. 207:753758.
Administration device: an 18-gage venflon (green) is suit- 11. Luboldt W, Straub J, Seemann M, Helmberger T, Reiser M. Effec-
able in all patients. In women or younger patients, the access tive contrast use in CT angiography and dual-phase hepatic CT per-
could be easier with a 20-gage venflon (pink), even though a formed with a subsecond scanner. Invest Radiol 1999;34:751760.
12. Bluemke DA, Fishman EK, Anderson JH. Effect of contrast concen-
reduced rate is recommended in this case (3.0 mL/s). tration on abdominal enhancement in the rabbit: spiral computed
Volume: with the 16-row generation scanners, 100 mL of tomography evaluation. Acad Radiol 1995;2:226231.
CM provides optimal enhancement if a bolus tracking tech- 13. Haage P, Schmitz-Rode T, Hubner D, Piroth W, Gunther RW.
nique is properly applied. With a lower number of detector Reduction of contrast material dose and artifacts by a saline flush
using a double power injector in helical CT of the thorax. AJR Am
rows, an increased volume is needed, down to 150 mL of CM
J Roentgenol 2000;174:10491053.
with 4-row generation scanners. 14. Hopper KD, Mosher TJ, Kasales CJ, TenHave TR, Tully DA,
Rate: a 4 mL/s rate provides optimal results. Reduced injec- Weaver JS. Thoracic spiral CT: delivery of contrast material pushed
tion rates (3.0 mL/s) are needed with smaller venflons and/or in with injectable saline solution in a power injector. Radiology 1997;
patients with small or fragile veins. 205:269271.
15. Bader TR, Prokesch RW, Grabenwoger F. Timing of the hepatic
Bolus chaser: the use of bolus chaser, when possible, is arterial phase during contrast-enhanced computed tomography of
strongly recommended to increase the amount of CM used the liver: assessment of normal values in 25 volunteers. Invest Radiol
during the scan and/or to reduce the overall amount of CM 2000;35:486492.
administered. 16. Sadick M, Lehmann KJ, Diehl SJ, Wild J, Georgi M. [Bolus tracking
and NaCl bolus in biphasic spiral CT of the abdomen] Bolustrig-
TIMING AND SYNCHRONIZATION TECHNIQUES
gerung und NaCl-Bolus bei der biphasischen Spiral-CT des Abdo-
Fixed delay technique: a delay of 18 s allows a good scan in mens. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 1997;
most of the patients. In patients with mildly or severely impaired 167:371376.
cardiac function, a delay of 25 s is recommended. 17. Fleischmann D, Rubin GD, Bankier AA, Hittmair K. Improved
Test bolus technique: calculate the tPME of the test bolus uniformity of aortic enhancement with customized contrast
medium injection protocols at CT angiography. Radiology 2000;
and then add 4 s for an optimal scan delay. 214:363371.
Bolus tracking technique: is safe and allows a tailored scan 18. Hittmair K, Fleischmann D. Accuracy of predicting and controlling
synchronization. The trigger threshold is set at 100 HU with a time-dependent aortic enhancement from a test bolus injection. J
transition delay of 4 s needed to give breath-hold instructions Comput Assist Tomogr 2001;25:287294.
19. Fleischmann D, Hittmair K. Mathematical analysis of arterial
to the patient. It requires proper training of the patient, and
enhancement and optimization of bolus geometry for CT angiogra-
of the technician. phy using the discrete Fourier transform. J Comput Assist Tomogr
1999;23:474484.
REFERENCES 20. Bae KT, Tran HQ, Heiken JP. Multiphasic injection method for
1. Cademartiri F, van der Lugt A, Luccichenti G, Pavone P, Krestin uniform prolonged vascular enhancement at CT angiography: phar-
GP. Parameters affecting bolus geometry in CTA: a review. macokinetic analysis and experimental porcine model. Radiology
J Comput Assist Tomogr 2002;26:596607. 2000;216:872880.
 CHAPTER 23 / CONTRAST MATERIAL INJECTION TECHNIQUES 245

21. Nakajima Y, Yoshimine T, Yoshida H, et al. Computerized tomo- 25. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral
graphy angiography of ruptured cerebral aneurysms: factors af- computed tomography angiography for determination of coronary
fecting time to maximum contrast concentration. J Neurosurg 1998; artery stenoses. Am J Cardiol 2001;88:11911194.
88: 663669. 26. Nieman K, Rensing BJ, van Geuns RJ, et al. Usefulness of multislice
22. Kaatee R, Van Leeuwen MS, De Lange EE, et al. Spiral CT angiog- computed tomography for detecting obstructive coronary artery dis-
raphy of the renal arteries: should a scan delay based on a test bolus ease. Am J Cardiol 2002;89:913918.
injection or a fixed scan delay be used to obtain maximum enhance- 27. Vogl TJ, Abolmaali ND, Diebold T, et al. Techniques for the detec-
ment of the vessels? J Comput Assist Tomogr 1998;22:541547. tion of coronary atherosclerosis: multi-detector row CT coronary
23. Nieman K, Oudkerk M, Rensig BJ, et al. Coronary angiography with angiography. Radiology 2002;223:212220.
multislice computed tomography. Lancet 2001;357:599603. 28. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama
24. Achenbach S, Ulzheimer S, Baum U, et al. Noninvasive coronary PM, de Feyter PJ. Reliable noninvasive coronary angiography with
angiography by retrospectively ECG-gated multislice spiral CT. Cir- fast submillimeter multislice spiral computed tomography. Circula-
culation 2000;102:28232828. tion 2002;106:20512054.
 CHAPTER 24 / VISUALIZATION TECHNIQUES FOR CTA 247

24 Visualization Techniques for Contrast-

Enhanced CT Angiography of Coronary Arteries

JEAN-LOUIS SABLAYROLLES, MD AND PASCAL GIAT, PhD

INTRODUCTION tricles; (2) double oblique coronal plane with multiplanar vol-
One of the critical components for effective cardiac com- ume reformation (MPVR) (45 mm) of the right coronary ar-
puted tomography (CT) application is a fully integrated and tery; and (3) double oblique axial plane with MPVR of left
optimized image visualization, postprocessing, and analysis common artery, left anterior descending (LAD) artery, and left
tool. Study of native axial images is not enough to assess all circumflex artery. Each phase is displayed successively and
cardiac structures, especially coronary arteries (Fig. 1). Post- the best phase is selected. Figure 3, taken from a study done
processing is critical to visualize the arterial system and heart on 87 exams, shows that the best phase, which depends on the
chambers. Specific tools have been designed and are still under patient, is usually between 75 and 80%.
development. They require a powerful postprocessing work- IDENTIFICATION OF RIGHT AND LEFT
station. A complete study of coronary arteries needs to follow CORONARY ARTERIES TOGETHER
a systematic protocol including various postprocessing tools. WITH COLLATERAL BRANCHES IN VOLUME DATA SETS
Locating the coronary arteries is based on a 3D image pro-
STUDY OF CORONARY ARTERIES cess fully dedicated to heart assessment. This 3D image pro-
SELECTION OF OPTIMUM PHASE WITH MULTIPHASE cessing enables mapping of the coronary arteries network with
REVIEW OF REFORMATTED IMAGES respect to other cardiac structures and mediastinal vessels. It
Acquisition synchronized with the electrocardiogram (ECG) allows the user to localize coronary artery lesions, and to diag-
makes it possible to eliminate motion blur caused by the heart- nose and map an occlusion or an aneurysm.
beats. After acquisition, the data are collected and processed For the coronary artery study, only the best phase is used.
along with the ECG recording. Slice thickness, increment, and For the same patient, cycle times may be different for the study
reconstruction window can be varied with respect to the dura- of the right coronary artery, the circumflex, and the left anterior
tion of the cardiac cycle. In general, diastole is centered at 80% descending artery. When misregistration occurs due to a vary-
of the cycle time when the coronaries and chambers show less ing heart rate, images are produced in different cardiac phases
motion and best coronary filling. For coronary artery examina- and combined using phase registration to improve image qual-
tion, it is sometimes necessary to reconstruct several stages of ity. In addition, the use of all phases together allows a morpho-
the cardiac cycle to obtain optimum image quality without logical study in diastole and systole of the left ventricle walls.
motion artifacts (Fig. 2). It is possible to reconstruct images Volume Rendering
of other phases in the cardiac cycle and to produce dynamic This 3D study is used to visualize and to identify the coro-
sequences not only with the basic axial images but also using nary arteries. Special software is required to automatically
reconstructed images. remove bone structures and some mediastinal structures, nota-
This is a first step in cardiac CT imaging that is crucial prior bly the pulmonary trunk and left atrium, by using the Autoselect
to any image postprocessing. In practice, data are reconstructed tool. This allows better visualization of the coronary ostia and
at different cardiac phases, typically every 10% of the RR cycle branches in any plane by erasing the left atrium and pulmonary
interval using 0.6-mm slice thickness every 0.6 mm. All slices trunk (Fig. 4).
of all phases are loaded on the workstation (Advantage Win- One can also rotate the 3D heart view by selecting a one-
dows, GE Medical Systems) simultaneously, using specific touch catheterization protocol that orientates the image to a
postprocessing tools (CardIQ Analysis). The selection of the typical cath lab view. Cine mode imaging of all predefined
best phase is performed on 2D reformatted images using three projections can be achieved on all axes. This 3D image pro-
planes: (1) short axis plane to check the regularity of both ven- cess provides detail on the localization of coronary arteries
with respect to other cardiac structures, especially in the case
of the anomalous origin of coronary arteries and coronary
From: Contemporary Cardiology: CT of the Heart: bypasses.
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

247
248 SABLAYROLLES AND GIAT

Fig. 1. Axial: Axial sectionsECG-synchronized volumetric acquisition with 0.6-mm sections.

3D Maximum Intensity Projection aorta), to give users a qualitative overview of coronary vessel
Three-dimensional maximum intensity projection (MIP) structures in MIP or volume rendering.
gives a MIP projection of coronary arteries with automatic STUDY OF BIFURCATIONS
fading of contrasted cardiac chambers. Images obtained can be WITHIN DIFFERENT PROJECTIONS
aligned in right anterior oblique view, left anterior oblique view, MPVR is an MIP with a thickness =5 mm. The different
spider view, or any view routinely used for conventional X-ray seg-ments of the coronary arteries can be differentiated. The
coronary angiography (Fig. 5). Three-dimensional MIP is used system makes use of the low contrast level of the fatty tissue sur-
to study the different branches of the right and left coronary rounding the vessels. Each coronary artery can be displayed, in
arteries in all projections and to isolate a given artery in order its entirety or segment by segment, using cine MPVR (Fig. 7).
to determine the potential area of lesions. This can be easily For example, this can be used to study arterial bifurcations, which
done when the stenosis is caused by soft plaque. However, this require special attention because the risk of stenosis or thrombo-
is more complex for stenosis caused by calcified plaque, which sis is even greater where the bifurcation can be divided into at
requires a 2D study to clearly separate the lumen from the cal- least two branches of the same size, and show an increasingly
cifications of the artery wall. sharp angle between the artery and its branches (Fig. 8).
Three-dimensional MIP that leads to real coronary artery STUDY OF VESSEL WALLS WITH CARDIQ ANALYSIS
mapping supports MPVR to 2D reformat. IN 2D, MPVR, OR VIRTUAL ENDOSCOPY
Both 3D techniques show advantages and disadvantages. Three-dimensional images allow the user to localize coro-
Three-dimensional MIP enables good visualization of the nary artery lesions but are insufficient to enable precise study
lumen and parietal wall calcifications. Three-dimensional vol- of lesions. Two-dimensional reformatting is necessary for the
ume rendering is more appropriate to specify the mapping of study and the quantification of stenosis, atheromatous plaques,
each coronary artery and its position relative to the surrounding and intra-stent lumen in both the short and long axis of the
structures (Fig. 6). vessel.
Coronary Vessel Tree MPVR With Autobone
A coronary vessel tree can be generated and displayed in a Based on 3D MIP reconstructed images with automatic fad-
semi-automatic process (from a single point in the ascending ing of contrasted cardiac cavities, all collaterals are analyzed
 CHAPTER 24 / VISUALIZATION TECHNIQUES FOR CTA 249

Fig. 3. Table and graph of cardiac phases. Optimal motion suppres-

sion typically occurs at 75% and 80% RR.

together with their distal branches without being hidden by the

other opacified structures (e.g., pulmonary artery and veins,
left atrium, and so on) (Fig. 9).
Curved 2D Reformatting
Curved 2D reformatting preserves the characteristics of the
basic axial slices; the lumen and the wall can be dissociated
(soft plaque, calcification, stent). This study can easily be made
using cardiac vessel analyzing (CVA), a dedicated postprocess-
ing procedure that automatically tracks, extracts, and measures
coronary artery vessels. This allows rapid and consistent imag-
ing and analysis of cardiac vessel pathologies. With single-
point deposit on each vessel, it automatically identifies and
tracks the vessel centerline of each branch. CVA automatically
produces vessel lumen view, MPVR, and 2D curved reforma-
tion (Fig. 10). It is then possible:
Fig. 2. Phase selection: Reconstruction of data in several phases. To analyze each branch in MPVR and 2D in different
Optimal phase = 80%. projections (Fig. 11).

Fig. 4. Volume-rendering technique (VRT) shows entire thorax (1). VRT after automatic suppression of thoracic structures (2). VRT after
suppression of pulmonary vessels and of left atrium by Autoselect (3).
250 SABLAYROLLES AND GIAT

Fig. 5. Maximum intensity projection with automatic suppression of opacified cavities. Routine projections in X-ray corony catheter:
frontal view (1), lateral view (2), spider view (3), left anterior oblique view with cranial angulation (4).

To assess and quantify a stenosis. CVA enables quanti- The fibrous-rich or atheromatous plaque is a well-defined,
tative analysis of the stenosis. The lumen view is an automatic homogenous plaque with intermediate density (50 HU)
unfolded view of the true lumen of the vessel with quantifica- (Fig. 15).
The eccentric combined plaque shows an internal compo-
tion of the mean diameter and the surface area at each point
nent that is hypodense together with a peripheral calcified
along the centerline (Fig. 12). component (Fig. 16).
Plaque Study With Curved 2D Reformatting Calcified plaque (density >130 HU) shows a density
Two-dimensional reformatting is the only reconstruction greater than the density of the lumen at the stenosis level
technique that is able to differentiate the vessel lumen from the with a blooming component that exaggerates its size. Cal-
atheromatous plaque that causes the stenosis, and to specify its cified plaque generates an irregular increase in the size of
nature. This approach could provide information for the iden- the artery owing to its external development.
The intimal hyperplasia of an intrastent restenosis is
tification of vulnerable plaque. At present, it is possible to
expressed by an intrastent hypodensity compared with the
define soft, combined, or calcified plaque and to characterize wall of the hyperdense stent, which causes a small irregu-
plaque composition and rupture-prone soft coronary lesions. lar lumen diameter (Fig. 17).
Based on mean CT attenuation, predominantly lipid-rich Stenosis Quantification
plaque can be differentiated from predominantly fibrous-rich
The measurements of diameters and surfaces allows the
plaque.
automatic quantification of the degree of stenosis. The accu-
The noncalcified lipid-rich plaque (mean density 20 HU)
racy of the contour detection allows the differentiation of the
shows a hypodense structure whose density value is greater
than the density of perivascular fat. At the present time, the calcified plaques from contrasted lumen. An objective estima-
current spatial and density resolution of a CT scanner does tion of the degree of stenosis can be achieved. Those measures
not allow a precise study of the plaques lipidic compo- depend on the quality of the acquisition, especially with respect
nent, ulceration, or dissection (Figs. 13 and 14). to the spatial resolution (Fig. 18).
 CHAPTER 24 / VISUALIZATION TECHNIQUES FOR CTA 251

Fig. 8. Bifurcations: areas at risk for stenosis.

Fig. 6. Maximum intensity projection (1) and volume rendering (VR) Fig. 9. Multiplanar volume reformation with Autobone from maxi-
(2) of coronary ostia. In VR, improved visualization of anomalous mum intensity projection reconstructions: Good visualization of main
origin of left coronary artery: ostium in right sinus, main trunk passes trunk, left circumflex (1), and left anterior descending (2) with excel-
ventral to the aorta. lent display of their distal branches.

Fig. 7. A cine multiplanar volume reformation centered on the left circumflex artery is used to visualize the ostia of each collateral.
252 SABLAYROLLES AND GIAT

Fig. 10. Coronary vessel analysis: automatic location (1) and identification (2) of the lumen of each left coronary artery branch.

Fig. 11. Coronary vessel analysis: study in multiplanar volume reformation (1) and in 2D (2) of left anterior descending and in 2D (3) of left
circumflex.

Navigator/Virtual Endoscopy adapted to their anatomy. For evaluation of the myocardium

The navigator or virtual endoscopy performs an endolu- wall, cardiac images may be reformatted in various orienta-
minal exploration of all arterial segments, particularly in the tions, such as short-axis and long-axis chamber views, using
case of anomalous origin of coronary arteries or bifurcations, integrated batch-processing tools. The reformatted images can
where stenosis or occlusions often occur (Fig. 19). be saved as separate image series for further processing, such
as cardiac cine review and functional analysis.
STUDY OF MYOCARDIUM As with other cross-sectional techniques such as echocardio-
WALL AND HEART CHAMBERS graphy or magnetic resonance imaging (MRI), the cardiac
MORPHOLOGIC IMAGING WITH SINGLE chambers can be studied according to the following views.
OR MULTIPHASE 2D REFORMATTING Compared to previous techniques mentioned above, 2D images
With the same data, volume acquisition allows excellent obtained by CT are reconstructed retrospectively using post-
assessment of the cardiac chambers, the myocardium, pericar- processing based on volume acquisition, and are not directly
dium, and heart valves on 2D reconstructions in different planes generated during the acquisition.
 CHAPTER 24 / VISUALIZATION TECHNIQUES FOR CTA 253

Fig. 12. Coronary vessel analysis: automatic measurements of lumen diameter and area.

Fig. 13. Coronary vessel analysis: maximum intensity projection and volume rendering: left anterior descending stenosis.
254 SABLAYROLLES AND GIAT

Fig. 14. Coronary vessel analysis: left anterior descending stenosis in 2D long axis and short axissoft plaque.

Fig. 15. Coronary vessel analysis: left circumflex stenosis on fibrous plaque in 2D long axis and short axis.

Fig. 16. Coronary vessel analysis: main trunk stenosis on mixed plaque in 2D long axis and short axis.
 CHAPTER 24 / VISUALIZATION TECHNIQUES FOR CTA 255

Fig. 17. Coronary vessel analysis: LAD stent restenosis in 2D LA and SA.

Two-chamber view in diastole and systole: the reference contour detection can be modified manually. Automatic
plane is perpendicular to the septum and runs through the detection of endocardium contours can be easily detected
middle of the mitral valve plane and the apex of the left automatically due to the difference in density between
ventricle. opacified cavities and myocardium. On the other hand,
Four-chamber view in diastole and systole: the reference automatic detection of epicardium needs more time and
plane of the long axis view runs through the middle of the should be performed manually.
mitral valve plane and the apex of the left ventricle. 2. Calculation of LV and RV end-diastolic and systolic vol-
Short-axis view in diastole and systole: cuts run perpendicu- ume, stroke volume, ejection fraction (EF), and myocar-
lar to the reference plane of the major axis view (Fig. 20). dial mass at end diastole. These 3D measurements are
It is possible to perform automatic measurement of volumes, currently being validated with comparison to conventional
techniques (Fig. 21).
ejection fraction, total or segmental myocardial mass, and per-
3. Generation of a volumetric 3D movie with the cardiac
fusion, with imaging of the surface of the lesion. contours (Fig. 21).
LEFT-VENTRICULAR WALL 4. Left-ventricular wall motion, thickness, and thickening are
MOTION AND FUNCTION ANALYSIS available in either a graphic or a bulls-eye format to aid in
The CardIQ function allows the user to display multiphase the detection of heart abnormalities (Fig. 22).
cardiac CT images and calculate heart function parameters. Practical experience and technological progress have made
CardIQ function measures, graphically plots, and tabulates a it possible to overcome the problems of acquisition and recon-
number of left- and right-ventricular (LV and RV) functions struction, with the aim of developing a simple method of non-
and wall motion parameters derived from reformatted short- invasive examination, on an out-patient basis, for patients
axis data sets. Function analysis comprises several steps: suffering from ischemic heart disease. This method would pro-
1. Automatic detection of endocardial and epicardial con- vide not only precise diagnostic possibilities but also informa-
tours of LV on short-axis reformated images. Automatic tion that is essential for the therapeutic decision.
256 SABLAYROLLES AND GIAT

Fig. 18. Coronary vessel analysis: quantification of a segment 1 left anterior descending (LAD) stenosissurfacic measurements: main trunk
= 13.6 mm, LAD stenosis = 2.5 mm, LAD downstream from the stenosis = 7.9 mm.

Fig. 19. Coronary vessel analysis: virtual endoscopy 1: main trunk stent for ostium stenosis. Near part of stent extends into aorta lumen 2:
anomalous origin of right and left coronary artery ostia in right sinus (arrows).
 CHAPTER 24 / VISUALIZATION TECHNIQUES FOR CTA 257

Fig. 20. Ventriculography: 2-chamber 2D slices and SA in systole and diastole.

Fig. 21. Function: function analysis and 3D movie.

258 SABLAYROLLES AND GIAT

Fig. 22. Function: left-ventricular wall motion, thickness, and thickening.

 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 259

25 CT Angiography for Assessment

of Coronary Artery Anomalies

STEFFEN C. FROEHNER, MD, MATTHIAS WAGNER, MD,

JUERGEN BRUNN, MD, AND RAINER R. SCHMITT, MD

INTRODUCTION tulae can be subject to such injuries as well, not to mention the
In Summer 2002, a 21-yr-old man was found dead one often hemodynamically significant shunt volume they show.
early morning in his apartment. He had been a healthy sports- Although sudden infant death syndrome (SIDS) is a cause
man, but sometimes he felt an atypical pressure in the left side for sudden infant death, other causes should be ruled out before
of his chest, independent of body stress. In all former medical diagnosing SIDS. Cardiac causes for sudden infant death
examinations (i.e., electrocardiogram [ECG], stress ECG, and include anomalies of the coronary arteries in many cases. The
echocardiography), particularly in the armed forces, no patho- BlandWhiteGarland syndrome has to be mentioned in the
logical diagnosis had been found. The gross specimen showed first place.
an anomaly of his coronary arteries: the right coronary artery Anomalies of the coronary arteries found in children may be
(RCA) had its origin out of the left sinus of Valsalva and associated with other congenital anomalies of the heart like
showed a so-called interarterial course, between aorta and Fallots syndrome, transposition of the great arteries, Taussig
pulmonary artery. Bing heart (double outlet right ventricle), or common arterial
This type of coronary artery anomaly is considered to run a trunk (6).
higher risk of myocardial ischemia or sudden death, especially Beside the hemodynamic effects, anomalies of coronary
in younger patients under the age of 35 yr (1,2). arteries are thought to be associated with a higher risk of arte-
Such coronary anomalies are identified only incidentally riosclerosis, particularly in vessels with aberrant courses (7).
during life, often because of insufficient clinical suspicion. In catheter angiography as a 2D imaging method, the
However, because abnormal coronary artery origin is now depiction of coronary artery anomalies may be very difficult.
amenable to surgical treatment, early clinical identification is The accurate course of these anomalous vessels often is not
crucial. With regard to congenital coronary artery anomalies in clearly displayed. In many cases a selective angiography of the
young competitive athletes, standard testing with ECG under aberrant vessels is impossible.
resting or exercise conditions is rather unlikely to provide clini- As a noninvasive method which has proved its higher sen-
cal evidence of myocardial ischemia, and would not be reliable sitivity, multislice computed tomography (MSCT) of the coro-
as a screening test in large athletic populations. Premonitory nary arteries, with its 3D approach, may well become the
cardiac symptoms often occur shortly before sudden death. method of choice for the detection and characterization of coro-
Therefore, it appears wise to exclude such an anomaly after a nary artery anomalies and related syndromes.
history of exertional syncope or chest pain. These observations This chapter describes the value of MSCT in detection of
also have important implications for the preparticipation anomalies of the coronary arteries; it also characterizes the
screening of competitive athletes (3). different types of coronary artery anomalies, including their
Under certain conditions, anomalies of the coronary arteries clinical relevance. Finally, diagnostic aspects and capabilities
may lead to severe complications in valve and bypass surgery, of MSCT in this clinical setting are discussed.
by accidental injury of the aberrant vessels during an operation Congenital anomalies of the coronary arteries are a very
(4). There are also reports describing a compression of an heterogeneous group of lesions, which can produce severe clini-
aberrant left circumflex coronary artery (LCx) following cal symptoms or be clinically silent, leading only incidentally
bioprosthetic valve ring implantation (5). Coronary artery fis- to a correct diagnosis. There may be malignant or nonma-
lignant forms, depending on their course. For that reason it is
of high prognostic and sometimes therapeutic importance to
define the correct origin and course of the aberrant artery and
to describe its course accurately.
From: Contemporary Cardiology: CT of the Heart:
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

259
260 FROEHNER ET AL.

EXAMINATION AND RECONSTRUCTION lies of the coronary arteries do not present any of them. The
TECHNIQUES FOR ASSESSMENT OF CORONARY etiology of coronary anomalies in primarily healthy persons is
ARTERY ANOMALIES WITH MSCT still not known and requires further embryological and genetic
research.
MSCT, in particular the now available 16- and 64-row scan-
ners, with their ability to generate approximately isotropic Clinical Importance of Coronary Artery Anomalies
voxels, represent a new and highly promising tool to visualize Anomalies of the course of coronary arteries can be catego-
coronary arteries with respect to their origin and course. Spe- rized intomalignant and nonmalignant forms. Malignant
cifically, the manifold reconstruction potentialsmultiplanar forms are combined with an increased risk of myocardial
reformations (MPR), multiplanar volume reformations ischemia or sudden death (1,16,17).
(MPVR), surface shading (SSD), vessel tracking, and volume These forms mostly show a course between pulmonary
rendering (VR)enable accurate visualization of coronary artery and aorta (interarterial). The most common case is an
arteries. Using these 3D complementary capabilities, the origin of the right coronary artery from the left sinus of
detection and characterization of anomalies of the coronary Valsalva, coursing between aorta and pulmonary artery. But
artery system is feasible with high precision and detail. anomalies of the left main artery (LM) or the left anterior
Examination protocols to depict anomalies of the coronary descending artery (LAD) arising from the right sinus of
arteries with MSCT do not differ from those presented in pre- Valsalva with this course are also associated with higher risk.
vious chapters to evaluate coronary artery disease. In special It has been suggested that myocardial ischemia and sudden
cases, such as anomalies with ectopic origin of coronary arter- death result from transient occlusion of the aberrant coronary
ies (see subheading Anomalies of Ectopic Origin and Course artery, caused by an increase of blood flow through the aorta
of Coronary Arteries), it is important to use a wider scan field and pulmonary artery during exercise or body stress. The rea-
to cover the whole course of these vessels. MPRs and 3D vol- son is either a kink at the sharp leftward or rightward bend at the
ume rendering may be very helpful to demonstrate these anoma- vessels ostium or a pinchcock mechanism between aorta and
lies and to provide an accurate depiction of the special vascular pulmonary artery. Up to 30% of these patients are at risk for
anatomy. sudden death (18). Long courses of anomalous coronary arter-
The disadvantage of this method is the time-consuming ies ventral to the pulmonary artery may sometimes be associ-
reconstruction of the data sets by an experienced physician. In ated with a higher risk of myocardial ischemia. Dilatation of the
particular, complex vessel anomalies have to be reconstructed pulmonary trunkfor example, as a result of pulmonary
accurately with full use of software and anatomic knowledge. hypertensioncan lead to a stretching of the aberrant vessel,
which causes lumen reduction and consequent myocardial
SYNOPSIS OF CORONARY ARTERY ANOMALIES ischemia.
CORONARY ARTERY ANOMALIES OF ORIGIN An origin of either the left or right coronary artery from the
AND COURSE pulmonary artery (BlandWhiteGarland syndrome for the left
Prevalence and Etiology coronary artery) also must be considered as malignant. This
Coronary artery anomalies constitute 13% of all congenital anomaly is frequently associated with myocardial ischemia and
malformations of the heart, but in approx 0.46 to 1% (8) of the sudden death in early childhood. A surgical intervention is
normal population, anomalies of the coronary arteries are absolutely required.
reported to be found incidentally during catheter angiography, Other courses do not lead to any clinical symptoms. They
echocardiography, or autopsy (9,10). are rather incidental findings in catheter angiography, CT of
The etiology of coronary artery anomalies is still uncertain. the coronary arteries, or autopsy, without any clinical rel-
A possibility of maternal transmission of some varieties is sug- evance. In case of planned cardiac surgery, these anomalies
gested, particularly for a single coronary artery (11). Familial nevertheless take on an increasing importance, because the
clustering is also reported for anomalies of an LCx originating risk of accidental surgical injuries, in particular because con-
from the right sinus of Valsalva, one of the most common genital anomalies of the heart (e.g., Fallots syndrome, trans-
anomalies (12). position of the great arteries, TaussigBing heart, or common
Anomalies of the coronary arteries may also be combined arterial trunk) are often combined with anomalies of the coro-
with Klinefelters syndrome, although both entities are very nary arteries (19).
rare (13). In trisomy 18 (i.e., Edwards syndrome), coronary Therefore, every anomaly of origin and course of the coro-
artery anomalies rarely can be found as well (14). nary arteries should be described accurately prior to cardiac
A raised plasma phenylalanine as in phenylketonuria of surgery. If the catheter angiography cannot provide detailed
pregnant women may induce several congenital pathologi- information (20) (which it usually does not), MSCT has to be
cal changes of the heart, including coronary artery anoma- performed to describe the exact courses of aberrant vessels.
lies, sometimes in combination with mental and physical Finally, aberrant coronary arteries seem to run a higher risk
retardation. Maternal phenylketonuria may be responsible of arteriosclerosis. In patients with anomalies of the coronary
for these malformations, which may be prevented by dietary arteries, the aberrant vessels where found to be more diseased
treatment (15). than the vessels with normal course (21). So it is necessary to
Although all these syndromes and predispositions may lead pay attention to these vessels in case of atypical chest pain or
to coronary artery anomalies, most of the patients with anoma- pathological ECG findings.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 261

Classification of Coronary Artery Anomalies Clinical relevance: these patients suffer from myocardial
of Origin and Course ischemia or sudden death in early childhood. A surgical inter-
The most frequent and clinically significant anomalies are vention is recommended as soon as possible.
detailed in Table 1. In the following subheadings, the different Origin of the RCA From the LCx
types of coronary artery anomalies are presented and the clini- Anatomy: a very rare but often described anomaly of the
cal relevance is discussed. right coronary artery is the agenesis of the right coronary ostium
Origin of the RCA From the Left Sinus of Valsalva with a continuation of the LCx as a right coronary artery and a
Anatomy: this anomaly of the RCA may present different retrograde perfusion of the RCAs supply area (23).
appearances. There may be a common ostium or two different Clinical relevance: this type of anomaly may lead to early
ostia. The right coronary artery can take three different courses. and severe myocardial ischemia resulting from already moder-
First, ventral to the pulmonary artery; second, between pulmo- ate stenosis of the LCx.
nary artery and the aorta (interarterial); and third, dorsal to the Anomalies of Origin and Course
aorta. These aberrant vessels lead to the right atrioventricular of the Left Main Coronary Artery
groove, where the RCA takes its normal course. Fig. 1AD SEPARATE OSTIA OR COMMON OSTIUM OF LAD AND LCX
demonstrates an origin of the right coronary artery from the left Anatomy: the most common anomaly of the left coronary
sinus of Valsalva and an interarterial course. The decrease of vessel system is the absence of the left main coronary artery.
the lumen in the proximal part and the sharp kink are clearly LAD and LCx have separate ostia in the left coronary cusp
depicted. without an anomaly of the course (Fig. 3AB). LAD and LCx
Clinical relevance: according to the literature (22), the most may also originate from one common ostium (Fig. 4AB).
common type of anomalous coronary arteries in our data was an Clinical relevance: no clinical relevance, but selective con-
origin of the right coronary artery from the left sinus of Val- trast injection during catheter angiography may be difficult,
salva or the left main artery. Approximately 34% of the anoma- causing insufficient angiographic.
lies detected showed this particular appearance. ORIGIN OF THE LEFT MAIN CORONARY ARTERY
An interarterial course has to be considered malignant as a FROM THE RIGHT SINUS OF VALSALVA
result of the possible pinchcock mechanism and the sharp right- Anatomy: the most common and important anomaly of ori-
ward kink of the vessel directly distal to the ostium. A higher gin of the left main coronary artery is an origin from the right
risk of myocardial ischemia and sudden death is likely. sinus of Valsalva. There may be a common ostium or an origin
Origin of the RCA From Branches from the right coronary artery. Three different courses are pos-
of the Left Coronary Artery sible: a course of the left main coronary artery between aorta
Anatomy: rarely, the right coronary artery can have its origin and pulmonary artery (Fig. 5), a course ventral to the pulmo-
from branches of the left coronary artery system, particularly nary artery (Fig. 6AE), and a course dorsal to the aorta to the
from the LAD with a course ventral to the pulmonary artery, or left side.
from the LCx with a course dorsal to the aorta. Clinical relevance: depending on their course, these anoma-
Clinical relevance: no clinical relevance; these anomalies lies can be malignant or nonmalignant. The first one has an
are only incidental findings. increased risk of myocardial ischemia or sudden death caused
Origin of the Right Coronary Artery by a sharp kink at the ostium and a possible compression of the
Fromthe Posterior Sinus of Valsalva vessel between aorta and pulmonary artery in case of higher
Anatomy: a very rare anomaly of the coronary arteries is an blood flow under exercise. The detailed description can be
origin of the right coronary artery from the posterior sinus of found under the subheading Clinical Importance of Coronary
Valsalva. This anomaly is detected only incidentally by cath- Artery Anomalies.
eter angiography or autopsy. ORIGIN OF THE LEFT MAIN CORONARY ARTERY
Clinical relevance: no clinical relevance. FROM THE RIGHT SINUS OF VALSALVA AND DIFFERENT COURSES
Separate Origins of Right Coronary OF LEFT ANTERIOR DESCENDING AND LEFT CIRCUMFLEX
Artery and Conus Artery Anatomy: another possibility of a left main arterys anomaly
Anatomy: the conus artery may arise from a separate ostium with an origin from the right sinus of Valsalva is an early divi-
in the right sinus of Valsalva, either as a small conus artery sion in LAD and LCx. The LAD then shows a course ventral to
providing only the conus area (Fig. 2A,B), or as a large conus the pulmonary artery, the LCx an interarterial course between
artery supplying a wide area of the genuine RCA area with a aorta and pulmonary trunk.
consecutive hypoplastic RCA. Clinical relevance: the interarterial course of the LCx is also
Clinical relevance: this aberrant conus artery is particularly combined with a higher risk of myocardial ischemia.
at risk to be injured by ventriculotomy or other maneuvers in TRANSSEPTAL COURSE OF THE ABERRANT LEFT MAIN ARTERY
heart surgery. Anatomy: a very rare anomaly of the left main coronary
Origin of the Right Coronary Artery artery is a transseptal course to the left side after originating
From the Pulmonary Trunk from the right sinus of Valsalva.
Anatomy: analogous to the BlandWhiteGarland syndrome Clinical relevance: the clinical relevance of these anomalies
with an origin of the left coronary artery from the pulmonary is still not known. In case of myocardial hypertrophy or volume
trunk, the right coronary artery also can originate from this. We overload of the ventricles, there may be a higher risk of myo-
may call it a reversed BlandWhiteGarland syndrome. cardial ischemia.
262 FROEHNER ET AL.

Table 1
Most Frequent and Clinically Significant Anomalies of the Coronary Arteries
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 263

Table 1 (Continued)
Most Frequent and Clinically Significant Anomalies of the Coronary Arteries

1: Normal anatomy of the coronary arteries.

2: Origin of the right coronary artery (RCA) from the left main artery, course dorsal to the aorta.
3: Origin of the right coronary artery from the left sinus of Valsalva, course between aorta
and pulmonary artery (interarterial course).
4: Origin of the right coronary artery from the left sinus of Valsalva, course ventral to the
pulmonary artery.
5: Origin of the right coronary artery from the pulmonary artery ("reversed" Bland-White-
Garland syndrome).
6: Dominant conus artery (CA) with a separate ostium, which supplies the main area of the
RCA. The RCA is a small vessel, with early division. A reverse case is also possible:
separate ostia of conus artery and RCA, but the conus artery is the smaller vessel.
7: Origin of the left main artery from the right coronary artery, course dorsal to the aorta.
8: Origin of the left main artery from the right coronary artery, course between aorta and
pulmonary artery (interarterial course).
9: Origin of the left main artery from the right sinus of Valsalva, course ventral to the
pulmonary artery.
10: Origin of the left main artery from the right sinus of Valsalva, the left descending coronary artery
(LAD) courses ventral to the aorta, the left circumflex shows a course interarterial between aorta and
pulmonary artery.
11: Origin of the left main artery from the pulmonary artery (Bland-White-Garland syndrome).
12: Origin of the LAD from the RCA, course between aorta and pulmonary artery
(interarterial course).
13: Origin of the LAD from the right sinus of Valsalva, course ventral to the pulmonary artery.
14: Origin of the circumflex branch from the right coronary artery, course dorsal to the aorta.
15: Origin of the circumflex branch from the first diagonal branch of the LAD.

Anomalies of origin of the left main coronary artery in the mortality from heart failure. The adult type develops myocar-
right sinus of Valsalva are frequently found in patients with dial infarction, arrhythmias, sudden cardiac death, or signs of
other congenital anomalies of the heart, particularly in cases of congestive heart failure. These patients can generally be pre-
Fallot's syndrome. vented from myocardial ischemia or heart failure only by a
ORIGIN OF THE LEFT MAIN CORONARY ARTERY surgical intervention (24). The risk of fatality from a congeni-
FROM THE PULMONARY ARTERY tal coronary abnormality far outweighs the smaller risk of a
(BLAND-WHITE-GARLAND SYNDROME) surgical intervention. Nevertheless, cases of a Bland-White-
Anatomy: one of the most important anomalies of coronary Garland syndrome have been reported in which the patient
arteries in children is an origin of the left coronary artery from survived to middle age with minimal cardiovascular prob-
the pulmonary trunk, known as Bland-White-Garland syn- lems. Good exercise tolerance and long-term survival may
drome. occasionally be possible even without surgery for patients
Clnical relevance: clinical symptoms result from myocar- with this anomaly (25). Some patients show only ECG alter-
dial ischemia caused by an arterio-venous shunt owing to ations, without any clinical symptoms such as angina, atypi-
reversed flow in the left main coronary artery to the pulmo- cal chest pain, or abnormal shortness of breath during or after
nary artery. The childhood type of this anomaly presents high exercise (26).
264 FROEHNER ET AL.

Fig. 1. Origin of the right coronary artery (RCA) from the left sinus
of Valsalva, showing a course between aorta and pulmonary artery.
The proximal part of the RCA shows a decrease of the lumen and a
sharp kink directly after the ostium. (A) Thin maximum intensity
projection reconstruction. (B) Volume rendering. (C) Vessel
view reconstruction, showing the vessel stretched longitudinally.
The decrease of the vessels lumen is also demonstrated in this
reconstruction method. PA, pulmonary artery; AO, aorta. (D) Vir-
tual angioscopy of the ostium of the RCA. View from aortic root into
the ostium of the RCA.

Anomalies of Origin and Course Clinical relevance: the latter course is combined with a
of the Left Anterior Descending Artery higher risk of myocardial ischemia.
The LAD can show similar anomalies of origin and course ORIGIN OF THE LEFT ANTERIOR DESCENDING
like the left main artery. FROM THE LEFT CIRCUMFLEX
Anatomy: a rare coronary artery anomaly is an origin of the
ORIGIN OF THE LEFT ANTERIOR DESCENDING ARTERY LAD from the LCx and a course leading to the anterior inter-
FROM THE RIGHT SINUS OF VALSALVA ventricular groove.
Anatomy: the LAD may originate from the right sinus of Clinical relevance: clinical symptoms are not known.
Valsalva either with a common ostium together with the RCA ORIGIN OF THE LEFT ANTERIOR DESCENDING
or separate ostia, or from the RCA itself. Two different courses FROM THE PULMONARY ARTERY
are possible: ventral to the pulmonary artery or between aorta Anatomy: similar to the BlandWhiteGarland syndrome,
and pulmonary trunk. the LAD can arise out of the pulmonary trunk.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 265

Fig. 2. Aberrant conus artery with an origin from the right sinus of Valsalva. (A) Thin maximum intensity projection reconstruction. (B)
Volume-rendering.

Fig. 3. Separate ostia of left anterior descending and left circumflex arteries in the left sinus of Valsalva. (A) Thin maximum intensity
projection reconstruction. (B) Volume-rendering.

Clinical relevance: the symptoms and clinical relevance for and the RCA or right sinus of Valsalva. The branch originating
this particular anomaly are similar to those mentioned for the from the RCA or the right aortic cusp may course interarterial
BlandWhiteGarland syndrome. or ventral to the pulmonary artery.
DOUBLE LEFT ANTERIOR DESCENDING CORONARY ARTERY Clinical relevance: this anomaly may be combined with an
Anatomy: a very rare anomaly is a double left anterior de- isolated transposition of the great arteries (27). An interarterial
scending coronary artery arising either from the left main artery course is accompanied by a higher risk of myocardial ischemia.
266 FROEHNER ET AL.

Fig. 4. Common ostium of left anterior descending and left circumflex arteries out of the right sinus of Valsalva. (A) Thin maximum intensity
projection reconstruction. (B) Volume rendering.

Fig. 5. Origin of the left main coronary artery out of the right sinus of Valsalva, course between aorta and pulmonary arterythin maximum
intensity projection reconstruction. Note the sharp kink of the left main coronary artery (arrow).

Fig. 6. (right) Origin of the left main artery out of the right sinus of Valsalva with a common ostium together with the right coronary artery
(RCA), course ventral of the pulmonary artery. (A) Thin maximum intensity projection (MIP) reconstruction. (B) Thin MIP reconstruction
of the long left main coronary artery. A first diagonal branch leaves the left main coronary artery (LMA) before left anterior descending (LAD)
and left circumflex (LCx) arise. (C) The volume rendering shows the common ostium and the early division. (D) The long course of the left
main coronary artery ventral of the pulmonary artery and the late division into LAD and LCx. Note the origin of the first diagonal branch before
the division into LAD and LCx. (E) Virtual angioscopy of the common ostium of left main coronary artery and RCA. The small ostium of the
sinus node artery (SNA) out of the proximal RCA is depicted clearly.
CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 267
268 FROEHNER ET AL.

Fig. 7. Origin of the left circumflex out of the right sinus of Valsalva, course dorsal of the aorta. (A) Thin maximum intensity projection
reconstruction. (B) Volume rendering.

TRANSSEPTAL COURSE OF THE ABERRANT ORIGIN OF THE LEFT CIRCUMFLEX CORONARY ARTERY
LEFT ANTERIOR DESCENDING FROM BRANCHES OF THE LEFT ANTERIOR DESCENDING
Anatomy: a transseptal course of the LAD originating from Anatomy: another appearance of anomalies of the LCx is an
the right sinus of Valsalva is also possible but extremely rare origin out of the first diagonal branch of the LAD or others of
(28). its branches leading likewise to the left atrioventricular groove.
Clinical relevance: as described under subheading Trans- Clinical relevance: a clinical relevance is not given; these
septal Course of the Aberrating Left Main Artery, such courses anomalies are only detected incidentally.
of coronary arteries are believed to be malignant. ORIGIN OF THE LEFT CIRCUMFLEX CORONARY ARTERY
Anomalies of Origin and Course FROM THE PULMONARY ARTERY
of the Left Circumflex Coronary Artery Anatomy: a rare anomaly of the LCx is an origin from the
In general, anomalies of the origin of the LCx are gener- pulmonary trunk or the right pulmonary artery. There may be
ally considered nonmalignant. However, a few reports of a large LCx arising from the right or left pulmonary artery or the
myocardial ischemia exist also in these anomalies (29). These pulmonary trunk which is filled in a retrograde way from a
anomalies are mostly identified incidentally in catheter dominant LAD (32).
angiography or autopsy, without causing any clinical symp- Clinical relevance: this type of anomaly leads to myocardial
toms. ischemia in the same manner as the BlandWhiteGarland
A comparison with Thallium-201 myocardial imaging in syndrome and is reported occasionally (33). A surgical correc-
patients with angina pectoris and anomalous origin of the tion is always required in early childhood to prevent myocar-
LCx concluded that an anomalous origin of the circumflex dial ischemia and consequent heart failure.
coronary artery did not cause impairment of myocardial per- ORIGIN OF THE LEFT CIRCUMFLEX CORONARY ARTERY
fusion unless it is the site of significant coronary arterial steno- FROM THE DISTAL RIGHT CORONARY ARTERY
sis (30). Anatomy: another possibility of anomalies of the LCx is an
ORIGIN OF THE LEFT CIRCUMFLEX CORONARY ARTERY origin of this vessel as a terminal extension of the RCA, leading
FROM THE RIGHT SINUS OF VALSALVA to the left atrioventricular sulcus (34).
Anatomy: the most common anomaly of the LCx is an ori- Clinical relevance: this type of anomaly may lead to early
gin from the right sinus of Valsalva or the RCA and a course and severe myocardial ischemia due to preexisting moderate
dorsal to the aorta leading to the left atrioventricular groove stenosis of the RCA.
(Fig. 7AB). Complex and Combined Anomalies of Origin
Clinical relevance: this anomaly is found in 0.7% of angiog- and Course of the Coronary Arteries
raphies. It is considered to be more frequent in patients with Among the numerous variations of anomalies of the coro-
congenital stenosis of the aortic valve (31). There is no higher nary arteries, combined and complex appearances also can be
risk of myocardial ischemia or sudden death. detected.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 269

Fig. 8. An aberrant vessel originating from the right coronary artery runs ventral to the pulmonary artery to the left side and feeds the middle
and distal part of the left anterior myocardium. The left anterior descending artery turns into a very thin vessel after 3 cm. (A) Thin maximum
intensity projection reconstruction. (B) Volume rendering.

ORIGIN OF ALL CORONARY ARTERIES Anomalies of Origin and Course

FROM ONE SINUS OF VALSALVA of Individual Branches of the Coronary Arteries
Anatomy: all of the three main coronary arteries can arise Variations of the normal anatomy as a ramus intermedius
either out of the right, left, or posterior sinus of Valsalva, pre- are often seen in catheter angiography, MSCT, or autopsy. The
senting common or separated ostia with all the courses delin- branches of the coronary artery system can present multiple
eated above. Combinations of the different anomalies described nonpathological variations. Because of its 2D projectional dis-
before are also possible. There may be several appearances play, catheter angiography may sometimes be unable to depict
possible which have not been reported yet. the accurate course of the variations. Therefore, some of these
Clinical relevance: the clinical relevance of these anomalies aberrations may mimic coronary artery fistulae in catheter
depends on the course of the vessel. Interarterial courses are angiography.
considered to be malignant, with a higher risk of myocardial ABERRANT SINUS NODE ARTERY
ischemia. Anatomy: A frequent anomaly is an aberrant sinus node
INTRACAVITARY COURSE OF THE RIGHT CORONARY ARTERY artery, originating from the LCx, finding its way dorsal to the aorta
and leading to the right side. This anomaly is one of the most
Anatomy: another complex course of coronary arteries con-
common (Fig. 9AB).
cerns the RCA. It can take an intracavitary course in the right
Clinical relevance: this particular anomaly of the coronary
atrium after a primarily epicardial course. Also, muscular
arteries has no clinical relevance and has therefore to be con-
branches of the RCA are rarely found to run subendocardially
sidered nonmalignant. In catheter angiography, this anomaly is
into the right atrium (35).
considered a normal finding, but it can mimic a coronary artery
Clinical relevance: the clinical relevance of this particular
fistula. Another possible anomaly of branches of the coronary
anomaly is still not known; further investigation is needed. A arteries is an aberrant conus artery.
higher pressure in the right atrium, for instance due to Anomalies of Ectopic Origin and Course
pulmanary embolism or tricuspid valve insufficiency, may lead of Coronary Arteries
to compression of the coronary artery. These extremely rare anomalies of the coronary arteries
ABERRANT ARTERY ORIGINATING FROM THE RIGHT CORONARY are mostly of no hemodynamic relevance. However, they are
ARTERY SUPPLYING THE LEFT ANTERIOR MYOCARDIUM important for the angiographer and the surgeonthe
Anatomy: a special case is shown in Fig. 8ABthe LAD angiographer may have difficulty visualizing the coronary
turns into a very thin vessel after the first 3 cm. In that case, an artery course by selective coronary angiography or aorto-
aberrant vessel originating from the RCA passes ventral to the graphy; the surgeon may injure an ectopic vessel during an
pulmonary artery to the left side and feeds the middle and distal operation.
parts of the anterior left myocardium. ECTOPIC ORIGIN OF CORONARY ARTERIES
Clinical relevance: in case of pulmonary hypertension, the ABOVE THE AORTIC CUSP
dilatation of the pulmonary artery may lead to a stretching of Anatomy: most common is an origin of coronary arteries
this aberrant vessel and subsequent myocardial ischemia. from the ascending aorta above the particular aortic cusp.
270 FROEHNER ET AL.

Fig. 9. Aberrant sinus node artery originating from the left circumflex with a course dorsal to the aorta. (A) Thin maximum intensity
projection reconstruction. The dorsal pathway of the aberrant vessel is seen clearly. (B) Multiplanar reformation.

Fig. 10. (above and on next page)Aberrant right coronary artery (RCA) with a ventral origin from the aorta above the right sinus of Valsalva
and an inter-arterial course between aorta and pulmonary artery. (A,B) Thin maximum intensity projection reconstruction. The sharp kink
of the aberrant vessel and the decrease of the lumen near the origin are clearly visible . The RCA originates leftwards above the right sinus of
Valsalva. (C,D) Volume rendering. The origin of the RCA leftwards above the right sinus of Valsalva is accurately shown. (E) Virtual
angioscopy of the ostia of left main coronary artery (LMA) out of the left aortic cusp (LC) and the aberrant origin of the RCA out of the ascendant
aorta above the right aortic cusp (RC).

Depending on the origin and course, these anomalies may Clinical relevance: as a result of the partial interarterial
produce clinical symptoms as well. Fig. 10AE shows an course this anomaly has to be considered malignant. Without
origin of the RCA out of the ascending aorta slightly above any evidence of coronary heart disease in CT angiography
the right aortic cusp with a sharp kink to the right and a partial (CTA), this particular patient had atypical angina symptoms
course between aorta and pulmonary artery. with exercise.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 271

Aberrant coronary arteries with an origin above the particu-

lar aortic cusp may also present a kink at the sharp downward
bend, which implies a higher risk of myocardial ischemia.
ECTOPIC ORIGIN OF CORONARY ARTERIES
ABOVE THE AORTIC CUSP IN COMBINATION
WITH A MALROTATION OF THE AORTIC ROOT
Anatomy: another complex anomaly is shown in Fig. 11AE.
The aortic root is rotated 90 clockwise. The RCA originating
from the right sinus of Valsalva arises ventrally and follows a
short pathway between aorta and pulmonary artery. The left
main coronary artery originates above the left sinus of Valsalva
with a sharp leftward kink.
Clinical relevance: this anomaly is thought to be malignant.
SINGLE CORONARY ARTERY
Anatomy: a single coronary artery may arise ectopically from
the innominate or the right carotid artery. Ectopic origin of a
coronary artery from the internal mammary artery, the left sub-
clavian artery, or from bronchial branches is extremely rare.
Clinical relevance: the clinical relevance of these extremely spite of an early therapy consisting of intravenous gamma
rare anomalies depends on their course. Their significance is globulin and aspirin, coronary involvement may develop dur-
not known. In such cases of long-distance aberrations, an addi- ing the first years after diagnosis, even if an early regression of
tional scan with a larger scan field is necessary in MSCT to the main underlying disease is found (36).
depict the whole course of the special vascular anatomy. Anatomy and Pathophysiology of Coronary Artery Fistulae
CORONARY ARTERY FISTULAE Coronary artery fistula is a rare condition in which a com-
Prevalence and Etiology munication exists between one or two coronary arteries and a
Coronary fistula are repeatedly described in case reports. cardiac chamber or a systemic vein. It causes a shunt from the
The general prevalence is still not known. In our study of more high-pressure coronary artery to a lower-pressure cardiac
than 1100 patients, two coronary artery fistula where found. chamber or vein in 90% of cases (37). Coronary artery fistulae
Kawasaki disease (see Chapter 26) as an acute vasculitis of resemble other aortic runoff lesions in which blood leaves the
unknown cause often affects children, and in 2025% of cases aorta through the fistula during diastole, in this instance enter-
it can cause coronary artery anomalies like coronary fistulae. In ing one of the cardiac chambers. This causes an obligatory
272 FROEHNER ET AL.

Fig. 11. Ninety-degree clockwise rotation of the aortic root. The right
coronary artery (RCA) arises from the right sinus (RS) of Valsalva;
the left main coronary artery arises above the left sinus (LS) of Val-
salva. Owing to the clockwise rotation of the aortic root, the RCA has
a short pathway between aorta and pulmonary artery. The left main
artery primarily leads toward dorsal, then undergoes a sharp leftward
kink. (A) Thin maximum intensity projection (MIP) reconstruction.
(B) Volume rendering: the 3D reconstruction clearly shows the origin
of the left main coronary artery above the left sinus of Valsalva and
the interarterial pathway of the RCA. (C,D) Thin MIP reconstruc-
tion and volume rendering. Clear depiction of the ventral origin of the
RCA and the dorsal origin of the left coronary artery resulting from
the clockwise rotation of the aortic root. (E) Thin MIP reconstruc-
tion of the aortic root with aortic valve showing the 90 clockwise
rotation.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 273

shunt, because it connects the high-pressure aorta with the low- coronary intratrabecular arterial network. A systolic coronary
pressure cardiac chamber. When the shunt leads into a right- artery nar-rowing is observed in 0.512% of patients under-
sided cardiac chamber, the hemodynamics resemble those of an going coronary arteriography. Although this entity is mostly
extracardiac left-to-right shunt. Shunt flow occurs during sys- benign, cases of acute myocardial ischemia, cardiogenic
tole and diastole except with fistulae to the left ventricle, which shock, and sudden death are reported (47,48). Myocardial
have largely diastolic flow. When the connection is to a left- bridges are a common finding in autopsy. The incidence at
sided cardiac chamber, the hemodynamics mimic those of an postmortem examination is about 3055% (49). In catheter
aortic insufficiency. Usually the volume of blood through the angiography, they are detected in less than 10% of the patients
fistula is small, but it may occasionally be as large as twice the on very careful review of high-quality angiograms.
cardiac output. When it is large, a wide aortic pulse pressure is Anatomy
found. Myocardial perfusion may be diminished for that por- Any left coronary branch may be involved. The midsegment
tion of the myocardium supplied by the abnormally connecting of the LAD is by far the most common localization. There is a
coronary artery. This represents a hemodynamic steal phenom- higher prevalence in males (70%). Myocardial bridges are also
enon, and may lead to myocardial ischemia (38,39). more common in patients with idiopathic left-ventricular
Because of the increased blood flow, the involved coronary hypertrophy.
artery is dilated and often tortuous. In some instances, the left- Normally the coronary arteries and their major branches
to-right shunt through the fistula may be small, but the afferent course in the epicardial fat, but occasionally they course beneath
coronary arteries may be greatly dilated. Focal saccular aneu- the myocardium for some distance.
rysms may develop, which eventually can calcify. Myocardial bridges usually cross the LAD and less fre-
Because the fistulae allow increased blood flow from the quently involve other left-ventricular muscular branches. Myo-
aorta into the cardiac chambers, the involved chambers may be cardial bridges affecting the right coronary artery system should
dilated in proportion to the volume of shunt. not cause systolic compression, because the right-ventricular
The fistulae more commonly terminate in the right side of the systolic pressure is lower than the aortic pressure.
heart and occasionally in the superior or inferior vena cava, right Clinical Relevance
atrium (40), coronary sinus, right ventricle (65%), or pulmonary The clinical significance of myocardial bridges is still unknown.
artery (17%). However, they also may terminate in the left side of The coronary artery is compressed during systole. This event can
the heart, either to the left atrium or the left ventricle. Communi- be demonstrated angiographically. Although coronary blood
cations to either peripheral pulmonary artery branches, pulmo- flow occurs primarily during diastole, blood flow may be sig-
nary vein, or mediastinal veins are extremely rare. Up to 50% of nificantly hampered by the bridges during tachycardia.
the fistulae arise from the right coronary artery. Most of the patients do not have typical symptoms of angina
Coronary artery fistulae can occasionally show abstruse and have negative exercise stress test results, suggesting that
courses up to a spider-like configuration and arise from more the myocardial bridge is normally not physiologically signifi-
than one coronary artery (41) (Fig. 12 AD). Numerous com- cant. The risk of myocardial ischemia, vasospasm (50), malig-
munications between the coronary artery and the terminating nant arrhythmia, and sudden cardiac death is nevertheless
localization can rule out surgical correction. increased (5154). Occasionally, typical findings of myocar-
Clinical Relevance dial ischemia are found in patients with a myocardial bridge.
Most of the patients with this condition are eventually rec- Such a patient may benefit from either stenting the vessel under
ognized by the appearance of a continuous and usually loud the bridge (55,56), operative division of the bridging muscle
murmur. On physical examination, there may be evidence of band (57), or coronary artery bypass grafting. Nowadays, mini-
cardiac enlargement and increased cardiac activity. The first mally invasive methods should be the first choice (58).
and second heart sounds may be normal or slightly increased if The presence of myocardial bridging distal to coronary
the flow through the heart is moderately increased. lesions should be considered seriously in preprocedural evalu-
Whether chest pain can be caused by small coronary artery ation of the lesions as a potential risk factor for intracoronary
fistulae is uncertain. Larger communications represent a hemo- thrombus formation. Although pathologists had long recog-
dynamic burden and may cause myocardial ischemia. Particu- nized that the epicardial coronary artery might on occasion
larly pediatric patients tend to be symptomatic, showing atypical course directly through a segment of cardiac muscle, the
and typical chest pain and myocardial ischemia associated physiological significance of this phenomenon was consid-
with ECG alterations (42). Almost all coronary artery fistulae can ered benign. This is partly based in traditional teaching that
be repaired by surgery (43) or embolization and stenting (44), coronary blood flow delivery to the left-ventricular myocar-
except for diffuse angiomatous communications. To prevent fur- dium occurs primarily during the diastolic phase of the car-
ther progression and complications such as aneurysms, even small diac cycle.
fistulae with a hemodynamically low significant shunt should be Unlike coronary angiography, MSCT is capable of depict-
closed (45). A spontaneous closure of an coronary artery fistula ing vessels as well as muscle. Using thin perpendicular MPR
is extremely rare, but has been reported (46). reconstructions or volume rendering, myocardial bridges and
MYOCARDIAL BRIDGES their intramyocardial pathway can be depicted accurately. Also,
Prevalence and Etiology thin MIP reconstructions are very useful to show these alter-
Myocardial bridging also appears to be a congenital ations (Fig. 13AC) . In comparison to catheter angiography,
anomaly, due to failure of externalization of the primitive even in diastole a lumen decrease sometimes is found. Whether
274 FROEHNER ET AL.

Fig. 12. Multiple spider-like fistulae from left anterior descending (LAD) artery and right coronary artery (RCA) to the pulmonary artery. (A,B)
Thin maximum intensity projection reconstruction. From RCA (A) and LAD (B), multiple fistulae lead to the pulmonary artery. (C,D)
Volume rendering. The spider-like fistulae are clearly depicted. There is a common confluence to the pulmonary artery.

this means a higher risk of myocardial ischemia or not has to be The prevalence of coronary artery anomalies is reported to
evaluated in future studies. A statistical survey about the be under 1%. These statistical data are based on catheter
sensitivity and specificity of MSCT detecting myocardial angiographic examinations. In our cohort of more than 1100
bridges still does not exist, but it promises to become a potent MSCT examinations of the coronary arteries, the prevalence
tool in investigations of myocardial bridges. Especially, recon- was 2.6%. The reason for this discrepancy may be that coro-
structions in systole open a new approach to this entity. nary artery anomalies are very difficult to depict in catheter
angio-graphy, and if they are visualized, their course is likely
SUMMARY to be described inaccurately. In our studies, only 36% of the
Anomalies of origin and course of the coronary arteries anomalies found in computed tomography could be displayed
and coronary artery fistulae are often difficult to detect on selectively in catheter angiography (Table 2). Based on our
catheter angiography. In addition, selective visualization may experience, it is not surprising that a high number of coronary
be impossible in many cases. artery anomalies was not detected angiographically before.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 275

Fig. 13. Myocardial bridge in the middle segment of the left anterior
descending (LAD) artery. (A) Thin maximum intensity projection
reconstruction. (B) Volume rendering. The LAD immerses into myo-
cardium and emerges after one and a half centimeters without lumen
reduction. (C) Multiplanar reconstruction: the LAD shows an intra-
muscular pathway.

and their origin can be determined exactly. Unlike catheter

angiography, where the configuration of fistulae often is dif-
ficult to depict, MSCT is able to show the individual feeders,
and particularly the confluence, which is essential for the
planning of surgical interventions. Most of the fistulae have
more than one feeder but only a single outlet into the cardiac
chamber or vein. For visualization of coronary artery fistulae,
volume rendering has proved to be the best reconstruction
method, showing the entire complexity of this anomaly in a
3D format.
The value of MSCT in depiction of myocardial bridges still
has to be evaluated; however, the first experiences look very
promising.
Owing to its 3D imaging capability, MSCT represents a MSCT angiography of the coronary arteries is emerging as
promising tool to depict coronary artery anomalies in spite of an essential imaging tool and the method of choice to detect and
a lower resolution in comparison with catheter angiography. characterize anomalies of the coronary arteries with high reso-
As a result of the systemic application of contrast medium, all lution and significance. In comparison with catheter angiogra-
coronary vessels will be displayed, unlike angiography, which phy, the capability to detect coronary artery anomalies is much
is hardly able to selectively depict anomalous origin vessels. higher. In conclusion, any suggested anomaly of the coronary
The anatomic relationship of aberrant vessels to neighboring arteries should be assessed with CTA to determinate its accu-
structures can also be described accurately, showing the real rate course and origin.
topographic course along the anatomic pattern. The topogra- REFERENCES
phy of aberrant vessels is clearly displayed in 3D images, as
1. Virmani R, Burke AP, Farb A. Sudden cardiac death. Cardiovasc
opposed to the 2D catheter angiography projection. Pathol 2001;10(6):275282.
MSCT furthermore provides clear visualization of coro- 2. Fornes P, Lecomte D. [Sudden death and physical activity and
nary artery fistulae. The courses can be delineated accurately sports]. Rev Prat 2001;51(12 Suppl):S31S35.
276 FROEHNER ET AL.

Table 2
Overview of the Coronary Anomalies Found in 1100 Multislice CT Studies,
Their Courses, and the Number Seen at Selective Catheterization

Type of anomaly CT Coronary angiography

Angiography Selective
Origin Course Cases found performed angiography possible

RCA from LSV interarterial 10 7 2

LCA from RSV ventral 2 1 1
interarterial 2 2 1
LAD from RSV ventral 1 1 0
LCx from RSV dorsal 3 2 2
LAD, LCx separately from LSV normal 4 1 1
Single coronary artery from RSV ventral 1 1 0
Sinus node artery from LCx dorsal 1 1 0
Conus artery from RSV normal 1 0 0
Malrotation of aortic root interarterial 1 0 0
Ectopic origin of RCA interarterial 1 1 0
Coronary artery fistula ventral 2 2 0
Totals 29 19 7 (=36%)

LAD, left anterior descending artery; LCA, left coronary artery; LCx, left circumflex artery; LSV, left sinus of Valsalva; RCA,
right coronary artery; RSV, right sinus of Valsalva.

3. Basso C, Maron BJ, Corrado D, Thiene G. Clinical profile of con- 16. Cox ID, Bunce N, Fluck DS. Failed sudden cardiac death in a patient
genital coronary artery anomalies with origin from the wrong aortic with an anomalous origin of the right coronary artery. Circulation
sinus leading to sudden death in young competitive athletes. J Am 2000;102(12):14611462.
Coll Cardiol 2000;35(6):14931501. 17. Le T, Laskey WK, McLaughin J, White C. Utility of magnetic reso-
4. Mikaeloff P, Loire R, Amiel M, et al. [Anomaly of the origin of the nance imaging in a patient with anomalous origin of the right coro-
circumflex artery. Effects on the risk of mitral and mitro-aortic valve nary artery, acute myocardial infarction, and near-sudden cardiac
replacement]. Arch Mal Coeur Vaiss 1979;72(8):895898. death. Cathet Cardiovasc Diagn 1997;42(2):205207.
5. de Marchena EJ, Russo CD, Wozniak PM, Kessler KM. Compres- 18. Page HL, Engel HJ, Campbell WB, et al. Anomalous origin of the
sion of an anomalous left circumflex coronary artery by a biopros- left circumflex coronary artery. Recognition, angiographic demon-
thetic valve ring. J Cardiovasc Surg (Torino) 1990;31(1):5254. stration and clinical significance. Circulation 1974;50:768773.
6. Felmeden D, Singh SP, Lip GY. Anomalous coronary arteries of 19. Brown JW, Park HJ, Turrentine MW. Arterial switch operation:
aortic origin. Int J Clin Pract 2000;54(6):390394. factors impacting survival in the current era. Ann Thorac Surg
7. Samarendra P, Kumari S, Hafeez M, Vasavada BC, Sacchi TJ. 2001;71(6):19781984.
Anomalous circumflex coronary artery: benign or predisposed to 20. Felmeden D, Singh SP, Lip GY. Anomalous coronary arteries of
selective atherosclerosis. Angiology 2001;52(8):521526. aortic origin. Int J Clin Pract 2000;54(6):390394
8. Harikrishnan S, Jacob SP, Tharakan J, et al. Congenital coronary 21. Samarendra P, Kumari S, Hafeez M, Vasavada BC, Sacchi TJ.
anomalies of origin and distribution in adults: a coronary Anomalous circumflex coronary artery: benign or predisposed to
arteriographic study. Indian Heart J 2002;54(3):271275. selective atherosclerosis. Angiology 2001;52(8):521526.
9. Click RL, Holmes DR, Jr., Vlietstra RE, Kosinski AS, Kronmal RA. 22. Garg N, Tewari S, Kapoor A, Gupta DK, Sinha N. Primary congeni-
Anomalous coronary arteries: location, degree of atherosclerosis tal anomalies of the coronary arteries: a coronary arteriographic
and effect on survivala report from the Coronary Artery Surgery study. Int J Cardiol 2000;74(1):3946.
Study. J Am Coll Cardiol 1989;13(3):531537. 23. Shammas RL, Miller MJ, Babb JD. Single left coronary artery with
10. Garg N, Tewari S, Kapoor A, Gupta DK, Sinha N. Primary congeni- origin of the right coronary artery from distal circumflex. Clin
tal anomalies of the coronary arteries: a coronary arteriographic Cardiol 2001;24(1):9092.
study. Int J Cardiol 2000;74(1):3946. 24. Cherian KM, Bharati S, Rao SG. Surgical correction of anomalous
11. Ohshima T, Lin Z, Sato Y. Unexpected sudden death of a 12-year- origin of the left coronary artery from the pulmonary artery. J Card
old male with congenital single coronary artery. Forensic Sci Int Surg 1994;9(4):386391.
1996;82(2):177181. 25. Nightingale AK, Burrell CJ, Marshall AJ. Anomalous origin of the
12. Rowe L, Carmody TJ, Askenazi J. Anomalous origin of the left left coronary artery from the pulmonary artery: natural history and
circumflex coronary artery from the right aortic sinus: a familial normal pregnancies. Heart 1998;80(6):629631.
clustering. Cathet Cardiovasc Diagn 1993;29(4):277278. 26. Flamm MD, Stinson EB, Hultgren HN, Shumway NE, Hancock EW.
13. Albazzaz SJ. Klinefelters syndrome with anomalous origin of left Anomalous origin of the left coronary artery from the pulmonary
main coronary artery. Cathet Cardiovasc Diagn 1990;20(4):241243. artery. Surgical treatment by ostial occlusion through pulmonary
14. Arizawa M, Nakayama M, Suehara N. [Clinical spectrum and con- arteriotomy. Circulation 1968;38(1):113123.
genital anomalies in trisomy 18]. Nippon Sanka Fujinka Gakkai 27. Salachas A, Achenbach K, Liberatos G, Hatzioannidis V,
Zasshi 1989;41(10):15451550. Goudevenos I, Sideris D. Isolated corrected transposition of great
15. Henglein D, Niederhoff H, Bode H. Origin of the left coronary artery arteries and double left anterior descending artery originating from
from the right pulmonary artery and ventricular septal defect in a the left and right coronary artery. A rate combination of coronary
child of a mother with raised plasma phenylalanine concentrations artery anomaly and congenital heart disease. Angiology 1996;47(1):
throughout pregnancy. Br Heart J 1990;63(3):180182. 6772.
 CHAPTER 25 / CTA TO ASSESS CORONARY ARTERY ANOMALIES 277

28. Roberts WC, Dicicco BS, Waller BF, et al. Origin of the left main 43. Kamiya H, Yasuda T, Nagamine H, et al. Surgical treatment of con-
from the right coronary artery or from the right aortic sinus with genital coronary artery fistulas: 27 years experience and a review of
intramyocardial tunneling to the left side of the heart via the ven- the literature. J Card Surg 2002;17(2):173177.
tricular septum. The case against clinical significance of myocar- 44. Atmaca Y, Altin T, Ozdol C, Pamir G, Caglar N, Oral D. Coronary-
dial bridge or coronary tunnel. Am Heart J 1982;104(2 Pt 1): pulmonary artery fistula associated with right heart failure: success-
303305. ful closure of fistula with a graft stent. Angiology 2002;53(5):
29. Silverman KJ, Bulkley BH, Hutchins GM. Anomalous left circum- 613616.
flex coronary artery: normal variant of uncertain clinical and 45. Bauer M, Bauer U, Alexi-Meskishvili V, et al. [Congenital coronary
pathologic significance. Am J Cardiol 1978;41(7):13111314. fistulas: the most frequent congenital coronary anomaly]. Z Kardiol
30. Molajo AO, Bray CL, Prescott MC, Testa HJ. Thallium-201 myo- 2001;90(8):535541.
cardial imaging in patients with angina pectoris and anomalous aortic 46. Muhler E, Keutel J, von Bernuth G. [Spontaneous occlusion of a
origin of the circumflex coronary artery. Int J Cardiol 1988;18(3): congenital coronary artery fistula]. Z Kardiol 1984;73(8):538540.
371381. 47. Nayar PG, Nyamu P, Venkitachalam L, Ajit SM. Myocardial infarc-
31. Levin DC, Fellows KE, Abrams HL. Hemodynamically significant tion due to myocardial bridging. Indian Heart J 2002;54(6):
primary anomalies of the coronary arteries. Angiographic aspects. 711712.
Circulation 1978;58(1):2534. 48. Arnau Vives MA, Martinez Dolz LV, Almenar BL, Lalaguna LA,
32. Ott DA, Cooley DA, Angelini P, Leachman RD. Successful surgical Ten Morro F, Palencia PM. [Myocardial bridging as a cause of acute
correction of symptomatic cor triatriatum dexter. J Thorac Cardio- ischemia. Description of a case and review of the literature]. Rev Esp
vasc Surg 1979;78(4):573575. Cardiol 1999;52(6):441.
33. Hernando JP, Oliva MJ, Zarauza MJ, et al. [Anomalous origin of the 49. Ferreira AG, Jr., Trotter SE, Konig B, Jr., Decourt LV, Fox K, Olsen
circumflex artery from the pulmonary artery in a patient with rheu- EG. Myocardial bridges: morphological and functional aspects.
matic mitral stenosis]. Rev Esp Cardiol 1995;48(5):359361. Br Heart J 1991;66(5):364367.
34. Sagkan O, Ornek E, Yesildag O. Left circumflex coronary artery 50. Kodama K, Morioka N, Hara Y, Shigematsu Y, Hamada M, Hiwada
arising as a terminal extension of right coronary artery. A case report. K. Coronary vasospasm at the site of myocardial bridgereport of
Angiology 1994;45(5):405408. two cases. Angiology 1998;49(8):659663.
35. Kolodziej AW, Lobo FV, Walley VM. Intra-atrial course of the 51. Fornes P, Lecomte D. [Sudden death and physical activity and
right coronary artery and its branches. Can J Cardiol 1994;10(2): sports]. Rev Prat 2001;51(12 Suppl):S31S35.
263267. 52. Arnau Vives MA, Martinez Dolz LV, Almenar BL, Lalaguna LA,
36. Sciacca P, Falsaperla R, Barone P, et al. [Cardiac involvement in Ten Morro F, Palencia PM. [Myocardial bridging as a cause of
Kawasaki disease. Our experience]. Minerva Pediatr 2001;53(2): acute ischemia. Description of a case and review of the literature].
8793. Rev Esp Cardiol 1999;52(6):441.
37. Braunwald E, Zipes D, Libby P, eds. Heart Disease, 6th ed. Saunders, 53. Agirbasli M, Hillegass WB, Jr., Chapman GD, Brott BC. Stent pro-
Philadelphia: 2001. cedure complicated by thrombus formation distal to the lesion within
38. Rubini G, Ettorre GC, Sebastiani M, Bovenzi F. [Evaluation of a muscle bridge. Cathet Cardiovasc Diagn 1998;43(1):7376.
hemodynamic significance of arteriovenous coronary fistulas: diag- 54. Agirbasli M, Martin GS, Stout JB, Jennings HS, III, Lea JW, Dixon
nostic integration of coronary angiography and stress/rest myocar- JH, Jr. Myocardial bridge as a cause of thrombus formation and
dial scintigraphy]. Radiol Med (Torino) 2000;100(6):453458. myocardial infarction in a young athlete. Clin Cardiol 1997;20(12):
39. Brigui M, Remadi F, Belkhiria N, Zemni J, Chabrak S, Gargouri R. 10321036.
[Angina and myocardial ischemia related to a fistula between the 55. Prendergast BD, Kerr F, Starkey IR. Normalisation of abnormal
anterior interventricular artery and the pulmonary artery in a 56-year- coronary fractional flow reserve associated with myocardial bridg-
old man]. Ann Cardiol Angeiol (Paris) 1995;44(2):8690. ing using an intracoronary stent. Heart 2000;83(6):705707.
40. Goldberg L, Mekel J. Congenital aneurysm of the left main coro- 56. Jeremias A, Haude M, Ge J, et al. [Emergency stent implantation in
nary artery with fistulous communication to the right atrium the area of extensive muscle bridging of the anterior interventricular
and coronary steal phenomenon. Cardiovasc J S Afr 2001;12(1): ramus after post-interventional dissection]. Z Kardiol 1997;86(5):
4851. 367372.
41. Lai MC, Chen WJ, Chiang CW, Ko YL. An unusual case of dual 57. Furniss SS, Williams DO, McGregor CG. Systolic coronary occlu-
coronary artery fistulas to main pulmonary artery. Chang Gung Med sion due to myocardial bridginga rare cause of ischaemia. Int J
J 2002;25(1):5155. Cardiol 1990;26(1):116117.
42. Wang NK, Hsieh LY, Shen CT, Lin YM. Coronary arteriovenous 58. Pratt JW, Michler RE, Pala J, Brown DA. Minimally invasive coro-
fistula in pediatric patients: a 17-year institutional experience. nary artery bypass grafting for myocardial muscle bridging. Heart
J Formos Med Assoc 2002;101(3):177182. Surg Forum 1999;2(3):250253.
 CHAPTER 26 / MDCT FOR KAWASAKI DISEASE 279

26 Multidetector-Row CT
for Assessment of Kawasaki Disease

TORU SAKUMA, MD AND KUNIHIKO FUKUDA, MD

KAWASAKI DISEASE to prevent coronary artery complications, and aspirin clears

GENERAL acute inflammatory symptoms (9). IVGG therapy has reduced
Kawasaki disease is an acute febrile illness, causing mu- both the morbidity of Kawasaki disease and the apparent inci-
cosal inflammation, skin rash, and cervical lymphadenopathy, dence of coronary artery abnormalities from approx 2025% to
recognized most often in children younger than 4 yr of age. It less than 5% at 68 wk after initiation of therapy (6). In the pre-
was first described by Dr. Tomisaku Kawasaki in Japanese IVGG treatment stage, the incidence of coronary aneurysm in
literature in 1967 (1), and then in English in 1974 (2). The acute Kawasaki disease was 25%, 55% of which showed
disease is of unknown etiology that produces a systemic vascu- regression on follow-up angiography. This is a characteristic
litis, which is most severe in the medium-sized arteries, and phenomenon in Kawasaki disease. Regression is likely to occur
especially prominent in the coronary arteries. This can be asso- within 1 or 2 yr after onset, and it is unlikely to occur more than
ciated with considerable morbidity and mortality, mostly the several years after onset (1012). During follow-up, ischemic
result of myocardial involvement and coronary artery compli- heart disease developed in 4.7% and myocardial infarction in
cations such as aneurysm, calcification, and stenosis. In Japan 1.9%. Death occurred in 0.8% (13). The standardized mortality
as well as in North America, Kawasaki disease is presently a ratio of Kawasaki disease was 1.25, but for the male group with
leading cause of acquired heart disease in children (3). cardiac sequelae was 2.35. So the mortality rate among male
DIAGNOSTIC CRITERIA patients with cardiac sequelae due to Kawasaki disease seemed
The diagnosis of Kawasaki disease is made according to the higher than that in the general population (14). The number of
guidelines prepared by the Committee on Rheumatic Fever, patients with cardiac lesions 1 mo after onset has decreased
Endocarditis, and Kawasaki Disease, Council on Cardiovascular year by year, and that is possibly attributable to the increase in
Disease in the Young, American Heart Association (Table 1) the proportion of patients treated with IVGG and an increase in
(4), and the Japan Kawasaki Disease Research Committee the dosage (15).
(5), because of the absence of a specific laboratory test. The Long-term management of Kawasaki disease, which is based
principal diagnostic criteria of Kawasaki disease are persistent on risk stratification, is also shown by the research committee,
fever, conjunctival injection, inflamed oropharyngeal mucosa, and risk level categories are summarized (3). Patients with a
changes in the peripheral extremities, erythematous rash, and normal coronary artery, evaluated by echocardiography or
cervical lymphadenopathy. At least five of these six principal CAG during the acute stage of Kawasaki disease, have little
diagnostic criteria should be satisfied to establish a diagnosis of morbidity in childhood, and no strict follow-up is usually required.
Kawasaki disease. Where coronary aneurysms are recognized
by transthoracic echocardiography or coronary angiography ASSESSMENT OF CORONARY ARTERY
(CAG), patients with four of the above diagnostic criteria can ECHOCARDIOGARPHY
also be diagnosed with Kawasaki disease. The initial evaluation and follow-up of the coronary arterial
TREATMENT AND MANAGEMENT lesions in Kawasaki disease are done by echocardiography,
Standard initial treatment involves intravenous administra- because it is considered to be the most useful and essential
tion of high-dose globulin and oral administration of aspirin. method to evaluate coronary aneurysms. Evaluation of coro-
Treatment is started preferably within the first 10 d from the nary artery morphology should include quantitative and quali-
onset of illness (68). Two mg/kg intravenous globulin tative assessment of the inner diameter of the vessels. When the
(IVGG) combined with at least 30 to 50 mg/kg/d aspirin pro- coronary artery diameter is larger than normal but a segmental
vides maximum protection against development of coronary aneurysm is not apparent, the vessel is described as ectatic.
abnormalities after Kawasaki disease. Gamma globulin helps Aneurysms are classified as small (less than 5 mm in internal
diameter), medium (5 to 8 mm in internal diameter), or giant
From: Contemporary Cardiology: CT of the Heart: (more than 8 mm in internal diameter). Many publications have
Principles and Applications demonstrated the high efficacy and accuracy of echocardio-
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

279
280 SAKUMA AND FUKUDA

Table 1
Diagnostic Criteria

Principal Clinical Findings a

Fever persisting at least 5 d b and the present of at least four of the following five principal features:
1. Changes in extremities:
a. Acute: erythema and edema of hands and feet
b. Convalescent: membranous desquamation of fingertips
2. Polymorphous exanthema
3. Bilateral, painless bulbar conjunctival injection without exudates
4. Changes in lips and oral cavity: erythema and cracking of lips, strawberry tongue, diffuse injection of oral and
pharyngeal mucosae
5. Cervical lymphadenopathy (=1.5 cm in diameter), usually unilateral
a Patients with fever and fewer than four principal symptoms can be diagnosed as having Kawasaki disease when coronary
artery disease is detected by 2D echocardiography or coronary angiography. Other diagnoses should be excluded. The physician
should be aware that some children with illness not fulfilling these criteria have developed coronary artery aneurysms.
b Many experts believe that in the presence of classic features, the diagnosis of Kawasaki disease can be made by experienced
observers before d 5 of fever.
Adapted from ref. 4.

graphy in detecting and characterizing coronary artery diseases aneurysm, had poor distensibility and was stiff on examination
(16,17). It has been reported that echocardiography has 98% with intracoronary isosorbide dinitrate (21). These findings
sensitivity and 95% specificity in the diagnosis of left main suggest that regressed ectasia and aneurysm has potential risk
coronary artery aneurysms in comparison with CAG. How- for later complications, even in cases of normal CAG. There-
ever, echocardiography is less sensitive in detecting right coro- fore, long-term follow-up is recommended in cases of asymp-
nary artery (RCA) lesions, as the middle part of the RCA is tomatic transient aneurysm, to screen for development of
located behind the sternum and the distal part of RCA is the premature coronary artery atherosclerosis in adulthood.
furthest away from the chest wall. In addition, it is impossible MAGNETIC RESONANCE IMAGING
to evaluate small aneurysms in the peripheral branches and the Electrocardiogram (ECG)-gated magnetic resonance imag-
lumens of calcified aneurysms by echocardiography. There is ing (MRI) has also been used for the noninvasive evaluation of
also a limitation in the accurate evaluation of the degree of the aortic root and the major proximal portion of the coronary
stenosis and the size of intramural thrombosis within aneu- arteries (22). Recent development of magnetic resonance (MR)
rysms. Furthermore, visualization and characterization of the hardware and rapid acquisition pulse sequences enables high
coronary arteries through transthoracic echocardiography resolution MR angiography (MRA) of the coronary arteries.
become progressively more difficult as the child grows. Sensitivity, specificity, accuracy, and negative predictive value
CORONARY ANGIOGRAPHY for patients with coronary artery disease by coronary MRA
CAG offers complete images of coronary arteries and were 94%, 45%, 75%, and 82% respectively, compared with
more detailed definition of coronary artery anatomy than CAG (23). Free-breathing 3D coronary MRA also accurately
echocardiography, making it possible to detect coronary artery defines coronary artery aneurysms in patients with Kawasaki
stenosis or thrombotic occlusion and to determine the extent of disease (24). However, several data acquisitions are needed
collateral artery formation. However, there are certain risks along the major axis of the arteries because of the limitation in
associated with this procedure as a result of its invasive nature the volume coverage of the scans. The scanning time of approx
and the exposure to ionizing radiation, as well as its being more 30 min is a relatively long time for young children and adoles-
expensive. Whether to use CAG should be determined on a cents. In addition, voxel size of 0.7 1.0 1.5 mm is not
risk-vs-benefit balance, including abnormal findings on adequate for 3D imaging in comparison with the isotropic data
echocardiography, symptoms or signs of ischemia, ausculta- obtained with multidetector-row CT (MDCT).
tory evidence of valvular regurgitation, evidence of cardiac
dysfunction, and the need for intracoronary thrombolytic ASSESSMENT OF CORONARY ARTERIES BY MDCT
treatment (3). INTRODUCTION
Whether or not regression may reverse and eventually Single-slice helical CT has a limited role in the management
develop into stenosis is uncertain at the present time. However, of Kawasaki disease because of cardiac motion artifacts due to
postmortem examinations of some patients with angiograph- long acquisition times and lack of ECG gating, although there
ically documented regression of coronary artery aneurysm have has been one report presenting two cases of Kawasaki disease
revealed intimal proliferation and fibrosis not apparent on the with unrecognized coronary artery aneurysms detected by
angiogram (18). On intravascular ultrasound imaging (IVUS), single-slice helical CT (25).
the regressed coronary aneurysms also demonstrated a marked Electron beam CT (EBCT) is an effective, noninvasive
thickening of the intima with or without calcification, which method of identifying coronary artery aneurysms. Frey et al.
bore a striking resemblance to early atherosclerotic lesions reported in their series that 9 of 10 coronary artery aneurysms
(19,20). The coronary artery wall, at the site of the regressed in Kawasaki disease were visualized with EBCT. In one false
 CHAPTER 26 / MDCT FOR KAWASAKI DISEASE 281

Table 2
Scan Protocol
SOMATOM Volume Zoom SOMATOM Sensation 16

Collimation 4 1 mm 12 0.75 mm
Gantry rotation time (ms) 500 420
Temporal resolution (ms) 125250 105210
Spacial resolution 0.7 0.7 1.25 0.7 0.7 0.8
Table feed (mm/rotation) 1.5 2.8
Tube voltage (kV) 150 120
Maximum tube current (mA) 500 500
Contrast agent (mL) 2.0 body weight (BW) 2.0 BW
Scan delay time (s) 20 bolus tracking
Scan time (s) 35 10

case, two contiguous aneurysms were mistaken for a single proximal and distal parts. The area within reach of
aneurysm. The inability to resolve the aneurysms as sepa- echocardiography was defined as the proximal part and the
rate structures was probably due to 8-mm section thickness in region beyond the reach of echocardiography as the distal part.
this case (26). Spatial resolution of EBCT is suboptimal for the The visibility of coronary arteries was 93.6% in the proximal
evaluation of coronary arteries compared with that of MDCT. part and 62.9% in the distal part. In the proximal part, 18 cal-
Z-axis collimation of EBCT is up to 3 mm, while in 16-row cifications, 24 aneurysms, 19 both calcification and aneurysms
MDCT it is 0.75 mm. Another problem with EBCT is that the and one stenosis were found. The sensitivity of MDCT was
prospectively ECG-triggered sequential method cannot obtain 94.3% in the detection of abnormalities within the proximal
phase-constant cardiac volume imaging during the scan, which segments of the coronary arteries, and the specificity was 100%
results in potential image degradation. when echocardiography was used as the gold standard. In the
MDCT with retrospective ECG gating is considered a well- distal part, where echocardiography cannot reach, 16 calcifica-
suited modality for noninvasive diagnosis of the coronary tions, 2 aneurysms, 8 both calcifications and aneurysms, and 4
arteries, as it can optimize temporal resolution, spatial resolu- stenoses were detected.
tion, and volume coverage. REPRESENTATIVE CASE STUDIES
METHOD Case 1: 19-Yr-Old Male
Three-dimensional volume data were obtained with MDCT This patient presented with continuous fever, conjunctival
(SOMATOM Volume Zoom and SOMATOM Sensation 16, injection, erythema and edema of hands and feet, and poly-
Siemens AG, Medical Solutions) (Table 2). Average scan range morphous exanthema. He was diagnosed with Kawasaki dis-
was 100 mm. All image acquisitions were performed in deep ease at 20 mo from birth. The patient was treated with aspirin
inspiratory breath-hold among patients who were able to do so, and dipyridamole, recovering within 25 d. Echocardiography
and those patients who could not do so were allowed to breath revealed coronary aneurysms in the proximal part of the RCA
freely under sedation with chloral hydrate to achieve an and left coronary artery (LCA). The initial CAG was per-
adequate study. The injection of 2 mL/kg of 370-mgI/mL non- formed 2 mo after the onset, because of abnormal Q wave and
ionic contrast medium (iopamidol) was performed via a peri- arrhythmias, and revealed a coronary aneurysm with stenosis
pheral intravenous route during one breath-hold, with and thrombus in the RCA and an aneurysm with stenosis in the
simultaneous registration of the ECG signal also being per- LCA. Repeat coronary angiographies were performed at the
formed. The injection speed ranged from 1 to 3 mL/s, depend- ages of 5, 8, 11, 14, and 19 yr. At the age of 19, left circumflex
ing on the patients age and size. The ECG signal was used to (LCx) obstruction was discovered, and thallium-201 myocar-
reconstruct the images at identical time points, about 400 ms dial perfusion scintigraphy revealed ischemia in the lateral
before the next R wave. For 3D reconstruction and visualiza- wall of the left ventricle. There were multiple calcified aneu-
tion of the coronary arteries, curved multiplanar reformation rysms in the proximal part of the coronary arteries, which
(MPR) and maximum intensity projection (MIP) techniques were all detected on 4-row MDCT (Fig. 1A,B). Although these
were applied (Heart View CT, Siemens AG, Medical Solu- aneurysms were followed up by echocardiography, intralu-
tions). Curved MPR images were rendered along the course of minal patency was impossible to evaluate because of heavy
each coronary artery from the axial data set, while MIP images mural calcification. MDCT showed large thrombosis within
were produced in the same way with 3-mm thickness. the giant calcified aneurysm of the coronary arteries. One of
DETECTABILITY BY MDCT the advantages of MDCT over echocardiography is that
MDCT findings in 70 cases with Kawasaki disease were MDCT can depict intraluminal information within the calci-
analyzed, where the coronary arteries were subdivided into fied wall.
282 SAKUMA AND FUKUDA

Fig. 1. Thin-slice maximum intensity projection images of proximal right coronary artery and left coronary artery reveal multiple calcified
aneurysms, which were successfully followed up by echocardiography. However, large thrombosis (arrow) within the giant aneurysm was not
depicted by echocardiography because of heavy calcification.

Fig. 2. Volume-rendered image of multidetector-row CT delineates Fig. 3. Volume-rendered image shows obstruction in left circumflex
heavily calcified aneurysm in segment 2, which was beyond the scope coronary artery (arrow).
of ultrasound.

Case 2: 20-Yr-Old Male row cardiac MDCT was performed on an out-patient basis in
The disease was diagnosed when the patient was admitted to order to evaluate the potential lesions in the coronary arteries.
our hospital suffering a high temperature at 6 yr of age. He was Cardiac CT confirmed all the abnormalities detected by CAG
treated with aspirin and dipyridamole, being discharged from and verified no new development (Figs. 2 and 3).
hospital after 38 d. CAG, performed at the age of 14, showed Case 3: 21-Yr-Old Female
LCx stenosis and calcification of the RCA. The patient was The disease was diagnosed when the patient was 2 yr of age,
followed up with treatment of aspirin and dipyridamole. At the when she was admitted to our hospital with a high temperature.
age of 18, he underwent further angiography, which showed She was discharged from the hospital after 48 d treatment with
LCx stenosis and RCA segment 2 calcified aneurysm. Sixteen- aspirin and dipyridamole. Echocardiography delineated RCA
 CHAPTER 26 / MDCT FOR KAWASAKI DISEASE 283

Fig. 4. Volume-rendered images of multidetector-row CT clearly delin-

eate giant aneurysm in segment 1, calcified giant aneurysm in segments
2 and 6.

Fig. 5. Automated segmentations of right coronary (A) and left anterior descending artery (B) show mural calcification and intraluminal
thrombosis in giant aneurysm, which are suitable for quantitative assessment of vessel diameters and stenosis degree. There is a focal area of
calcification in the proximal left circumflex, which is in normal caliber (arrow). Findings are in keeping with regressed aneurysm. Such lesions
carry the potential risk to develop luminal stenosis, and therefore careful and close follow-up is required. Multidetector-row CT is able to detect
small calcifications, which cannot be shown by coronary angiography.

and LCA aneurysms, which were also detected in CAG per- 201 myocardial perfusion scintigraphy. Echocardiography
formed at the ages of 3, 4, 6, and 7. At the age of 10, she revealed a giant aneurysm of the distal RCA and calcification
underwent further angiography, which showed giant aneurysms of the left main trunk. Cardiac CT on 16-row MDCT was per-
in segment 1, segments 23, and segments 56, small aneurysm formed on an out-patient basis in order to evaluate the potential
in segment 11, and calcification in segments 23 and segment lesions in the coronary arteries. Cardiac CT confirmed all the
5. The patient was followed up with treatment of aspirin, dipy- abnormalities detected by CAG (Fig. 4A). Furthermore, intra-
ridamole, ticlopidine HCl, and ubidecarenone. At the age of 19, mural thrombosis in calcified giant aneurysms and small calci-
pharmacological stress electrocardiography showed ST-T fications in LCx, which were not evaluated by CAG, were also
changes in II, III, aVF, but no ischemia detected on thallium- clearly delineated (Figs. 5A,B and 6A).
284 SAKUMA AND FUKUDA

Fig. 6. Quantitative analysis of short-axis multiplanar reformation images of large calcified aneurysms, using Vessel-View soft ware (Siemens
AG, Medical Solutions), shows accurate diameter of aneurysms.

SUMMARY Heart Association. Diagnostic Guidelines for Kawasaki Disease.

Circulation 2001;103:335336.
MDCT enables evaluation of coronary arteries, which can- 5. The Japan Kawasaki Disease Research Committee. Diagnostic
not be reached by echocardiography, such as areas behind Guidelines of Kawasaki Disease (in Japanese). The 5th Revised
calcification and the distal part of coronary arteries. MDCT is Edition, February 2002.
6. Newburger JW, Takahashi M, Burns JC, et al. The treatment of
considered to be highly efficient in the delineation of abnor- Kawasaki syndrome with intravenous gamma globulin. N Engl J
malities in coronary arteries with Kawasaki disease. Med 1986;315:341347.
Coronary artery bypass grafting using internal thoracic and 7. Newburger JW, Takahashi M, Beiser AS, et al. A single intravenous
gastroepiploic arteries is increasing in Japan and the United infusion of gamma globulin as compared with four infusions in the
States, because of the long patency of arterial grafts. The actu- treatment of acute Kawasaki syndrome. N Engl J Med 1991;324:
16331639.
arial graft patency rate for arterial grafts has been reported 8. Dajani AS, Taubert KA, Gerber MA, et al. Diagnosis and therapy of
as 77.1% 1.1% (27). Several types of catheter intervention Kawasaki disease in children. Circulation 1993;87:17761780.
have also been performed in the management of coronary 9. Terai M, Shulman ST. Prevalence of coronary artery abnormalities
stenosis caused by Kawasaki disease (28,29). Therefore, in Kawasaki disease is highly dependent on gamma globulin dose
but independent of salicylate dose. J Pediatr 1997;131:888893.
MDCT is also thought to be a promising non-invasive modal- 10. Nakano H, Ueda K, Saito A, Nojima K. Repeated quantitative
ity for the assessment of the patency of bypass grafts or angiograms in coronary arterial aneurysm in Kawasaki disease. Am
coronary stenosis after interventional treatment instead of J Cardiol 1985;56:846851.
conventional CAG. 11. Takahashi M, Mason W, Lewis AB. Regression of coronary aneu-
rysms in patients with Kawasaki syndrome. Circulation 1987;75:
387394.
REFERENCES 12. Akagi T, Rose V, Benson LN, Newman A, Freedom RM. Outcome
1. Kawasaki T. Acute febrile mucocutaneous syndrome with lymphoid of coronary artery aneurysms after Kawasaki disease. J Pediatr 1992;
involvement with specific desquamation of the fingers and toes in 121:689694.
children.(in Japanese) Jpn J Allergy 1967;16:178222. 13. Kato H, Inoue O, Kawasaki T, Fujiwara H, Watanabe T, Toshima H.
2. Kawasaki T, Kosaki F, Okawa S, Shigematsu I, Yanagawa H. A new Long-term consequences of Kawasaki disease. Circulation 1996;94:
infantile acute febrile mucocutaneous lymph node syndrome 13791385.
(MLNS) prevailing in Japan. Pediatrics 1974;54:271276. 14. Nakamura Y, Yanagawa H, Harada K, Kato H, Kawasaki T. Mortal-
3. Dajani AS, Taubert KA, Takahashi M, et al. Guidelines for long- ity among persons with a history of Kawasaki disease in Japan: the
term management of patients with Kawasaki disease: report from fifth look. Arch Pediatr Adolesc Med 2002;156:162165.
the Committee on Rheumatic Fever, Endocarditis, and Kawasaki 15. Yanagawa H, Nakamura Y, Yashiro M, et al. Incidence survey
Disease, Council on Cardiovascular Disease in the Young, Ameri- of Kawasaki disease in 1997 and 1998 in Japan. Pediatrics 2001;
can Heart Association. Circulation 1994;89:916922. 107:e33.
4. Council on Cardiovascular Disease in the Young, Committee on 16. Capannari TE, Daniels SR, Meyer RA, Schwartz DC, Kaplan S.
Rheumatic Fever, Endocarditis, and Kawasaki Disease, American Sensitivity, Specificity and predictive value of two-dimensional
 CHAPTER 26 / MDCT FOR KAWASAKI DISEASE 285

echocardiography in detecting coronary artery aneurysms in 23. Kim WY, Danias PG, Stuber M, et al. Coronary magnetic reso-
patients with Kawasaki disease. J Am Coll of Cardiol 1986;7: nance angiography for the detection of coronary stenosis. N Engl
33560. J Med 2001;345:18631869.
17. Hiraishi S, Misawa N, Takeda N, Horiguchi Y. Transthoracic ultra- 24. Greil GF, Stuber M, Botnar RM, et al. Coronary magnetic reso-
sonic visualisation of coronary aneurysm, stenosis, and occlusion in nance angiography in adolescents and young adults with Kawasaki
Kawasaki disease. Heart 2000;83:400405. disease. Circulation 2002;105:908911.
18. Naoe S, Takahashi K, Masuda H, Tanaka N. Coronary findings post 25. Hamada R, Yano I, Fujiwara M, et al. CT screening for unrecognized
Kawasaki disease in children who died of other causes. Prog Clin coronary sequel of Kawasaki disease. Acta Paediatr Jpn 1995;
Biol Res 1987;250:341346. 37:416418.
19. Suzuki A, Yamagishi M, Kimura K, et al. Functional behavior and 26. Frey EE, Matherne GP, Mahoney LT, Sato Y, Stanford W, Smith
morphology of the coronary artery wall in patients with Kawasaki WL. Coronary artery aneurysms due to Kawasaki disease: diagnosis
disease assessed by intravascular ultrasound. J Am Coll Cardiol 1996; with ultrafast CT. Radiology 1998;167:725726.
27:291296. 27. Kitamura S, Kameda Y, Seki T, et al. Long-term outcome of myo-
20. Sugimura T, Kato H, Inoue O, Takagi J, Fukuda T, Sato N. cardial revascularization in patients with Kawasaki coronary artery
Vasodilatory response of the coronary arteries after Kawasaki dis- disease. A multicenter cooperative study. J Thorac Cardiovasc Surg
ease: evaluation by intracoronary injection of isosorbide dinitrate.J 1994;107:663673.
Pediatr 1992;121:684688. 28. Akagi T, Ogawa S, Ino T, et al. Catheter interventional treat-
21. Sugimura T, Kato H, Inoue O, et al. Congenital heart disease: intravas- ment in Kawasaki disease: a report from the Japanese Pediatric
cular ultrasound of coronary arteries in children: assessment of the Interventional Cardiology Investigation Group. J Pediatr 2000;137:
wall morphology and the lumen after Kawasaki disease. Circulation 181186.
1994;89:258265. 29. Ishii M, Ueno T, Ikeda H, et al. Sequential follow-up results of cath-
22. Bisset III GS, Strife JL, KcCloskey J. MR imaging of coronary eter intervention for coronary artery lesions after Kawasaki disease:
artery aneurysms in a child with Kawasaki disease. AJR 1989;152: quantitative coronary artery angiography and intravas-cular ultra-
805807. sound imaging study. Circulation 2002;105:30043010.
 CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY 287

27 Multidetector-Row CT of the Coronary

Arteries for Planning of Minimally
Invasive Bypass Surgery

CHRISTOPHER HERZOG, MD, SELAMI DOGAN, MD,

AND THOMAS J. VOGL, MD

INTRODUCTION placement, since any intraoperative collision of the arms must

Minimally invasive coronary artery bypass surgery is gain- strictly be avoided. The main task of preoperative
ing increasing clinical importance as an alternative procedure multidetector-row CT (MDCT) examinations, therefore, is to
to conventional open-chest techniques (1,2). Recent technical provide additional morphologic information of the surgical
developments in the field of computer-enhanced technology target site and to portray the exact topographic relationship
have markedly reduced surgical access, now enabling the clini- between the heart and the surrounding tissue.
cal use of entirely closed-chest procedures such as totally
MINIMALLY INVASIVE BYPASS SURGERY
endoscopic coronary artery bypass grafting (TECABG) (3).
However, one of the technical drawbacks in TECABG pro- OVERVIEW OF MINIMALLY INVASIVE
cedures is the lack of tactile feedback. The dissection and SURGICAL TECHNIQUES
exposition of intramural coronary arteries or vessels hidden The philosophy behind all types of keyhole procedures
deep inside the epicardiac fatty tissue relies mainly on visual is to perform cardiac surgery without cardio-pulmonary bypass
information and thus is technically much more challenging than (CPB) in order to avoid inflammatory whole-body responses
conventional surgery (3,4). Unfortunately, preoperative coro- or intra-/postoperative embolic infarction (7). In addition, the
nary angiography (CAG) often does not provide all relevant surgical trauma is reduced, thus distinctly improving the
visual information for these minimally invasive surgical proce- cosmetic and functional postoperative outcome. Until recently,
dures. Thus, if collateral circulation is absent, the coronary this was mainly achieved through an antero-lateral minithor-
territory distal to a total vascular occlusion often cannot be acotomy, a surgical approach that combined the advantages of
displayed properly, consequently rendering impossible a suffi- reduced surgical trauma with the benefits of off-pump surgery
cient evaluation of wall quality, plaque composition, and vas- (810). However, minimally invasive direct coronary artery
cular diameter of the target anastomotic site. In addition, even bypass grafting (MIDCABG), as this technique is called, in
if visualization of the diseased vessel and its distal segments is general is limited to the revascularization of a maximum of two
possible, angiograms often lack other important morphological target vessels owing to restricted access to the surgical target
information, such as the vessels position relative to the sur- site.
rounding cardiac fatty tissue or its exact cardiac course (5). The For a multivessel revascularization, CPB still is mandatory,
latest developments in TECABG procedures, such as interven- nowadays performed by using new but less biocompatible
tions on the beating heart (6) or the introduction of a fourth technologies such as the Port Access System (Heartport Inc.)
swivel arm, even more necessarily require exact preoperative (1113). This CPB system consists of a femoro-femoral car-
analysis of the target site. Thus, in beating-heart procedures, diopulmonary bypass and an endoaortic balloon clamp.
the distal anastomosis can only be performed on coronary With the recent advent of robotically enhanced telemani-
arteries that are temporarily pinched off. Suture and explora- pulation, a new, powerful tool has been created to further mini-
tion times therefore have to be as short as possible in order to mize surgical access, for the first time allowing true closed
avoid any ischemic damage to the mycardiac tissue distal to this chest totally endoscopic procedures (Figs. 1, 2) (2,1416).
ligature. A fourth swivel arm distinctly aggravates correct TECABG surgery initially also was restricted to the use of Port
Access systems for CPB. However, the latest developments in
surgical techniques render possible TECABG procedures on
the beating heart and therefore help to avoid undesirable side
From: Contemporary Cardiology: CT of the Heart:
Principles and Applications effects of extracorporeal circulation (7), refine procedural flow,
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ and shorten the operative time.

287
288 HERZOG, DOGAN, AND VOGL

Fig. 1. Equipment for totally endoscopic coronary artery bypass grafting procedures. On the left, the main console (lower row), consisting of
binoculars that allow three-dimensional endoscopic visualization of the surgical field and scissors-like handles that allow telemanipulation of
the robots swivel arms (upper row). In the middle, the slave unit, with its swivel arms, fixing the endoscopic camera and all surgical
instruments (upper row) within the patients body and directly transferring even the slightest movements of the scissor-like handles. On the
right is a schematic drawing of the endoscopic instruments, ports position respectively, within the patients thorax (lower row) and a picture
of the endoscopic camera for 3D visualization of the surgical field (upper row).

OPERATIVE TECHNIQUE FOR TECABG PROCEDURES

For a typical left-sided approach for left internal mammary
artery (LIMA) to LCA bypass grafting, the patient is placed on
the operating table in a supine position with the left chest
elevated by about 3040. Usually the fifth intercostal space
(ICS) close to the anterior axillary line is identified, and after
deflation of the left lung a camera port is placed bluntly to

Fig. 2. Schematic drawing, view from above, showing the intraopera-

tive setting of a totally endoscopic coronary artery bypass grafting
(TECABG) procedure. The surgeon is sitting in front of the console
(1) allowing 3D endoscopic visualization of the intraoperative field,
and telemanipulates the robot (2) with scissors-like handles (a). The
robot itself consists of three swivel arms, one carrying the camera,
two holding the surgical instruments (newest prototypes now consist
of four swivel arms). Perioperative monitoring is provided through
the anesthesiologist (3). Though many TECABG procedures still are
performed with assistance of a heart-lung machine (4), latest techni-
cal developments such as the Port Access System also render pos-
sible procedures on the beating heart. Reprinted from ref. 35, with
permission from the Radiological Society of North America.
 CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY 289

Fig. 3. Four steps of a totally endoscopic coronary artery bypass grafting procedure, left anterior oblique projection. The black arrows point
to the coronary artery, the white arrows mark the bypass. After dissection of the epicardium, the surgeon first has to identify the relevant cardiac
structures. As shown in this case, epicardiac fatty tissue (F) very often prevents a direct visualization of the target vessel (upper row left).
Without further morphological information, only careful dissection of the fatty tissue allows exact localization of the vessels position (upper
row right). The target vessel then is incised over a distance of 6 to 7 mm (lower row left) and subsequently grafted in an end-to-side technique
with the left internal mammary artery, which previously has been mobilized from the chest wall (lower row right). Reprinted from ref. 35, with
permission from the Radiological Society of North America.

avoid left-ventricular injury. The chest is insufflated with System. After initiation of CPB and venting the right ven-
warm CO2 (37C). After insertion of the endoscope, two ports tricle via the pulmonary artery, the heart is decompressed and
are placed under visual control to accommodate the two robot endoscopic pericardiotomy is performed safely. Identifica-
arms usually in the third and seventh ICS (Fig. 1). The LIMA tion of the target vessel is performed prior to cardiac arrest
is mobilized from the subclavian artery down to the distal using direct endoscopic visualization. Infusion of antegrade
bifurcation using a 30-degree endoscope angled upwards. The cristalloid St. Thomas cardioplegia delivered to the aortic root
distal end of the LIMA is skeletonized for grafting, and a soft via the Port Access endoclamp provides reliable cardiac
bulldog clamp (Scanlan Int.) is placed (2,3). arrest. A 67-mm arteriotomy is carried out, and the left
The left femoral artery and vein are dissected with a 3-cm anterior descending artery (LAD) is grafted with the LIMA
oblique incision in the groin. After heparinization, CPB is in an end-to-side technique (Fig. 3) with a running 7.0 prolene
instituted by femoro-femoral perfusion using the Port Access suture (7.5 cm length, Fumalene, Fumedica Medizintechnik).
290 HERZOG, DOGAN, AND VOGL

Table 1
Scanning and Image Reconstruction Parameters.
Comparison Between 4- and 16-Row Multidetector-Row CT (MDCT)

4-Row MDCT 16-Row MDCT

Plain Series
Scanning Parameters
Rotation time (ms) 420 500
Temp. Resolution min 105 125
Temp. Resolution max 210 250
kV 120 120
mAs 300 300
Collimation 4 2.5 16 1.5
Table feed/rotation 3.8 5.7
Image Reconstruction

FOV (mm) 220 220

Kernel B 35 B 35
Slice Tickness (mm) 3 2
Increment (mm) 1.5 1
Contrast enhanced series
Scanning Parameters
Rotation time (ms) 420 500
Temp. Resolution min 105 125
Temp. Resolution max 210 250
kV 120 120
mAs 300 300
Collimation 41 16 0.75
Table feed /rotation 1.5 2.8
Contrast medium (mL) 150 (+ 30 /Test Bolus) 120
Flow Rate (mL/s) 3.5 4
Image Reconstruction
FOV (mm) 220 220
Kernel B 35 B 35
Slice Tickness (mm) 1.25 0.75
Increment (mm) 0.6 0.3

FOV, field of view.

After completion of the anastomosis, the aortic endoclamp is and fast scan times, streak artifacts caused through undiluted
deflated and the patient is weaned from CPB. Venous and contrast medium flowing into the right atrium are distinctly
arterial cannulae are removed and two chest tubes are placed reduced.
through the camera port and an instrument port incision (2,3). Patients with heart rates higher than 65 bpm should previ-
ously receive a short-lasting -blocker (Brevibloc, 100 mg,
MDCT EVALUATION 1 mL/10 kg BW) in order to obtain rates of 60 bpm or less.
SCANNING PARAMETERS For proper assessment of calcified plaques, both contrast-
All MDCT examinations should be performed using ECG- enhanced and plain examinations are mandatory. Scanning
gating. In our studies we used both 4-row (SOMATOM Plus 4 is done on both scanners with 120 kV and 300 mAs. For the
VolumeZoom, Siemens) and 16-row (SOMATOM Sensation 4-row scanners, scanning parameters were 500 ms rotation
16, Siemens) CT scanners. Sixteen-row technology allowed time, 4 2.5 collimation and 3.8-mm table feed per rotation
sufficient evaluation even of small vessels (>1 mm) and resulted for the plain series, 4 1 mm slice collimation and 1.5-mm
in less motion artifacts at elevated heart rates (>75 beats per table feed per rotation for the contrast-enhanced series
minute [bpm]) as compared to four-row scanners, thus in gen- respectively (Table 1). All patients receive 150 mL of a non-
eral leading to better image quality. In addition, as a result of ionic contrast medium (Ultravist, Schering Inc.) through
the favorable combination of high contrast media flow rates an 18G intravenous antecubital catheter infused with a flow
 CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY 291

rate of 3.5 mL/s. Start delay is calculated using Test Bolus Criteria that need to be analyzed are: (a) the atherosclerotic
Technique with a region of interest (ROI) placed in the plaque load of each single coronary artery segment; (b) the
ascending aorta (30 mL contrast medium at a flow rate of 3.5 composition of the individual plaques; and (c) the epicardial/
mL/s) (Table 1). intramural course of the vessel.
For the 16-row scanner, scanning parameters are 420 ms Plaque load and plaque composition are best determined
rotation time, 16 1.5-mm collimation and 5.7-mm table feed on transverse scans and MPR. The cardiac course of the coro-
per rotation for the plain series, 16 0.75 mm slice collima- nary vessels can be additionally assessed on 3D reformations,
tion and 2.8-mm table feed per rotation for the contrast-en- which also serve for surgical morning-round demonstrations
hanced series, respectively (Table 1). In these patients, only and intra-operative orientation for the surgeon.
120 mL of a nonionic contrast medium is infused with a flow To grade the segmental plaque load, we use the classifica-
rate of 4.0 mL/s. Using Care Bolus Technique, no additional tions of the American Heart Association (AHA), which subdi-
contrast medium is necessary to determine the start delay. vide the coronary artery territory into 15 segments and
After contrast medium injection scanning starts automatically distinguish between six different degrees of atherosclerosis
as soon as a certain trigger point is reached (usually 160 (26): (I) irregular wall outline with <25% stenosis; (II) slight
Hounsfield units [HU]) in a ROI that is placed in the ascend- stenosis (2550%); (III) moderate stenosis (5174%); (IV)
ing aorta (Table 1). hemodynamically relevant stenosis (7589%); subtotal steno-
IMAGE RECONSTRUCTION sis (9099%) and (V) vascular occlusion (100%). The prevail-
ing degree of occlusion is ascertained by using an automated
Image reconstruction is performed using retrospective ECG distance-measuring tool (Plus4VZ Wizzard, Siemens).
gating, a technique that allows continuous image reconstruc- Regarding the composition of the plaque, it is classified as
tion from volume data sets during any phase of the cardiac calcified and non-calcified plaques. Plaques with a mean
cycle (17,18). Reconstruction parameters for 4-row and 16- attenuation >130 HU are graded as calcified, while plaques
row examinations are 220-mm FOV, kernel B35, a medium with a mean attenuation <130 HU are graded as noncalcified
soft-tissue kernel, 3-mm effective slice thickness, and 1.5-mm (27). Calcified plaques were identified on plain scans,
increment for the plain series (Table 1). For the contrast- noncalcified plaques on contrast-enhanced scans.
enhanced series, image reconstruction is done on the 4-row The cardiac course of each coronary artery is assessed by
scanner with 1.25-mm effective slice thickness and 0.6-mm distinguishing a possible epicardiac or intramural course of the
increment, with 1.0-mm effective slice thickness and 0.5-mm prevailing arterial segment (AHA 115).
increment on the 16-row scanner respectively (Table 1). Finally, taking into consideration the results of all three ref-
For image reconstruction on both scanners, the adaptive ormation techniques, a conclusive recommendation is given
cardiac volume reconstruction algorithm (ACV) is used, regarding which coronary segment is best suited for allocation
which is provided with the software (19). Taking into account of the distal bypass-anastomosis, and the coronary diameter
basic cardiac physiology, sufficient image reconstruction in this region is calculated. In general only segments showing
seems to be feasible solely if performed between the late sys- no calcified plaques and a rather near-surface course are
tole (i.e., ascending t wave) and the late diastole (i.e., begin- regarded as suited for distal bypass touchdown.
ning of p wave) (20,21). Each data set is consequently
reconstructed at multiple time points within this interval, dif- RESULTS OF ACTUAL STUDIES
fering from each other by 50 ms. Owing to a decrease in the Regarding the value of 4- and 16-row MDCT before
length of the t wavep wave interval at increasing heart rates TECABG, until now only two studies have been performed.
(HRs), the number of possible reconstruction time points per Both are briefly described in the following.
patient ranges between 10 (low HR) and 6 (elevated HR). It
4-ROW MDCT
is favorable at each time point to perform an antegrade and
absolute (i.e., in milliseconds) image reconstruction in rela- Methods
tion to the R peak (21). Subsequently for each patient and each In a comparative study performed between July 2000 and
main coronary artery (RCA, LCA, LCx) separately, one spe- January 2002, 36 consecutive patients (22 males and 14
cific reconstruction should be determined showing the least females) scheduled for TECABG underwent both invasive
motion artifacts, therefore allowing proper image interpreta- CAG and four-row MDCT before surgery. CAG was per-
tion (2225). formed in a routine manner to identify potential candidates
for surgery, followed by an additional MDCT examination
IMAGE REFORMATION AND EVALUATION less than 48 h before the operation. The mean time between
MDCT image evaluation usually is done on both transverse CAG and MDCT amounted to 17 d (range 7 to 23 d). The
scans and secondary reformations, i.e., MPR and 3D reforma- patients average age was 60.1 yr (range 41 to 78 yr), their
tions. We analyze transverse scans, MPR, and 3D reformations mean heart rate amounted to 62.9 bpm (range 47 to 86 bpm).
on a separate workstation (VZ Leonardo, Siemens). Three- All patients presented with single-vessel coronary heart dis-
dimensional reformations are produced using volume-render- ease. Thirty-two patients had a hemodynamically relevant
ing technique. Transverse scans and MPR are displayed on a (>75%) stenosis of the proximal LCA, one patient presented
512 512 matrix, three-dimensional reformations on a 256 with a >75% stenosis of both the proximal LCA and the first
256 matrix. diagonal branch, and four with a total occlusion of the proxi-
292 HERZOG, DOGAN, AND VOGL

mal RCA. Thirty-four operations were performed on the ar- ment for both techniques was calculated by means of a
rested heart using a heart-lung machine (1); in two patients, binominal confidence interval for . The agreement between
surgery was performed on the beating heart in total off-pump the methods and investigators was interpreted according to the
coronary artery bypass (TOPCAB) technique (2). Assess- value or value respectively, as poor (<0.20), fair (0.21
ment criteria for CAG and 4-row MDCT were: visibility and 0.40), moderate (0.410.60), good (0.610.80), very good
cardiac course of coronary arteries, localization and degree (0.810.90), or excellent (0.911.00). A 95% CI, calculated by
of stenoses, composition of atherosclerotic plaques, and vas- a standard method, was assigned to each calculated and
cular diameter at the site of anastomosis. Finally, both tech- value.
niques were used separately to recommend an appropriate Sensitivity, specificity, positive and negative predictive
site for distal bypass anastomosis. All MDCT evaluations value of MDCT and coronary angiography in the detection of
were performed according to the criteria described in detail hemodynamically relevant stenoses was determined by subdi-
(see Preoperative MDCT Evaluation) and were analyzed vision of AHA groups IVI into two different groups: (1)
by two independent observers. All results were calculated patients without hemodynamically relevant stenoses (< 75%),
relative to the prevailing results from CAG and surgery. i.e. AHA groups I, II, and III; and (2) patients with hemody-
Coronary angiograms had been performed in different tech- namically relevant stenoses (>75%), i.e., AHA groups IV, V,
nical systems using the Judkins technique. At least four views and VI. Values for MDCT were calculated relative to all 15
of the left and two views of the right coronary artery system AHA segments (reference standard: coronary angiograms) as
were analyzed by two observers in consensus, who were both well as relative to those segments that were intraoperatively
trained in this technique. In order to avoid recall bias, neither explored (reference standard: surgery). Values for invasive
of them had any knowledge of the CT results, and evaluation angiograms were calculated only with surgery serving as the
was undertaken only on data sets that had previously been reference standard.
blinded. Sensitivity and specificity as well as positive and negative
On angiograms, plaque composition and visibility of coro- predictive values in the detection of calcified plaques were
nary artery segments on CAG was described as either calcified calculated for MDCT and CAG relative to the surgical find-
or noncalcified, visible or nonvisible, respectively. Calcified ings. The agreement between investigators for this evaluation
plaques were searched for on both plain and contrast-enhanced was assessed by means of the statistic as described above. A
series. 95% CI, calculated by a standard method, was assigned to the
The prevailing degree of stenosis and the segmental diam- calculated value.
eter at the planned site of distal anastomosis was determined The validity of both imaging methods in the correct identi-
using a stenosis grading tool with automatic distance and scale fication of a segments cardiac courseon the epicardiac
calibration (Osiris, Digital Imaging UnitUHGE). surface, deep within the fatty tissue, or within the myocar-
As with MDCT, finally a conclusive recommendation was diumwas calculated as proportional to the actual surgical
given regarding the optimal place for distal bypass-anastomo- findings.
sis, and the segmental diameter in this region was calculated. The correspondence rate of MDCT and coronary angiogra-
All results were intraoperatively evaluated by one of the phy regarding the allocation of distal bypass-anastomosis and
surgeons, using 3D reformations and angiograms for compari- the segmental diameter at the site of anastomosis was also cal-
son. As a result of a restricted field of vision, only those seg- culated proportional to the surgical results. A segmental diam-
ments were compared that could be visualized through the eter was regarded to be of equal size if the measured difference
operation microscope, i.e., the stenotic/occluded segment itself, amounted to less than 1 mm. Bypass allocation was rated as
the segments before and behind this occlusion/stenosis, any correct if segmental allocation corresponded to surgery. Results
surrounding side branches, and possible crossing veins. All concerning the segments cardiac course, the allocation of
differences as well as conformities were noted and additionally bypass-anastomosis, and the segmental diameter at the site of
documented on videotape, thus allowing a second inspection anastomosis were obtained by both observers in consensus.
after the operation. Results
The number of coronary segments that could be evaluated On 4-row MDCT scans, 80.4% (434/540) of all coronary
on MDCT scans and coronary angiograms was determined in segments could be evaluated by observer 1 and 78.3% (423/
proportion to both the total number of segments and the seg- 540) by observer 2. The value between investigators
ments that could be visualized during surgery. Agreement amounted to 0.870 (CI: 0.786, 0.954), thus showing a very good
between investigators for this evaluation was calculated by agreement.
means of the statistic, interpreting the results according to the On the average, 100% (108/108) of all proximal segments
value as poor (<0.20), fair (0.210.40), moderate (0.410.60), (1, 5, and 11), 94.5% (68/72) of the medial RCA (segments 2, 3),
good (0.610.80), very good (0.810.90), or excellent (0.91 90.3% (65/72) of the medial LCA (segments 6, 7), 83.3%
1.00). A 95% confidence interval (CI), calculated by a standard (30/36) of the medial LCx (13), 91.7% (33/36) of the distal
method, was assigned to the calculated value. Possible differ- RCA (segment 4), 80.6% (29/36) of the distal LCA (segment
ences between both methods were tested for significance using 8), and 55.6% (20/36) of the distal LCx (segment 15) were
a comparison of Poisson frequencies. classified as visible. Side branches of the LCA (segments 9, 10)
Regarding the detection and grading of atherosclerosis and were regarded as visible in 70.8 (51/72), those of the LCx (seg-
with reference to all 15 AHA segments, a coefficient of agree- ments 12, 14) in 41.7% (30/72) (Table 2).
 Table 2
Average Visibility of Coronary Arteries (%):
Correlation Among 4-Row Multidetector-Row CT (MDCT), 16-Row MDCT Coronary Angiography, and Surgical Findings

4-row MDCT 16-row MDCT Angiography 4-row MDCT/Surgery 16-row MDCT/surgery Angiography/surgery
% absolute % absolute % absolute % absolute % absolute % absolute

RCA proximal (S 1)1 100 (36/36) 100 (12/12) 100 (36/36) 100 (5/5) 100 (5/5)
medial (S 23) 94.5 (68/72) 83.3 (10/12) 97.2 (70/72) 50 (5/10) 60 (6/10)
distal (S 4) 91.7 (33/36) 83.3 (10/12) 94.4 (34/36)
LCA proximal (S 5) 100 (36/36) 100 (12/12) 100 (36/36) 100 (22/22) 100 (10/10) 100 (22/22)
medial (S 67) 90.3 (65/72) 100 (24/24) 90.3 (65/72) 88.5 (54/61) 100 (24/24) 90.2 (55/61)

293
distal (S 8) 80.6 (29/36) 87.5 (21/24) 91.7 (33/36) 78.1 (25/32) 91.7 (11/12) 90.6 (29/32)
diagonal2 (S 910) 70.8 (51/72) 83.3 (20/24) 80.6 (58/72) 67.2 (43/64) 91.4 (10/14) 85.9 (55/64)
LCx proximal (S 11) 100 (36/36) 100 (12/12) 100 (36/36)
medial (S 13) 83.3 (30/36) 83.3 (10/12) 100 (36/36)
distal (S 15) 55.6 (20/36) 58.3 (7/12) 100 (36/36)
marginalc (S 12/14) 41.7 (30/72) 66.7 (8/12) 98.6 (71/72)
total 80.4 (434/540) 87.7 (158/180) 94.6 (511/540) 79.4 (154/194) 91.7 (55/60) 88.7 (172/194)

RCA, right coronary artery; LCA, left coronary artery; LCx, left circumflex artery.
aBrackets indicate AHA segments (S 115).
CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY

bDiagonal branches.
cMarginal branches.
293
294 HERZOG, DOGAN, AND VOGL

On coronary angiograms, 94.6% (511/540) of all segments tions: two due to a low experience curve at the beginning of the
could be evaluated (Table 2). Segments which were not dis- investigation, three because the intraoperative situation
played properly were all located behind total vascular occlusions. required a different bypass technique (switch to jump-graft
Considering only segments that were intraoperatively technique in one case, venous bypass instead of IMA-bypass
explored (194/540), on 4-row MDCT scans, 79.4% (154/194) grafting in two cases), and four because of incorrect identifica-
could be evaluated by observer 1 and 76.8% (149/194) by tion of either fatty tissue (1), an intramural course (e.g., myo-
observer 2. The value between investigators amounted to cardial bridging) (1) or the vascular segments themselves (2)
0.925 (CI: 0.784, 1.000), thus representing an excellent agree- (Table 4).
ment. On angiograms, 88.7% (172/194) were visible to the CAG resulted in 11 wrong allocations: two owing to a low
observer (Table 2). Both methods showed no significant differ- experience curve at the beginning, three because intraopera-
ences in overall visualization, either for observer 1 (p = 0.346) tive findings required a different surgical approach, and six
or for observer 2 (p = 0.219). because the target vessel intraoperatively either was hidden
The coefficient of agreement () for both techniques regard- deep inside the epicardial fatty tissue (3) or showed an intra-
ing detection and grading of atherosclerosis (AHA IVI) mural course (3) (Fig. 5).
amounted to 0.759 (CI: 0.721, 0.795) for observer 1 and 0.731 Bridging of coronary artery segments through either myo-
(CI: 0.691, 0.768) for observer 2, thus showing a good agree- cardium or epicardial fatty tissue thus was identified more re-
ment for both. On the average, 4-row MDCT overestimated liably on MDCT scans than on angiograms (Fig. 5): 80% (4/5)
17.6% (29/165) and underestimated 6.5% (11/165) of all as compared to 20% (1/5) for myocardial bridging, 66.7% (2/3)
stenoses. as compared to 0% (0/3), respectively, for bridging through
Considering all 15 AHA segments, sensitivity and specific- epicardial fatty tissue (Table 4).
ity for MDCT and observer 1 in the identification of hemody- The segmental diameter at the site of distal anastomosis was
namically relevant stenoses (>75%) amounted to 76.4% (CI: measured correctly on MDCT scans in 72.0% (18/25), on
62.9%, 86.7%) (42/55) and 99.6% (CI: 98.5%, 99.9%) (483/ angiograms in 80.0% (20/25) (Table 4). Measurements were
485). For observer 2 a sensitivity of 70.9% (CI: 57.1%, 82.4%) undertaken only in the last 25 patients.
(39/55) and a specificity of 98.4% (CI: 96.8%, 99.3%) (477/ On MDCT scans, seven segments were not assessed prop-
485) was obtained. Positive predictive value (PPV) was 95.5% erly: three due to moving artifacts, three because of missing
(CI: 84.5, 99.4) (42/44) for observer 1 and 82.9% (CI: 69.2%, contrast media enhancement due to a proximal total vascular
92.4%) (39/47) for observer 2. Negative predictive value occlusion, and one because of underestimation of the actual
(NPV) amounted to 97.4% (CI: 95.6%, 98.6%) (483/496) for diameter.
observer 1 and to 96.8% (CI: 94.8%, 98.1%) (477/493) for On angiograms, all five segments that were not visualized
observer 2 (Table 3). The value between investigators was adequately were located behind a total vascular occlusion.
0.9518 (CI: 1.000, 0.867), thus equivalent to an excellent 16-ROW MDCT
agreement. This study is still in progress, and concentrates on the poten-
Considering only those segments that were intraopera- tial benefit of 16-row MDCT in patients scheduled for
tively explored from the endoluminal side (n = 37), hemody- TECABG procedures on the beating heart (2) and being oper-
namically relevant stenoses (>75%) were identified on 4-row ated on with assistance of a newly introduced, fourth swivel
MDCT scans with 91.9% sensitivity (34/37) (CI: 78.1, 98.0) arm. In these procedures, bypass anastomosis can be achieved
by observer 1 and with 89.2% sensitivity (33/37) (CI: 74.6, only through short-time ligature of the prevailing coronary
97.0) by observer 2. Sensitivity for coronary angiograms artery. Since exploration and suture time must not exceed 20 min
amounted to 100% (37/37) (Table 3). Specificity, PPV, NPV, in order to prevent any ischemic damage to the myocardium,
and were not calculated because of a lack of negative con- fast and correct identification of both the prevailing vessel and
trols. the suited region for anastomosisi.e., a rather superficially
Both observers detected calcified plaques on 4-row MDCT located segment which is definitively free of calcified
scans with 100% sensitivity (CI: 84.7%, 100%) (18/18). For plaquesis highly important for the postoperative outcome. In
CAGs, sensitivity was 83.3% (CI: 58.6%, 96.4%) (15/18) addition, a fourth swivel arm distinctly aggravates correct plan-
(Fig. 4, Table 3). Specificity was 100% (CI: 85.4, 100%) (19/19) ning of the port placement for the endoscopic instruments: not
for both techniques and both MDCT observers respectively. only must each patients anatomy be considered individually,
For 4-row MDCT, PPV (18/18) (CI: 84.7%, 100%) and NPV but also any intraoperative collision of the arms must be
(19/19) (CI: 85.4%, 100%) amounted to 100% for each. For avoided.
coronary angiography, PPV was 100% (CI: 81.9%, 100%) (15/15) Methods
and NPV 86.4% (CI: 65.1%, 97.1%) (19/22) (Table 3). The The study was initiated in September 2002. Until now, 12
value between MDCT investigators was 1.00, thus showing an consecutive patients (9 males and 3 females) were included,
excellent agreement. undergoing both invasive coronary angiography (CA) and
The appropriate allocation for distal bypass anastomosis was 16-row MDCT. The mean time between CA and MDCT
correctly identified on 4-row MDCT scans in 75.1% (28/37), amounted to 21 d (range 9 to 30 d). The patients average age
on coronary angiograms in 70.3% (26/37) (Table 4, Fig. 5). was 58.1 yr (range 52 to 69 yr), their mean heart rate amounted
4-row MDCT led to nine wrong preoperative bypass alloca- to 68.3 bpm (range 56 to 112 bpm). Eight patients presented
 Table 3
4-row Multidetector-Row CT (MDCT) Correlated to Coronary Angiography and Surgery
(sensitivities, specificities, positive [PPV], and negative [NPV] predictive values for both MDCT observers listed separately)

Sensitivity Specificity PPV NPV

% absolute % absolute % absolute % absolute

4-row MDCT high grade stenoses (>75%) Observer 1 76.4 (42/55) 99.6 (483/485) 95.5 (42/44) 97.4 (467/496)
(considering AHA segments 115 / Observer 2 70.9 (39/55) 98.4 (477/485) 82.9 (39/47) 96.8 (477/493)
reference standard: angiography)
high grade stenoses (>75%) Observer 1 91.9 (34/37) n.c. n.c. n.c. n.c. n.c. n.c.
(considering only segments explored intraop. / Observer 2 89.2 (33/37) n.c. n.c. n.c. n.c. n.c. n.c.
reference standard:surgery)
calcified plaques (>130 HU) Observer 1 100 (18/18) 100 (19/19) 100 (18/18) 100 (19/19)
(considering only segments explored intraop. / Observer 2 100 (18/18) 100 (19/19) 100 (18/18) 100 (19/19)
reference standard: surgery)

16-row MDCT high grade stenoses (>75%) Observer 1 80.0 (12/15) 96.8 (153/158) 70.6 (12/17) 98.1 (153/156)

295
(considering AHA segments 115 / Observer 2 73.3 (11/15) 94.9 (150/158) 61.9 (11/19) 98.7 (150/154)
reference standard: angiography)
high grade stenoses (>75%) Observer 1 91.7 (11/12) n.c. n.c. n.c. n.c. n.c. n.c.
(considering only segments explored intraop. / Observer 2 91.7 (11/12) n.c. n.c. n.c. n.c. n.c. n.c.
reference standard:surgery)
calcified plaques (>130 HU) Observer 1 100 (13/13) 100 (5/5) 100 (13/13) 100 (5/5)
(considering only segments explored intraop. / Observer 2 100 (13/13) 100 (5/5) 100 (13/13) 100 (5/5)
reference standard: surgery)
Angiography
high grade stenoses (>75%) 100 (37/37) n.c. n.c. n.c. n.c. n.c. n.c.
CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY

(considering only segments explored intraop. /

reference standard: surgery)
calcified plaques (>130 HU) 61.5 (8/13) 100 (5/5) 100 (8/8) 50 (5/10)
(considering only segments explored intraop./
reference standard: surgery)

n.c., Not calculated owing to lack of negative controls.

295
296 HERZOG, DOGAN, AND VOGL

Fig. 4. Comparison between invasive angiography (left), mirror-inverted 15 right anterior oblique projection and multidetector-row CT
(MDCT) (right), left anterior oblique. Both techniques correctly identified the 100% stenosis (white arrowhead) within segment 6 of the left
coronary artery (LCA), but only MDCT revealed the marked calcifications of segment 7 of the LCA (delineated by white arrows), rendering
bypass grafting in this region rather difficult. Based upon this information the surgical access was adapted in order to reach the more distal
segment 8 and bypass grafting was performed successfully. D1 designates the first diagonal branch, M1 the first marginal branch, and LCX
the left circumflex artery. reprinted from ref. 35, with permission from the Radiological Society of North America.

Table 4
Preoperative Assessment of Segmental Diameter, Allocation
of Distal Bypass, and Detection of Morphological Traps (comparison between
4- and 16-row multidetector-row CT [MDCT] and invasive angiography)

Correct rating

% absolute

4-row MDCT
segmental diameter 72.0 (18/25)
(reference standard: surgery)
bypass allocation 75.1 (28/37)
(reference standard: surgery)
myocardial bridging 80 (4/5)
(reference standard: surgery)
epicardiac fatty tissue 66.7 (2/3)
(reference standard: surgery)
16-row MDCT
bypass allocation 91.7 (11/12)
(reference standard: surgery)
myocardial bridging 100 (2/2)
(reference standard: surgery)
epicardiac fatty tissue 100 (3/3)
(reference standard: surgery)

Coronary
angiography
segmental diameter 80.0 (20/25)
(reference standard: surgery)
(continued)
 CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY 297

Fig. 5. Comparison between invasive angiography (upper row), mirror-inverted 15 right anterior oblique projection and multidetector-row
CT (MDCT) (lower row), left anterior oblique. Both techniques correctly identified the 80% stenosis (white arrowhead) within segment 6 of
the left coronary artery (LCA), but only MDCT revealed the intramural course of segment 7 of the LCA (delineated by the white arrows). Based
upon this information, the surgical access was altered in order to reach the more distal segment 8 and bypass grafting was performed
successfully. Note the vein (V) that accompanies the intramural course of the LCA on the myocardial surface. D1 designates the first diagonal
branch, M1 the first marginal branch, and LCX the left circumflex artery. Reprinted from ref. 35, with permission from the Radiological Society
of North America.

Table 4 (Continued)
Preoperative Assessment of Segmental Diameter, Allocation of Distal Bypass,
and Detection of Morphological Traps (comparison between 4- and 16-row
multidetector-row CT [MDCT] and invasive angiography)

Correct rating

% absolute

bypass allocation 70.3 (26/37)

(reference standard: surgery)
myocardial bridging 20 (2/5)
(reference standard: surgery)
epicardiac fatty tissue 0 (0/3)
(reference standard: surgery)
298 HERZOG, DOGAN, AND VOGL

with a hemodynamically relevant (>75%) stenosis of the Sixteen-row MDCT identified bridging of coronary seg-
proximal LCA, three with an additional stenosis of the first ments through either myocardium (2/2) or epicardial fat (3/3)
diagonal branch, and one with a stenosis of the LCA, the first in all cases and correctly determined the distal site for bypass
diagonal, and the first obtuse marginal branch. Assessment touchdown in 91.6% (11/12).
criteria for CAG and 16-row MDCT were visibility and car-
diac course of coronary arteries, localization and degree of CONCLUSION AND FUTURE DIRECTIONS
stenoses, composition of atherosclerotic plaques, and vascu- TECABG is a new method for coronary artery bypass graft-
lar diameter at the site of anastomosis. Finally, both tech- ing, which can be used in selected patients to achieve single or
niques were interpreted separately to recommend an double arterial bypass grafting with internal thoracic arteries
appropriate site for distal bypass anastomosis. All 16-row (1,3,28). Intrathoracic orientation and identification of the
MDCT examinations and evaluations were performed accord- target vessel is more challenging than in open procedures.
ing to the methods in detail described above under the heading Because of the lack of tactile feedback, it is difficult to evaluate
MDCT Evaluation. All results were calculated relative to the quality of the target vessel just by visual control. The suc-
the prevailing results from CAG and surgery, which were cess rate of TECABG on the arrested heart (intention to treat a
obtained from the same observers and in the same manner as patient in closed chest technique without conversion to
described under the subheading Four-Row MDCT: Meth- minithoracotomy) varies between 80% and 95 % depending on
ods. Statistical calculations were also done following the factors such as patient selection and experience of the team (4).
approach used for 4-row MDCT. However, mainly as a result of problems with the identification
Results of the target vessel, inadequate exposition, or cases in which
Preliminary results show that 16-row MDCT properly successful realization of endoscopic anastomoses does not
displayed 87.7% (158/150) of all coronary segments. 100% seem feasible, many TECABG procedures intraoperatively
(36/36) of all proximal segments (1, 5, and 11), between 83.3% must be converted to a left side minithoracotomy (MIDCABG).
(segments 2, 3, and 13) and 100 % (segments 6, 7) of all medial In most of these cases, the target vessel is either confused with
segments, between 58.3% (segment 15) and 87.5% (segment 8) neighboring vessels, hidden deep inside the epicardial fatty
of all distal segments, 83.3% of all diagonal (segment 9, 10), tissue, bridged by myocardial tissue, or heavily calcified. Our
and 66.7% of all marginal branches (segment 12, 14) were results show that MDCT can identify such morphological traps
classified as visible (Table 2). very often and therefore could improve the success rate of
The value between investigators amounted to 0.79 (CI: TECABG procedures. First, the selection process of patients is
0.723, 1.000). The coefficient of agreement () for 16-row facilitated, thus rendering possible preoperative switching
MDCT and CA regarding detection and grading of atheroscle- towards alternative surgical techniques. Second, virtual evalu-
rosis (AHA IVI) amounted to 0.793 (CI: 0.762, 0.813) for ation of the operative site using 3D reformations before surgery
observer 1 and to 0.752 (CI: 0.736, 0.775) for observer 2, thus allows exact identification of the target vessel, its morphologi-
showing a good agreement for both. On the average, MDCT cal appearance, and relation to surrounding structures. The
underestimated 24.1% (7/29) and overestimated 13.8% (4/29) surgeon may preoperatively even simulate the surgical
of all stenoses. approach and the exploration of the vessel. Although our results
Hemodynamically relevant stenoses were identified with indicate that invasive coronary angiography more sensitively
16-row MDCT with 80.0% (CI: 51.9%, 95.7%) (12/15) sensi- identified and displayed vascular occlusions, MDCT neverthe-
tivity by observer 1, 73.3% (CI: 44.9%, 92.2%) (11/15) sensi- less often detected morphological traps that were not visual-
tivity respectively by observer 2 (Table 3). PPV was 70.6% (CI: ized with CA. In several patients, MDCT demonstrated that the
44.0, 89.7) (12/17) for observer 1 and 61.9% (CI: 38.4%, target vessel was hidden deep inside the epicardial fat or showed
81.9%) (11/19) for observer 2. NPV amounted to 98.1% (CI: an intramural course, and therefore clearly outperformed inva-
94.5%, 99.6%) (153/156) for observer 1 and to 98.7% (CI: sive coronary angiography, which identified these characteris-
95.3%, 99.8%) (150/154) for observer 2 (Table 3). The value tics in fewer patients. The distal bypass touchdown segment
between investigators was 0.831 (CI: 1.000, 0.782), thus often could be predicted correctly only by MDCT, whereas
equivalent to a good agreement. CAG failed to do so. In two patients, only 3D MDCT reforma-
Considering only those segments that were intraoperatively tions showed the fatal combination of a prominent superficial
explored from the endoluminal side (n = 12), hemodynamically diagonal branch and an intramural LCA. In such cases, acci-
relevant stenoses (>75%) were identified on 16-row MDCT dental grafting of the wrong vessel (diagonal branch versus
scans with 91.7% sensitivity (11/12) (CI: 61.5, 99.8) by both LCA) is a common risk of endoscopic surgery, because spatial
observers. orientation is difficult, the field of view restricted, and manual
Both observers detected calcified plaques on 16-row MDCT palpation of the vessel not possible (Fig. 6). Moreover, the
scans with 100% sensitivity (CI: 79.4%, 100%) (13/13) (Table composition of atherosclerotic plaques often cannot be pre-
3) and 100% specificity (CI: 54.9, 100%)(5/5). PPV (13/13) dicted sufficiently on angiograms (5). Thus, calcified plaques
(CI: 79.4%, 100%), and NPV (5/5) (CI: 54.9%, 100%) for both were detected with a markedly lower sensitivity with CAG than
observers also amounted to 100% (Table 3). The value on MDCT scans, which always identified these plaques in
between MDCT investigators was 1.00, thus showing an excel- 100%. If collateral circulation is absent, behind a vascular
lent agreement. occlusion on angiograms neither the calcified nor the
 CHAPTER 27 / MDCT FOR PLANNING BYPASS SURGERY 299

Fig. 6. Comparison between invasive angiography (left), mirror-inverted 15 right anterior oblique projection and multidetector-row CT
(MDCT) (right), left anterior oblique, showing how different coronary segments during totally endoscopic coronary artery bypass grafting
procedures may easily be confused as a result of the restricted field of view and lack of manual palpation. The patientthe same as shown in
Fig. 2presents a large diagonal branch (D1) running nearly parallel to the left coronary artery (LCA). Both vessels are hidden deep inside
the epicardiac fat (compare Fig. 2), and the LCA additionally nestles against the pulmonary trunk. Thus, in cases like this, accidental grafting
of the diagonal branch 1 is a common risk. Three-dimensional MDCT reformations provided additional morphological information of the
surrounding area, and therefore helped the surgeon to gain a better orientation. D1 designates the first diagonal branch, M1 the first marginal
branch, and LCX the left circumflex artery. Reprinted from ref. 35, with permission from the Radiological Society of North America.

noncalcified plaque will be displayed. MDCT on the contrary However, future developments of surgical devices will make
will at least detect calcified lesions very precisely (2934). The preoperative MDCT evaluation of the coronary arteries even
results shown above support this assertion. However, such cal- more necessary. Thus, new prototypes for TECABG proce-
cifications very often render successful bypass grafting impos- dures now link transparent flat panels that may be swung over
sible in the region of choice, and thus either prolong operation the patients body with the TECABG main console and slave
times or imply intraoperative switching towards alternative unit. This combination allows pre- and intraoperative virtual
surgical procedures. projection of the patients heart and, if multiple phase image
Four-row MDCT in comparison to 16-row MDCT showed reconstruction has previously been performed, even of the
no differences in the detection of calcified plaques. However, beating heart. With this technique, not only will preopera-
the latter provided increased in-plane and temporal resolu- tive planning of port placement be markedly facilitated, but
tion, and thus showed less susceptibility to motion artifacts, also intraoperative monitoring of the hearts position, the
allowed more coronary artery segments to be visualized, and intrathoracic course of the IMA, the localization of the surgical
thus facilitated preoperative allocation of the suited distal instruments, the position of the target vessel, and the suited
bypass touchdown segment. However the data for 16-row bypass touchdown segment.
scanners are only preliminary and need to be validated through At the moment, MDCT technique nevertheless is able to pro-
larger studies. vide extended information on the coronary target site, and there-
In general MDCT, because of several profound restric- fore should be regarded as an ideal additive planning tool for
tions, must still be regarded solely as a complementary but not complex minimally invasive procedures such as TECABG or
alternate technique to invasive angiography. Despite a tem- MIDCABG. Combining the advantages of invasive coronary
poral resolution of up to 105 ms, it remains highly susceptible angiographye.g., high temporal and spatial resolution, blood
to motion artifactsparticularly at elevated heart ratesand flow information, assessment of functional parameters and col-
still offers only poor visualization of coronary artery seg- lateral circulationwith the gross morphological superiority of
ments <1 mm. MDCT thus can be of great benefit to the operative outcome.
300 HERZOG, DOGAN, AND VOGL

REFERENCES 20. Luisada AA, MacCanon DM. The phases of the cardiac cycle. Am
Heart J 1972;83:705711.
1. Falk V, Diegeler A, Walther T, et al. Endoscopic coronary artery
21. Herzog C, Abolmaali N, Balzer JO, et al. Heart-rate-adapted image
bypass grafting on the beating heart using a computer enhanced
reconstruction in multidetector-row cardiac CT: influence of physi-
telemanipulation system. Heart Surg Forum 1999;2:199205.
ological and technical prerequisite on image quality. Eur Radiol
2. Falk V, Diegeler A, Walther T, Jacobs S, Raumans J, Mohr FW.
2002;12:26702678.
Total endoscopic off-pump coronary artery bypass grafting. Heart
22. Fischbach R, Opitz C, Juergens KU, et al. Einfluss des Rekonstruk-
Surg Forum 2000;3:2931.
tionsintervalls auf die Qualitt der Koronariendarstellung mit
3. Dogan S, Aybek T, Westphal K, Mierdl S, Moritz A, Wimmer-
Mehrschicht-Spiral CT und retrospektiver EKG-Synchronisation.
Greinecker G. Computer-enhanced totally endoscopic sequential
Fortschr Roentgenstr 2001;173:54.
arterial coronary artery bypass. Ann Thorac Surg 2001;72:610611.
23. Achenbach S, Ropers D, Holle J, Muschiol G, Daniel WG, Moshage
4. Mohr FW, Falk V, Diegeler A, et al. Computer-enhanced robotic
W. In-plane coronary arterial motion velocity: measurement with
cardiac surgery: experience in 148 patients. J Thorac Cardiovasc
Surg 2001;121:842853. electron-beam CT. Radiology 2000;216:457463.
5. Chocron S, Etievent JP, Dussaucy A, et al. Inter-observer reliability 24. Hong C, Becker CR, Huber A, et al. ECG-gated reconstructed multi-
in the interpretation of coronary angiograms. Eur J Cardiothorac detector row CT coronary angiography: effect of varying trigger
Surg 1996;10:671675. delay on image quality. Radiology 2001;220:712717.
6. Dogan S, Graubitz K, Aybek T, et al. How safe is the port access 25. Kopp AF, Ohnesorge B, Flohr T, et al. [Cardiac multidetector-row
technique in minimally invasive coronary artery bypass grafting? CT: first clinical results of retrospectively ECG-gated spiral with
Ann Thorac Surg 2002;74:15371543; discussion 1543. optimized temporal and spatial resolution]. Rofo Fortschr Geb
7. Kirklin JK, Westaby S, Blackstone EH, Kirklin JW, Chenoweth DE, Rontgenstr Neuen Bildgeb Verfahr 2000;172:429435.
Pacifico AD. Complement and the damaging effects of cardiopul- 26. ACC/AHA guidelines for cardiac catheterization and cardiac cath-
monary bypass. J Thorac Cardiovasc Surg 1983;86:845857. eterization laboratories. American College of Cardiology/American
8. Talwalkar NG, Cooley DA. Minimally invasive coronary artery Heart Association Ad Hoc Task Force on Cardiac Catheterization.
bypass grafting: a review. Cardiol Rev 1998;6:345349. J Am Coll Cardiol 1991;18:11491182..
9. Duhaylongsod FG. Minimally invasive cardiac surgery defined. 27. Kopp AF, Schroeder S, Baumbach A, et al. Non-invasive characteri-
Arch Surg 2000;135:296301. sation of coronary lesion morphology and composition by multislice
10. Vassiliades TA, Jr., Rogers EW, Nielsen JL, Lonquist JL. Mini- CT: first results in comparison with intracoronary ultrasound. Eur
mally invasive direct coronary artery bypass grafting: intermediate- Radiol 2001;11:16071611.
term results. Ann Thorac Surg 2000;70:10631065. 28. Aybek T, Dogan S, Andressen E, et al. Robotically enhanced totally
11. Groh M, Grossi EA. Multivessel coronary bypass grafting with endoscopic right internal thoracic coronary artery bypass to the right
minimal access using cardiopulmonary bypass. Curr Cardiol Rep coronary artery. Heart Surg Forum 2000;3:322324.
1999;1:331334. 29. Gutfinger DE, Leung CY, Hiro T, et al. In vitro atherosclerotic plaque
12. Wimmer-Greinecker G, Matheis G, Dogan S, et al. Complications of and calcium quantitation by intravascular ultrasound and electron-
port-access cardiac surgery. J Card Surg 1999;14:240245. beam computed tomography. Am Heart J 1996;131:899906.
13. Glower DD, Komtebedde J, Clements FM, Debruijn NP, Stafford- 30. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium
Smith M, Newman MF. Direct aortic cannulation for port-access measurement: agreement of multirow detector and electron beam
mitral or coronary artery bypass grafting. Ann Thorac Surg 1999;68: CT. AJR Am J Roentgenol 2001;176:12951298.
18781880. 31. Schmermund A, Baumgart D, Gorge G, et al. Coronary artery
14. Boehm DH, Reichenspurner H, Gulbins H, et al. Early experience calcium in acute coronary syndromes: a comparative study of elec-
with robotic technology for coronary artery surgery. Ann Thorac tron-beam computed tomography, coronary angiography, and intra-
Surg 1999;68:15421546. coronary ultrasound in survivors of acute myocardial infarction and
15. Kappert U, Schneider J, Cichon R, et al. Wrist-enhanced instrumen- unstable angina. Circulation 1997;96:14611469.
tation: moving toward totally endoscopic coronary artery bypass 32. Schroeder S, Kopp AF, Baumbach A, et al. Non-invasive characteri-
grafting. Ann Thorac Surg 2000;70:11051108. sation of coronary lesion morphology by multi-slice computed
16. Loulmet D, Carpentier A, dAttellis N, et al. Endoscopic coronary tomography: a promising new technology for risk stratification of
artery bypass grafting with the aid of robotic assisted instruments. patients with coronary artery disease. Heart 2001;85:576578.
J Thorac Cardiovasc Surg 1999;118:410. 33. Janowitz WR, Agatston AS, Viamonte M. Comparison of serial
17. Ohnesorge B, Flohr T, Becker C, et al. [Cardiac imaging with rapid, quantitative evaluation of calcified coronary artery plaque by
retrospective ECG synchronized multilevel spiral CT]. Radiologe ultrafast computed tomography in persons with and without obstruc-
2000;40:111117. tive coronary artery disease. Am J Cardiol 1991;68:16.
18. Ohnesorge B, Flohr T, Becker C, et al. Cardiac imaging by means of 34. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M,
electrocardiographically gated multisection spiral CT: initial expe- Detrano R. Quantification of coronary artery calcium using ultrafast
rience. Radiology 2000;217:564571. computed tomography. J Am Coll Cardiol 1990;15:827832.
19. Flohr T, Ohnesorge B. Heart rate adaptive optimization of spatial 35. Herzog C, Dogan S, Diebold T, et al. Multi-detector row CT versus
and temporal resolution for electrocardiogram-gated multislice spi- coronary angiography: preoperative evaluation before totally endo-
ral CT of the heart. J Comput Assist Tomogr 2001;25:907923. scopic coronary artery bypass grafting. Radiology 2003;229:200208.
 CHAPTER 28 / CTA FOR CORONARY BYPASS GRAFTS 301

28 CT Angiography for Assessment

of Coronary Bypass Grafts

MARCELLO DE SANTIS, MD

With the advent of subsecond rotation combined with pro- (SVG) by DeBakey and Garrett in 1964 salvaged a complicated
spective electrocardiogram (ECG) triggering or retrospective left anterior descending (LAD) coronary endoarterectomy (2).
ECG gating, conventional computed tomography (CT) with In the 1960s, Mason Sones showed the feasibility of selec-
spiral capability and superior general image quality has chal- tive coronary arteriography and collected a large library of
lenged electron beam CT (EBCT) in the domain of cardiac cineangiograms that were studied in depth by Rene Favaloro
imaging. The introduction of multislice CT scanning with the (3). Sones and Favaloro formed an innovative team that
Siemens SOMATOM Volume Zoom with 4 simultaneously demonstrated the efficacy and safety of SVG interposition
scanned slices, half-second rotation, and 250-ms maximum and aortocoronary SVGs for single-vessel, left main, and
temporal resolution has recently opened new horizons for car- multivessel coronary disease, thus favoring the worldwide
diac CT imaging. ECG-gated multislice spiral CT represents a application of this approach. Ironically, with demonstration of
leap in image quality of CT angiography (CTA) of the coronary the dramatic benefits obtainable by saphenous vein grafting
arteries. The fast volume coverage allows scanning the heart came recognition of the ultimately palliative nature of the
with 1-mm slice collimation within a single breath-hold (10 cm operation, as a result of the accelerated atherosclerosis that devel-
in 2530 s) for high-resolution imaging. Three-dimensional ops within the grafted saphenous vein conduits. During the first
reconstruction with approx 1-mm slice width and submillime- year after bypass surgery, up to 15% of venous grafts occlude;
ter increment provide data of unique quality for visualization of between 1 and 6 yr the graft attrition rate is 1% to 2% per year,
the coronary arteries and of the arterial/venous grafts utilized and between 6 and 10 yr it is 4% per year. By 10 yr after surgery,
for surgical revascularization. This chapter will be focused on only 60% of vein grafts are patent and only 50% of patent vein
the clinical background, the previous CT applications in this grafts are free of significant stenosis (46).
setting, and the recent results of multidetector-row CT (MDCT) Although the left IMA initially fell from favor as a result of
evaluation of coronary artery bypass grafting (CABG). early, ill-founded concerns regarding low flow rates and tech-
nical difficulties in implantation, today it is recognized that
CORONARY BYPASS GRAFTINGBACKGROUND
selection of the left IMA rather than a saphenous vein as the
Surgical revascularization for atherosclerotic heart disease initial conduit is the single most important factor in improved
is one of the milestones in medical history. Relief of angina survival, freedom from cardiac events, and long-term graft
after revascularization, improvement in exercise tolerance, and patency after coronary bypass surgery (8590% after 10 yr).
the global benefit on survival have attended this approach since The favorable effects on mortality and morbidity are observed
the early stages of development (Fig. 1). After many surgical irrespective of age, gender, or left-ventricular function, and are
efforts to relieve angina pectoris, including the direct implan- particularly evident if the left IMA is implanted into a proxi-
tation of the internal mammary artery (IMA) into the myocar- mally stenosed LAD, in view of the large area of myocardium
dium (Vineberg procedure), coronary surgery moved into the subtended by this native vessel (7) (Fig. 2). The profound and
modern era in the 1950s. The first direct surgical approach to sustained benefits afforded by the IMA grafting have given
the coronary circulation in a patient was likely performed by impetus to both the utilization of other arterial conduits as coro-
Mustard in 1953 using a carotid-to-coronary bypass, but the nary bypass grafts (right IMA, right gastroepiploic artery, radial
first clinical use of the IMA to graft a coronary vessel followed artery, inferior epigastric artery, and so on) and the develop-
an intraoperative misadventure of William Longmire in 1958 ment of minimally invasive coronary artery bypass grafting
after disintegrating a right coronary artery (1). Similarly, the (MICABG). This innovative technique, first proposed by
first successful clinical aortocoronary saphenous vein graft Benetti and colleagues in 1994, does not involve the use of
cardiopulmonary bypass, or of a median sternotomy. Instead,
through a small left thoracotomy, the left IMA is harvested with
From: Contemporary Cardiology: CT of the Heart: or without the aid of a thoracoscope, the pericardium is opened,
Principles and Applications and the arterial conduit is grafted to the LAD (8). At present,
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ single-vessel coronary artery disease involving the LAD is the

301
302 DE SANTIS

Fig. 1. Schematic drawing of conventional surgical revascularization for coronary artery disease.

primary indication for MICABG. Wider application of the tech- nary angiography (CAG), conventional CT scanning with
nique, including right IMA or gastroepiploic artery grafting to contrast enhancement proved to have the potential to image
the right coronary artery, is currently under evaluation. saphenous vein grafts (9,10), but preliminary promising results
on this respect were confined to simple patency assessment
CORONARY BYPASS GRAFTING
(occluded vs not occluded) (11,12). With spiral CT technology,
PREVIOUS RESULTS OF CT EVALUATION it became possible to scan the entire heart during the arterial
(CONVENTIONAL SPIRAL CT AND EBCT) phase of contrast enhancement in a single breath-hold, thus
Like any other vascular diagnostic field, coronary bypass significantly reducing both cardiac and respiratory motion
grafting was subjected to CT evaluation early after its introduc- artifacts; early spiral CT results, obtained in small popula-
tion in clinical practice. In comparison with conventional coro- tions of CABG patients in comparison with conventional CAG,
 CHAPTER 28 / CTA FOR CORONARY BYPASS GRAFTS 303

showed good sensitivity/specificity in terms of venous graft

patency, but failed to define both arterial IMA graft and distal
anastomosis site patencies (13).
More recent generations of spiral CT single-detector-row
scanners with subsecond acquisition (0.75 s) were successfully
applied to the contrast-enhanced 3D visualization of venous as
well as arterial IMA grafts; in a series of 134 bypass grafts (42
IMA and 92 venous) double-blind evaluated by subsecond spiral
CT and conventional CAG, Engelmann et al. found a CT-deter-
mined overall sensitivity of 92% for patency with an accuracy
of 88% (IMA) and 96% (venous, p = NS) (14). In addition, the
newer capabilities of spiral CT imaging allowed the introduc-
tion in this setting of a new parameter of CABG patency assess-
ment, i.e., graft flow, which can be qualitatively extracted
through the graft length as obtained from multiple 3D recon-
struction images at varying Hounsfield unit (HU) thresholds
(faster flows correlate with longer graft lengths on most 3D
reconstruction thresholds) (15). These refinements in 3D spiral
CT angiography of coronary bypass grafting represented a
definite step forward in CT vascular imaging and gave rise to
the current competition in this respect with EBCT, whose
application for CABG patency assessment has scored similar
results in two recently reported prospective studies, with over-
all better evaluation of distal anastomosis sites due to the ECG-
triggered EBCT acquisition mode (16,17).
The recent introduction of cardiosynchronization in spiral CT
acquisition mode has followed the need to completely suppress
cardiac motion when millimeter/submillimeter spatial resolu-
Fig. 2. Schematic visualization of internal mammary artery graft tion has to be coupled with high temporal resolution, as recom-
implantation to LAD in the presence of coronary artery disease. mended in coronary vessel imaging. The crucial importance of
this aspect has been recently demonstrated by using an ECG-
triggered acquisition approach for spiral CT evaluation of
proximal anastomoses in CABG patients (18); despite the sub-
optimal temporal resolution, optimized timing of scanning as
provided by the ECG-triggered mode allowed good visualiza-
tion of proximal graft anastomoses in all patients, with assess-
ment not only of overall patency (occluded vs not occluded)
but also of graft disease, i.e., stenosis detection and grading
(Figs. 3,4).
This innovative diagnostic capability, as offered by the
cardiosynchronization of CT scanning, explains well the
potential of the newer generation of MDCT scanners.
CORONARY BYPASS GRAFTING
PREOPERATIVE APPLICATIONS OF MDCT
The results of several meta-analysis reports on graft fail-
ure as well as the spreading application of the minimally inva-
sive approach for CABG surgery have ultimately led to the
need for adjunctive preoperative information regarding both
the native vessels and the arterial conduits. This information,
which can be only partially obtained by using conventional
CAG, includes:
1. precise localization of the LAD course (extra- or intramyo-
cardial) and its relationship to adjacent structures
2. emergence, course, size, and branching of the IMAs (and
Fig. 3. Source image with clear depiction of the emergence from the of any other arterial conduit if planned)
ascending aorta of multiple venous grafts. 3. emergence, course, and size of the left subclavian artery.
304 DE SANTIS

Fig. 5. Sagittal thin maximum intensity projection reconstruction of

the entire course of left internal mammary artery.

Fig. 4. Source image with visualization of the emergence and proxi- cant anomalies, eventually correlated to graft failure, in 79/262
mal portion of a venous graft with a significant stenosis at this level. patients (30%) including common origin of another large artery,
large side branches, tortuosity, atypical origin/course, hypo-
plasia, and atherosclerotic lesions, resulting in more difficult
IMA preparation in 68/262 patients (26%) and complete modi-
Intramyocardial LAD course is an absolute contraindication fication of surgical strategy in 11/262 patients (4%) (20). These
to MICABG, whereas deep, nonpalpable extramyocardial LAD findings, coupled with the above-mentioned LAD anatomic
is associated with a more difficult MICABG approach for cor- assessment and added to the left subclavian artery evaluation
rect minithoracotomy positioning; in this respect, EBCT with for suspected extensive brachiocephalic atherosclerosis,
ECG-triggered acquisition has already been applied to visual- strongly recommend pre-MICABG fast contrast-enhanced
ize LAD anatomy before MICABG surgery with successful MDCT ungated thoracic examination for optimal surgical
results in a small population of patients (19). The application of planning, as already suggested by our experience in this setting
MDCT technology at higher pitches can be of value in this (21) (Fig. 5).
setting by means of fast contrast-enhanced thoracic examina- Another fundamental preoperative application of MDCT is
tion (3-mm collimation), without prospective triggered or ret- represented by redo CABG surgery. From an epidemiologic
rospective gated ECG acquisition, and with excellent anatomic point of view, further revascularization, either reoperative
depiction of LAD. bypass surgery or percutaneous intervention, is required in 4%
The proximal origin of the internal mammary artery, either of patients by 5 yr, 19% of patients by 10 yr, and 31% of patients
right or left, is on the concavity of the subclavian artery, just by 12 yr after initial bypass surgery (22). Despite the increasing
opposite to the thyrocervical trunk, which is the second branch numbers of patients undergoing second and third reoperations,
on the convexity of the subclavian artery (the first branch is the repeat revascularization has considerable limitations, taking
vertebral artery). After crossing the subclavian veins, the IMAs into account the perioperative morbidity and mortality, which
line the sternum on both sides for a distance of approx 12 cm escalate further as the clinical benefits diminish. As compared
from the sternal border, and they are accompanied in general by with initial surgery, reoperation carries a higher mortality rate
one or two internal mammary veins. After a proximal, medial (3 to 7%), with a high rate of perioperative myocardial infarc-
thymic branch, the IMA anastomoses with the intercostal arter- tion (4 to 11.5%). Redo surgery is also associated with less
ies beyond each rib until it reaches the sixth intercostal space, complete relief of angina and with reduction in saphenous vein
where it divides into two major branches: the craniocaudal graft patency as compared with initial bypass surgery (23,24).
branch enters the sheath of the musculus rectus abdominis and Extremely meticulous planning of surgical approach is there-
anastomoses with the superior epigastric artery, whereas the fore mandatory in this subgroup of redo CABG patients and
major lateral branch follows the cartilaginous arch of the ribs. must be defined taking into account several parameters, both
In a series of 262 consecutive patients undergoing cardiac cath- cardiovascular (native CAD, venous-IMA graft patencies, and
eterization prior to CABG, Bauer et al. found surgically signifi- courses) and thoracic (sternotomy, rib cage, mediastinum, lung
 CHAPTER 28 / CTA FOR CORONARY BYPASS GRAFTS 305

Fig. 6. Volume-rendering 3D reconstruction of multiple venous

bypass grafts.

parenchyma). In this respect, ungated MDCT has already been

successfully applied as a preoperative tool in the case of redo
CABG surgery (25); 3D visualization of cardiovascular as well
as thoracic structures, as offered by the high-resolution MDCT
technology, allows user-friendly appreciation of complex post-
surgical anatomy and therefore a more confident approach for
the surgeon.
CORONARY BYPASS GRAFTING
POSTOPERATIVE APPLICATIONS OF MDCT
To date, few studies have been prospectively addressed to
the noninvasive evaluation of coronary bypass grafts by means
of ECG-gated MDCT acquisition (Fig. 6). Previous MDCT
reports on this respect (25) were both restricted to small popu-
lations of patients and, much more important, ungated, thus
without visualization of distal anastomosis sites. Cardio- Fig. 7 (A,B). Volume-rendering 3D reconstruction views of a single
synchronization for MDCT evaluation of CABG was first venous graft to the posterior descending artery in a patient with sig-
proposed by von Smekal et al. by means of an ECG-triggered nificant right coronary artery disease.
approach providing sufficient volume coverage (120- to
140-mm scan range) within a single breath-hold with 0.5-s
rotation and 4 2.5-mm collimation. However, due to the lack collimation for 1.25-mm slice width, 120 kV, 300 mA, and a
of spatial resolution with 2.5-mm slice width, the distal anas- spiral pitch between 1.5 and 2.0, a scan range of 120 mm can
tomoses and distal patency of native coronary vessels cannot be be covered within a 30- to 35-s breath-hold. Owing to the spiral
evaluated with this approach (26). pitch higher than 1.5, reconstruction is restricted to the single-
Retrospective ECG gating applied to contrast-enhanced segment multislice cardiac volume (MSCV) reconstruction
MDCT scanning, as already showed in the technical part of this algorithm (27) with a fixed temporal resolution of 250 ms for
book, is undoubtedly the method of choice for an optimal heart rates up to approx 74 bpm; this temporal resolution is
cardiosynchronization of CTA raw data collection; this usually sufficient for motion-free MDCT imaging of coronary
approach, coupled with extensive volume coverage and sub- bypass grafts and their distal anastomoses, whose motion
millimeter spatial resolution, provides unique visualization of amplitude is significantly lower than that of native vessels
arterial-venous grafts almost completely free of cardio-respira- (Figs. 8 and 9). Patient preparation and image postprocessing
tory motion along their entire course at the expense of increased with 3D reconstruction algorithms are obviously similar to those
radiation exposure (Fig. 7 A,B). Using 0.5-s rotation, 4 1 mm usually employed for coronary MDCT imaging protocols.
306 DE SANTIS

Fig. 8. Source image showing the distal anastomosis site of a left internal mammary artery implantation to left anterior descending artery.

Regarding the preliminary results, Nieman et al. first artifacts represent the major causes for impaired image quality
reported 88% (15/17) of evaluable bypass grafts in a restricted in this setting, and further improvements of MDCT technology
population of four patients, with overall detection of 4/5 graft are expected to reduce in the next future the percentage of
lesions in comparison to conventional CAG (28). Moreover, a unevaluable grafts and to optimize visualization and evaluation
wider population of 65 patients with a total of 182 bypass grafts of distal anastomosis sites.
was prospectively evaluated by Ropers et al., comparing MDCT
REFERENCES
and conventional CAG results in order to assess overall accu-
1. Shumacker HB. The Evolution of Cardiac Surgery. Indiana Univer-
racy in detecting graft occlusion and stenosis (29); higher sen- sity Press, Bloomington, IN: 1992.
sitivity (98%) and specificity (99%) than ever before were 2. Garrett HE, Dennis EW, DeBakey ME. Aortocoronary bypass
obtained by MDCT in terms of bypass graft patency assess- with saphenous vein graft: seven-year follow-up. JAMA 1973;223:
ment, whereas a satisfactory diagnostic accuracy for the detec- 792794.
3. Favaloro RG. Critical analysis of coronary artery bypass graft
tion of high-grade bypass stenoses (sensitivity 75%, specificity
surgery: a 30-year journey. J Am Coll Cardiol 1998;31:1B63B.
92%) was achieved when grafts were imaged with sufficient 4. Campeau L, Enjalbert M, Lesperance J, et al. The relation of risk
quality (Figs. 10 and 11). factors to the development of atherosclerosis in saphenous vein
Distal anastomoses visualization represents one of the cru- bypass grafts and the progression of disease in the native circula-
cial aspects of bypass grafting MDCT imaging, and its full tion: a study 10 years after aortocoronary bypass surgery. N Engl
J Med 1984;311:13291332.
appreciation depends on the rigorous application of the acqui- 5. Bourassa MG. Fate of venous grafts: the past, the present and the
sition technique as well as on the correct utilization of 3D future. J Am Coll Cardiol 1991;5:10811083.
reconstruction algorithms usually available, in order to obtain 6. Fitzgibbon GM, Kafka HP, Leach AJ, et al. Coronary bypass graft
evidence-based images (Figs. 1214). In addition, MDCT tech- fate and patient outcome: angiographic follow-up of 5065 grafts
related to survival and reoperation in 1388 patients during 25 years.
nology is generally well suited for stenting evaluation and there-
J Am Coll Cardiol 1996;28:616626.
fore may be successfully applied even in those infrequent cases 7. Loop FD. Internal-thoracic-artery grafts: biologically better coro-
of graft disease percutaneous treatment (Fig. 15). nary arteries. N Engl J Med 1996;334:263265.
The major drawback of such noninvasive approach seems at 8. Calafiore AM, Angelini GD, Bergsland J, et al. Minimally inva-
the moment to be the significant number of unevaluable grafts; sive coronary artery bypass grafting. Ann Thorac Surg 1996;62:
15451548.
as stressed in the study of Ropers et al. (29), only 62% of the 9. Brundage BH, Lipton MJ, Herfkens RJ, et al. Detection of patent
patent bypass grafts could be evaluated for the presence or bypass grafts by computed tomography: a preliminary report. Circu-
absence of high-grade stenoses. In general, metal and motion lation 1980;61:826831.
 CHAPTER 28 / CTA FOR CORONARY BYPASS GRAFTS 307

Fig. 9. Source image with evidence of venous graft anastomosis to the second diagonal branch.

10. Ullyot DJ, Turley K, McKay CR, et al. Assessment of saphenous 16. Ha JW, Cho SY, Shim WH, et al. Noninvasive evaluation of coro-
vein graft patency by contrast-enhanced computed tomography. nary artery bypass graft patencyusing three-dimensional angiogra-
J Thorac Cardiovasc Surg 1982;83:512518. phy obtained with contrast-enhanced electron beam CT. AJR Am J
11. Godwin JD, Califf RM, Korobkin M, et al. Clinical value of coronary Roentgenol 1999;172:10551059.
bypass graft evaluation with CT. AJR Am J Roentgenol 1983;140: 17. Lu B, Dai RP, Jing BL, et al. Evaluation of coronary artery bypass graft
649655. patency using three-dimensional reconstruction and flow study on elec-
12. Daniel WG, Dohring W, Stender HS, et al. Value and limitations of tron beam tomography. J Comput Assist Tomogr 2000;24:663670.
computed tomography in assessing aortocoronary bypass graft 18. von Smekal A, Lachat M, Wildermuth S, et al. Proximal anastomoses
patency. Circulation 1983;67:983987. of aortocoronary bypasses. Evaluation with ECG-triggered single-
13. Tello R, Costello P, Ecker CP, et al. Spiral CT evaluation of coronary slice computerized tomography. Radiologe 2000;40:130135.
artery bypass graft patency. J Comput Assist Tomogr 1993;17: 19. Ohtsuka T, Takamoto S, Endoh M, et al. Ultrafast computed
253259. tomography for minimally invasive coronary artery bypass grafting.
14. Engelmann MG, von Smekal A, Knez A, et al. Accuracy of spiral J Thorac Cardiovasc Surg 1998;116:173174.
computed tomography for identifying arterial and venous coronary 20. Bauer EP, Bino MC, von Segesser LK, et al. Internal mammary
graft patency. Am J Cardiol 1997;80:569574. artery anomalies. Thorac Cardiovasc Surg 1990;38:312315.
15. Tello R, Hartnell GG, Costello P, et al. Coronary artery bypass graft 21. De Santis M, Quagliarini F, Leonetti C, et al. MDSCT pre-operative
flow: qualitative evaluation with cine single-detector row CT and evaluation of internal mammary arteries (IMAs) in patients candi-
comparison with findings at angiography. Radiology 2002;224: date to minimally invasive coronary artery bypass grafting
913918. (MICABG) (abs). ECR, Vienna, Austria: 2003.
308 DE SANTIS

Fig. 12. Volume-rendering 3D reconstruction of a patent distal anas-

tomosis between internal mammary artery graft and left anterior
descending artery.

Fig. 10. Axial visualization of a diseased venous graft on the left side
of the pulmonary trunk in a patient with internal mammary artery
implantation.

Fig. 13. Volume-rendering 3D reconstruction of a patent internal

mammary artery graft implanted to left anterior descending artery.

22. Weintraub WS, Jones EL, Craver JM, et al. Frequency of repeat
coronary bypass or coronary angioplasty after coronary artery by-
pass surgery using saphenous venous grafts. Am J Cardiol
1994;73:103112.
Fig. 11. Axial image showing a completely occluded graft closer to 23. Loop FD, Lytle BW, Cosgrove DM, et al. Reoperation for coronary
a small patent right coronary artery and an occluded circumflex artery atherosclerosis: changing practice in 2509 consecutive patients. Ann
in the presence of patency of the corresponding graft. Surg 1990;212:378386.
 CHAPTER 28 / CTA FOR CORONARY BYPASS GRAFTS 309

Fig. 15. Multiplanar reformation 3D reconstruction of a patent stent

Fig. 14. Volume-rendering 3D reconstruction of a patent distal anas-
for venous graft disease.
tomosis between venous graft and obtuse marginal.

24. Cameron A, Kemp HG Jr, Green GE. Reoperation for coronary artery 27. Ohnesorge B, Flohr T, Becker CR, et al. Cardiac imaging by means
disease: 10 years of clinical follow-up. Circulation 1988;78 (Suppl I): of electrocardiographically gated multisection spiral CT: initial
I/158I/162. experience. Radiology 2000;217:564571.
25. Yamaguchi A, Adachi H, Ino T, et al. Three-dimensional computed 28. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography
tomographic angiography as pre-operative evaluation of a patent with multi-slice computed tomography. Lancet 2001;357 (9256):
internal thoracic artery graft. J Thorac Cardiovasc Surg 2000;120: 599603.
811812. 29. Ropers D, Ulzheimer S, Wenkel E, et al. Investigation of aorto-
26. von Smekal A. The potential of cardio-computed tomography. coronary artery bypass grafts by multislice spiral computed tomo-
Multislice CT: a practical guide. Proceedings of the 5th International graphy with electrocardiographic-gated image reconstruction. Am J
Somatom CT User Conference. Zurich, June 2000. Cardiol 2001;88:792795.
 CHAPTER 29 / EBCT AND MDCT TO EVALUATE CORONARY STENT PATENCY 311

29 Contrast-Enhanced Electron Beam CT

and Multidetector-Row CT
in the Evaluation of Coronary Stent Patency

HEIKO PUMP, MD, STEFAN MHLENKAMP, MD, RAIMUND ERBEL, MD,

AND RAINER SEIBEL, MD

INTRODUCTION we systemically explored the potential of fast-CT imaging to

High-pressure stent implantation is an established tech- visualize coronary stent morphology (25,26) (Fig. 1) and
nique to maintain luminal integrity following interventional coronary stent patency using contrast-enhanced flow studies
revascularization in native coronary arteries and bypass grafts. (2729). Technical advances in spatial and temporal resolution
Large controlled randomized trials demonstrated superior in both EBCT and MDCT technology have led to an image
long-term patency in comparison to percutaneous transluminal quality that may allow broad clinical use of noninvasive
coronary angioplasty (PTCA) alone (813). The beneficial contrast-enhanced fast-CT scanning to assess coronary stent
effects of stent implantation on restenosis can be attributed to morphology and patency.
larger acute lumen dimensions compared to balloon angio-
APPROACHES TO CORONARY
plasty, and the elimination of vessel recoil after intervention
(14,15). All currently available stents are made of metal, and
STENT IMAGING USING CT
they induce significant intimal hyperplasia in some patients. Depending on the aim of the study, it is reasonable to distin-
Improvements in antithrombotic and anticoagulation therapy guish between morphologic imaging to visualize stent localiza-
have substantially reduced the rate of in-stent thrombosis (16 tion (its relation to side branches and stent-related stenosis) on
18). Brachytherapy and drug-eluting stents have been success- the one hand, and functional imaging (to [semi-] quantitatively
fully used in in-stent restenosis (1924). evaluate the functional impact of luminal narrowing) on the
Despite these substantial improvements in interventional other. Unattenuated studies have been used in the past to visu-
revascularization, the rate of restenosis remains in the range of alize coronary stents, which at the time was hardly possible
>1020% in routine clinical practice. Considering the still with any other imaging modality. However, the diagnostic
increasing rate in the use of stents, this represents a significant information contained within the resulting images is limited to
proportion of cases each year. Stent-related restenosis is prima- localize the stent and to distinguish coronary calcification from
rily based on two mechanisms: elastic recoil with luminal nar- coronary stents, which in some cases remains difficult even in
rowing shortly after stent deployment and the induction of studies with thin sections. Noninvasive contrast-enhanced
intimal cell proliferation caused by the controlled vessel-wall coronary CT angiography (CTA) can be used to assess stent
trauma during stent implantation. Optimal strut expansion and patency as per contrast enhancement in the course of the stented
stent localization in relation to side branches may have an artery, where an unenhanced distal coronary artery lumen usu-
impact on the long-term outcome. ally reflects significant in-stent restenosis. Recent improve-
A noninvasive tool to rule out stent-related stenosis would ments in spatial resolution now provide an image quality where
be of substantial clinical value if only a fraction of recatheteri- visualization of nonocclusive in-stent neointimal hyperplasia
zations could be avoided. seems within reach. Using contrast-enhanced flow studies at
Coronary stents are barely visible with imaging techniques rest and after stress, e.g., with infusion of adenosine, may per-
such as high-resolution digital fluoroscopy. In contrast, elec- mit noninvasive assessment of coronary flow reserve. These
tron beam computed tomography (EBCT) and multidetector- approaches may be used in combination, but the additional
row computed tomography (MDCT) provide superior spatial diagnostic information must be carefully weighed against
resolution that allows visualization of stents both in native additional radiation and use of contrast agent.
coronary arteries and bypass grafts. In our initial investigations ELECTRON BEAM COMPUTED TOMOGRAPHY
From: Contemporary Cardiology: CT of the Heart: In 1994, Eldredge et al. were the first investigators to assess
Principles and Applications
stent patency in native coronary arteries by using EBCT (30).
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

311
312 PUMP ET AL.

Fig. 1. In vitro study of two stent types after dilatation. Comparison of electron beam CT images, photos, and schematic drawings.

Table 1
Coronary Arterial Stent Patency:
Assessment With Electron Beam CT depicted the absence of high-grade stenoses in 189 of 193
stented vessels (specificity, 98%). The interpretation was false-
No. of stented vessels 221
No. of patent vessels at coronary artery (CA) 189 positive in four vessels (positive predictive value, 82% [18/22
and correctly detected at electron beam CT (EBCT) vessels]) and false-negative in five (negative predictive value,
No. of stent stenosis at CA 23 97% [189/194 vessels]). In 1999, we published the results of a
No. of stent stenosis at CA 18 prospective study with 44 patients. In correlation with coronary
and correctly detected at EBCT angiography, we found sensitivities of 65% and 84%, respec-
Senstivity for detection of stenoses 78% tively, and positive and negative predictive values of 69% and
with EBCT
Specificity for depiction of the absence 98%
82%, respectively (25), which is in the same range as other
of stenoses with EBCT noninvasive tests such as stress ECG.
Negative predictive value 97%
Positive predictive value 82% EBCT SCAN PROTOCOLS
Each examination consists of a single-section volume study
and a contrast-enhanced multisection flow study. EBCT in the
single-slice mode is performed, beginning at the pulmonary
They performed a contrast-enhanced multisection EBCT and trunk, with 3045 contiguous 3-mm sections triggered at 80%
evaluated stent patency with the cineloop and gamma variate- of the RR interval, and using a 100-ms acquisition time. The
based time density analysis. In 1996 Schmermund et al. reported localization of the stent and of the calcified plaques is impor-
the assessment of coronary PalmazSchatz stents in 22 patients tant for the planning of the contrast-enhanced studies. The
(27). Twenty of 22 patients (91%) could be analyzed by using depiction of each stent and of calcifications with respect to their
cineloop evaluation, and in 17 (85%) of these 20 patients, a location is based on typical landmarks, and the analysis is lim-
gamma variate curve was fitted to the segment distal to the ited to that of major coronary arteries (i.e., proximal, middle,
stent. In 1998 we reported our findings using various stent types: and distal segments) and major side branches (i.e., first diago-
we found a higher rate of cine loop evaluation (94.8%) and a nal and/or marginal branches). The table position is adjusted so
lower rate of successful distal gamma variate fits (49%) as that the stent is positioned between the scanning planes and
compared with the study by Schmermund et al. (28). Overall images can be obtained proximal and distal to the stent. In
accuracy was 94.3% with a sensitivity of 77%, which is consis- addition, an incorrect attenuation measurement on the contrast
tent with findings in a larger cohort (29) (Table 1). Compared studies caused by calcified plaques can be avoided. Initial cir-
with coronary angiography, EBCT permitted the detection of culation time assessment is followed by a multislice flow study,
18 of 23 high-grade stenoses (sensitivity, 78%) and correctly to obtain time-density images from the coronary artery lumen.
 CHAPTER 29 / EBCT AND MDCT TO EVALUATE CORONARY STENT PATENCY 313

This study is performed with a bolus injection of 20 mL of technology, with 1-mm or even submillimeter section thick-
iopromide (370 mg of iodine per milliliter). In the multislice ness and dedicated ECG-gated image reconstruction algorithms
flow study, one bolus injection of 50 mL of iopromide (370 mg provide an improved spatial resolution compared with EBCT.
of iodine per milliliter) with an injection rate of 710 mL/s are The minor temporal resolution of MDCT compared with EBCT
performed, which may be followed by a second study during remains a drawback today, and in some cases it is necessary
adenosine infusion. In a few cases, the scan volume may not to treat patients with oral or intravenous metoprolol prior to
encompass the entire length of the stented segment, especially imaging to achieve a heart rate less than 60 beats per minute
in patients with multiple stents, so that dual injections may be (bpm). The increased gantry rotation speed (420 ms) of the
required to image proximal and distal segments. The multislice second-generation MDCT may have the potential to expand
flow study protocol consists of 8-mm slices, eight levels,10 the range of heart rates.
times per level, electrocardiogram triggered at 80% of the RR In contrast to the prospective ECG-gated EBCT scan proto-
interval on every second heartbeat, and a 50-ms acquisition time. col, the MDCT examination is performed as a spiral scan com-
In the second step, the multisection flow study results are bined with simultaneously digitized ECG. To our knowledge,
analyzed and the images obtained at each level are combined in the performance of multisection flow studies with densitomet-
a cineloop for qualitative assessment based on the examiners ric measurements and cineloop evaluations are not considered
visual inspection of the images. In the third step, integrated as standard tools in the evaluation of the coronary arteries, and
scanner software or off-line assessment can be used for densi- in most cases, software options are not routinely available com-
tometric evaluation in a time-attenuation analysis of the con- mercially.
trast-enhanced studies using a gamma variate, which can be
4-ROW MDCT
fitted to the data. The slope of the curve in comparison to that
in the aorta may be used for semi-quantitative estimation of Initially, a 4 2.5-mm collimation with simultaneous ECG
epicardial blood flow. recording is used to obtain retrospectively reconstructed 3-mm
sections of the whole heart to localize the coronary stents and
SUGGESTED EBCT CRITERIA FOR PATENCY, to determine coronary calcifications.
STENOSIS, AND OCCLUSION After determination of the scanning delay with a bolus injec-
Initially, coronary stent patency was assessed only under rest- tion of 15 mL contrast agent, a contrast-enhanced 4-row MDCT
ing conditions. The degree of opacification of the segment distal (130150 cc contrast agent, flow rate 4 cc/s) is performed by
to the stent was used to distinguish occluded from patent vessels. using a 4 1-mm collimation and a 500 ms rotation time. The
The coronary stent was defined to be patent if distal vessel opaci- axial images are reconstructed retrospectively in 1-mm slice
fication was similar to that in control vessels of similar size in the thickness with a 50% slice overlap. Usually, maximum inten-
cineloop. An automated fit of a gamma variate was used as an sity projection (MIP) reconstructions, 3D reconstructions in
additional observer-independent tool to define stent patency, volume-rendering technique (VRT), and the axial source
because in most cases a gamma variate can be fitted only if the images are used for the evaluation of stent patency.
time-density curve shows a typical flow profile (Fig. 2A,B).
Further, in a nonobstructed coronary artery, the delay in opaci- 16-ROW CT
fication after aortic root enhancement is usually brief, which can A 1.5-mm collimation with simultaneous ECG recording
be used as an additional criterion for stent patency. and a 420-ms rotation time is used to obtain retrospectively
The criteria of absence of hyperattenuation distal to the stent reconstructed 3-mm sections of the whole heart for stent local-
in the cineloop or hyperattenuation proximal to the stent but ization and determination of coronary calcium (Fig. 4).
weak contrast enhancement distally and no fitted gamma variate The 0.75-mm collimation with a 420-ms rotation time is
curve raise the suspicion of luminal narrowing or vessel occlu- used for the contrast-enhanced study. The optimal scan delay
sion (Fig. 3). is determined by direct visualization and measurement of the
In most cases, cineloop evaluation is sufficient to confirm hyperattenuation in the vessel lumen during the contrast-agent
stent patency. A comparison with the other coronary arteries injection by placing a region of interest in the ascending aorta.
usually leads to the correct diagnosis. The time-attenuation A total of 130150 cc contrast agent with a flow rate of 5 cc/s
analysis and gamma variate curve application are important is injected. Cross-sectional images are reconstructed with a
additional elements in confirming the obstruction in cases in slice thickness of 0.75 mm in 0.5-mm intervals. On the basis of
which there was weak contrast enhancement distal to the stent. the cross-sectional images, sliding thin-slab MIP, and 3D
Prolonged circulation time with insufficient contrast enhance- reconstructionsVRTthe coronary arteries are evaluated.
ment, arrhythmia, inadequate breath-holding, and calcifications
SUGGESTED MDCT CRITERIA FOR PATENCY,
in distal segments are frequent reasons for insufficient image
STENOSIS, AND OCCLUSION
quality.
The coronary stent is considered to be patent if the examiner
MULTIDETECTOR-ROW CT visualizes the hyperattenuated vessel proximal and distal to the
The development of mechanical MDCT systems with stented target vessel on the axial slices and the generated
increased scan speed and the introduction of retrospective gat- 2- and 3D reconstructions compared with the other coronaries
ing were very important steps in the noninvasive visualization (Figs. 5AC). Weak contrast enhancement distal to the stent
of the heart and the coronary vessels (31). The multirow CT or in the stent lumen are signs of hemodynamically relevant
314 PUMP ET AL.
 CHAPTER 29 / EBCT AND MDCT TO EVALUATE CORONARY STENT PATENCY 315

Fig. 2. (A) (opposite page) Time attenuation measurement obtained proximal to a patent stent in the right coronary artery with electron beam
CT. The gamma variate curve demonstrates a typical arterial curve. The arrow indicates the deposition of the region of interest. (B) Time
attenuation measurement obtained distal to the patent stent in the same patient with gamma variate curve. The arrow indicates the depostition
of the region of interest in the distal vessel region of the right coronary artery.

Fig. 3. Transverse electron beam CT scan shows a high-grade left anterior descending arterial stenosis distal to the stent. The region of interest
(B) is placed in a hypoattenuating distal vessel region (arrows). The unconnected dots below the 100-HU mark (B) indicate the ROI placed
in the hypoattenuating distal vessel region. A, gamma variate curve of the ascending aorta.
316 PUMP ET AL.

Fig. 4. Transverse CT scan (16-row CT) of stents placed in the left anterior descending artery and the first diagonal branch.

stenosis or occlusion (Figs. 6A,B). The visualization of in-stent SUMMARY

stenosis remains a problem caused by high-contrast artifacts of Currently, exact quantification of stenosis is possible only
the stainless steel stent struts, which make an exact differentia- with coronary angiography. The time-attenuation analysis
tion between struts noncalcified plaques, and calcified plaques allows only the differentiation between a patent stent, high-
impossible. The newest development of dedicated reconstruc- grade stenosis, and stent occlusion, but no exact quantification
tion algorithms and improved in-plane resolution might reduce of the coronary arterial flow.
these artifacts (32), and it may be possible to differentiate Overlapping contrast-enhanced single-section EBCT exami-
between stent struts and neoproliferation. nation is widely used in the noninvasive visualization of the
 CHAPTER 29 / EBCT AND MDCT TO EVALUATE CORONARY STENT PATENCY 317

Fig. 5. (A) Maximum intensity projection of a 16-row CT dataset. Visualization of the proximal and middle segment of the right coronary artery
with patent stents (arrows). (B) Maximum intensity projection of the same patient. Visualization of the middle and distal segments of the right
coronary artery with patent stents (arrows) and a patent distal vessel segment (arrowheads). (C) Volume-rendering technique of the same
patient. The white arrowheads show the stents in the proximal and middle vessel region of the right coronary artery. Visualization of a side
branch (arrows) and of the left anterior descending (black arrowheads).

coronary system, and several studies have demonstrated the The major drawback of MDCT is the significant increase of
high sensitivity and specificity in the detection of high-grade the effective radiation dose. Hunold et al. measured doses in the
stenosis in native coronary arteries (33,34), but the detection of range from 6.7 to 10.9 mSv for male patients and 8.1 to 13.0
in-stent stenosis with EBCT remains an unsolved problem. mSv for female patients in contrast-enhanced MDCT, and 1.5
MDCT technology, especially the acquisition of a 3D and 2.0 mSv for male and female patients in electron beam
data set with isotropic voxels, scan protocols developed for CTA, respectively (35). These data are consistent with reports
stent imaging, and postprocessing modalities may in future from others. The development of new scan techniques might
lead to a better visualization of the stent lumen. lead to a significant decrease of the radiation dose.
318 PUMP ET AL.

Fig. 6. (A) Maximum intensity projection of a stent occlusion in the left anterior descending artery. Weak contrast enhancement proximal to
the stent. There is no attenuation in the vessel region distally of the stent. Arrowheads indicate the stent. (B) Volume-rendering technique of
the same patient. The arrow indicates the vessel region proximal of the stent, the arrowheads indicate the stent.

Noninvasive imaging of coronary artery stents, stent 7. Serruys PW, de Jaegere P, Kiemeneji I, et al. A comparison of bal-
patency, and in-stent restenosis is a clinically attractive loon-expandable stent implantation with balloon angioplasty in
patients with coronary artery disease: Benestent Study Group. N
application of fast computed tomography. Currently, spatial Engl J Med 1994;331:489495.
resolution, especially in submillimeter distances from the 8. Serruys PW, Emanuelsson H, van der Giessen W, et al. Heparin-
stainless steel struts, is not always sufficient to quantitate coated Palmaz-Schatz stents in human coronary arteries: early
in-stent stenosis with an accuracy high enough for clinical outcome of the Benestent-II pilot study. Circulation 1996;93:
412422.
purposes. However, improvements in scanner technology are 9. Erbel R, Haude M, Hpp HW, et al. Coronary-artery stenting com-
in sight that might overcome this problem. Further, func- pared with balloon angioplasty for restenosis after initial balloon
tional imaging at the coronary artery or myocardial level (36) angioplasty: Restenosis Stent (REST) Study Group. N Engl J Med
using indicator dilution principles and additional infusion 1998;339:16721678.
of adenosine may further enhance diagnostic accuracy of non- 10. Antoniucci D, Valenti R, Santoro GM, et al. Restenosis after
coronary stenting in current clinical practice. Am Heart J 1998;135:
invasive assessment of coronary stent patency. 510518.
11. DiMario C, Reimers B, Almagor Y, et al. Procedural and follow-up
REFERENCES results with a new balloon expandable stent in unselected lesions.
1. Sigwart U, Puel J, Mirkovitch V, Joffre F, Kappenberger L. Intra- Heart 1998;79:234241.
vascular stents to prevent occlusion and restenosis after transluminal 12. De Gregorio J, Kobayashi Y, Albiero R, et al. Coronary artery
angioplasty. N Engl J Med 1987;316:701706. stenting in the elderly: short-term outcome and long-term
2. Haude M, Erbel R, Straub U, Dietz U, Schatz R, Meyer J. Results on angiographic and clinical follow-up. J Am Coll Cardiol 1998;32:
intracoronary stents for management of coronary dissection after 577583.
balloon angioplasty. Am J Cardiol 1991;67:691969. 13. Bittl JA. Advances in coronary angioplasty. N Engl J Med
3. Roubin GS, Cannon AD, Agrawal SK, et al. Intracoronary stenting 1996;335:12901302.
for acute and threatened closure complicating percutaneous 14. Kuntz RA, Gibson CM, Nobuyoshi M. Generalized model of
transluminal coronary angioplasty. Circulation 1992;85:916927. restenosis after conventional balloon angioplasty, stenting and
4. De Feyter PJ, De Scheerder I, van den Brand M, Laarman GJ, diretional atherectomy.J Am Coll Cardiol 1993;21:1525.
Suryarpranata H, Serrruys PW. Emergency stenting for refractory 15. Kuntz RA, Safian RD, Carozza JP Jr., et al. The importance of acute
acute coronary artery occlusion during coronary angioplasty. Am J luminal diameter in determining restenosis after coronary
Cardiol 1990;66:11471150. artherectomy or stenting. Circulation 1992;86:18271835.
5. Herman HC, Buchbinder M, Cleman MW, et al. Emergent use of 16. Pan M, Suarez de Lezo J, Velasco F, et al. Reduction of thrombotic
balloon-expandable coronary artery stenting for failed percutaneous and hemorrhagic complications after stent implantation. Am Heart
transluminal coronary angioplasty. Circulation 1992;86:812819. J 1996;132(6):11191126.
6. Fischman DL, Leon MB, Baim DS, et al. A randomized comparison 17. Spinler S, Cheng J. Antithrombotic therapy after intracoronary
of coronary-stent placement and balloon angioplasty in patients with stenting. Pharmacotherapy 1997;17(1):7490.
coronary artery disease: Stent Restenosis Study Investigators. N 18. Leon MB, Baim DS, Popma JJ, Gordon PC, et al. A clinical trial
Engl J Med 1994;331:496501. comparing three antithrombotic-drug regimens after coronary-
 CHAPTER 29 / EBCT AND MDCT TO EVALUATE CORONARY STENT PATENCY 319

artery stenting. Stent Anticoagulation Restenosis Study Investiga- 28. Pump H, Mhlenkamp, C. Sehnert, et al. Electron beam CT in the
tors. N Engl J Med 1998;339(23):16651671. noninvasive assessment of coronary stent patency. Acad Radiol
19. Teirstein PS, Massullo V, Jani S, et al. Catheter-based radiotherapy 1998;5:858862.
to inhibit restenosis after coronary stenting. N Engl J Med 1997;336: 29. Pump H, Mhlenkamp S, Sehnert, et al. Coronary arterial stent
16971703. patency: assessment with electron-beam CT. Radiology 2000;214:
20. Leon MB, Teirstein PS, Moses JW, et al. Localized intracoronary 447452.
gamma-radiation therapy to inhibit the recurrence of restenosis after 30. Elredge WJ, Santolan EC, Taylor MA, et al. Noninvasive assess-
stenting. N Engl J Med 2001;344:250256. ment of intracoronary stent patency using ultrafast computed
21. Serruys PW, Emanuelsson H, van der Giessen W, et al. Heparin- tomography (abstract). Am J Card Imaging 1994;8 (Suppl):12.
coated Palmaz-Schatz stents in human coronary arteries. Early out- 31. Achenbach S, Ulzheimer S, Baum U. Noninvasive coronary angiog-
come of the Benestent-II Pilot Study. Circulation 1996;93(3): raphy by retrospectively ECG-gated multislice spiral CT. Circula-
412422. tion 2000;102:2823.
22. Morice MC, Serruys PW, Sousa JE, et al. A randomized comparison 32. Flohr T, Bruder H, Stierstorfer, et al. New technical developments in
of a sirolimus-eluting stent with a standard stent for coronary multislice CT, part 2: sub-millimeter 16-slice scanning and increased
revascularization. N Engl J Med 2002;346(23):17731780. gantry rotation speed for cardiac imaging. Fortschr Rntgenstr 2002;
23. Fattori R, Piva T. Drug-eluting stents in vascular intervention. Lan- 174:10221027.
cet 2003 (18);361(9353):247249. 33. Moshage EL, Achenbach S, Seese B, et al. Coronary artery stenoses:
24. Verin V, Popowski Y, de Bruyne B, et al. Endoluminal beta-radia- three-dimensional imaging with electrocardiographically triggered
tion therapy for the prevention of coronary restenosis after balloon contrast agent-enhanced, electron-beam CT. Radiology 1995;196:
angioplasty. The Dose-Finding Study Group. N Engl J Med 2001; 707714.
344:243249. 34. Achenbach S, Moshage W, Ropers D, et al. Value of electron beam
25. Mhlenkamp S, Pump H, Baumgart D, et al. Minimally invasive computed tomography for the noninvasive detection of high-grade
evaluation of coronary stents with electron beam computed tomog- coronary artery stenoses and occlusions. N Engl J Med 1998;339:
raphy: in vivo and in vitro experience. Cath Cardiovasc Interv 19641971.
1999;48(1):3947. 35. Hunold P, Vogt FM, Schmermund A, et al. Radiation exposure dur-
26. Mhlenkamp S, Schmermund A, Pump H, et al. Characterization of ing cardiac CT: effective doses at multi-detector row CT and
balloon expandable coronary stent morphology using electron beam electron-beam CT. Radiology 2003;226:145152.
computed tomography. Materialwiss Werkst 1999;30:793800. 36. Mhlenkamp S, Behrenbeck TR, Lerman A, et al. Coronary
27. Schmermund A, Haude M, Baumgart D, et al. Non-invasive assess- microvascular functional reserve: quantification of long-term
ment of coronary Palmaz-Schatz stents by contrast enhanced elec- changes with electron-beam CT preliminary results in a porcine
tron beam computed tomography. Eur Heart J 1996;17:15461553. model. Radiology 2001;221(1):229236.
 CHAPTER 30 / CTA TO DETECT CORONARY STENOSIS 321

30 CT Angiography for the Detection

of Coronary Artery Stenosis

KOEN NIEMAN, MD AND FILIPPO CADEMARTIRI, MD

INTRODUCTION tion of contrast provides functional flow information. Cath-

In the past decade we have witnessed the development of eter-based angiography can be complemented with advanced
noninvasive coronary imaging using different imaging modali- coronary imaging techniques (such as intra-coronary ultra-
ties. Computed tomography (CT) and magnetic resonance sound [ICUS] or optical coherence tomography [OCT]) and
imaging (MRI) modalities have been applied for the quantifi- flow or pressure measurements to determine the functional
cation of coronary calcium, detection of coronary and bypass severity of a coronary obstruction. Competing with a technique
graft occlusion, and most recently the characterization of of such quality and versatility seems impossible. However,
noncalcified plaque material. However, the decisive applica- there are also a number of practical disadvantages to catheter-
tion of noninvasive coronary CT or MRI, which determines based coronary imaging. In addition to the considerable costs,
whether it will find widespread clinical application, will be the this invasive procedure involves a certain amount of patient
detection of coronary stenosis. The first comparative study discomfort and a small but not negligible risk of morbidity and
between MRI and conventional coronary angiography was mortality. Therefore, conventional coronary angiography is
published in 1993 (1), and numerous studies followed using applied only when adequate supportive evidence has been
various data acquisition techniques (2). Since 1997 a number of obtained, such as by exercise testing, and when either percuta-
studies have compared electrocardiogram (ECG)-triggered neous intervention or bypass surgery is anticipated.
electron beam computed tomography (EBCT) and conventional REQUIREMENTS FOR NONINVASIVE
coronary angiography, also with promising results (39). In CORONARY ANGIOGRAPHY
1999, 4-slice multislice spiral computed tomography (MSCT) In order to visualize the coronary arteries and find coronary
was introduced, and the first comparative publications appeared artery disease at an earlier stage, and to follow up patients with
in 2001 (1016). In 2002 the first results were published using known disease, a noninvasive and preferably less costly imag-
16-slice MSCT scanners with a submillimeter slice thickness ing technique would be desirable. To ensure sufficient image
and rotation time of less than a half second (17). In this chapter, quality, noninvasive techniques require a high spatial resolu-
we will discuss the practical considerations, diagnostic value, tion to image small coronary arteries, high temporal resolution
and remaining limitations of MSCT coronary imaging for the to acquire motion-free images, adequate contrast to distin-
detection of coronary stenosis. The clinical utility and future guish the coronary lumen from the vessel wall, and, particu-
developments will be discussed, as well as a comparison with larly in the case of computed tomography, a short scan time to
other noninvasive imaging techniques. acquire all data within the duration of a comfortable breath-
hold. Finally, to maintain the advantage of a lower health risk,
IMAGING REQUIREMENTS OF NONINVASIVE the noninvasive method should minimize the use of radiation
CORONARY ANGIOGRAPHY and contrast media, which particularly applies to computed
CATHETER-BASED SELECTIVE X-RAY CORONARY tomography.
ANGIOGRAPHY Besides attempting to approximate the diagnostic quality of
Conventional X-ray coronary angiography with selective conventional angiography, computed tomography and mag-
contrast enhancement of the coronary arteries remains the gold netic resonance imaging provide additional information regard-
standard for the in vivo detection and quantification of coro- ing the cardiac and three-dimensional coronary anatomy, and
nary artery stenosis. Multiple high-contrast projections with a the composition of the vessel wall.
0.1 0.1-mm image resolution are acquired each heart cycle.
Besides accurate quantitative assessment, the dynamic injec- ACQUISITION, RECONSTRUCTION,
POSTPROCESSING, AND EVALUATION
DATA ACQUISITION AND IMAGE
From: Contemporary Cardiology: CT of the Heart: RECONSTRUCTION
Principles and Applications A detailed description of the general data-acquisition and
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ image-reconstruction considerations for MSCT coronary

321
322 NIEMAN AND CADEMARTIRI

Fig. 1. Postprocessing techniques. A stenotic lesion (arrow) in the proximal left anterior descending coronary artery (LAD), displayed using
multiplanar reformation (A), thin-slab maximum intensity projection (B), curved maximum intensity projection (C), and volume rendering
(D,E). RCA, Right coronary artery; D, diagonal; IM, intermediate; CX, circumflex; and S, septal branch.

angiography can be found in the previous chapters. For the of reconstructions at a slightly varying reconstruction window
assessment of the coronary arteries, the highest image quality position, the optimal data set is used for image analysis. Par-
is required. This means the fastest X-ray source rotation speed ticularly at higher heart rates, different vessel segments are
and narrowest detector collimation possible, while still allow- optimally visualized during different phases.
ing complete scanning of the entire heart within a single breath- POSTPROCESSING AND DATA ANALYSIS
hold. Images are routinely reconstructed during the diastolic The coronary lumen can be assessed using different post-
cardiac phase to minimize motion artifacts, and out of a number processing techniques (Fig. 1). Generally, maximum intensity
 CHAPTER 30 / CTA TO DETECT CORONARY STENOSIS 323

Table 1
Diagnostic Performance of Multislice Spiral CT to Detect Coronary Stenosis,
Using Conventional Coronary Angiography As the Standard of Reference

N Assess. D Prev Excl Se Sp PPV NPV Se a

Nieman (10) 4 31 Segment 50% 0.9 27% 81% 97% 81% 97% 68%
Achenbach (11) 4 64 Branch 50% 1.1 32% 85% 76% 56% 93% 55%
70% 0.9 32% 91% 84% 59% 98% 58%
Knez (12) 4 43 Segment 50% 1.2 6% 78% 98% 84% 96% 51%
Vogl (13) 4 + 64 Segment 50% NR 28% 75% 99% 92% 98% NR
Kopp (14) b 4 102 Segment 50% 1.5 15% 86% 96% 76% 98% 86%
93% 97% 81% 99% 93%
Giesler (15) c 4 + 100 Branch 70% 1.0 29% 91% 89% 66% 98% 49%
Nieman (16) c 4 78 Segment 50% 0.9 32% 84% 95% 67% 98% 63%
Nieman (17) 16 + 58 Branch 50% 1.1 0 95% 86% 80% 97% 95%
Ropers (18) 16 + 77 Branch 50% 1.0 12% 92% 93% 79% 97% 73%

, Use of -blockers; Assess., method of assessment; D, diameter reduction considered significantly stenosed;
Prev, number of stenotic vessel per patient; Excl, percentage of excluded segments/branches. Se,Sensitivity; Sp,
specificity; PPV, positive; and NPV, negative predictive value; regarding the assessable segments/branches.
a Sensitivity including missed lesions in nonassessable segments/branches.
b Results by two observers, without consensus reading.
c Studies include patients from earlier publications.

projections (MIP) of thin slabs, or multiplanar reformations including all >2.0-mm branches (11,15), while the other stud-
(MPR), oriented parallel to the coronary of interest, are used in ies assessed a predefined number of proximal, middle, and
addition to the axial source images. MIP provides smooth, high- distal coronary segments, according to the ACC/AHA guide-
contrast images, and has the advantage that longer sections of lines, regardless of the vessel diameter (10,1214). One study
a vessel can be visualized in the same plane. However, when considered all coronary segments, including side-branches,
calcifications or stents are present, the coronary lumen becomes with a minimal diameter of 2.0 mm (16). All studies were
obscured by these high-attenuating structures, in which case performed at single centers, and all but one were based on a
MPR or the axial source images are more suitable. For presen- consensus reading by two blinded observers. In most of the
tation purposes and three-dimensional orientation of the refer- early studies, -receptor blocking medication, to reduce the
ring physician, advanced post-processing techniques are patients heart rate during the data acquisition, was not rou-
available such as curved MPR, volume rendering, and virtual tinely used (1012,14,16).
angioscopy. IMAGE INTERPRETABILITY
The percentage of coronary arteries or coronary segments
DIAGNOSTIC PERFORMANCE OF 4-SLICE MSCT with a subjectively adequate image quality varied between 6%
MSCT COMPARED TO CONVENTIONAL and 32% in the different studies. Reasons for nonassessability
CORONARY ANGIOGRAPHY mainly related to motion artifacts caused by residual cardiac
The first multislice spiral computed tomography scanners motion, severe coronary calcification, and voluntary patient
(MSCT), introduced in 1999 and equipped with four parallel movement, but also include the inability to discriminate the
detector arrays, were the first mechanical CT scanners that coronary artery lumen from adjacent contrast-enhanced struc-
allowed assessment of the coronary artery lumen. The first tures, technical scanner failure, and insufficient scan range
studies that compared MSCT coronary angiography and the (Fig. 2). With 4-slice MSCT scanners, the entire heart can be
gold standard of conventional X-ray coronary angiography covered within the duration of a long breath-hold of 3545 s.
were published in early 2001, with promising results (Table 1). However, a scan time of up to 40 s proves too long in a substan-
The studies were performed with comparable techniques but tial number of patients. Other incidental causes for reduced
varied in study design. The patient population varied between interpretability are beam-hardening artifacts from the high
31 and 102, and consisted of patients with suspected obstruc- concentration of contrast medium in the superior caval vein, or
tive coronary artery disease and an indication for conventional pacemaker wires. While MSCT in patients with arrhythmia is
coronary angiography. Therefore, the average number of dis- generally discouraged because of the end-diastolic volume
eased vessels varied between 0.9 and 1.5 per patient. Most variation, occasional premature contractions occur in many
studies considered a lumen diameter reduction of 50% as sig- patients and can adversely affect the assessment. All studies
nificant (10,1214,16), while others used a 70% cut-off point show that the LM can nearly always be evaluated. Of the
(11,15). In two studies, the main coronary branchesleft main remaining main branches, the LAD is least affected by motion
(LM), left anterior descending (LAD), left circumflex (LCx), artifacts. The most important cause for non-assessability in
and right coronary artery (RCA)were evaluated as a whole, case of the LAD is the presence of extensive calcifications.
324 NIEMAN AND CADEMARTIRI

Fig. 2. Image artifacts. The accuracy of Multislice CT can be affected by a number of artifacts. Consecutive slabs are reconstructed during
different cardiac phases in the occurrence of irregular heartbeats, which results in stair-step artifacts or in complete discontinuity of the artery
(arrow), as in this case (A,B). The presence of extensive calcification of the left anterior descending (LAD) and circumflex coronary artery
(CX) limits the assessability as a result of partial-volume artifacts and beam hardening (C,D). At faster heart rates, image-degrading motion
artifacts occur (E). Particularly in overweight patients, the attenuation by the surrounding tissue, such as the liver, can decrease the signal-
to-noise ratio. In this case, the distal right coronary artery, which runs at the same level as the liver, is poorly assessable (F).

The RCA, which has a large motion radius and short motion- related to diameter size. Both the middle segment of the RCA
sparse period during the diastole, is a branch that is most as well as the LCx have a larger motion radius compared to the
affected by motion artifacts caused by residual cardiac motion proximal segments. The more distal branches and side branches
during the image reconstruction interval. Visualization of the are most difficult to visualize. In a study that compared the
LCx also suffers from motion artifacts. Occasionally it can be assessability of the different coronary segments with a minimal
difficult to distinguish the small LCx from the adjacent con- diameter of 2.0 mm, the proximal RCA was evaluable in 88%,
trast-enhanced cardiac vein. The fact that proximal vessels are compared to 61% of the middle and 60% of the distal segments.
better visualized than more distal branches is only partially Assessment of the proximal LAD was possible in 89%, com-
 CHAPTER 30 / CTA TO DETECT CORONARY STENOSIS 325

Fig. 3. Lesions in the left main bifurcation. A significant lesion (arrow), consisting of two partially calcified plaques, is situated at the distal
part of the left main coronary artery, obstructing both the left anterior descending (LAD) as well as the left circumflex branch (CX) (A and
D). The cross-section of the vessel confirms the distinct configuration of the lesions (C). Additionally, more non-calcified plaque material
(arrowhead) can be observed in the proximal part of the left main artery (arrowheads) (A). Diagnonal (D) and marginal (M) branch.

pared to 77% of the middle and 75% of the distal segments. In pared to >50% lesions, Achenbach et al. found a higher diag-
this study, the proximal and the middle segments of the LCx nostic performance for the detection of lesions with a lumen
were equally difficult to evaluate57% and 56%, respec- diameter reduction of at least 70% (11). The results show that
tively(19). exclusion of disease in a normal vessel is less challenging than
DETECTION OF CORONARY the classification of a diseased vessel as significantly stenosed
OBSTRUCTION or not, particularly in the presence of extensive calcification.
When considering only the coronary arteries or segments The apparent size of the calcium deposits causes overestima-
that were imaged with sufficient image quality, the sensitivity tion of the total plaque size, which results in false-positive
of 4-slice MSCT to detect significant coronary obstruction, assessments. According to some investigators, MSCT coro-
defined as 50% lumen diameter reduction, ranges between nary angiography may be most valuable as a tool to exclude
75% and 95%, and the specificity between 76% and 99% (10 significant lesions in patients with a relatively low pretest like-
16). As can be expected, the sensitivity and specificity are lihood for the presence of stenoses, and not for the staging of
inversely correlated, as well as the number of excluded seg- patients with very high likelihood and expectedly advanced
ments and the diagnostic performance in the assessable seg- coronary artery degeneration. Figs. 38 are examples of MSCT
ments. The positive and negative predictive value ranged imaging of coronary obstruction with corresponding conven-
between 56% and 99%, and 93% and 99%, respectively. Com- tional X-ray angiograms.
326 NIEMAN AND CADEMARTIRI

Fig. 4. Significant lesion in the left circumflex coronary artery. A significant lesion (arrow) was found in the mid-segment of the left circumflex
branch, just proximal of the bifurcation of a major marginal branch (M), both by multislice CT (A) and conventional angiography (B).

Fig. 5. Diffuse coronary artery disease. This right coronary artery shows extensive atherosclerotic degeneration, with calcified and noncalcified
plaque material along the entire length of the proximal inner curve, and separate lesions more distally (A). Although the MSCT shows no severe
stenosis, the absence of significant lesions was more straightforward on the conventional angiogram.
 CHAPTER 30 / CTA TO DETECT CORONARY STENOSIS 327

Fig. 6. Occlusion of the left anterior descending coronary artery. Three-dimensional reconstruction (A) and curved multiplanar reconstruction
(B) of a CT coronary angiogram, showing an occluded (arrow) left anterior descending coronary artery (LAD) (B), which was confirmed by
conventional coronary angiography (C). A sagittal cross-section shows in detail the different plaque components, both calcified and noncalcified,
as well as some residual contrast enhancement within the obstructed segment (D). Right coronary artery (RCA), diagonal (D), intermediate
(IM), and circumflex branch (CX).

In a study comparing the diagnostic accuracy of MSCT in

relation to the proximity of the coronary segment, a sensitivity
and specificity of 92% and 96% were found in the largest proxi-
mal segments (RCA1, LM, LAD6), 85% and 90% for the middle
segments (RCA2, LAD7, LCx11), 71% and 94% for the distal
segments (RCA3, LAD8, LCx13), and 50% and 89% for the
side branches (19). In the study by Kopp et al., the sensitivity
to detect stenotic lesions significantly increased, to 97/99%, by
selectively assessing the proximal and middle coronary seg-
ments (14).
The segments and vessels that were excluded from analysis
because of inadequate image quality contained a substantial
number of undetected lesions. If these lesions in nonevaluable
vessels are included in the analysis as false-negative interpre-
tations, the sensitivity is much lower in most studies, between
49% and 93% (1016).
16-SLICE MSCT CORONARY ANGIOGRAPHY
To date two studies have been published comparing MSCT
and conventional coronary angiography using 16-slice MSCT
(17,18). The advantages offered by this new technique are a
faster rotation time of 0.42 s, an extended number of narrower
detector rows (0.75 mm), and shorter total scan time of approx
20 s. For the ECG-gated protocol, the 12 central detector rows
were applied. To optimize the image quality, consistent heart Fig. 7. Occlusion of the left circumflex coronary artery. Curved maxi-
rate control was incorporated into both protocols. In the study mum intensity projection of a predominantly noncalcified occlusion
by Nieman et al., patients with a prescan heart rate over 65 beats (arrow) of the left circumflex branch (A), confirmed by conventional
coronary angiography (B). The distal segment is filled collaterally by
the left anterior descending coronary artery.

Fig. 8. (oppositer page, bottom left) Stenosis of the distal right coronary artery. Using thin-slab maximum intensity projection (B) and volume
rendering (C), a stenotic lesion (arrow) is demonstrated in the distal right coronary artery. Also minor wall irregularities, caused by small
calcified lesions (arrowheads), can be observed, and were confirmed by conventional angiography (A).
328 NIEMAN AND CADEMARTIRI

Table 2
Patient-Based Diagnostic Performance of Multislice CT (MSCT) Coronary Angiography

4-slice MSCT (19) 16-slice MSCT (17)

Segment-based (N = 53) Vessel-based (N = 58)

Accuracy Predictive value Accuracy Predictive value

No lesions 9/14 (64%) 9/19 (47%) 7/7 (100%) 7/8 (88%)
Single lesion/vessel 9/23 (39%) 9/15 (60%) 12/16 (75%) 12/20 (60%)
Multiple lesions/vessel 11/16 (69%) 11/19 (58%) 26/35 (74%) 26/30 (87%)
Overall 29/53 (55%) 29/53 (55%) 45/58 (78%) 45/58 (78%)

per minute (bpm) were given an oral dose of 100 mg metoprolol none with significant coronary stenoses were falsely evaluated
1 h prior to the examination, decreasing the average heart rate as normal, without exclusion of nonassessable segments.
to 57 bpm. Ropers et al. used 50 mg of atenolol to decrease the Ropers et al. correctly assessed 85% as having one or more
heart rates of all patients with >60 bpm, down to an average lesions (18).
heart rate of 62 bpm. Only 7% and 12% of the coronary branches
CONSIDERATIONS AND LIMITATIONS
contained sections with a poor image quality. Compared to the
other branches, the RCA was still most vulnerable to image TEMPORAL RESOLUTION AND THE HEART RATE
quality degradation. Nieman et al., irrespectively of the image Coronary Motion
quality, evaluated all branches with a minimal lumenal diam- The coronary arteries are in constant motion, and therefore
eter of 2.0 mm, and found a sensitivity and specificity of 95% an infinitely short acquisition or reconstruction time is required
and 86% to detect significantly stenosed branches. The positive to acquire completely motionless images. Angiography studies
and negative predictive value were 80% and 97% (17). All four have shown that during diastole a brief moment of near-immo-
missed lesions were located in the LCx and marginal branches, bility occurs (20). The moment and duration of this window of
no lesions were missed in the LM, LAD, or RCA. The 20 over- imaging opportunity varies per person and per vessel, but
estimations included 7 lesions with a subsignificant (4049%) always shortens at higher heart rates. Generally, the RCA moves
diameter reduction, according to quantitative coronary angiog- at a wider radius and has a shorter motion-sparse period com-
raphy. Including only the evaluable (88%) vessels (minimal pared to the left coronary artery. The RCA is therefore most
diameter 1.5 mm), Ropers et al. found a sensitivity and speci- vulnerable to motion artifacts caused by cardiac motion (Fig. 2).
ficity of 92% and 93% to detect significant stenoses. Without Temporal Resolution
exclusion of nonevaluable lesions, the sensitivity was 73% (18). The temporal resolution of 4-slice MSCT scanners is 250 ms
at low heart rates. At a heart rate of 50 bpm the duration of the
PATIENT-BASED ASSESSMENT heart cycle measures 1200 ms, of which 21% is required for the
Understanding that different methods of data analysis and reconstruction of a set of axial slices. At a heart rate of 80 bpm
presentation were used is important to compare the results of the ratio is 33%, and at 120 bpm the ratio is 50%, thereby
the previously mentioned studies. For instance, by using an increasing the occurrence of motion artifacts. Considering the
evaluation based on the individual coronary segments, the rela- fact that the shortening of the diastolic phase at higher heart
tive and absolute number of nondiseased segments is much rates is more substantial than the shortening of the systolic
larger, compared to the number of nondiseased branches in phase, the negative effect of a fast heart rate on the image qual-
case of a main branchbased analysis. One of the consequences ity is even more profound. At higher heart rates, multi-segment
is that for the segment-based studies, the specificity is often reconstruction algorithms can improve the effective temporal
better compared to the those of the branch-based studies. Per- resolution by combining data from consecutive heart cycles.
haps a more comparable indicator of the clinical applicability However, this reduction is highly dependent on the actual heart
of MSCT coronary angiography is the diagnostic accuracy rate, and does not always result in an improvement of image
based on the individual patients. Using 4-slice MSCT, Giesler quality. Using a bi-segmental reconstruction algorithm at a
et al. showed that in 39 out of 100 patients (39%), all vessels rotation time of 500 ms, a heart rate of 68 bpm results in an
could be evaluated (15). In the study by Knez et al., the accu- effective temporal resolution of 125 ms, while at 80 bpm the
racy to detect no, single, double, or triple vessel disease was noncomplementary configuration of the X-ray source and
74% (32/43 patients) after exclusion of the nonassessable seg- detector array allows no reduction at all, maintaining a 250-ms
ments (12). Nieman et al. reported a 56% (45/78 patients) effective temporal resolution. Up to approx 75 bpm, the rela-
accuracy to distinguish no, single, or multivessel disease, with- tive temporal resolution, the ratio of the image reconstruction
out the exclusion of nonassessable segments (16). A high heart interval to the RR interval, is less than 30%. The use of more
rate affects also the diagnostic accuracy per patient in a nega- than two segments for reconstruction of a set of slices, which
tive way (16). By 16-slice MSCT, Nieman et al. reported a potentially reduces the duration of the image reconstruction
patient-based accuracy to distinguish no, single, or multivessel interval per cycle to less than 100 ms, requires a very slow table
disease of 78% (Table 2) (17). In this study, the number of advance, resulting in an increased radiation exposure to the
diseased vessels was overestimated in a number of patients, but patient.
 CHAPTER 30 / CTA TO DETECT CORONARY STENOSIS 329

Motion Artifacts editing of the ECG. This can be useful to correct for inappro-
Motion artifacts are probably the most important limitation priate interpretation of the ECG by the reconstruction algo-
of MSCT coronary angiography, and lead to substantial num- rithm, and provides an opportunity to manually insert R-wave
bers of nonassessable investigations. The high number of non- indicators in case of ECG noise. Editing of the ECG will not
assessable vessels reduces the clinical applicability of the suffice in patients with continuous arrhythmia. For instance,
technique. Two studies evaluated the diagnostic accuracy of during atrial fibrillation, the end-diastolic volume constantly
MSCT in relation to the heart rate of the patient. Giesler et al. varies because of the alternating filling time. Thereby, the heart
divided 100 patients into four groups and showed that in pa- will be displaced and have a different shape and position at each
tients with a heart rate below 60 bpm, motion artifacts occurred consecutive heart cycle and acquisition. Apart from cardiac
in only 8% of the coronary arteries, compared to 18% at a heart motion artifacts, this results in severe and noncorrectable
rate between 61 and 70 bpm, 41% at a heart rate between 71 and interslice discontinuity and noninterpretable results. On the
80 bpm, and 22% at a heart rate of more than 80 bpm. The other hand, nonsinus rhythm, delayed conduction, or other-
respective percentage of nonassessable vessels (22%, 23%, wise unusual configuration of the ECG is no contra-indication
50%, and 24%) resulted in a degrading overall sensitivity to for MSCT, as long as the RR interval variation is within an
detect >70% coronary diameter narrowing67%, 55%, 35%, acceptable range.
and 22%, for the respective heart-rate groups (15). In a study by
Nieman et al., 78 patients were equally divided into 3 groups HIGHLY ATTENUATING MATERIAL
according to the average heart rate during MSCT coronary STENTS AND SURGICAL MATERIAL
angiography. In the low-heart-rate group (564 bpm), inter- Material with strong X-ray attenuating characteristics, such
mediate-heart-rate group (673 bpm), and high-heart-rate as metal and calcium, cause beam hardening and partial volume
group (829 bpm), the number of assessable segments were artifacts. Because stents are positioned within the coronary,
78%, 73%, and 54%, resulting in an overall sensitivity to detect adjacent to the lumen, assessment of the lumen diameter is
>50% lumenal stenosis of 82%, 61%, and 32%, respectively. impaired. In patients who underwent bypass grafting, sternal
The accuracy of MSCT to classify patients as having no, single, wires and vascular clips can cause streak artifacts that hamper
or multivessel disease, without exclusion of nonassessable seg- proper assessment of the bypass grafts as well as coronary
ments, was 73%, 54%, and 42%, for each respective group (16). arteries. Both patient groups will be discussed in more detail in
Based on these and other experiences, many centers have intro- following chapters. Occasionally, pacemaker wires in the right
duced the administration of antichronotropic medication, such heart can cause identical artifacts, obscuring the RCA. In case
as -blockers, particularly in patients with higher heart rates, to of biventricular pacing systems, with wires positioned in the
reduce the occurrence of motion artifacts and improve the ac- cardiac veins, assessment of the left circumflex and the left
curacy of MSCT coronary angiography. anterior descending coronary arteries is severely hampered.
State-of-the-art scanners now have a rotation time below Calcifications
500 ms, and combined with sufficient heart-rate control, the Calcium deposits also cause a strong attenuation of the
reliability of MSCT has substantially improved (Table 1). It X-ray and are the most frequent cause of high-density artifacts,
needs to be established up to what heart rate these faster scan- such as partial-volume and beam-hardening artifacts. Partial-
ners can acquire motion-sparse images, but it seems unlikely volume artifacts are directly but not solely related to the size of
that a rotation time of 400 ms or more provides sufficient image the voxel, or three-dimensional image elements. The direct
quality in the majority of patients with a heart rate over 80 bpm, result is that calcified plaque material appears larger than it
when the coronary arteries are concerned. actually is, thereby increasing the apparent severity of the lumen
RESPIRATION AND THE SCAN TIME narrowing and complicating accurate assessment (Fig. 2). By
Respiratory motion is suppressed during scanning by main- experience, most reviewers will take the overestimation into
taining an inspiratory breath-hold. Using 4-slice scanners the account when assessing the lumen diameter of a calcified lesion.
relatively long scan time of 3545 s that is required to scan the Nevertheless, an accurate semi-quantitative assessment of
entire heart at a narrow collimation can be too long in a substan- coronary arteries with extensive coronary calcification remains
tial number of patients. In addition to respiratory motion arti- less reliable, and patients with suspected or known advanced
facts, the long breath-hold increases the patients heart rate, coronary artery disease are therefore not the most suitable can-
resulting in an increased occurrence of cardiac motion artifacts. didates for CT coronary angiography.
The new generation MSCT scanners are equipped with up to 16 All studies comparing MSCT and conventional coronary
slices and have a faster rotation, which results in a total scan angiography report that extensive calcification of the coro-
time below 20 s. A breath-hold of 20 s can be performed by nary arteries prevented assessment of a substantial number of
most patients, and does not result in a significant acceleration segments and resulted in a number of false-positive or false-
of the heart rate. negative interpretations of significant stenoses (1019). In
ARRHYTHMIA order to avoid contrast-enhanced CT angiography in these
Inappropriate ECG-synchronization results in interslice dis- individuals, some proposed to perform a low-dose nonen-
continuity (Fig. 2). Contrary to prospectively ECG-triggered hanced scan in all patients prior to angiography, to determine
modalities, such as EBCT and most MRI sequences, ECG- the amount of calcium in the coronary arteries and exclude
gated spiral CT image reconstruction allows for retrospective unsuitable candidates.
330 NIEMAN AND CADEMARTIRI

Table 3
Diagnostic Performance of Electron Beam CT to Detect Coronary Stenosis,
Using Conventional Coronary Angiography As the Standard of Reference

A N Assess. D Prev Excl. Se Sp PPV NPV Sea

Nakanishi (3) 3.0 37 Vessel 50% 0.8 74% 94% 68% 93% 74%
Reddy (4) 3.0 23 Vessel 50% 1.3 10% 88% 78% 65% 94% 77%
Schmermund (5) 3.0 + 28 Segment 50% 1.1 28% 82% 88% 57% 96% 70%
Rensing (6) 1.5 + 37 Segment 50% 1.1 81% 77% 94% 73% 95% 63%
Achenbach (7) 3.0 125 Vessel 70% 0.8 25% 92% 94% 78% 98% 70%
Budoff (8) 3.0 52 Vessel 50% 1.1 11% 78% 91% 78% 91% NR
Achenbach (9) 3.0 36 Vessel 75% 1.0 20% 92% 94% 85% 92% NR

A, use of atropine; Assess., method of assessment; D, diameter reduction considered significantly stenosed; Prev,
number of stenotic vessel per patient; Excl., percentage of excluded segments/branches. Sensitivity (Se), specificity
(Sp), positive (PPV) and negative predictive value (NPV) regarding the assessable segments/branches; NR, not reported.
aSensitivity including missed lesions in nonassessable segments/branches.

EBCT AND MR CORONARY ANGIOGRAPHY studies, many of the EBCT studies were limited to the proximal
ELECTRON-BEAM COMPUTED TOMOGRAPHY and middle coronary segments (6,7).
EBCT Coronary Angiography Future Developments
In 1997 the first studies comparing EBCT with conventional A recently introduced generation of EBCT scanners acquires
angiography were published. All but one study were performed two slices simultaneously, and is capable of scanning several
using an 1.0-mm overlapping 3.0-mm slice thickness (35, times during one heart cycle, which increases the radiation
79). In one study, a nonoverlapping 1.5-mm detector collima- exposure, but allows for retrospective selection of the data set
tion was applied (6). The nonmechanical EBCT is a sequential at the optimal phase. The slice acquisition time has been
CT scanner. Prospectively triggered by the patients electro- decreased to 50 ms, which will further improve the image qual-
cardiogram (ECG), the scanner acquires a single slice, after ity with regard to motion artifacts. Whether the use of thinner
which the table advances to the next slice position. The acqui- slices and faster scan times has a negative effect on the contrast-
sition is performed during the diastolic phase, and the exact to-noise ratio is currently unknown. Future studies will need to
timing of the electron generation is based on the preceding determine the incremental diagnostic value of this new EBCT
heart cycles. Because of the lack of mechanically rotating ele- technology.
ments, the slice acquisition time is very short: 100 ms. The one- MAGNETIC RESONANCE IMAGING
slice sequential scan design requires a long scan time and As a result of the lack of X-radiation and use of less harmful
breath-hold to cover the entire heart. To increase the scan cov- and optional contrast media, MRI is an attractive modality for
erage, atropine can be injected to increase the heart rate, and noninvasive imaging, including coronary angiography. Vari-
consequently the number of slices that can be acquired within a ous scanning techniques and data-acquisition sequences have
certain breath-hold time (6). been explored and compared to conventional coronary angiog-
Comparative Studies Against Conventional Angiography raphy. The first experiences in 1993 were very promising,
Table 3 lists the results from the comparative publications reporting a sensitivity and sensitivity to detect significant coro-
between contrast-enhanced EBCT and conventional coronary nary stenosis of 90% and 92% in 39 patients (1). Currently,
angiography for the purpose of the detection of significant respiratory-gated volumetric acquisitions of the entire heart are
coronary artery obstruction. The study populations ranged from acquired, or smaller targeted volumes are acquired during a
23 to 125 patients. The use of atropine to increase the number single breath-hold. Intravenous injection of contrast media is
of acquisitions per breath-hold was reported in two studies (6,7). possible but is not mandatory to image the coronary artery
At low heart rates, breath-hold durations of >50 s were reported. lumen. Many comparative studies using different 2- and 3D
In one study no segments were excluded (3); in the others, techniques have been published, with widely varying results
between 10% and 28% of the segments and vessels were (2). Between 4% and 48% of the vessels and segments needed
excluded due to impaired image quality. Similar to MSCT, to be excluded because of insufficient image quality. The sen-
noninterpretability was caused by motion artifacts and exten- sitivity and specificity to detect significant coronary obstruc-
sive calcification. Considering the assessable segments and tion ranged from 38% to 93% and 54% to 97%, respectively (2).
vessels, the sensitivity to detect significant lumenal narrowing A recent multicenter trial using a respiratory-gated free-breath-
ranged from 74% to 92%. The specificity ranged from 63% to ing scan protocol and a study population of 109 patients was
94%. The positive and negative predictive value ranged from published by Kim et al. in 2001 (21). The investigators reported
57% to 85% and 91% to 98%, respectively. If lesions in non- a sensitivity of 93% and specificity of 42% to detect significant
assessable vessels were included as false-negative results, the lesions in the assessable (84%) proximal and middle coronary
overall sensitivity decreased to 63% and 77%, although these segments. The overall sensitivity, including lesions in non-
figures were reported in only five studies. Contrary to the MSCT assessable segments, was not reported.
 CHAPTER 30 / CTA TO DETECT CORONARY STENOSIS 331

Despite its benign nature, MR coronary angiography is com- Table 4

plicated by a relatively low 3D image resolution of rarely less Potential Applications of CT Coronary Angiography
than 1 mm3, long scanning time, and inconsistent image qual- Early detection of stenoses in nonsymptomatics
ity. Compared to CT, the acquisition of the coronary MR is
Exclusion of coronary disease:
time-consuming and requires dedicated scanners and image
high-risk patients
sequences as well as experienced operators. prior to major (noncardiac) surgery
DISCUSSION Detection and/or exclusion of stenoses:
Atypical (unstable) chest pain
MSCT is currently the most accurate noninvasive Nonconclusive stress tests
angiographic modality for the detection of coronary stenosis. Substitution for diagnostic X-ray coronary angiography
Despite the use of radiation and contrast media, MSCT coro- Prior to percutaneous coronary intervention
nary angiography is a relatively safe and simple procedure. All High risk patients: aortic disease
data can be acquired within 20 s, often providing predictable Adjuvant to coronary angiography:
image quality, depending on the heart rate and the coronary Plaque characterization
calcium status of the patient. The contrast-to-noise ratio is high Complicated coronary intubation
and the 3D resolution of the current generation scanners is less Follow-up:
than 0.3 mm3. This high spatial resolution allows imaging of Percutaneous coronary intervention
small branches, often neglected in the MR and EBCT studies. Bypass surgery
CLINICAL IMPLEMENTATION
OF MSCT CORONARY ANGIOGRAPHY
However, within the foreseeable future, MSCT coronary
angiography will not replace coronary angiography as the ref- visualized well, and the value of plaque characterization by
erence coronary imaging tool. Conventional angiography is MSCT is currently being investigated and will be discussed in
performed without severe complications in the vast majority of the following chapters (22). Furthermore, MSCT presents a 3D
patients. Conventional angiography consistently provides high- depiction of the coronary arteries, which can be useful when a
quality data, with an excellent spatial resolution that allows coronary anomaly is suspected (23). Diseased vessels can eas-
quantitative assessment of the severity of the stenotic lesion. ily be related to an infarcted or perfusion-depleted myocardial
Apart from motion artifacts, image noise, or calcium-related segment. Besides the coronary arteries, the MSCT scan includes
artifacts, with a slice thickness between 0.5 and 1.0 mm, MSCT high-quality volumetric information on the entire heart and
can not be expected to provide comparable quantitative assess- lower lungs, resulting in (accidental) early detection of abnor-
ments. Conventional angiography can also be complemented malities, including pericardial disease, intra-cardiac thrombi,
by functional flow assessment and advanced plaque-imaging morphologic valvular disease (calcifications, thickening), lung
techniques. Finally, conventional angiography can immedi- tumors, and so on. Finally, the raw MSCT data can be used for
ately be followed by a percutaneous interventional procedure reconstruction of different cardiac phases, to evaluate the ven-
to treat the obstructive problem. tricular performance: ventricular cavity volumes, ejection frac-
In patients with a modest heart rate, MSCT could, however, tion, and regional myocardial wall thickening (24).
provide a useful and reliable alternative to diagnostic catheter- FURTHER IMPROVEMENT
based angiography for the initial detection and localization of To further improve the quality and quantitative potential of
coronary stenoses. Additionally, because of its noninvasive MSCT coronary angiography, fundamental characteristics such
nature, MSCT coronary angiography can be introduced into the as the spatial and temporal resolution need to be further opti-
diagnostic work-up of patients with anginal complaints at an mized. Evaluation of three-dimensional MSCT angiograms
earlier stage, when catheter-based angiography is not yet indi- may become more efficient and better reproducible with dedi-
cated. Potential applications are the exclusion of an acute coro- cated post-processing tools. More sophisticated tools that com-
nary obstruction in patients with atypical chest pain at the bine an accurate, reproducible assessment with presentable
emergency ward, coronary artery stenosis in patients who need overviews are currently being developed, and will improve the
major (noncardiac) surgery, or obstructive disease in patients clinical implementation of MSCT coronary angiography as a
with inconclusive stress test. MSCT may also be valuable when noninvasive tool to localize obstructive coronary artery disease.
repeated angiographic follow-up is indicated, or after percuta-
REFERENCES
neous coronary intervention or coronary artery bypass surgery
(Table 4). 1. Manning WJ, Li W, Edelman RR, et al. A preliminary report com-
paring magnetic resonance coronary angiography with conventional
ADDITIONAL VALUE OF MSCT angiography. N Engl J Med 1993;328832.
CORONARY ANGIOGRAPHY 2. Fayad ZA, Fuster V, Nikolaou K, et al. Computed tomography and
Besides being a noninvasive alternative to conventional magnetic resonance imaging for noninvasive coronary angiography
coronary angiography, MSCT provides additional and possi- and plaque imaging. Current and potential future concepts. Circula-
tion 2002;106:20262034.
bly valuable information with respect to the coronary artery 3. Nakanishi T, Ito K, Imazu M, et al. Evaluation of coronary artery
wall that is not provided by standard X-ray coronary angiogra- stenoses using electron-beam CT and multiplanar reformation.
phy (Figs. 1, 3, and 5). Nonstenotic atherosclerotic material is J Comput Assist Tomogr 1997;21:121127.
332 NIEMAN AND CADEMARTIRI

4. Reddy G, Chernoff DM, Adams JR, et al. Coronary artery stenoses: 15. Giesler T, Baum U, Ropers D, et al. Noninvasive visualization of
assessment with contrast-enhanced electron-beam CT and axial coronary arteries using contrast-enhanced multidetector CT: influ-
reconstructions. Radiology 1998;208:167172. ence of heart rate on image quality and stenosis detection. AR Am
5. Schmermund A, Rensing BJ, Sheedy PF, et al. Intravenous electron- J Roentgenol 2002;179:911916.
beam computed tomographic coronary angiography for segmental 16. Nieman K, Rensing BJ, van Geuns RJ, et al. Non-invasive coronary
analysis of coronary artery stenoses. J Am Coll Cardiol 1998;31: angiography with multislice spiral computed tomography: impact of
15471554. heart rate. Heart 2002;88:470474.
6. Rensing BJ, Bongaerts A, van Geuns RJ, et al. Intravenous coronary 17. Nieman K, Cademartiri F, Lemos PA, et al. Reliable noninvasive
angiography by electron beam computed tomography: a clinical coronary angiography with fast submillimeter multislice spiral com-
evaluation. Circulation 1998;98:25092512. puted tomography. Circulation 2002;106:20512054.
7. Achenbach S, Moshage W, Ropers D, et al. Value of electron-beam 18. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery
computed tomography for the noninvasive detection of high-grade stenoses with thin-slice multi-detector row spiral computed tomog-
coronary-artery stenoses and occlusions. N Engl J Med 1998;339: raphy and multiplanar reconstruction. Circulation 2003;107(5):
19641971. 664666.
8. Budoff MJ, Oudiz RJ, Zalace CP, et al. Intravenous three-dimen- 19. Nieman K, Rensing BJ, van Geuns RJM, et al. Usefulness of
sional coronary angiography using contrast-enhanced electron beam multislice computed tomography for detecting obstructive coronary
computed tomography. Am J Cardiol 1999;83:840845. artery disease. Am J Cardiol 2002;89:913918.
9. Achenbach S, Ropers D, Regenfus M, et al. Contrast enhanced elec- 20. Wang Y, Watts R, Mitchell I, et al. Coronary MR angiography:
tron beam computed tomography to analyse the coronary arteries in selection of acquisition window of minimal cardiac motion with
patients after acute myocardial infarction. Heart 2000;84:489493. electrocardiography-triggered navigator cardiac motion prescan-
10. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography ninginitial results. Radiology 2001;218:580585.
with multi-slice computed tomography. Lancet 2001;357:599603. 21. Kim WY, Daniel PG, Stuber M, et al. Coronary magnetic resonance
11. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary angiography for the detection of coronary stenoses. N Engl J Med
artery stenoses by contrast-enhanced, retrospectively electrocardio- 2001;345:18631869.
graphically-gated, multislice spiral computed tomography. Circula- 22. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection
tion 2001;103:25352538. and evaluation of atherosclerotic coronary plaques with multislice
12. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral computed tomography. J Am Coll Cardiol 2001;37:14301435.
computed tomography angiography for determination of coronary 23. Ropers D, Gehling G, Pohle K, et al. Anomalous course of the left
artery stenoses. Am J Cardiol 2001;88:11911194. main or left anterior descending coronary artery originating from the
13. Vogl TJ, Abolmaali ND, Diebold T, et al. Techniques for the detec- right sinus of Valsalva. Identification of four common variations by
tion of coronary atherosclerosis: multi-detector row CT coronary electron beam tomography. Circulation 2002;105:e42e43.
angiography. Radiology 2002;223:212220. 24. Dirksen MS, Bax JJ, de Roos A, et al. Usefulness of dynamic
14. Kopp AF, Schrder S, Kttner A, et al. Non-invasive coronary multislice computed tomography and left ventricular function in
angiography with high resolution multidetector-row computed to- unstable angina pectoris and comparison with echocardiography.
mography: results in 102 patients. Eur Heart J 2002;23:17141725. Am J Cardiol 2002;90:11571160.
 CHAPTER 31 / COMPLEMENTARITY OF CALCIUM SCORING AND CTA 333

31 Complementary Use of Coronary

Calcium Scoring and CT Angiography

ALEXANDER W. LEBER, MD

INTRODUCTION ing, is more demanding, requiring the administration of con-

Electron-beam computed tomography (EBCT) and multi- trast agent and additional radiation exposure.
slice CT (MSCT) are used for the detection and quantification
CALCIUM SCORING COMBINED WITH CTA
of coronary calcium, which is a good indicator of coronary
atherosclerosis and total plaque burden (16). It is suggested
TO DETERMINE PLAQUE BURDEN
that the coronary calcium score can be used to predict future AND PLAQUE MORPHOLOGY
coronary events in asymptomatic patients (37). However, Despite advances in our understanding of the pathogenesis
acute coronary events are initiated by rupture or superficial of atherosclerosis, coronary heart disease is still the leading
erosion of vulnerable coronary plaques, and these plaques are cause of death in Western societies. Approx 50% of all myocar-
not necessarily calcified (811). Therefore, for purposes of risk dial infarctions occur in patients with no prior symptoms. It is
stratification, the assessment of noncalcified plaques seems to well established that the risk for plaque rupture is predicted by
be necessary as well. Besides estimation of plaque burden and plaque burden and plaque composition. Reliable and accurate
risk stratification, evidence from several studies suggests that assessment of the composition of coronary atherosclerosis is
the CT-derived calcium score can be helpful in identifying currently achieved mainly by invasive methods like intra-
coronary artery disease (CAD) in symptomatic patients (12,13). coronary ultrasound or angioscopy (21). Since these are
In most studies, coronary calcium score cut-off points have invasive procedures, they are not suitable for preventive investi-
been presented, revealing high sensitivities and specificities to gations in asymptomatic patients. EBCT and MSCT enable an
diagnose or to exclude coronary stenoses (12,13). On an indi- accurate identification and quantification of calcified coronary
vidual basis, however, the estimation of luminal obstruction plaques noninvasively (14). The extent of coronary calcium
based on a calcium score cut-off point remains difficult. There- roughly reflects total plaque burden, and it is suggested that the
fore the consensus of most studies is that the use of age- and coronary calcium score is a powerful predictor of future coro-
gender-related percentiles is preferable (1214). Patients with nary events (413). However, myocardial infarction is initiated
scores in the highest percentile have a high likelihood for a by rupture or superficial erosion of vulnerable coronary
relevant stenosis, and in the lower percentiles, stenoses can be plaques, and these plaques are not necessarily calcified, as
excluded. According to these findings, scores in the interme- calcium is considered to be a frequent feature of stable lesions.
diate percentiles have no diagnostic value (14). Furthermore, Furthermore, calcified plaques reflect only the peak of the entire
recent studies have reported that, especially in young patients atherosclerotic burden, and the ratio of calcified to noncalcified
and in patients with unstable coronary artery disease, coronary plaques is highly variable among individuals (711). In animal
stenoses may be present in the absence of any coronary calcium studies and in a study using intravascular ultrasound, it has
(14,15). Both EBCT and MSCT offer the opportunity to visu- been demonstrated that an increase of plaque density in
alize coronary arteries directly after the administration of con- noncalcified plaques is part of a stabilization processes under
trast agent (1618). It has been demonstrated that noncalcified drug therapies (i.e., statin therapy) (22,23). For a more precise
plaques and coronary stenoses can be detected by CT angiog- determination of plaque burden and to monitor the course of
raphy (CTA) (1620). Therefore, the complementary use of coronary atherosclerosis, it is therefore necessary to determine
calcium scoring and CTA provides incremental information also noncalcified plaques. Recently it has been demonstrated
with regards to stenosis detection, determination of plaque that besides the detection of calcified plaques, after the admin-
burden, and plaque morphology. However, it has to be kept in istration of contrast agent, MSCT also allows the identification
mind that noninvasive angiography, in contrast to calcium scor- of noncalcified plaques with high accuracy (Fig. 1). Further-
more, noncalcified plaques can be further differentiated on the
basis of their CT density (19,20). In a first clinical study, coro-
From: Contemporary Cardiology: CT of the Heart: nary calcium scores in patients with myocardial infarction were
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
significantly lower than in stable CAD, whereas the number of

333
334 LEBER

Fig. 1. Detection of noncalcified plaques by multislice CT in relation to intravascular ultrasound (IVUS).

noncalcified plaques was higher, resulting in a similar plaque revealing high sensitivities and specificities to diagnose or to
burden among both groups. This study demonstrated that by exclude coronary stenoses. However, on an individual basis the
determining only calcified lesions, total plaque burden was estimation of luminal obstruction based on a calcium score cut-
significantly underestimated in patients with myocardial off point remains difficult. Therefore, the consensus of most
infarction (15). In another study, Schroeder and colleagues studies is that the use of age- and gender-related percentiles
found that in a set of 68 asymptomatic patients with a distinc- (score ranges) is preferable. A review of the literature reveals
tive risk profile, 53% of patients revealed calcified and that on average patients with scores in the highest percentile
noncalcified lesions, and 12.5% of patients revealed only (>75th percentile) have a high likelihood (>75% of cases) for a
noncalcified lesions (24). These initial results indicate that the relevant stenosis, and in the lower percentiles (< 25th) stenoses
possibility to identify and quantify noncalcified plaques may can be excluded in more than 85% of cases. According to these
improve future attempts to detect and monitor coronary athero- findings, scores in the intermediate percentiles (25th75th)
sclerosis in its preclinical stage. However the prognostic impact reveal an accuracy that is statistically not different from chance.
of noncalcified plaques is unknown so far, and it has to be By using intravenous contrast-enhanced EBCT angiography
determined in future large-scale studies. Therefore, contrast- and MSCT angiography, which allows the visualization of
enhanced scans for atherosclerosis screening should be per- coronary stenoses in the proximal and mid portions of the coro-
formed only in the setting of research trials. To date there is no nary arteries, calcium scoring can effectively be supported (14).
evidence to implement coronary plaque imaging by MSCT in EBCT angiography has been evaluated by a number of inves-
the clinical practice. tigators (18,2530). In most studies, high sensitivities (range
9378%) and high specificities (range 9888%) for the detec-
CALCIUM SCORING AND CTA tion of significant coronary stenoses have been reported after
TO EVALUATE PATIENTS excluding 2030% of coronary segments because of nondiag-
WITH SUSPECTED CORONARY STENOSES nostic image quality. Similar results have been found for
Evidence from several studies suggests that CT-derived 4-slice CT (18,17). With the development of 16-slice CT tech-
calcium scores can be helpful in identifying CAD in patients nology, noninvasive coronary angiography has become clini-
with chest pain. However, the estimation of luminal obstruc- cally more robust and more accurate, allowing the assessment
tion based on the calcium score remains difficult (17). In most of almost all coronary segments with diagnostic image quality
studies, calcium score cut-off points have been presented, (16,31). However, heavily calcified coronary plaques, which
 CHAPTER 31 / COMPLEMENTARITY OF CALCIUM SCORING AND CTA 335

Fig. 2. High-grade stenosis of the left anterior descending artery and the first diagonal branch in a patient with no coronary calcium.

on the one hand can obscure coronary stenoses and on the other observation to recent 16-slice CT studies, it is likely that this
hand may cause hardening artifacts that can resemble stenoses, technology with improved spatial and temporal resolution will
are still a major limitation impairing diagnostic accuracy of all provide a significant improvement of diagnostic accuracy.
CT technologies. Especially this limitation might be overcome
CT ANGIOGRAPHY IN PATIENTS
by a combined approach of calcium scoring and CTA.
WITH NO CORONARY CALCIUM
CALCIUM SCORING AND CTA IN PATIENTS OR LOW CALCIUM SCORES
WITH HIGH CALCIUM SCORES The results of the available studies indicate that exclusion of
There is only one study evaluating the complementary use any coronary calcium provides an extremely high negative
of EBCT calcium scoring and EBCT angiography to diagnose predictive value ranging from 90 to 99% to rule out coronary
significant coronary stenoses (14). In this study it was demon- stenoses. Almost the same values are obtained for calcium
strated that in the case of high coronary calcium scores the scores below the 25th percentile. However, there is evidence
diagnostic accuracy of EBCT angiography is severely affected. that especially in young patients, in smokers, and in patients
Of 31 patients with scores >310, EBCT angiography revealed with acute chest pain, the diagnosis of a significant luminal
diagnostic image quality in 27 patients. Accuracy for detecting obstruction is not a rare finding (32). In a set of patients with
coronary stenoses was 77% and was statistically not different atypical chest pain, 4 of 31 patients revealed a significant steno-
from calcium scoring alone. While on the one hand extensive sis despite a low or zero calcium score (14). In 3 of these 4
calcifications hinder the evaluation of CT angiograms, on the patients, EBCT angiography identified a high-grade stenosis.
other hand they indicate advanced atherosclerosis with a high In another study, 2 of 21 patients with myocardial infarction
probability for the presence of at least one high-grade stenosis. had no coronary calcium, but CTA by 4-slice CT identified the
Therefore, it is reasonable to avoid contrast-enhanced scans in culprit stenoses (Fig. 3) (15). Therefore, an additional CTA
patients with severe calcium scores. will add incremental diagnostic safety in symptomatic patients
who are younger than 40 yr, who are smokers, and who prob-
CTA IN PATIENTS WITH INTERMEDIATE ably report from a more unstable course of chest pain.
BORDERLINE CALCIUM SCORES
As already mentioned, intermediate calcium scores are very POSSIBLE FUTURE CLINICAL ALGORITHM FOR
difficult to handle, and in the clinical setting in the individual THE COMPLEMENTARY USE OF CALCIUM:
patient they have no diagnostic value. In a first study compar- SCORING AND CTA
ing calcium scoring and EBCT angiography, intermediate cal- Owing to the lack of prospective studies in large patient
cium scores (sensitivity <85% and specificity <80%) were cohorts, to date there exists neither a clinical recommendation
found in 33% of patients. In a review of the literature, equivocal nor a clinical guideline to routinely perform coronary calcium
calcium scores occur in up to 50% of patients (17). In these scoring or CTA in general practice. The algorithm and related
patients, CTA is of particular benefit. In a recent study, EBCT flowchart represented by Fig. 4 provides a better understanding
angiography guided the correct diagnosis in 80% of patients of the utility of the combined CT modalities for future diagnos-
with equivocal calcium scores (Fig. 2). By transferring this tic strategies.
336 LEBER

Fig. 3. Proximal 65% stenosis of the left anterior descending artery (arrows) in a patient with a borderline calcium score (176). (A) selective
coronary angiography. (B) Electron beam CT (EBCT) angiography, volume-rendering technique. (C) EBCT angiography, shaded surface
rendering technique. Figure taken from Leber et al. (14).

REFERENCES 8. Virmani R, Kolodgie F, Burke A, et al. Lessons from sudden

1. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcifica- coronary death: a comprehensive morphological classification
tion: pathophysiology, epidemiology, imaging methods, and clini- scheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol
cal implications. A statement for health professionals from the 2000;20(5):12621275.
American Heart Association Writing Group. Circulation 1996;94: 9. Kragel A, Reddy S, Wittes J, et al. Morphometric analysis of the
11751192. composition of atherosclerotic plaques in the four major epicardial
2. Knez A, Becker C, Becker A, et al. Determination of coronary cal- coronary arteries in acute myocardial infarction and in sudden coro-
cium with multi-slice spiral computed tomography: a comparative nary death. Circulation 1989;80(6):17471756.
study with electron-beam CT. Int J Cardiovasc Imaging 2002;18: 10. Virmani R, Burke A, Farb A. Coronary risk factors and plaque
295303. morphology in men with coronary disease who died suddenly. Eur
3. Wayhs R, Zelinger A, Raggi P. High coronary artery calcium scores Heart J 1998;19:678680.
pose an extremely elevated risk for hard events. J Am Coll Cardiol 11. Virmani R, Burke AP, Kolodgie FD, et al. Prognostic value of coro-
2002;39:225230. nary calcification and angiographic stenoses in patients undergoing
4. Wong ND, Budoff MJ, Pio J, Detrano RC. Coronary calcium and coronary angiography. J Am Coll Cardiol 1996;27:285290.
cardiovascular event risk: evaluation by age- and sex-specific 12. Guerci AD, Spadaro LA, Popma JJ, et al. Relationship of coronary
quartiles. Am Heart J 2002;143:456459. calcium score by EBCT to arteriographic findings in asymptomatic
5. Detrano RC, Doherty TM, Davies MJ, Stary HC. Predicting coro- and symptomatic adults. Am J Cardiol 1997;79:128133.
nary events with coronary calcium: pathophysiologic and clinical 13. Haberl R, Becker A, Leber A, et al. Correlation of coronary calcifi-
problems. Curr Probl Cardiol 2000;25:374402. cation and angiographically documented stenoses in patients with
6. Pohle K, Ropers D, Maffert R, et al. Coronary calcifications in young suspected coronary artery disease: results of 1,764 patients. J Am
patients with first, unheralded myocardial infarction: a risk factor Coll Cardiol 2001;37(2):451457.
matched analysis by electron beam tomography. Heart 2003;89: 14. Leber A, Knez A, Mukherjee R, et al. Usefulness of calcium scoring
625628. using electron beam computed tomography and noninvasive coro-
7. OMalley PG, Taylor AJ, Jackson JL, Doherty TM, Detrano RC. nary angiography in patients with suspected coronary artery disease.
Prognostic value of coronary electron-beam computed tomography Am J Cardiol 2001;88(3):219223.
for coronary heart disease events in asymptomatic populations. Am 15. Leber AW, Knez A, White CW, et al. Composition of coronary
J Cardiol 2000;85:945948. atherosclerotic plaques in patients with acute myocardial infarction
 CHAPTER 31 / COMPLEMENTARITY OF CALCIUM SCORING AND CTA 337

Fig. 4. Proposed algorithm for the complementary use of calcium scoring and CT coronary angiography.

and stable angina pectoris determined by contrast-enhanced 25. Moshage WE, Achenbach S, Seese B, Bachmann K, Kirchgeorg M.
multislice computed tomography. Am J Cardiol 2003;91:714718. Coronary artery stenoses: three-dimensional imaging with electro-
16. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama cardiographically triggered, contrast agent-enhanced, electron-beam
PM, de Feyter PJ. Reliable noninvasive coronary angiography with CT. Radiology 1995;196:707714.
fast submillimeter multislice spiral computed tomography. Circula- 26. Chernoff DM, Ritchie CJ, Higgins CB. Evaluation of electron
tion 2002;106:20512054. beam CT coronary angiography in healthy subjects. AJR Am J
17. Achenbach S, Ulzheimer S, Baum U, et al. Noninvasive coronary Roentgenol 1997;169:9399.
angiography by retrospectiveely gated multislice spiral CT. Circu- 27. Reddy GP, Chernoff DM, Adams JR, Higgins CB. Coronary artery
lation 2000;102:28232828. stenoses: assessment with contrast-enhanced electron-beam CT and
18. Leber AW, Knez A, Becker C, et al. Non-invasive intravenous coro- axial reconstructions. Radiology 1998;208:167172.
nary angiography using electron beam tomography and multislice 28. Schmermund A, Rensing BJ, Sheedy PF, Bell MR, Rumberger JA.
computed tomography. Heart 2003;89:633639. Intravenous electron-beam CT coronary angiography for segmental
19. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection analysis of coronary artery stenoses. J Am Coll Cardiol 1998;31(7):
and evaluation of atherosclerotic coronary plaques with multislice 15471554.
computed tomography. J Am Coll Cardiol 2001;37:1430435. 29. Achenbach S, Moshage W, Ropers D, Nossen J, Daniel WG. Value
20. Leber AW, Knez A, Becker A, et al. Accuracy of multi detector of electron beam computed tomography for the noninvasive detec-
spiral CT in identifying and differentiating the composition of coro- tion of high-grade coronary-artery stenoses and occlusions. N Engl
nary atherosclerotic plaques. A comparative study with intracoro- J Med 1998;339:19641971.
nary ultrasound. J Am Coll Cardiol 2004;43:12411247. 30. Rensing BJ, Bongaerts A, van Geuns RJ, van Ooijen P, M
21. Waxman S. Characterization of the unstable lesion by angiography, Oudkerk, PJ de Feyter. Intravenous coronary angiography by elec-
angioscopy, and intravascular ultrasound. Cardiol Clin 1999;17: tron beam computed tomography: a clinical evaluation. Circulation
295305. 1998;98(23):2509-2512.
22. Schartl M, Boksch W, Koschyk D, et al. Use of intravascular ultra- 31. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery
sound to compare effects of different strategies of lipid-lowering stenoses with thin-slice multi-detector row spiral computed tomo-
therapy on plaque volume and composition in patients with coronary graphy and multiplanar reconstruction. Circulation 2003;107:
artery disease. Circulation 2001;104(4):387389. 664666.
23. Gomberg-Maitland, Fuster V, Fayad Z, et al. Statins and plaque 32. Schmermund A, Baumgart D, Adamzik M, et al. Comparison of
stability. J Cardiovasc Risk 2003;10(3):161167. electron-beam computed tomography and intracoronary ultrasound
24. Schroeder S, Kopp A, Kuettner A, et al. Prevalence of noncalcified in detecting calcified and noncalcified plaques in patients with acute
plaques in asymptomatic patients with a high atherosclerotic risk coronary syndromes and no or minimal to moderate angiographic
profile. Z Kardiol 2002;Suppl. coronary artery disease. Am J Cardiol 1998;81:141146.
 CHAPTER 32 / CT VS MRI FOR CORONARY ARTERIES 339

32 CT vs Magnetic Resonance
for Imaging of the Coronary Arteries

ARMIN HUBER, MD

Coronary arteries are small tortuous vessels with physiologi- during the RR cycle and therefore reduction of motion artifacts.
cal motion by respiratory motion and cardiac contraction. The In order to achieve longer rest time of the coronary arteries
noninvasive imaging of normal and diseased coronary arteries during diastole, -blockers are used to reduce heart rate.
by magnetic resonance (MR) angiography has undergone Multislice CT allows for reliable detection and quantification
numerous technical improvements since its introduction in of coronary calcifications in unenhanced data sets acquired
1993. Early coronary MR techniques used a combination of with low-dose ECG triggering or in contrast-enhanced CTA
segmental acquisition of the k-space data and breath-holding. data sets acquired with ECG gating. As coronary CTA is pos-
More recent techniques allow use of both electrocardiogram sible with higher image quality compared to EBCT, calcium
(ECG) triggering and respiratory gating, eliminating the need screening is possible with comparable diagnostic accuracy
for patient breath-holding. Thus the acquisition time can be (1,2), and multislice CTs are widely available and useful for
longer than a single breath-hold, and spatial resolution can be many other indications than cardiac imaging, multislice CT
increased substantially. Three generations of MR coronary replaced EBCT in the majority of centers.
angiography techniques have been described and can be used MR angiography (MRA) is possible with high diagnostic
with or without contrast agents. accuracy comparable to that of digital subtraction angiography
Potential clinical applications of MR imaging (MRI) of the in many vascular territories, such as renal arteries, carotid
coronary arteries include coronary lesion detection, delinea- arteries, and the peripheral vascular tree (3,4). The technique
tion of congenital coronary artery anomalies, characterization that shows the highest diagnostic accuracy and short acquisi-
of previously known coronary lesions with imaging of the ves- tion times is contrast-enhanced 3D MRA with T1-weighted 3D
sel wall, coronary anatomy after transplantation, and quantifi- gradient echo sequences. For coronary MRA, it is still not clear
cation of coronary flow reserve by velocity encoding MR of the which MRA technique will be the technique of choice in future.
coronaries at rest and under adenosine stress. Pioneering work with first-generation coronary MRA
Electron beam computed tomography (EBCT) has been the started 1993. The MRA technique was a 2D gradient-echo tech-
method of choice to image coronary arteries by computed nique, which means that one slice is acquired during a single
tomography (CT), as the temporal resolution of 100 ms was breath-hold with ECG triggering and segmented k-space
much higher than that of helical CT. Unenhanced EBCT of the acquisition. The initial enthusiasm of the first very positive and
coronary arteries is useful to detect and quantify coronary cal- promising preclinical reports could not be reproduced by other
cifications. Variable studies have been reported about contrast- investigators, who found more negative results. Initial clinical
enhanced CT angiography (CTA) of the coronary arteries with testing of the first-generation coronary MRA technique was
EBCT (Fig. 1). One of the major limitations is the slice thick- reported by Manning et al. (5). They examined 39 patients and
ness of 3 mm or the limited volume that can be imaged during determined the diagnostic accuracy of MRA compared to con-
a breath-hold when the slice thickness is reduced. Another limi- ventional coronary angiography for detection of stenosis and
tation is limited signal-to-noise ratio (SNR) that can be achieved occlusions. In this first study, coronary MR had a sensitivity of
with EBCT. 90% and a specificity of 92% for correctly identifying indi-
Since introduction of multislice CT with retrospective ECG vidual vessels with 50% or greater angiographic stenosis. The
gating, contrast-enhanced data sets of the coronary arteries can corresponding positive and negative predictive values were
be acquired with high spatial resolution, thin slice thickness, 0.85 and 0.95. The overall sensitivity and specificity for cor-
and high SNR. The drawback of lower temporal resolution rectly classifying individual patients as having or not having
compared to electron beam CTA is compensated by retrospec- serious coronary disease were 97% and 70%. Subsequent
tively chosen optimal temporal window for data reconstruction attempts of imaging the coronary arteries by MRA by other
investigators to reproduce these results have not been success-
From: Contemporary Cardiology: CT of the Heart: ful. Duerinckx et al. found that a sensitivity of 90% for lesion
Principles and Applications detecting by using first-generation coronary MRA was opti-
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ mistic and could not be reproduced in a double-blinded pro-

339
340 HUBER

Fig. 1. (A) Thin-slice maximum intensity projection reconstruction of electron beam CT angiography data set shows high-grade stenoses of
the left anterior descending artery (arrow). (B) Right anterior oblique projection of conventional coronary angiography shows high-grade
stenoses of the left anterior descending artery (arrow).

spective study (6). In some of the patients, very promising data acquisition time is too long compared to the first pass time of
sets with spectacular image quality could be acquired; how- the contrast agent. Moreover, the navigator techniques usually
ever, first-generation coronary MRA was not consistent enough are based on a time-of-flight MRA technique, which means the
to reproduce the results first reported by Manning et al. Post degree of stenosis is overestimated by flow voids and the SNR
et al. found a sensitivity for detection of significant coronary is limited. Recent techniques allow the combination of 3D
artery lesions of 36% in an initial study and results of 33% to TrueFISP and 3D Flash techniques (Figs. 5,6) with respiratory
75% in an follow-up study (7). An important practical problem navigator gating. For respiratory gating, different techniques
of the first-generation 2D coronary MRA is the potential of can be used. First, retrospective respiratory gating with con-
overreading or underreading coronary artery lesions. As only one tinuous data acquisition and retrospective selection of suitable
slice can be acquired during a single breath-hold, the course of data at the same diaphragm position; second, prospective res-
tortuous coronary arteries is not strictly in plane. Thus more piratory gating with acquisition of data only at the correct dia-
slices than one have to be acquired during different breath- phragm position; and third, slice correction algorithms with
holds to assess a longer part of one vessel, and partial-volume correction of data position dependent on the diaphragm posi-
effects can decrease diagnostic accuracy as well as different tion, with a correction factor of 0.3 to 0.6. For second-genera-
respiratory positions. tion techniques with 3D acquisition and respiratory gating,
To overcome the problems of singleslice, single-breath- various values for diagnostic accuracy were found. Mller et al.
hold acquisitions, the second-generation techniques introduce found a sensitivity of 87% and a specificity of 97%. Huber
respiratory gating. With the combination of respiratory gating et al. found a sensitivity of 73% and a specificity of 50% (8),
and ECG triggering, a continuous high-resolution data set of and Nikolaou et al. found a sensitivity and specificity of and
the entire heart can be acquired with a 3D technique. The res- 71%/72% and 53%/60% for navigator echo MRA with and
piratory gating can be performed with bellows, respiratory without the use of a slice interpolation technique (9). One of the
belts, and with noninvasive navigator pulses positioned on the major problems is that excellent image quality can be achieved
dome of the right diaphragm (Fig. 2). Most studies used non- in some of the patients, but image quality is variable and cannot
invasive navigator pulses. The advantages of these techniques be reproduced reliably at a high standard. Therefore, diagnostic
are the ability to image the coronaries despite their tortuous accuracy decreases in larger patient populations. In contrast to
course continuously in a high-resolution data set (Figs. 3,4). the already reported studies, Lethimonnier used a navigator
The spatial resolution can be increased, as the scan time for one technique with prospective instead of retrospective respiratory
data set is not restricted to the length of a single breath-hold. A gating (10). This technique reduces the acquisition time, but
disadvantage is that the entire examination time can be very diagnostic accuracy could not be improved compared to the
long because of the double trigger/gating technique (respira- studies with retrospective respiratory gating.
tory and ECG). Another disadvantage is that usual extravascular Kim et al. examined 109 patients scheduled for X-ray coro-
contrast agents are not suitable to increase SNR because the nary angiography with free-breathing coronary MRA (11).
 CHAPTER 32 / CT VS MRI FOR CORONARY ARTERIES 341

Fig 2. (A) Slab of navigator pulse positioned on the dome of the right diaphragm. (B) Noninvasive prospective detection of the position of the
diaphragm during respiratory cycle.

Fig. 3. Circumflex artery: multiplanar reformation reconstruction of Fig. 4. Left anterior descending artery: multiplanar reformation
high-resolution magnetic resonance (MR) coronary angiography data reconstruction of high-resolution magnetic resonance (MR) coronary
set acquired with retrospectively gated navigator echo MR angiogra- angiography data set acquired with retrospectively gated navigator
phy technique. Modified from ref. 8. echo MR angiography technique. Modified from ref. 8.

They found that 84% of the proximal and middle segments of quence technique is similar to a contrast-enhanced T1w 3D
coronary arteries were interpretable on MRA; 83% of clinically gradient-echo technique, which is used for many other vascu-
significant lesions (more than 50% reduction in diameter on lar territories with high image quality and high diagnostic
X-ray angiography) were identified. However, the sensitivity, accuracy. However, with the combination of ECG triggering
specificity, and accuracy for patients with disease of the left and the reduction of scan time to a single breath-hold, spatial
main coronary artery or three-vessel disease were 100%, 85%, resolution is limited and the acquisition time in the RR cycle
and 87%. The negative predictive values for any coronary artery is relatively long (250 ms). Compared to multislice CT, the
disease and for left main artery or three-vessel disease were acquisition window during the RR cycle is similar; however,
81% and 100%. the prospective ECG-triggering technique leads to a lower
The third-generation techniques allow acquisition of a whole image quality concerning cardiac motion artifacts than the
3D slab during a single breath-hold with the use of strong gra- retrospectively ECG-gated technique, which is used for CTA
dient systems and fast T1- or T2-weighted gradient echo pulse in multislice CTA of the coronary arteries. Wielopolski et al.
sequence techniques, either 3D Flash or 3D TrueFISP. introduced the volume coronary artery targeted scan (VCAT)
Kessler et al. used a single transverse 3D slab that covers technique (13). This technique uses seven to nine small 3D
only the proximal coronary artery tree (12). The pulse-se- slabs to cover the entire coronary artery tree. With the use of
342 HUBER

Fig. 5. Drawing shows segmented data acquisition during RR interval with prospective respiratory gating and prospective electrocardiogram
gating with 3D Flash technique.

Fig. 6. Drawing shows segmented data acquisition during RR interval with prospective respiratory gating and prospective electrocardiogram
gating with 3D TrueFISP technique.
 CHAPTER 32 / CT VS MRI FOR CORONARY ARTERIES 343

Fig. 7. (A) Magnetic resonance (MR) coronary angiography. 3D TrueFISP breath-hold technique has a lower spatial resolution but a shorter
scan time than the respiratory triggered technique. (B) MR coronary angiography. 3D TrueFISP respiratory triggered technique has a higher
spatial resolution but a longer scan time than the respiratory triggered technique.

small 3D slabs instead of a single proximal 3D slab, spatial Continuous gantry rotation and table movement causes the
resolution can be improved significantly. The disadvantage of projection data to be obtained along a spiral or helical path. A
this technique is that a usual extravascular contrast agent can- new generation of MDCT permits working with 16 detector
not be used for seven or nine injections. Usually, at least a rows in a single gantry rotation and therefore improves spatial
single dose of Gd-DTPA (0.1 mmol/L) is necessary to perform resolution and reduces scan time of the entire heart by 50%.
contrast-enhanced MRA successfully. Thus, the number of With helical CT, it was not possible to image the coronary
contrast-material injections is restricted to three, which is not arteries with high spatial resolution and reduction of cardiac
enough for seven or nine 3D slabs. Therefore, the VCAT tech- motion by ECG triggering. The introduction of helical CT with
niques can be used unenhanced with limited signal-to-noise multirow detectors (4 rows) allows for a coverage of the entire
ratio or with the use of an intravascular contrast agent. The T2 heart during a single breath-hold (40 s) with high spatial reso-
weighted 3D TrueFISP pulse sequence is another method to lution and thinner slice thickness (1.25 mm) than is possible
image the coronary without contrast material, either with a with EBCT (3 mm). However, the acquisition window during
breath-hold approach or with a respiratory triggered technique the RR cycle is limited by a rotation time of 500 ms. This is
(Fig. 7). The advantage of this technique is that fluid and blood five times longer than with EBCT. The first step to reduce the
has an intrinsic high contrast in TrueFISP sequences without acquisition window is to use a 180-degree sector for image
the use of contrast material. reconstruction instead of a 360-degree sector. That means the
The heart can be imaged with the use of two different CT acquisition window is reduced to 250 ms. The second step to
modalities: one employs nonmechanical movement of the improve image quality is to choose the right delay time in
X-ray source (i.e., EBCT) and the other involves the motion of the RR cycle retrospectively after the scan to reconstruct the
the X-ray source and table, combined with multiple detectors images at the time of minimal motion during the RR interval.
to acquire the data in spiral or helical fashion (i.e., multi- Hong et al. showed that the rest time of each coronary artery
detector-row CT [MDCT]). In order to freeze cardiac motion, is different in each individual and is variable in different indi-
a cardiac-dedicated CT system was developed in 1982 on the viduals (14). Furthermore, the length of the rest time is depen-
basis of a nonmechanical movement of the X-ray source and dent on the patients heart rate. That means high image quality
fixed detector arrays. The EBCT uses a single, curved anode can be achieved in patients with low heart rates (under 65 beats
with four tungsten targets underneath the patient, and a focused per min [bpm]). Therefore, the third step to cope with the prob-
electron beam that is rapidly swept across these targets to pro- lem of the long acquisition window is to decrease the patients
duce an X-ray fan beam detected by two detector rows above heart rate with a -blocker. If it is not possible to decrease the
the patient. patients heart rate, a segmented algorithm can be used to ac-
Mechanical MDCT systems were introduced in 1998, and quire the data for one 180-degree sector reconstruction over
allow for scanning with one X-ray tube and 4 detector rows in two RR intervals. However, this method can cause additional
a single gantry rotating twice per second around the patient. motion artifacts.
344 HUBER

Fig. 8. Maximum intensity projection reconstruction of a multislice Fig. 9. 3D reconstruction of a multislice CT angiography (16 row)
CT angiography (16 row) data set shows normal anatomy of right data set shows normal anatomy of left anterior descending artery and
coronary artery. a diagonal branch.

Compared to EBCT, multislice CT has a higher spatial reso- Vogel et al. analyzed different CT reformation techniques in
lution, especially a smaller slice thickness and a higher SNR assessing the coronary arteries examined with multislice CT,
and contrast-to-noise ratio (CNR), as it is possible to apply a including 3D reformations, virtual endoscopic reformations,
higher radiation dose. In spite of the fact that the acquisition multiplanar reformations, and transverse source images (25).
window of multislice CTA of coronary arteries is much longer The highest diagnostic accuracy was found at transverse scanning.
than with coronary MRA or EBCT, cardiac motion artifacts can In comparison to conventional coronary angiography, sev-
be suppressed sufficiently in the majority of patients with retro- eral studies show that the sensitivity of multislice CTA in
spective ECG gating, where patients heart rate can be reduced detection of coronary artery stenosis cannot replace invasive
under 65 bpm (Figs. 8,9). angiography. However, the high negative predictive value in
CTA using EBCT was performed in various studies with all studies suggests use of the method in patients with a low
similar imaging parameters. In all studies, a single-slice mode pretest probability for having coronary artery disease, who
was used with a high in-plane matrix (512 512); the temporal would be scheduled for coronary angiography to exclude coro-
resolution was 100 ms in all studies (Table 1). The percentage nary heart disease. One limitation in detecting coronary artery
of depicted vessels with adequate image quality ranged from stenosis is the difficulty of correctly determining the degree of
72% to 90%i.e., in all studies a certain number of vessel narrowing of the luminal diameter. Another limitation is the
segments with reduced image quality had to be excluded from high density of coronary artery wall calcifications. In regions
further evaluation (1520). of strongly calcified vessels, the assessment of coronary artery
Nieman et al. found a sensitivity and specificity for a com- stenosis is very difficult. Compared to CT of the coronaries,
parison of multislice CTA (4 slices) with conventional coro- coronary MRA has the advantage that calcified plaques show
nary angiography of 82% and 93% and positive and negative low signal intensity and do not decrease diagnostic accuracy
predictive value of 66% and 97% in detection of stenosis for the assessment of coronary artery stenosis. The advantage
(=50%) (21). Knez et al. found a sensitivity and specificity of of multislice CTA is that calcified plaques can be identified and
83% and 78% for the detection of stenosis (=50%) and 67% for quantified in the same data set that is acquired after contrast-
the detection of occlusions, a specificity of 98% and a negative material application for CTA (26).
predictive value of 96% (22). Achenbach et al. excluded 32% Table 1 shows sensitivity and specificity of different
of the examined vessels and found a sensitivity and specificity noninvasive coronary angiography methods in comparison to
of 85% and 76% (23). Becker et al. found a sensitivity, speci- conventional coronary angiography, and the number of
ficity, and negative predictive value for the detection of stenoses excluded patients or vessel segments from the evaluation
(>50%) with multislice CTA of 81%, 90%, and 97%, respec- because of low image quality. The results show that none of the
tively (24). However, the agreement for determining the degree three methodsMRA, contrast-enhanced EBCT, or contrast-
of stenoses with multislice CT was only moderate ( = 0.58). enhanced MDCTcan replace conventional invasive coronary
 CHAPTER 32 / CT VS MRI FOR CORONARY ARTERIES 345

Table 1
Sensitivity and Specificity of Coronary CT and Magnetic Resonance (MR)
Angiography Studies As Compared With Conventional Angiography (Stenosis =50%)

No. of
No. of patients, vessels, or
Techniques patients segments excluded (%) Sensitivity % Specificity %

CT angiography with electron-beam CT

Nakanishi et al. 1997 (16) 37 4 patients (11% ) 74 94
Achenbach et al. 2000 (17) 36 29 vessels (20%) 92 94
Schmermund et al. 1998 (18) 28 93 segments (28%) 82 88
Reddy et al. 1998 (19) 23 7 vessels (10%) 88 63
Buddoff et al. 1999 (20) 52 23 vessels (11%) 78 91
Achenbach et al. 1998 (21) 125 124 vessels (25%) 92 94

CT angiography with multidetector CT

Niemann et al. 2002 (22) 53 55 segments (30%) 82 93
Knez et al. 2001 (23) 44 29 segments (6%) 78 98
Achenbach et al. 2001 (24) 64 82 vessels (32%) 85 76
Becker et al. 2002 (25) 28 81 90
Vogl et al. 2002 (26) 64 38 segments (19%) 75 99

MR angiography, first generation, 2D breath-hold

Manning et al. 1993 (5) 39 9 vessels (6%) 90 92
Post et al. 1997 (7) 35 15 vessels (11%) 63 89

MR angiography second generation, 3D retrospective respiratory gating

Post et al. 1996 (38) 20 3 vessels (4%) 38 95
Mueller et al. 1997 (39) 35 35 segments (14%) 83 94
Sandstede et al. 1999 (40) 30 7 patients (23%) 81 89
Huber et al. 1999 (9) 20 45 segments (31%) 79 54
Sardanelli et al. 2000 (41) 42 39 segments (14%) 82 89
Gonschior et al. 2001 (42) 20 45 segments (31%) 79 54
Nikolaou et al. 2001 (10) 40 75 segments (27%) 72 60

3D prospective respiratory gating

Lethimonnier 1999 (11) 20 3 patients (15%) 65 93
Kim et al. 2001 (12) 109 123 segments (16%) 93 42

MR angiography third generation, 3D breath-hold

Van Geuns et al. 2000, 38 85 segments (31%) 68 97
unenhanced (43)
Regenfus et al. 2000, 50 82 segments (23%) 86 91
contrast-enhanced (44)

angiography. However, multislice CTA is a promising method Fuster and Fayad showed that the coronary artery wall can
in excluding coronary artery disease in patients with a low be imaged with MRI with T1, T2, and PD weighted TSE pulse
pretest probability of having coronary artery disease, as all sequences, which allow imaging of atherosclerotic plaques with
studies show a high negative predictive value and a high speci- different contrast depending on the pulse sequence (29,30).
ficity. The study performed by Kim et al. shows that coronary High-resolution MR has emerged as the potential leading
MRA has the potential role to detect or exclude three-vessel noninvasive in vivo imaging modality for atherosclerotic
disease or stenoses of the left main coronary artery (11). plaque characterization. MR differentiates plaque components
Moreover, contrast-enhanced multislice CTA can detect not on the basis of biophysical and biochemical parameters such as
only calcified plaques but also noncalcified plaques (27,28) chemical composition and concentration, water content, physi-
and indicate atherosclerosis in an early stage. That means cal state, molecular motion, or diffusion. MR provides imaging
multislice CTA, as a result of its high CNR, can demonstrate without ionizing radiation and can be repeated over time. In
not only the reduction of luminal diameter by noncalcified vivo MR plaque imaging and characterization have been per-
plaques but also noncalcified plaques in regions where no sig- formed utilizing a multicontrast approach with high-resolution
nificant stenosis is located . Owing to their risk for plaque rup- black blood spin echo and fast spin echobased MR sequences.
ture, first attempts were made to classify noncalcified plaques The signal from the blood flow is rendered black through pre-
by their density using multislice CTA. paratory pulses (e.g., radio-frequency spatial saturation or
346 HUBER

inversion recovery pulses) to better image the adjacent vessel REFERENCES

wall. High-resolution black-blood MR of both normal and ath- 1. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium
erosclerotic human coronary arteries was performed for direct measurement: agreement of multirow detector und electron beam
assessment of coronary wall thickness and the visualization of CT. AJR Am J Roentgenol 2002;176:12951298.
focal atherosclerotic plaque in the wall. The difference in maxi- 2. Knez A, Becker C, Becker A, et al. Determination of coronary
calcium with multi-slice spiral computed tomography: a compara-
mum wall thickness between the normal subjects and patients tive study with electron-beam CT. Int J Cardiovasc Imaging
(>40% stenosis) was statistically significant. The coronary MR 2002;18:295303.
plaque imaging study by Fayad et al. was performed during 3. Huber A, Scheidler J, Wintersperger B, et al. Moving-table MR
breath-holding to minimize respiratory motion, with a spatial angiography of the peripheral runoff vessels: comparison of body
coil and dedicated phased array coil systems. AJR Am J Roentgenol
resolution of 0.46 0.46 2.0 mm3. To alleviate the need for 2003;180:13651373.
the patient to hold his or her breath, Botnar et al. (31) have 4. Schoenberg SO, Essig M, Bock M, et al. Comprehensive MR evalu-
combined the black-blood fast spin echo (FSE) method and a ation of renovascular disease in five breathholds. J Magn Reson
real-time navigator for respiratory gating and real-time slice Imaging 1999;10:347356.
position correction, and achieved a near isotropic spatial reso- 5. Manning WJ, Li W, Edelman RR. A preliminary report comparing
magnetic resonance coronary angiography with conventional
lution (0.7 0.7 1 mm3). This method provided a quick way angiography. N Engl J Med 1994;328:828832.
to image a long segment of the coronary artery wall, and may 6. Duerinckx AJ, Urman MK. Two-dimensional coronary MR angiog-
be useful for rapid coronary plaque burden measurement. raphy: analysis of initial clinical results. Radiology 1994;193(3):
Future studies need to address the possibilities. 731738.
7. Post JC; van Rossum AC, Hofman MBM, et al. Clinical utility of
As shown in animal experimental studies, MR is a powerful two-dimensional magnetic resonance angiography in detecting coro-
tool to serially and noninvasively investigate the progression nary artery disease. Eur Heart J 1997;18:426433.
and regression of atherosclerotic lesions in vivo (32,33). In 8. Huber A, Nikolaou K, Gonschior P, et al. Navigator echo-based
asymptomatic, untreated, hypercholesterolemic patients with respiratory gating for three-dimensional MR coronary angiography:
carotid and aortic atherosclerosis, Corti et al. (34) showed that results from healthy volunteers and patients with proximal coronary
artery stenoses. AJR Am J Roentgenol 1999;173:95101.
MR can be used to measure the efficacy of lipid-lowering 9. Nikolaou K, Huber A, Knez A, et al. Navigator echo-based respira-
therapy with statin. Atherosclerotic plaques were assessed with tory gating for 3D-MR coronary angiography: reduction of scan
MR at different times after lipid-lowering therapy. Significant time using a slice-interpolation technique. J Comput Assist Tomogr
regression of atherosclerotic lesions was observed, despite the 2001;25:378387.
10. Lethimonnier F, Furber A, Morel O, et al. Three-dimensional coro-
early and expected hypolipidemic effect of the statins. Changes nary artery MR imaging using prospective real-time respiratory
in the vessel wall were observed for 12 mo. As with previous navigator and linear phase shift processing: comparison with con-
experimental studies, there was a decrease in the vessel wall ventional coronary angiography. Magn Reson Imaging 1999;17:
area and no alteration in the lumen area after 12 mo (35,36). 11111120.
Recently, a case-controlled study demonstrated substantially 11. Kim YW, Danias PG, Stuber M, et al. Coronary magnetic resonance
angiography for the detection of coronary stenoses. Circulation
reduced carotid plaque lipid content without change in overall 2001;247:18631869.
lesion area in patients treated for 10 y with an aggressive lipid- 12. Kessler W, Laub G, Achenbach S, Ropers D, Moshage W, Daniel
lowering regime compared with untreated control (37). WG. Coronary arteries: MR angiography with fast contrast-en-
Comparing multislice CTA and MRI of the vessel wall, CTA hanced three-dimensional breath-hold imaginginitial experience.
Radiology 1999;210:566572.
is the imaging method that allows to examine the coronary 13. Wielopolski PA, van Geuns RJ, de Feyter PJ, Oudkerk M. Breath-
arteries lumen, and detect calcified plaques and noncalcified hold coronary MR angiography with volume-targeted imaging.
plaques during a single breath-hold (2628). However MRI has Radiology 1998;209:209219.
the potential to classify the plaques that are already identified 14. Hong C, Becker CR, Huber A, et al. ECG-gated, retrospectively
with different pulse sequence and different contrast at a defined reconstructed multidetector-row CT coronary angiography: effect
of varying trigger delay on image quality. Radiology 2001;220:
region. In contrast to multislice CTA, the scan time is much 712717.
longer with MRI. Therefore, for a screening approach, 15. Nakanishi T, Ito K, Imazu M, et al. Evaluation of coronary artery
multislice CTA is superior to MRI in detection of vessel wall stenoses using electron-beam CT and multiplanar reformations.
plaques. J Comput Assist Tomogr 1997;21:121127.
16. Achenbach S, Ropers D, Regenfus M, et al. Contrast enhanced elec-
Coronary artery stents are a major problem for noninvasive tron beam computed tomography to analyse the coronary arteries in
imaging methods. EBCT, multislice CT, and MRI show arti- patients after acute myocardial infarction. Heart 2000;84:489493.
facts at the regions where stents are implanted. Therefore, 17. Schmermund A, Rensing BJ, Sheedy PF, Bell MR, Rumberger JA.
in-stent re-stenosis cannot be assessed directly, as the stent Intravenous electron-beam computed tomographic coronary angiog-
lumen is not visible without artifacts. One method to identify raphy for segmental analysis of coronary artery stenoses. J Am Coll
Cardiol 1998;31(7):1547-1554.
in-stent re-stenosis is velocity encoding MRI of the coronary 18. Reddy GP, Chernoff DM, Adams JR, et al. Coronary artery stenoses:
arteries. Nagel et al. performed a study comparing intracor- assessment with contrast-enhanced electron-beam CT and axial
onary flow wire examinations at rest and under adenosine stress, reconstructions. Radiology 1998;208:167172.
and determined the coronary flow reserve both with MRI and 19. Budoff MJ, Oudiz RF, Zalace CP, et al. Intravenous three-dimen-
sional coronary angiography using contrast enhanced electron beam
with invasive measurements (44). They found very promising computed tomography. Am J Cardiol 1999;83:840854.
results, which allow identification of in-stent re-stenosis 20. Achenbach S, Moshage W, Ropers D, et al. Value of electron-beam
noninvasively and determine their hemodynamical relevance. computed tomography for the noninvasive detection of high-grade
 CHAPTER 32 / CT VS MRI FOR CORONARY ARTERIES 347

coronary-artery stenoses and occlusions. N Engl J Med 1998;339: 33. Helft G, Worthley SG, Fuster V, et al. Progression and regression of
19641971. atherosclerotic lesions monitoring with serial noninvasive magnetic
21. Niemann K, Rensing BJ, van Geuns RJ, et al. Usefulness of resonance imaging. Circulation 2002;105:993998.
multislice computed tomography for detecting obstructive coronary 34. Corti R, Fayad ZY; Fuster V, et al. Effects of lipid-lowering by
artery disease. Am J Cardiol 2002;89:913918. simvastation on human atherosclerotic lesions: a longitudinal study
22. Knez A, Becker CR, Leber A, et al. Usefulness of multislice spiral by high-resolution, noninvasive magnetic resonance imaging. Cir-
computed tomography angiography for determination of coronary culation 2001;104:249252.
artery stenoses. Am J Cardiol 2001;88:11911194. 35. Zhao XQ, Yuan C, Hatsukami TS, et al. Effects of prolonged inten-
23. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary sive lipid-lowering therapy on the characteristics of carotid athero-
artery stenoses by contrast-enhanced, retrospectively electrocardio- sclerotic plaques in vivo by MRI: a case-controlled study.
graphically-gated, multislice spiral computed tomography. Circula- Arterioscler Thromb Vasc Biol 2001;21(10):1623-1629.
tion 2001;103:25352538. 36. Manning WJ, Li W, Edelman RR. A preliminary report comparing
24. Becker CR, Knez A, Leber A, et al. Detection of coronary artery magnetic resonance coronary angiography with conventional
stenoses with multislice helical CT angiography. J Comput Assist angiography. N Engl J Med 1993;328:828832.
Tomogr 2002;26(5):750755. 37. Post JC; van Rossum AC; Hofman MBM, et al. Three-dimensional
25. Vogl TJ, Abolmaali ND; Dibold T, et al. Techniques for the detec- respiratory-gated MR angiography of coronary arteries: comparison
tion of coronary atherosclerosis: multi-detector row CT coronary with conventional coronary angiography. AJR Am J Roentgenol
angiography. Radiology 2002;223:212220. 1996;166:13991404.
26. Hong C, Becker CR, Schoepf UJ, et al. Coronary artery calcium: 38. Mueller MF, Fleisch M, Kroeker R, et al. Proximal coronary artery
absolute quantification in nonenhanced and contrast-enhanced stenosis: three-dimensional MRI with fat saturation and navigator
multi-detector row CT studies. Radiology 2002;223:474480. echo. J Magn Reson Imaging 1997;7:644651.
27. Becker CR, Knez A, Leber A, Treede H, Haberl R, Reiser MF. 39. Sanstede JJ, Pabst T, Beer M, et al. Three-dimensional MR coronary
[Angiography with multi-slice spiral CT. Detecting plaque, angiography using the navigator technique compared with conven-
before it causes symptoms.] MMW Fortschr Med 2001;143(16): tional coronary angiography. AJR Am J Roentgenol 1999;172:135
3032. 139.
28. Leber AW, Knez A, White CW, et al. composition of coronary 40. Sardanelli F, Molinari G, Zandrino F, et al. Three-dimensional,
atherosclerotic plaques in patients with acute myocardial infarc- navigator-echo MR coronary angiography in detecting stenoses of
tion and stable angina pectoris determined by contrast-enhanced the major epicardial vessels, with conventional coronary angiog-
multislice computed tomography. Am J Cardiol 2003;91:714 raphy as the standard of reference. Radiology 2000;214:808814.
718. 41. Gonschior P, Pragst I, Valassis G, et al. High-resolution MR
29. Fayad ZA, Nahar T, Fallon JT, et al. In vivo MR evaluation of angiography: results in diseased arteries. J Invasive Cardiol 2001;
atherosclerotic plaques in the human thoracic aorta: a comparison 13:151157.
with TEE. Circulation 2000;101:25032509. 42. van Geuns RJ, Wielopolski PA, de Bruin HG, et al. MR coronary
30. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human angiography with breath-hold targeted volumes: preliminary clini-
coronary artery lumen and wall imaging using black-blood mag- cal results. Radiology 2000;217:270277.
netic resonance imaging. Circulation 2002;102:506510. 43. Regenfus M, Ropers D, Achenbach S, et al. Noninvasive detection
31. Botnar RM, Kim WY, Bornert P, et al. 3D coronary vessel wall of coronary artery stenosis using contrast-enhanced three-dimen-
imaging utilizing a local inversion technique with spiral image sional breath-hold magnetic resonance coronary angiography. J Am
acquisition. Magn Reson Med 2001;46:848854. Coll Cardiol 2000;36:4450.
32. McConnel MV; Aikawa M, Maier SE, et al. MRI of rabbit athero- 44. Nagel E, Thouet T, Klein C, et al. Noninvasive determination of coro-
sclerosis in response to dietary cholesterol lowering. Arterioscler nary blood flow velocity with cardiovascular magnetic resonance in
Thromb Vasc Biol 1999;19:19561959. patients after stent deployment. Circulation 2003;107:17381743.
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 349

CONTRAST-ENHANCED CT
OF THE HEART:
PRINCIPLES OF ATHEROSCLEROSIS VI
AND VESSEL WALL IMAGING
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 351

33 Pathology and Pathophysiology

of Coronary Atherosclerotic Plaques

RENU VIRMANI, MD, ALLEN P. BURKE, MD, FRANK D. KOLODGIE, PhD,

ANDREW FARB, MD, ALOKE V. FINN, MD, AND HERMAN GOLD, MD

INTRODUCTION noncellular components include lipid, proteoglycans, collagen,

The death rate from coronary artery disease has declined in elastic fibers, calcium, iron (hemosiderin), and blood compo-
the past few decades through greater understanding of risk fac- nents, including fibrin and factor VIII. These various compo-
tors of coronary heart disease as well as through better treat- nents help to form the various lesions that are recognized as part
ment, including the creation of coronary care units. However, of the atherosclerotic process. These include adaptive intimal
because of the lack of an animal model of unstable plaque, our thickening (AIT), intimal xanthomas, pathologic intimal thick-
understanding of atherosclerotic plaque morphology comes ening (PIT), fibroatheroma, thin-cap fibroatheroma, plaque
only from static histology of lesion morphology in patients rupture, plaque erosion, calcified nodule, fibrocalcific plaque,
dying of acute coronary syndromes (1). Although transgenic healed plaque rupture, and fibrous plaque (either from healed
models of atherosclerosis have markedly enhanced our under- plaque erosion or propagated thrombus). The best-known of
standing of certain aspects of plaque progression and regres- the classifications is the American Heart Association report by
sion, they have failed thus far to explain the relationship of the Stary et al. (2,3). The various lesions of atherosclerosis will be
coagulation parameters and plaque morphology that precipi- described and the plaque morphologies that lead to coronary
tate coronary thrombosis (1). Until we are able to create a better thrombosis along with the precursor lesion of plaque rupture,
model or study plaque morphology prospectively and deter- also known as thin-cap fibroatheroma or vulnerable plaque.
mine the mechanisms and the anatomic markers of progres-
ADAPTIVE INTIMAL THICKENING (AIT)
sion, we will make progress very slowly. This review is based
on examination of human coronary artery pathology in patients AIT occurs in most arteries once flow is established in utero
dying a sudden coronary death, in order to ascertain the patho- or soon after birth, consisting primarily of smooth muscle cells,
logic lesion morphologies that are linked to plaque progression which are strongly -actin positive and surrounded by a
and thrombosis, which will be necessary for us to be able to proteoglycan-rich matrix (Fig. 1A). Macrophages are rarely
recognized by invasive or noninvasive means the prospective detected. These lesions are most prominent at branch points and
lesions that are likely to produce symptoms. are considered by many to be precursor lesions of atherosclero-
sis. Kim et al. have shown in hypercholesterolemic swine that
CLASSIFICATION OF ATHEROSCLEROSIS atherosclerotic lesions arise almost exclusively from intimal cell
Our understanding of atherosclerosis has been enhanced by masses. Similarly, we have observed that lesions of pathologic
the development of the various classifications that have come intimal thickening with lipid pools arise at sites of branch points
from the insights of scientists like Virchow, who was a pioneer where adaptive intimal thickening is prominent, and therefore
in pathology. Systematic studies of lesion development, are most likely the precursor lesions of atherosclerosis.
described by the giants of atherosclerosis including Robert
INTIMAL XANTHOMAS (FATTY STREAK)
Wissler, Herbert Stary (Table 1), Henry McGill, and Michael
Davies in the last century, have made enormous contributions We have used the microscopic term intimal xanthoma for
to the better understanding of early lesions, the influence of risk fatty streak, the corresponding designation for the lesion as
factors on plaque progression, and the role of plaque rupture in grossly seen in the aorta. We believe intimal xanthoma is not a
the occurrence of luminal thrombosis. lesion of atherosclerosis, as it eventually regresses in humans.
The cellular participants of atherosclerosis include smooth The Pathobiologic Determinants of Atherosclerosis in Youth
muscle cells, endothelial cells, macrophages, T- and B-lym- (PDAY) study showed that fatty streaks are prominent in the
phocytes, red cells, platelets, neutrophils, and basophils. The third decade in the thoracic aorta, but regress later in life.
The site of advanced lesions is generally the abdominal aorta,
From: Contemporary Cardiology: CT of the Heart:
Principles and Applications where fatty streaks are relatively uncommon. Fatty streaks are
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ rich in macrophages (hence the term xanthoma) but lack
necrotic cores or lipid pools (Fig. 1B).
351
352 VIRMANI ET AL.

Table 1
Atherosclerotic Plaque Classifications

Virmani et al.
Initial Progression
Early plaques Type I: Microscopic detection Intimal thickening None
of lipid droplets in intima
and small groups of macrophage
foam cells
Type II: Fatty streaks visible on gross Intimal xanthoma None
inspection, layers of foam cells,
occasional lymphocytes and mast cells
Type III (intermediate): Extracellular lipid Pathologic intimal Thrombosis (erosion)
pools among layers of smooth muscle cells thickening
Intermediate plaque Type IV: Well-defined lipid core; may Fibrous cap atheroma Thrombosis (Erosion) a
develop surface disruption (fissure)
Late lesions Type Va: New fibrous tissue overlying Thin cap fibroatheroma Thrombus (rupture)
lipid core (multilayered fibroatheroma) b Healed plaque rupture, Repeated rupture or erosion
erosion with or without total
occlusion
Type Vb: Calcification c Fibrocalcific plaque (with
or without necrotic core)
Type Vc: Fibrotic lesion with minimal lipid
(could be result of organized thrombi)
Miscellaneous/ Type VIa: Surface disruption Calcified nodule Thrombus (usually nonocclusive)
complicated features Type VIb: Intraplaque hemorrhage
Type VIc: Thrombosis
a May further progress with healing (healed erosion).
b May overlap with healed plaque ruptures.
c Occasionally referred to as type VII lesion. Modified from ref. 26.

PATHOLOGIC INTIMAL THICKENING (PIT) and lymphocytes. The thickness of the fibrous cap varies; a thin
PIT lesion is also rich in smooth muscle cells in a cap is believed to impart instability, and therefore warrants a
proteoglycan matrix (Fig. 2A). However, the area close to the separate designation of thin cap fibroatheroma.
media shows loss of smooth muscle cells and an accumulation We believe that the necrotic core originates from foam cell
of fat as a lipid pool, which stains positive with oil red O. A infiltration of the lipid pool and eventual breakdown of the
significant number of smooth muscle cells close to the media foamy cell, which occurs as a result of apoptosis or necrosis.
show accumulation of intracellular fat and may appear foamy, How the necrotic core enlarges and becomes rich in free cho-
best appreciated by transmission electron microscopy. The lesterol is poorly understood. Hemorrhage into a plaque, either
smooth muscle cells may appear as ghosts with surrounding through the breakdown of the vasa vasorum or possibly through
thickened basement membrane, which stains strongly Periodic plaque fissuring, may play an important role. It is possible that
Acid Schiff (PAS) positive (3a). The surrounding matrix also the vasa vasorum become leaky, allowing fibrinogen to leak
stains positive for oil red O and often shows the presence of into the plaque if pressure within them rises, as they have a
monohydrate cholesterol, which may appear as cholesterol poorly developed basement membrane and often lack pericytes
crystals (lipid pools). There may be macrophage infiltration in and smooth muscle cells. In support of this concept is the fre-
the superficial regions of the plaque, but these macrophages are quent finding of fibrin, by immunohistochemical methods,
not in proximity to the lipid pools. It is believed that smooth within the necrotic core.
muscle cells are undergoing apoptosis, degeneration, and cal-
THIN FIBROUS CAP ATHEROMA
cification (22). We believe that pathologic intimal thickening
is the precursor lesion of the fibroatheroma. The original description of Stary et al. did not mention the
thin fibrous cap atheroma (TCFA) as a precursor lesion of
FIBROUS CAP ATHEROMA plaque rupture. Partly based on their morphologic resemblance
Fibrous cap atheromas are lesions with a necrotic core and to acute ruptures, we believe that they are the precursor lesions.
a thick fibrous cap (Fig. 2B). The necrotic core is rich in acel- The TCFA is defined as a lesion with a necrotic core and an
lular debris and varies in size. The fibrous cap covers the overlying fibrous cap that is <65 m and infiltrated by mac-
necrotic core completely, is rich in smooth muscle cells within rophages (>25 per high power magnification [0.03-mm diam-
a collagen/proteoglycan matrix, and may contain macrophages eter field]) (4) (Fig. 3). The thickness criterion of TCFA was
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 353

Fig. 1. Adaptive intimal thickening (A) and intimal xanthoma (B). Lesions uniformly present in all populations, although intimal xanthomas
are more prevalent with exposure to a Western diet. Intimal xanthomas are commonly produced in animal models; however, they usually do
not develop into progressive atherosclerotic lesions. Both lesions occur soon after birth; the intimal xanthoma (otherwise known as a fatty
streak) is known to regress. Intimal thickening consists mainly of smooth muscle cells in a proteoglycan-rich matrix, while intimal xanthomas
primarily contain macrophage-derived foam cells, T lymphocytes, and varying degrees of smooth muscle cells. Reproduced from ref. 1, with
permission.

chosen because 95% of fibrous caps adjacent to acute plaque plaque rupture (the precursor of which is presumably the
rupture measured <65 m, with a mean cap thickness of 23 TCFA), erosion, and calcified nodule (Fig. 5). The most fre-
19 m. We have also compared other morphologic characteris- quent cause of thrombosis being plaque rupture (over 60%),
tics between plaque rupture and TCFA. The necrotic core is plaque erosion is the second most frequent (about 35%),
smaller than that seen in rupture (Table 2). In addition, the whereas calcified nodule is a rare cause of thrombosis (<5%).
percent area occupied by the necrotic core is greater in rupture These three types of thrombosis occur, respectively, in the set-
than TCFA, although we were unable to determine any signifi- ting of a TCFA, fibroatheroma or pathologic intimal thicken-
cant differences in length (Table 2).The cross-sectional lumi- ing, and a calcified plaque with or without a necrotic core (Fig. 1).
nal narrowing of TCFA is less than that of acute rupture; nearly PLAQUE RUPTURE
80% of TCFAs had luminal narrowing <75% in contrast to Plaque rupture is defined as a disruption, discontinuous
rupture (Fig. 4). fibrous cap, underlying necrotic core, and superimposed lumi-
nal thrombus. Ruptured plaques that result in occlusive luminal
LESIONS WITH THROMBI thrombi almost invariably have a large necrotic core, which
From studies carried out in patients dying sudden coronary occupies =25% of the plaque area in 80% of cases. Plaque
death, we have observed three main causes of luminal thrombi: rupture is the major underlying mechanism of luminal throm-
354 VIRMANI ET AL.

Fig. 2. Pathological intimal thickening (A) vs atheroma (B). Pathological intimal thickening is a poorly defined entity sometimes referred to
in the literature as an intermediate lesion. True necrosis is not apparent, and there is no evidence of cellular debris; some lipid may be present
deep in the lesion near the elastic lamina (EL), but it is dispersed. The fibrous cap overlying the areas of lipid is rich in smooth muscle cells
and proteoglycans. Some scattered macrophages and lymphocytes may also be present, but are usually sparse. The more definitive lesion or
fibrous cap atheroma classically shows a true necrotic core (NC), containing cholesterol esters, free cholesterol, phospholipids, and triglyc-
erides. The fibrous cap consists of smooth muscle cells in a proteoglycan-collagenous matrix, with a variable number of macrophages and
lymphocytes. The media underneath the plaque is often thin. Reproduced from ref. 1, with permission.

Table 2
Approximate Sizes of Necrotic Core in Fibroatheroma, Thin Cap Atheroma, and Acute Rupture

Plaque type

Dimension Fibroatheroma Thin cap atheroma Acute plaque rupture

Length, mm, mean/range 6 mm (range 118 mm) 8 mm (range 217 mm) 9 mm (range 2.5 22)
Cross sectional area, mm2 1.2 2.2 1.7 1.1 4.1 5.5
% cross sectional area 15 20% 23 17% 34 17%

Reproduced from ref. 27 with permission.

bosis in sudden death (60%), acute myocardial infarction the cap and release matrix metalloproteinases, which are
(75%), and unstable angina (70%), and is rarely present in stable responsible for the breakdown of the collagens (8,9). There is
angina (57). The mechanism of rupture is poorly understood; also evidence that macrophage myeloperoxidase may be
however, it is generally believed that macrophages infiltrate responsible for the disruption of the fibrous cap by producing
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 355

Fig. 3. Morphologic variants of the thin-cap atheroma. Thin caps may emerge in fibroatheromas with insignificant plaque burden and insig-
nificant luminal narrowing, in fibroatheromas with large cores that are eccentric or eccentric, and frequently in plaques with evidence of prior
rupture. (A) In this plaque with insignificant plaque burden, there is an area of proteoglycan-rich smooth muscle cells (arrowheads) suggestive
of a prior rupture, and multiple areas of thin caps (arrows). (B) The necrotic core (NC) is large, with an area of thinned cap (arrow). (C) The
necrotic core (NC) is concentric with an extensive area of cap thinning (arrow). (D) There may be a healed rupture (arrows) with a proteoglycan-
rich smooth muscle cell reparative layer (arrowhead).

Fig. 4. The distribution frequency of plaque ruptures (A,B) and thin-cap atheromas (C,D) by size of lipid core or lipid core as a percent of plaque
area (x axis). The majority of plaque ruptures occur when lipid core area forms 2550% of plaque area, or 13 mm2 lipid core area. In the case
of thin-cap atheromas, the degree of cross-sectional area luminal narrowing and area of necrotic core is shifted to the left (lesser or smaller)
as compared to plaque ruptures. Reproduced from ref. 27, with permission.
356 VIRMANI ET AL.

Fig. 5. Lesions with thrombi. Ruptured plaques are thin fibrous-cap atheromas with luminal thrombi (Th). These lesions usually have an
extensive necrotic core containing large numbers of cholesterol crystals, and a thin fibrous cap (<65 m) infiltrated by foamy macrophages
and a paucity of T-lymphocytes. The fibrous cap is thinnest at the site of rupture and consists of a few collagen bundles and rare smooth muscle
cells. The luminal thrombus is in communication with the lipid-rich necrotic core. Erosions occur over lesions rich in smooth muscle cells and
proteoglycans. Luminal thrombi overly areas lacking surface endothelium. The deep intima of the eroded plaque often shows extracellular lipid
pools, but necrotic cores are uncommon; when present, the necrotic core does not communicate with the luminal thrombus. Inflammatory
infiltrate is usually absent, but if present, it is sparse and consists of macrophages and lymphocytes. Calcified nodules are plaques with luminal
thrombi showing calcific nodules protruding into the lumen through a disrupted thin fibrous cap (FC). There is absence of endothelium at the
site of the thrombus, and inflammatory cells (macrophages, T-lymphocytes) are absent. Reproduced from ref. 1, with permission.

the pro-oxidant species, hypochlorous acid. We have shown There are relatively few macrophages and lymphocytes adja-
that up to 15% of macrophages within ruptured fibrous caps are cent to the thrombus in the majority of plaque erosions. Plaque
rich in myeloperoxidase. We have observed that the numbers of erosion accounts for 20% of all sudden coronary deaths and
plaque hemorrhages in the coronary arteries of patients dying 35% of coronary thrombi in patients dying suddenly with coro-
suddenly with acute plaque rupture are greater than in patients nary thrombosis (1,6). The mean age of patients dying with
with severe coronary disease dying with other forms of throm- plaque erosion is less than that of patients dying with acute
bosis or without coronary thrombosis. These data suggest that rupture, and there is less severe narrowing at sites of thrombo-
hemorrhage, due to plaque fissure or rupture of vasa vasorum, sis in plaque erosion. Plaque erosion accounts for over 80% of
may be precursors to plaque rupture. Davies has shown that thrombi occurring in women less than 50 yr of age. The lesions
11% of hearts have plaque fissure with intra-intimal thrombus, tend to be eccentric and are infrequently calcified (6). Plaque
2% have intimal thrombus as the only mechanism of sudden erosions tend to embolize more frequently than plaque rup-
death, while rupture with thrombosis was identified in 74%, ture (74%, vs 40%, respectively) (6). The risk factors for ero-
and only 5% had no acute coronary lesion (10). We have sion are poorly understood; similar to acute rupture, there is an
observed vasa vasorum to be more prominent in patients dying increased risk among smokers, but there appears to be a lesser
suddenly during exercise with plaque rupture, compared to association with dyslipidemia, as compared to plaque rupture.
patients dying at rest. Also, hemorrhage into a plaque was more The underlying plaque in erosions is generally pathologic inti-
frequent in patients dying during exercise than in patients dying mal thickening or fibrous cap atheroma. The most frequent
at rest (11). The rupture site has been reported to be the shoulder location for both erosion and rupture is the proximal left ante-
region in a majority of cases. However, we observed the most rior descending coronary artery (nearly two-thirds of patients)
frequent site to be the central region of the cap (42%), followed followed by the right coronary artery and the left circumflex.
by shoulder (36%) and circumferential (4%). In 18% of rup- However, patients dying with plaque erosion frequently do not
tures, it was not possible to determine the exact location because have extensive disease, unlike patients dying with stable plaque
of extensive destruction of the rupture site (11). or plaque rupture. Over one-half of deaths are seen in patients
PLAQUE EROSION with single-vessel disease, and one-quarter with double-vessel
Plaque erosion is defined as an acute thrombus in direct disease. The etiology of plaque erosion is poorly understood;
contact with intimal plaque rich in smooth muscle cells within however, it is believed that coronary vasospasm may be
a proteoglycan matrix, and an absence of endothelium (1,6). involved.
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 357

Fig. 6. Healed plaque rupture. (A) demonstrates areas of intraintimal lipid-rich core with hemorrhage and cholesterol clefts. (B) shows a higher
magnification of the looser smooth muscle cell formation within a collagenous proteoglycanrich neointima showing a clear demarcation with
the more fibrous regions of the old plaque to the right. (C) and (D) demonstrate the layers of collagen by Sirius red staining. (C), note the area
of dense dark red collagen surrounding the lipid hemorrhagic cores seen in corresponding view in panel A. (D) demonstrates an image taken
with polarized light. The dense collagen (Type 1) which forms the fibrous cap is lighter reddish-yellow and is disrupted (arrow), with the newer
greenish Type III collagen on the right and above the rupture site. (A and B, Movat pentachrome). Reproduced from ref. 28, with permission.

CALCIFIED NODULE thrombus. Healed ruptures in the coronary vascular bed are
The least frequent lesion that results in coronary luminal readily detected microscopically by the identification of the
thrombus is the calcified nodule, a plaque that contains calci- breaks in the fibrous cap, which is rich in type I collagen, and
fied plates at the site of luminal disruption. There may or may an overlying repaired lesion, which is richer in type III col-
not be necrotic core within the plaque, but the characteristic lagen. The healed lesion can be more easily recognized by
features are the breakage of calcified plates, often with a fibrotic Movat stain (healed site identified by the brilliant blue-green
reaction, interspersed fibrin, a disrupted surface, and an over- color of the proteoglycan-rich matrix) and confirmed by
lying thrombus. Occasionally, bone formation with osteoblasts picrosirius red staining and polarized microscopy. When
and osteoclasts is present (1). The lesion is located in the mid- viewed under polarized light, this stain highlights the breaks in
right coronary artery in 50% of cases, generally in a heavily the fibrous cap, which is rich in collagen type I, as yellow-red
calcified segment. It is more common in older male individuals birefringence with an underlying necrotic core (Fig. 6). The
than women. We have observed these lesions in the carotid plaque overlying the fibrous cap consists of smooth muscle
arteries as well as the coronaries, and speculate that their devel- cells in a proteoglycan matrix, which is rich in type III collagen
opment may be related to the repeated episodes of plaque hem- and has green birefringence under polarized light (sirius red
orrhage. stained) (11).
We and others believe that healing of a disrupted plaque is
HEALED PLAQUE RUPTURES the main stimulus for plaque progression and is a major factor
Not all plaque ruptures result in symptoms; therefore, there in causing chronic high-grade coronary stenosis once a late
should be morphologic hallmarks that could be recognized atherosclerotic necrotic core with a thin fibrous cap has formed
within plaques that are representative of a previous site of (12). This mechanism would explain the phasic rather than
358 VIRMANI ET AL.

frequency is as high as 40% when there are no other sites with

acute plaque rupture. Healed myocardial infarcts are seen in
90% of patients with a total occlusion (1). The lumen is com-
posed of dense collagen and/or proteoglycans with interspersed
capillaries, arterioles, smooth muscle cells, and macrophages.
In the case of total occlusions secondary to thrombi, the proxi-
mal and distal ends organize initially, followed by the middle
segment, which may demonstrate entrapped red cells and fibrin
without cellular ingrowth for long periods. Sites of total occlu-
sion are often associated with negative remodeling of the ves-
sel, probably related to collagen replacement of the thrombus
with eventual crosslinking, resulting in artery shrinkage. There
is often little calcification within total occlusions, probably
because the plaque has formed via organized thrombus, as
opposed to successive ruptures of necrotic cores with preserved
blood flow.
FIBROUS AND FIBROCALCIFIC PLAQUES
A subset of coronary artery plaques have little evidence of
lipid pool or necrotic core formation, and are designated as
fibrous or fibrocalcific plaques, depending on the presence of
calcification. The mechanism of plaque enlargement of such
plaques is unknown, but may be in some cases related to propa-
gated thrombus or healed plaque erosions (Fig. 8). These lesions
are rich in type I collagen, but type III collagen and
proteoglycans may also be present. The calcium deposition
may be related to calcified apoptotic smooth muscle cells in the
absence of a significant influx of macrophages and other
inflammatory cells. The term fibrous plaque has not been typi-
cally used in formal classifications of coronary artery athero-
sclerosis, partly because it is a general category that may
represent an etiologically diverse set of lesions.
PLAQUE HEMORRHAGE
As mentioned above, the interrelationship between
intraplaque hemorrhage, fibrin, necrotic core, and vasa vasorum
has not been fully explained. Most vasa vasorum originate
from the parent artery and ramify plexogenically in the adven-
Fig. 7. Total occlusion. Plaques with total occlusion from prior titia (14). There is a clear relationship between numbers of vasa
thrombi contain mostly smooth muscle cells in a collagen- vasorum (Fig. 9) and blood flow through them, and mass of
proteoglycan-rich matrix with capillaries and inflammatory cell infil- coronary plaque (15). In addition, increased vasa vasorum are
trate. This section shows a necrotic core (NC), although this is not
always present, and the lumen is filled with an organized thrombus
associated with atherosclerotic plaque expansion (14), and
(orgTh) with multiple capillary channels. plaque regression is paralleled by a decrease in blood flow
through the vasa vasorum (16). Further evidence for a dynamic
relationship between vasa vasorum and the atherosclerotic pro-
linear progression of coronary disease observed in angiograms cess are the fact that they respond to vasoconstrictor stimuli,
carried out annually in patients with chronic ischemic heart and that their density is associated with exertional plaque rup-
disease (12). However, these are speculations that need to be ture (11).
proven when we can prospectively recognize the sites of vul- The mechanisms by which vasa vasorum contribute to
nerability and follow up the lesional morphologies with plaque growth are poorly understood. Plaque iron content is a
advanced imaging techniques. We have also shown that these surrogate marker of preceding hemorrhage, and correlates with
lesions are the most heavily calcified as compared to acute plaque neovascularization (14). The frequency of iron deposi-
rupture or thin-cap fibroatheroma, suggesting that healing of a tion in the coronary plaques is higher in patients with acute
ruptured plaque also results in greater extent of calcification (13). coronary syndrome than in those dying of noncardiac causes
(17). Since plaque hemorrhage is a frequent phenomenon,
TOTAL OCCLUSIONS (see FIG. 7) especially in advanced plaques, it is not surprising that some of
Among patients dying sudden coronary death without prior the free cholesterol may have its origin from the red cell mem-
symptoms, total occlusion is a frequent finding at autopsy. The brane (18). The increase in iron content correlated with
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 359

there is little if any calcium present. However, pathologic inti-

mal thickening and fibrous plaques have a high propensity for
microscopic calcification, especially when there is evidence of
smooth muscle cell apoptosis. The biology of intimal calcifica-
tion within atherosclerotic regions is complex, and involves
smooth muscle cell matrix vesicles (related to apoptotic bod-
ies), bone-associated proteins, lipids, and inflammation in the
formation of extracellular intimal calcifications (22). Almost
all necrotic cores with apoptotic smooth muscle cells contain
calcification (at least 90%), and those cores with predominantly
macrophage infiltrates are less likely to initially calcify. The
nature of the calcifications differs by cell type, in that fibrous
cores contain finely granular calcifications, occasionally coa-
lescing into masses, whereas macrophage necrotic cores
often contain larger crystalline deposits. Calcification of
fibrous plaques in which there is smooth muscle cell death
cores progresses to plate-like sheets of calcification and often
pipestem-like arteries. However, calcification of necrotic,
inflamed, macrophage-rich cores tends to be more irregular,
resulting in a radiographic appearance of irregular calcifica-
tion. Ultrastructural study of coronary plaques will demonstrate
calcifications in close contact with dying smooth muscle cells,
macrophages, and surrounding cholesterol clefts (Fig. 10). In
less than 5% of plaques, actual ossification may occur, with
osteoblasts, osteoclasts, and even marrow formation (Fig. 11).
Calcification does not in itself appear to play a role in the
thrombotic process, except under the unusual circumstances of
nodular calcification. Successive ruptures or fissures of the
fibrous cap result in layering of the plaque, and the deeper
layers close to the internal elastic lamina typically demonstrate
more severe calcification than the surface layers.
We have performed quantitative analysis of calcified matrix
in intermediate and later plaque stages, including fibro-
atheromas, acute and healed plaque ruptures, total occlusions,
Fig. 8. Healed erosion shows deep multilayering of collagen sepa- and plaque erosion. In contrast to rupture, plaque erosion does
rated by elastin layers (arrowheads) and a paucity of smooth muscle not expose the contents of the necrotic core to the lumen. Eroded
cells. The superficial plaque is rich in smooth muscle cells, collagen, plaques, perhaps partly as a function of their pathogenesis, do
and proteoglycans.
not typically occur in calcified arteries, and demonstrate far
less calcified matrix than acute or healed ruptures. The degree
of calcification by plaque type is illustrated in Fig. 12.
intraplaque hemorrhage as identified by glycophorin A stain- The relationship between coronary plaque calcification and
ing, a red-cell specific anion-exchange protein. In a recent study plaque instability has been debated. Biomechanical studies
of pulmonary artery atherosclerotic plaques by Arbustini et al., have calculated stress at different regions of the plaque. Math-
glycophorin A positive erythrocyte membranes were shown to ematical models using large-strain finite element analysis have
be a major component of the necrotic core in patients with shown that increased lipids are associated with areas of weak-
chronic thromboembolic pulmonary hypertension (19). ness of the fibrous cap, but not calcification (23). Although
Rupture of vasa vasorum is not the sole etiology of calcification is a good marker for plaque burden, absolute cal-
intraplaque hemorrhage. As Constantinides originally sug- cium scores do not indicate plaques that are unstable or prone
gested, plaque hemorrhage may occur from cracks or fissures to result in clinical events. It has been stated that calcification
originating from the lumen (20). Intraplaque hemorrhage, is a disease marker as opposed to a process marker, unlike
whether from vasa vasorum or the lumen, may be critical to the markers of inflammation(24). These findings are corroborated
lesion progression, as it may provide a significant source of by autopsy studies that demonstrate a good correlation between
nonmetabolic cholesterol (21). plaque size and morphometric analysis of calcification, but no
correlation between residual lumen and calcification (Table 3).
CORONARY CALCIFICATION Few studies have correlated the radiologic pattern of calcifica-
In early coronary atherosclerotic plaques, coronary calcifi- tion with plaque instability (25), but there is some suggestion
cation is identified using sensitive histochemical stains, such as that speckled or fragmented calcification patterns as determined
the von Kossa stain. When such tests are applied to fatty streaks, radiologically are most likely associated with unstable plaque
360 VIRMANI ET AL.

Fig. 9. Vasa vasorum within plaque. (A) At low magnification, there is thinning of the media with dilated spaces within the plaque. (B) A higher
magnification of the boxed area in (A) shows dilated channels to be blood vessels (arrows). (C) A higher magnification of the boxed area in
B shows a muscular artery feeder vessel traversing the media. (D) A different segment of near-total occlusion stained with Ulex europaeus lectin
highlighting neovessels brown within the adventitia and intima.

types (Fig. 13). It remains to be determined whether imaging of sification scheme for atherosclerotic lesions. Arterioscler Thromb
calcification morphology (instead of simply burden) will be Vasc Biol 2000;20:1262275.
2. Stary HC, Chandler AB, Glagov S, et al. A definition of initial, fatty
beneficial in assessing patient risk for coronary events. streak, and intermediate lesions of atherosclerosis. A report from the
Committee on Vascular Lesions of the Council on Arteriosclerosis,
REFERENCES American Heart Association. Circulation 1994;89:24622478.
1. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons 3. Stary HC, Chandler AB, Dinsmore RE, et al. A definition of
from sudden coronary death: a comprehensive morphological clas- advanced types of atherosclerotic lesions and a histological classi-
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 361

Fig. 10. Ultrastructural images of coronary atherosclerotic plaques show extracellular calcifications adjacent to an apoptotic smooth muscle
cell (SMC) (A), spherical calcification associated with degenerating smooth muscle cell organelles (B), and small calcifications surrounding
cholesterol clefts in area of macrophage degeneration (C).

fication of atherosclerosis. A report from the Committee on Vascu- 16. Williams JK, Armstrong ML, Heistad DD. Vasa vasorum in athero-
lar Lesions of the Council on Arteriosclerosis, American Heart As- sclerotic coronary arteries: responses to vasoactive stimuli and
sociation. Circulation 1995;92:13551374. regression of atherosclerosis. Circ Res 1988;62:515523.
3a. Kockx MM, De Meyer GRY, Muhring J, Jacob W, Bult H, Herman 17. Virmani R, Roberts WC. Extravasated erythrocytes, iron, and fibrin
AG. Apoptosis and related proteins in different stages of human in atherosclerotic plaques of coronary arteries in fatal coronary heart
atherosclerotic plaques. Circulation 1998;97:23072315. disease and their relation to luminal thrombus: frequency and
4. Burke AP, Farb A, Malcom GT, Liang YH, Smialek J, Virmani R. significance in 57 necropsy patients and in 2958 five mm segments
Coronary risk factors and plaque morphology in men with coronary of 224 major epicardial coronary arteries. Am Heart J 1983;
disease who died suddenly. N Engl J Med 1997;336:12761282. 105:788797.
5. Arbustini E, Dal Bello B, Morbini P, et al. Plaque erosion is a major 18. Alberts B, Bray D, Lewis J, Raff M, Roberts K, Watson JD (eds).
substrate for coronary thrombosis in acute myocardial infarction. Molecular Biology of the Cell. Garland, Inc., New York: 1983;260.
Heart 1999;82:269272. 19. Arbustini E, Morbini P, DArmini AM, et al. Plaque composition in
6. Farb A, Burke AP, Tang AL, et al. Coronary plaque erosion without plexogenic and thromboembolic pulmonary hypertension: the criti-
rupture into a lipid core. A frequent cause of coronary thrombosis in cal role of thrombotic material in pultaceous core formation. Heart
sudden coronary death. Circulation 1996;93:1354363. 2002;88:177182.
7. Roberts WC, Kragel AH, Gertz SD, Roberts CS, Kalan JM. 20. Constantinides P. Plaque fissures in human coronary thrombosis
The heart in fatal unstable angina pectoris. Am J Cardiol 1991;68: J Atheroscler Res 1966;6:117.
22B27B. 21. Pasterkamp G, Virmani R. The erythrocyte: a new player in athero-
8. Libby P. Molecular bases of the acute coronary syndromes. Circu- matous core formation. Heart 2002;88:115116.
lation 1995;91:28442850. 22. Proudfoot D, Shanahan CM. Biology of calcification in vascular
9. Libby P, Sukhova G, Lee RT, Galis ZS. Cytokines regulate vascular cells: intima versus media. Herz 2001;26(4):245251.
functions related to stability of the atherosclerotic plaque. 23. Huang H, Virmani R, Younis H, Burke AP, Kamm RD, Lee RT. The
J Cardiovasc Pharmacol 1995;25 Suppl 2:S912. impact of calcification on the biomechanical stability of atheroscle-
10. Davies MJ, Thomas A. Thrombosis and acute coronary-artery rotic plaques. Circulation 2001;103:10511056.
lesions in sudden cardiac ischemic death. N Engl J Med 24. Hunt ME, OMalley PG, Vernalis MN, Feuerstein IM, Taylor AJ.
1984;310:11371140. C-reactive protein is not associated with the presence or extent
11. Burke AP, Farb A, Malcom GT, Liang Y, Smialek JE, Virmani R. of calcified subclinical atherosclerosis. Am Heart J 2001;141:
Plaque rupture and sudden death related to exertion in men with 206210.
coronary artery disease. JAMA 1999;281:921926. 25. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary
12. Mann J, Davies MJ. Mechanisms of progression in native coronary calcification: insights from sudden coronary death victims. Z Kardiol
artery disease: role of healed plaque disruption. Heart 1999;82: 2000;89:4953.
265268. 26. Virmani R, Burke AP, Kolodgie FD, Farb A. Vulnerable plaque: the
13. Burke AP, Weber DK, Kolodgie FD, Farb A, Taylor AJ, Virmani R. pathology of unstable coronary lesions. J Intervent Cardiol
Pathophysiology of calcium deposition in coronary arteries. Herz 2002;15:439446.
2001;26:239244. 27. (No authors listed.) Can atherosclerosis imaging techniques improve
14. Kumamoto M, Nakashima Y, Sueishi K. Intimal neovascularization the detection of patients at risk for ischemic heart disease? Proceed-
in human coronary atherosclerosis: its origin and pathophysiologi- ings of the 34th Bethesda Conference. Bethesda, MD, USA. October
cal significance. Hum Pathol 1995;26:450456. 7, 2002. J Am Coll Cardiol. 2003;41(11):18561917.
15. Heistad DD, Armstrong ML. Blood flow through vasa vasorum of 28. Burke AP, Kolodgie FD, Farb A, et al. Healed plaque ruptures and
coronary arteries in atherosclerotic monkeys. Arteriosclerosis sudden coronary death : evidence that subclinical rupture has a role
1986;6:326331. in plaque progression. Circulation 2001;103(4):934940.
362 VIRMANI ET AL.

Fig. 11. Ossification in coronary plaque. (A) demonstrates a low-magnification image of a coronary artery section (proximal left anterior
descending artery) in the left anterior descending coronary artery of an obese elderly woman with hypertensive atherosclerotic heart disease.
(A) is a low magnification of the cross-section of the artery; contrast material (black) was injected into the artery postmortem. (B) demonstrates
the bony trabeculae and (C) a higher magnification of lacunae containing osteoblasts.
 CHAPTER 33 / PATHOLOGY OF CORONARY ATHEROSCLEROTIC PLAQUES 363

Fig. 12. Amount of calcification by plaque morphology. Morphometric measurements were made of calcified matrix in several hundred
coronary lesions classified by plaque type. (A) Healed ruptures demonstrated the greatest mean area of calcification; plaque erosions dem-
onstrate almost no calcification. (B) When expressed as a percentage of the plaque area, total occlusion and plaque ruptures have relatively
little calcium, and plaque erosions the least. Fibroatheroma and healed ruptures demonstrate the greatest amount of calcified matrix. Repro-
duced from ref. 13, with permission.

Table 3
Relationship Between Calcification
and % Luminal Narrowing and Residual Lumen, by Arterial Section

Calcified area vs
Calcified area vs % luminal narrowing residual lumen area
Artery (simple regression) (simple regression)
Men and women
Men Women (combined)

ra p DF ra p DF rb p

LM .20 .20 30 .15 .59 14 .14 .58

PLAD .56 <.0001 122 .15 .27 53 .07 .53
MLAD .45 <.0001 91 .50 <.0001 66 .10 .29
DLAD .64 .0003 27 .31 .11 24 .18 28
LD .59 .001 26 .75 .0004 17 .13 .71
PLC .30 .01 78 .25 .07 51 .11 .32
MLC .02 .90 35 .31 .29 13 .21 .40
LOM .60 .0001 34 .72 0.01 11 .13 .30
PRC .12 .15 143 .09 .54 54 .09 .38
MRC .29 .001 127 .63 <.0001 68 .09 .37
DRC .32 .0004 116 .36 .06 27 .17 .29
aT values are all positive (positive correlation).
bT values are negative (plad, mrc, drc, ld) or positive (lm, lom, mlad, plc, mlc, prc).
Abbreviations: LM, left main; PLAD, proximal left anterior descending; MLAD, mid left anterior
descend-ing; DLAD, distal left anterior descending; LD, left diagonal; PLC, proximal left circumflex;
MLC, mid left circumflex; LOM, left obtuse marginal; PRC, proximal right coronary; MRC, mid right
coronary; DRC, distal right coronary; df, degrees of freedom.
Segments are bolded showing r values > 0.4 and a significant correlation p < 0.05. Reproduced from
ref. 13, with permission.
364 VIRMANI ET AL.

Fig. 13. Relationship between plaque morphology and radiographic calcification. Plaque erosions (A) were exclusively present in areas with
stippled or no calcification. Plaque ruptures (B) were most frequently seen in areas of speckled calcification, but were also present in blocked
or diffuse calcification. Curiously, there were no ruptures in segments devoid of any calcification. Thin-capped atheromas were most frequently
present in areas of speckled calcification (C), but were also seen in heavily calcified or uncalcified areas, suggesting that calcification pattern
is not helpful in diagnosing these lesions. Healed ruptures are almost always seen in areas of calcification, and most frequently in diffusely
calcified areas (D). Reproduced from ref. 13, with permission.
 CHAPTER 34 / VULNERABLE ATHEROSCLEROTIC PLAQUE PATHOGENESIS 365

34 Pathogenesis of the Vulnerable

Atherosclerotic Plaque

MASANORI AIKAWA, MD, PhD

INTRODUCTION members of the matrix metalloproteinase (MMP) family

Vascular inflammation plays a major role in the develop- (11,12). Among them, interstitial collagenases (MMP-1/colla-
ment of chronic coronary atherosclerosis and the onset of acute genase-1, MMP-8/collagenase-2, and MMP-13/collagenase-3)
coronary syndromes, including myocardial infarction, a lead- found in human atheroma can cleave interstitial collagen mac-
ing cause of death in the United States (1). Macrophage accu- romolecules (Fig. 3) (1315).
mulation and the consequent expression of proteolytic and Macrophage-derived collagenases of the MMP family may
thrombogenic molecules critically contribute to the formation cause formation of a collagen-poor, thin fibrous cap overlying
of vulnerable plaque prone to disruption and thrombus for- macrophage- and lipid-rich plaques, another feature typical of
mation, leading to fatal obstruction of the coronary arteries. A the vulnerable plaque. Collagen, which tolerates much greater
number of clinical investigations suggest hypercholesterolemia tensile stress than elastin, usually determines the stability and
as one of the major coronary risk factors. Preclinical studies durability of a wide variety of tissues (16). Using computer
demonstrate that hypercholesterolemia promotes accumulation modeling, Loree et al. demonstrated the inverse relationship
of oxidatively modified low-density lipoprotein (oxLDL) in between cap thickness and peak circumferential stress in the
the arterial wall, inducing endothelial cell (EC) activation and, plaque (17,18). Burk et al. from Virmanis group reported that
in turn, invasion of inflammatory cells such as macrophages thin fibrous caps measure approx 6070 m thick, and mea-
(Fig. 1) (13). Recent advances in vascular medicine provide surement near the site of rupture can be as little as 25 m (19).
evidence that lipid-lowering therapy prevents acute coronary Therefore, the collagen-poor, thin fibrous cap caused by col-
complications in patients, most likely by functionally improv- lagenolytic activity is likely responsible for instability and
ing inflammation of atheroma (1,4). This chapter will discuss consequent physical disruption of atherosclerotic plaques.
the biology of vascular inflammation and acute coronary com- Macrophages in atheroma also overexpress tissue factor, a
plications, as well as evidence of mechanisms by which lipid strong initiator of the blood coagulation cascade, and plasmi-
lowering stabilizes the vulnerable plaque. nogen activator inhibitor 1 (PAI-1), which inhibits fibrinolysis
(20,21). The fibrous cap usually separates thrombogenic mac-
FEATURES TYPICAL OF THE VULNERABLE rophages from various coagulation factors in circulating blood.
PLAQUE PRONE TO ACUTE MYOCARDIAL However, physical disruption of the cap causes direct contact
INFARCTION (see FIG. 2) between blood containing coagulation factors and macroph-
Our classical view suggested that myocardial infarction age-derived tissue factor, possibly accelerating thrombus for-
usually occurs in a critically stenosed coronary artery, detect- mation. Simultaneously, anti-fibrinolytic PAI-1 stabilizes the
able by angiography. However, pathologic and angiographic clot, which may favor massive, obstructive thrombosis in the
studies in the 1980s determined that fissure or rupture of the coronary arteries.
thin fibrous cap in coronary atheroma containing preserved
lumen often triggers acute fatal thrombosis (57). Several fac-
OXIDATIVE STRESS AND EC
tors contribute to the physical instability and thrombogenicity ACTIVATION MEDIATE FORMATION
of the atherosclerotic plaque (Fig. 1). Atheroma with a higher OF THE VULNERABLE PLAQUE
risk for disruption and thrombosis often contain a prominent Clinical and epidemiologic studies suggest that elevated
accumulation of macrophages, key players in any inflamma- plasma levels of low-density lipoprotein (LDL) increase the
tory diseases (810). Lesional macrophages in such vulnerable risk of acute coronary syndromes, including acute myocardial
plaques express high levels of various proteinases, including infarction, unstable angina, and cardiac sudden death (2224).
Beginning early in the 20th century, a number of preclinical
investigations using animal models demonstrated that hyperc-
From: Contemporary Cardiology: CT of the Heart: holesterolemia induces experimental atheroma formation
Principles and Applications
(2528). In vitro studies, however, long suggested that native
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

365
366 AIKAWA

Fig. 1. Monocyte recruitment and macrophage accumulation in atheroma. Oxidative stress induces endothelial cell activation and enhances
expression of adhesion molecules such as vascular cell adhesion molecule-1, which binds monocytes in the circulating blood (1). Monocytes
then migrate into the intima in response to chemokines including monocyte chemoattractant protein-1 (MCP-1) (2). Oxidative stress induces
MCP-1 expression. Monocytes differentiate into macrophages in the intima (3). Several molecules such as macrophage-colony stimulating
factor induce macrophage differentiation and activation. Activated macrophages express a number of molecules related to atherogenesis,
plaque instability and thrombogenicity. Macrophage proliferation likely plays an important role in formation of the atherosclerotic plaques rich
in this cell type (4). Oxidative stress also induces macrophage activation and proliferation. Adapted from ref. 3, with permission.

Fig. 2. Diagram demonstrating pathological and biological features typical of the vulnerable atherosclerotic plaque. The so-called vulnerable
plaque usually contains prominent macrophage accumulation and lipid pool in its atheromatous core, underlying a thin, collagen-poor fibrous
cap. Macrophages produce matrix metalloproteinases (MMPs) including collagenases, which may weaken the fibrous cap and promote physical
disruption. Tissue factor and plasminogen activator inhibitor 1 (PAI-1) produced by macrophages may accelerate formation of massive
thrombus and inhibit fibrinolysis at the sites of rupture, and likely contribute to fetal obstruction of the coronary artery. Smooth muscle cells
(SMCs) in the fibrous cap tend to undergo apoptosis and are likely responsible for collagen loss. Intimal SMCs also express MMPs and tissue
factor. Activated endothelial cells promote monocyte recruitment and macrophage accumulation in atheroma.
 CHAPTER 34 / VULNERABLE ATHEROSCLEROTIC PLAQUE PATHOGENESIS 367

Fig. 3. Macrophage expression of collagenases of the MMP family in human atheroma. Immunohistochemistry detected MMP-1/collagenase-
1, MMP-8/collagenase-2, and MMP-13/collagenase-3 in atheroma of the human carotid artery. Both collagenases colocalize with macrophages
detected by CD68 antibody. Adapted from refs. 1315, with permission.

LDL itself does not induce features associated with atheroma The excess LDL that accumulates in the artery wall as a
(e.g., transformation of monocyte/macrophages into lipid- result of hypercholesterolemia can be oxidatively modified
laden foam cells). In the late 1980s, the oxidation hypothesis (31,32). Human and experimental atheroma both contain
finally linked LDL with atherogenesis by introducing evidence oxLDL (3335). Atherosclerotic lesions intrinsically produce
that oxLDL is proinflammatory (29). Since then, vascular biol- excess reactive oxygen species (ROS) such as superoxide anion
ogy has provided much evidence that hypercholesterolemia (O2), promoting further oxidative modification of LDL (32).
induces oxidative stress, leading to vascular inflammation and OxLDL instigates an inflammatory response on arterial ECs. In
atherosclerosis (30). vitro studies suggest that oxLDL and its derivatives increase
368 AIKAWA

Fig. 4. Macrophage growth in human atheroma. In situ hybridization detected histone mRNA expression, a sensitive marker for DNA
replication, in the intima of a human carotid artery. mRNA signals colocalized with macrophages detected by immunohistochemistry for CD68
on a serial section. Scale bar: 50 m. Original magnification: 400. Adapted from ref. 45, with permission.

expression of cell-adhesion molecules, including vascular cell- enous inhibitors of plasminogen activators (i.e., PAI-1) (57).
adhesion molecule 1 (VCAM-1) and chemokines such as mono- Such fibrinolytic imbalance may stabilize clots and favor fatal,
cyte chemoattractant protein 1 (MCP-1), both of which occlusive thrombus formation.
critically contribute to leukocyte recruitment into the intima
PLAQUE STABILIZATION:
(Fig. 1) (3638). Preclinical studies employing genetically
altered mice demonstrate in vivo evidence supporting the role MECHANISMS BY WHICH LIPID LOWERING
of VCAM-1 and MCP-1 in the formation of macrophage-rich BY HMG-COA REDUCTASE INHIBITORS
atheroma (3941). In rabbits, high-cholesterol diet induces EC PREVENTS ACUTE CORONARY SYNDROMES
expression of VCAM-1 (35,42). Since the 1990s, clinical studies have established that lipid-
Macrophage proliferation may play an important role in the lowering therapy by HMG-CoA reductase inhibitors (statins)
development of vascular inflammation (Fig. 4) (4345). OxLDL prevents the onset of acute coronary events despite no or mod-
induces macrophage proliferation in vitro (46) and also activates est reduction in angiographic stenoses (3,58,59). This discrep-
nuclear factor-B (NF-B), a key transcription factor that regu- ancy raised the concept that lipid lowering may modify the
lates a wide variety of atherosclerosis-associated genes (47). vulnerable plaque in a functional manner (stabilization)
Furthermore, the discovery of macrophage scavenger receptors rather than simply shrinking the lesion (regression) (60,61).
that selectively take up oxLDL and induce foam-cell formation A number of animal studies previously focused on regression
improved our understanding of the role hypercholesterolemia of atheroma by lipid lowering (62,63); however, more recent
plays in vascular inflammation and atherogenesis (48). preclinical studies, including our own, improved mechanistic
Nitric oxide (NO) exerts antiatherogenic actions by inhibit- understanding of the clinical benefits of lipid lowering and
ing monocyte adhesion to ECs or suppressing expression of furnished the plaque stabilization hypothesis (1,4). Clinical
VCAM-1 and MCP-1 by EC (4952). Apparently normal ECs and preclinical studies also suggest that statins have various
constitutively express endothelial NO synthase (eNOS), which direct or pleiotopic effects independent of lipid lowering (1,64).
produces NO from L-arginine. However, eNOS expression Our studies on plaque stabilization have employed mainly
decreases in human atherosclerotic arteries (53). Oxidative rabbit models of atherosclerosis (28). Beginning early in the
stress decreases eNOS expression by cultured ECs as well as 20th century, accumulating knowledge has suggested that ath-
bioavailability of NO (54). erosclerotic lesions of hypercholesterolemic rabbits mimic
Superficial erosion of the plaque due to EC detachment may chronic coronary atherosclerosis in humans (25,26,28). We
also contribute to the onset of acute coronary syndromes (55). found that rabbit aortic atherosclerosis created by both
Rajavashisth et al. reported that oxLDL induces EC expression mechanical injury and high-cholesterol diet contained a num-
of membrane type 1matrix metalloproteinase (MT1-MMP), ber of features associated with vulnerable plaques in humans:
an activator of MMP-2 (56). MMP-2 cleaves type IV collagen, prominent macrophage accumulation, a thin and collagen-poor
a major component of the basement membrane underlying the fibrous cap, and expression of MMP and tissue factor (65,66).
endothelium, thus suggesting that EC activation by oxidative Potent inflammatory mediators CD154 (or CD40 ligand) and
stress may cause plaque erosion. CD40, which induce MMP and tissue factor expression, also
Additionally, activated ECs may affect the fibrinolytic bal- occur in rabbit atheroma (66). To address the effects of lipid
ance in atheroma. Relatively normal ECs express the tissue- lowering itself without direct vascular effects of statins, we
type plasminogen activator (t-PA), a fibrinolytic molecule. On examined the way these features change in hypercholester-
the other hand, activated ECs increase the production of endog- olemic rabbits during lipid lowering by diet alone (Table 1)
 CHAPTER 34 / VULNERABLE ATHEROSCLEROTIC PLAQUE PATHOGENESIS 369

Table 1
Effects of Lipid Lowering on Atheroma of Hypercholesterolemic Rabbits
Diet Statins
Macrophage accumulation Decreased Decreased
Macrophage proliferation ND Decreased
Macrophage apoptosis ND No change
MMP expression/activity Decreased Decreased
Collagen content Increased Increaseda
Tissue factor expression/activity Decreased Decreased
PAI-1 expression ND Decreased
CD154 (CD40L) expression Decreased Decreased
PDGF- expression Decreased ND
SMC maturity Increased ND
SM1 and SM2, a marker for mature SMCs Increased ND
SMemb, a nonmuscle or embryonic myosin Decreased ND
ROS production (O2) Decreased ND
LDL accumulation Decreased Decreased
oxLDL accumulation Decreased Decreased
Plasma autoantibody for oxLDL Decreased ND
VCAM-1 Decreased ND
MCP-1 Decreased ND
eNOS Increased ND
Microvessels Decreased ND
ND, not determined; MMP, matrix metalloproteinase; PAI-1, plasminogen activator inhibitor 1;
PDGF-, platelet-derived growth factor beta; SMC, smooth muscle cells; ROS, reactive oxygen
species; oxLDL, oxidatively modified low-density lipoprotein; VCAM-1, vascular cell-adhesion
molecule 1; MCP-1, monocyte chemoattractant protein 1; eNOS, endothelial NO synthase.
aNo change with fluvastatin.

(1,4,35,6567). Dietary cholesterol lowering in rabbit atheroma tained a prominent accumulation of oxLDL underlying acti-
reduced macrophage accumulation expressing MMPs includ- vated ECs that expressed VCAM-1 (Fig. 7A, top panels). In
ing MMP-1/collagenase-1 and, in parallel, increased intersti- contrast, few if any ECs in atheroma expressed eNOS, in agree-
tial collagen content, a key determinant of plaque stability (Fig. ment with a previous study on human atheroma (53). However,
5) (65). A rabbit study by Kockx et al. and a clinical study by lipid lowering by diet alone reduced ROS production to levels
Crisby et al. reported a similar finding of increased collagen similar to those found in normal aortas. Dietary lipid lowering
content in atherosclerotic plaques by lipid lowering (68,69). reduced oxLDL accumulation and VCAM-1 expression con-
These studies suggest that lipid lowering prevents acute coro- comitantly (Fig. 7A bottom panels, and 7B) (35). In contrast,
nary syndromes in part by increasing the mechanical strength lipid lowering by diet increased eNOS expression substantially.
of atheroma. ECs, SMCs, and macrophages in human atheroma express
Lipid lowering also reduced expression and activity of tissue MCP-1, a potent chemokine that induces monocyte recruit-
factor, suggesting decreased thrombogenic potential, another ment into the arterial wall (72). In aortic atheroma of choles-
major contributor to acute coronary events (66). Additionally, terol-fed rabbits, ECs as well as SMCs and macrophages
inflammatory mediators CD154 and CD40 decreased substan- contained immunoreactive MCP-1 (35). However, this
tially during cholesterol lowering (66). More recently, we chemokine was almost undetectable after cholesterol lowering.
reported that lipid lowering by statin treatment also reduces After lipid lowering, aortic ECs also exhibited a more normal
expression of MMPs, tissue factor, and their inducer CD154 in ultrastructure compared with those of the atherosclerotic in-
hypercholesterolemic rabbits (Fig. 6) (45,70). tima, which had features typical of activated ECs (Fig. 8). Taken
Smooth muscle cells (SMCs) in the fibrous cap of rabbit together, these pre-clinical and clinical results suggest that
atheroma exhibited immature phenotype compared to those in cholesterol lowering reduces oxidative stress and ameliorates
apparently normal media found in human atheroma (67,71). EC dysfunction, favoring stabilization of the vulnerable plaque.
However, dietary lipid lowering promoted accumulation of
SMCs of more mature phenotype in the fibrous cap, expressing CLINICAL SIGNIFICANCE
fewer MMPs and less tissue factor than those found in baseline OF PLEIOTROPIC EFFECTS OF STATINS
lesions (66,67). Accumulating preclinical evidence using animal models or
We furthermore tested the hypothesis that cholesterol low- in vitro systems demonstrates that statins possess effects other
ering reduces oxidative stress and EC activation in atheroma than reduction of cholesterol synthesis, including anti-inflam-
(Table 1) (35). Atherosclerotic aortas of cholesterol-fed rabbits matory effects on vascular cells (1,64). We reported that ceri-
produced high levels of ROS, including superoxide anion vastatin inhibited M-CSF-induced macrophage growth in vitro
(O2). Baseline lesions in hypercholesterolemic rabbits con- with clinically achievable concentrations (45). This treatment
370 AIKAWA

Fig. 5. Lipid lowering by diet alone reduces MMP-1/collagenase-1 expression and increases interstitial collagen content in rabbit atheroma.
Top panels: The rabbit aortic lesion after 4 mo on a high-cholesterol diet (baseline lesion) contained high level of MMP-1/collagenase-1
expression, a potent collagenolytic enzyme. Picrosirius red staining with polarization barely detected accumulation of interstitial collagen on
a serial section of this collagenase-rich plaque. Bottom panels: 16 mo of dietary lipid lowering decreased MMP-1/collagenase-1 expression
and, in parallel, increased collagen content, a key determinant of plaque stability. Original magnification: 10. From ref. 65, with permission.

also reduced macrophage activation as examined by decreased lipophilic statins suppress proliferation or induce apoptosis of
MMP-9 activity and tissue factor expression (45). Previous cultured SMCs more effectively than do hydrophilic statins
studies employing fluvastatin and simvastatin demonstrated (70,76,77). Clinical trials for all statins, however, have shown
similar findings regarding macrophage proliferation and acti- benefits somewhat similar to the prevention of acute coronary
vation in vitro (7375). These cholesterol-independent, syndromes. Insufficient cell permeability of hydrohilic
antimacrophage effects should favor stabilization of the vul- pravastatin is not reflected by significant risk reduction in clini-
nerable plaque. Thus far, however, no unambiguous clinical cal trials. Moreover, both lipophilic and hydrophilic statins
evidence suggests that such pleiotropic effects substantially possess anti-inflammatory effects independent of LDL reduc-
increase the benefits of statin treatment in patients. tion in patients, as determined by plasma levels of hsCRP
Like other classes of drugs, statins have variable lipophilic (78,79). Thus, the mechanisms by which all statins with differ-
properties. Unlike hydrophilic compounds, lipophilic com- ent lipophilic properties produce cholesterol-independent, anti-
pounds are generally cell permeant, and their effects on periph- inflammatory effects remain obscure. Concentrations of statins
eral tissues or cultured cells are usually more direct. For instance, used in in vitro investigations often exceed those achievable in
 CHAPTER 34 / VULNERABLE ATHEROSCLEROTIC PLAQUE PATHOGENESIS 371

Fig. 6. Lipid lowering by statin treatment reduces expression of CD154 and tissue factor in atheroma of hypercholesterolemic rabbits. The aortic
intima of 34-mo-old Watanabe heritable hyperlipidemic rabbits, a model of endogenous hypercholesterolemia, exhibited expression of tissue
factor, a strong activator of blood coagulation, and its inducer CD154 (or CD40 ligand). With cerivastatin treatment for 32 mo, expression of
both CD40L and TF decreased substantially. The arrow indicates the tunica media. Scale bar: 200 m. From ref. 45, with permission.

patients, further stirring controversy about the clinical impor- sis. Based on clinical and our own preclinical evidence, this
tance of their pleiotropic effects. On the other hand, animal chapter has also underscored the importance of lipid manage-
studies including our own clearly suggest that lipid lowering by ment in the prevention of acute coronary syndromes. New
diet alone improves a number of features associated with vas- targets for antiinflammatory therapies beyond lipid lowering
cular inflammation and activation. Clinical statin trials also may include inhibition of the renin-angiotensin-aldosterone
demonstrate that aggressive cholesterol lowering produces a system and activation of peroxisome proliferator-activated
greater reduction of inflammatory burden than does more receptors (PPARs) (4,8083). Interestingly, inhibitors for
moderate therapy. These preclinical and clinical data suggest these pathways, like statins, have seemingly pleiotropic effects
the importance of lowering lipids by diet therapy, lifestyle on vascular inflammation as well. In addition to screening con-
changes, and, if necessary, statins. ventional coronary risk factors (e.g., LDL levels), sensitive but
inexpensive biomarkers for new risk factors (e.g., hsCRP) may
FUTURE PERSPECTIVES identify currently unattended high-risk patients. Moreover,
In the last two decades, advances in clinical and basic car- novel imaging modalities will probably detect not only luminal
diovascular medicine forged several missing links between stenoses but also more qualitative features, including the
coronary risks and the onset of acute thrombotic complications. biological processes typical of the vulnerable plaque (84).
In particular, discoveries regarding the role of oxidative stress These new approaches should provide more effective and indi-
in vascular inflammation substantially improved our mecha- vidualized strategies for the prevention of acute thrombotic
nistic understanding of the pathogenesis of coronary thrombo- complications of atherosclerosis.
372 AIKAWA

Fig. 7. Dietary lipid lowering reduces oxidatively modified low-density lipoprotein (oxLDL) accumulation and vascular cell-adhesion mol-
ecule 1 (VCAM-1) expression in rabbit atheroma. Top panels: OxLDL epitopes (MDA-lysine) accumulated in the aortic intima beneath
activated endothelial cells (ECs) expressing VCAM-1 in hypercholesterolemic rabbits which consumed the atherogenic diet for 4 mo (baseline).
Bottom panels: OxLDL epitopes (MDA-lysine) and VCAM-1 expression by ECs were almost undetectable in the intima of rabbit aorta after
16 mo of lipid lowering, while CD31, an EC marker, indicated an intact monolayer. The atheroma from control animals that continued the
atherogenic diet for 16 mo sustained oxLDL and VCAM-1. Scale bar: 50 m. Original magnification 400. From ref. 35, with permission.

ACKNOWLEDGMENT 7. Ambrose JA, Tannenbaum MA, Alexopoulos D, et al. Angiographic

progression of coronary artery disease and the development of myo-
The author acknowledges a number of colleagues and col- cardial infarction. J Am Coll Cardiol 1988;12:5662.
laborators for their tremendous contributions to the study con- 8. Davies MJ, Richardson PD, Woolf N, Katz DR, Mann J. Risk of
cepts and experiments that this chapter mentioned. This work thrombosis in human atherosclerotic plaques: role of extracellular
was supported in part by grants from the National Institutes of lipid, macrophage, and smooth muscle cell content. Br Heart J
1993;69:377381.
Health, National Heart, Lung, and Blood Institute (PO1 9. van der Wal AC, Becker AE, van der Loos CM, Das PK. Site of
HL48743 and Merit Award to Dr. Peter Libby, SCOR intimal rupture or erosion of thrombosed coronary atherosclerotic
P50HL56985 to Drs. Libby and Aikawa, and R01HL66086 to plaques is characterized by an inflammatory process irrespective of
Dr. Aikawa), and Japan Heart Foundation (to Dr. Aikawa). We the dominant plaque morphology. Circulation 1994;89:3644.
10. Moreno PR, Falk E, Palacios IF, Newell JB, Fuster V, Fallon JT.
also acknowledge Karen E. Williams for her excellent editorial
Macrophage infiltration in acute coronary syndromes. Implications
assistance. for plaque rupture. Circulation 1994;90:775778.
11. Nagase H, Woessner JF Jr. Matrix metalloproteinases. J Biol Chem
REFERENCES 1999;274:2149121494.
1. Aikawa M, Libby P. The vulnerable atherosclerotic plaque; patho- 12. Brinckerhoff CE, Matrisian LM. Matrix metalloproteinases: a tail
genesis and therapeutic approachhh. Cardiovasc Pathol 2004;13(3): of a frog that became a prince. Nat Rev Mol Cell Biol 2002;3:
125138. 207214.
2. Munro JM, Cotran RS. The pathogenesis of atherosclerosis: athero- 13. Galis ZS, Sukhova GK, Lark MW, Libby P. Increased expression of
genesis and inflammation. Lab Invest 1988;58:249261. matrix metalloproteinases and matrix degrading activity in vulner-
3. Libby P, Aikawa M. Evolution and stabilization of vulnerable ath- able regions of human atherosclerotic plaques. J Clin Invest 1994;
erosclerotic plaques. Jpn Circ J 2001;65:473479. 94:24932503.
4. Libby P, Aikawa M. Stabilization of atherosclerotic plaques: new 14. Sukhova G, Schoenbeck U, Rabkin E, et al. Evidence of increased
mechanisms and clinical targets. Nature Medicine 2002;8:1257 collagenolysis by interstitial collagenases-1 and -3 in vulnerable
1262. human atheromatous plaques. Circulation 1999;99:25032509.
5. Davies MJ, Thomas AC. Plaque fissuringthe cause of acute myo- 15. Herman MP, Sukhova GK, Libby P, et al. Expression of neutrophil
cardial infarction, sudden ischaemic death, and crescendo angina. collagenase (matrix metalloproteinase-8) in human atheroma: a
Br Heart J 1985;53:363373. novel collagenolytic pathway suggested by transcriptional profil-
6. Hackett D, Davies G, Maseri A. Pre-existing coronary stenoses in ing. Circulation 2001;104:18991904.
patients with first myocardial infarction are not necessarily severe. 16. Lee RT, Libby P. The unstable atheroma. Arterioscler Thromb Vasc
Eur Heart J 1988;9:13171323. Biol 1997;17:18591867.
 CHAPTER 34 / VULNERABLE ATHEROSCLEROTIC PLAQUE PATHOGENESIS 373

Fig. 8. Ultrastructure of activated endothelial cells (ECs) in rabbit atheroma and its normalization by lipid lowering. Top panel: ECs in aortic
atheroma of cholesterol-fed rabbits showed a cuboidal or rounded structure, typical of an activated phenotype. Monocytic cell (indicated by
the arrowhead) appears to be entering the intima through the gap between ECs. The circular cells containing abundant lipid particles in
subendothelial space of the aorta of the baseline lesion are probably macrophage-derived foam cells. Bottom panel: The aortic ECs of the treated
animals had a more squamous morphology. The size, density, and amount of cytoplasmic organelles of ECs of the atheroma of the treated
animals are substantially smaller than those in the hypercholesterolemic animals. Accumulation of organized collagen fibrils was more
prominent in the intima of the treated animals than in atheroma of the high-cholesterol-fed rabbits. Original magnification: 3000. From ref.
35, with permission.

17. Loree HM, Kamm RD, Stringfellow RG, Lee RT. Effects of fibrous 22. Stamler J, Shekelle R. Dietary cholesterol and human coronary heart
cap thickness on peak circumferential stress in model atheroscle- disease. The epidemiologic evidence. Arch Pathol Lab Med 1988;
rotic vessels. Circ Res 1992;71:850858. 112:10321040.
18. Loree HM, Tobias BJ, Gibson LJ, Kamm RD, Small DM, Lee RT. 23. Libby P, Aikawa M, Schonbeck U. Cholesterol and atherosclerosis.
Mechanical properties of model atherosclerotic lesion lipid pools. Biochim Biophys Acta 2000;1529:299309.
Arterioscler Thromb 1994;14:230234. 24. LaRosa JC, Hunninghake D, Bush D, et al. The cholesterol facts. A
19. Burke AP, Farb A, Malcom GT, Liang YH, Smialek J, Virmani R. summary of the evidence relating dietary fats, serum cholesterol,
Coronary risk factors and plaque morphology in men with coro- and coronary heart disease. A joint statement by the American Heart
nary disease who died suddenly. N Engl J Med 1997;336:1276 Association and the National Heart, Lung, and Blood Institute. The
1282. Task Force on Cholesterol Issues, American Heart Association.
20. Wilcox JN, Smith KM, Schwartz SM, Gordon D. Localization of Circulation 1990;81:17211733.
tissue factor in the normal vessel wall and in the atherosclerotic 25. Anitschkow N, Chalatow S. On experimental cholesterin steatosis
plaque. Proc Natl Acad Sci USA 1989;86:28392843. and its significance in the origin of some pathological processes
21. Lupu F, Bergonzelli GE, Heim DA, et al. Localization and produc- (1913). Reprinted in Arteriosclerosis 1983;3:178182.
tion of plasminogen activator inhibitor-1 in human healthy and ath- 26. Vesselinovitch D. Animal models and the study of atherosclerosis.
erosclerotic arteries. Arterioscler Thromb 1993;13:10901100. Arch Pathol Lab Med 1988;112:10111017.
374 AIKAWA

27. Armstrong ML, Heistad DD. Animal models of atherosclerosis. 50. De Caterina R, Libby P, Peng HB, et al. Nitric oxide decreases
Atherosclerosis 1990;85:1523. cytokine-induced endothelial activation. Nitric oxide selectively
28. Aikawa M, Fukumoto Y, Rabkin E, Libby P. Rabbit models of reduces endothelial expression of adhesion molecules and
atherosclerosis. In: Simon DI, Rogers C (eds), Vascular Disease proinflammatory cytokines. J Clin Invest 1995;96:6068.
and Injury: Preclinical Research. Humana Press, Totowa, NJ: 2000; 51. Khan BV, Harrison DG, Olbrych MT, Alexander RW, Medford
175191. RM. Nitric oxide regulates vascular cell adhesion molecule 1 gene
29. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. expres-sion and redox-sensitive transcriptional events in human
Beyond cholesterol. Modifications of low-density lipoprotein that vascular endothelial cells. Proc Natl Acad Sci USA 1996;93:
increase its atherogenicity. N Engl J Med 1989;320:915924. 91149119.
30. Libby P. Inflammation in atherosclerosis. Nature 2002;420:868874. 52. Tsao PS, Wang B, Buitrago R, Shyy JY, Cooke JP. Nitric oxide
31. Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. regulates monocyte chemotactic protein-1. Circulation 1997;96:
Beyond cholesterol. Modifications of low-density lipoprotein that 934940.
increase its atherogenicity. N Engl J Med 1989;320:915924. 53. Oemar BS, Tschudi MR, Godoy N, Brovkovich V, Malinski T,
32. Berliner JA, Navab M, Fogelman AM, et al. Atherosclerosis: basic Luscher TF. Reduced endothelial nitric oxide synthase expression
mechanisms. Oxidation, inflammation, and genetics. Circulation and production in human atherosclerosis. Circulation 1998;97:
1995;91:24882496. 24942498.
33. Haberland ME, Fong D, Cheng L. Malondialdehyde-altered protein 54. Liao JK, Shin WS, Lee WY, Clark SL. Oxidized low-density lipo-
occurs in atheroma of Watanabe heritable hyperlipidemic rabbits. protein decreases the expression of endothelial nitric oxide syn-
Science 1988;241:215218. thase. J Biol Chem 1995;270:319324.
34. Yla-Herttuala S, Palinski W, Rosenfeld ME, et al. Evidence for the 55. Farb A, Burke AP, Tang AL, et al. Coronary plaque erosion without
presence of oxidatively modified low density lipoprotein in athero- rupture into a lipid core. A frequent cause of coronary thrombosis in
sclerotic lesions of rabbit and man. J Clin Invest 1989;84:10861095. sudden coronary death. Circulation 1996;93:13541363.
35. Aikawa M, Sugiyama S, Hill C, et al. Lipid lowering reduces oxida- 56. Rajavashisth TB, Liao JK, Galis ZS, et al. Inflammatory cytokines
tive stress and endothelial cell activation in rabbit atheroma. Circu- and oxidized low density lipoproteins increase endothelial cell
lation 2002;106:13901396. expression of membrane type 1-matrix metalloproteinase. J Biol
36. Kume N, Cybulsky MI, Gimbrone MA Jr. Lysophosphatidylcholine, Chem 1999;274:1192411929.
a component of atherogenic lipoproteins, induces mononuclear leu- 57. Schneiderman J, Sawdey MS, Keeton MR, et al. Increased type 1
kocyte adhesion molecules in cultured human and rabbit arterial plasminogen activator inhibitor gene expression in atherosclerotic
endothelial cells. J Clin Invest 1992;90:11381144. human arteries. Proc Natl Acad Sci USA 1992;89:69987002.
37. Khan BV, Parthasarathy SS, Alexander RW, Medford RM. Modi- 58. Gotto AM, Farmer JA. Lipid-lowering trials. In: Braunwald E, Zipes
fied low density lipoprotein and its constituents augment cytokine- DP, Libby P (eds), Heart Disease: A Text Book of Cardiovascular
activated vascular cell adhesion molecule-1 gene expression in Medicine. W.B. Saunders, Philadelphia: 2001;126146.
human vascular endothelial cells. J Clin Invest 1995;95:12621270. 59. Libby P, Aikawa M. Mechanisms of plaque stabilization with statins.
38. Cushing SD, Berliner JA, Valente AJ, et al. Minimally modified low Am J Cardiol 2003;91:4B8B.
density lipoprotein induces monocyte chemotactic protein 1 in 60. Brown BG, Zhao XQ, Sacco DE, Albers JJ. Lipid lowering and
human endothelial cells and smooth muscle cells. Proc Natl Acad plaque regression. New insights into prevention of plaque disrup-
Sci USA 1990;87:51345138. tion and clinical events in coronary disease. Circulation 1993;87:
39. Cybulsky MI, Iiyama K, Li H, et al. A major role for VCAM-1, but 17811791.
not ICAM-1, in early atherosclerosis. J Clin Invest 2001;107: 61. Libby P. Molecular bases of the acute coronary syndromes. Circu-
12551262. lation 1995;91:28442850.
40. Gu L, Okada Y, Clinton SK, et al. Absence of monocyte 62. Small DM. George Lyman Duff memorial lecture. Progression and
chemoattractant protein-1 reduces atherosclerosis in low density regression of atherosclerotic lesions. Insights from lipid physical
lipoprotein receptor-deficient mice. Mol Cell 1998;2:275281. biochemistry. Arteriosclerosis 1988;8:103129.
41. Boring L, Gosling J, Cleary M, Charo IF. Decreased lesion forma- 63. Wissler RW, Vesselinovitch D. Can atherosclerotic plaques regress?
tion in CCR2/ mice reveals a role for chemokines in the initiation Anatomic and biochemical evidence from nonhuman animal mod-
of atherosclerosis. Nature 1998;394:894897. els. Am J Cardiol 1990;65:33F40F.
42. Li H, Cybulsky MI, Gimbrone MA Jr., Libby P. An atherogenic diet 64. Liao JK. Isoprenoids as mediators of the biological effects of statins.
rapidly induces VCAM-1, a cytokine-regulatable mononuclear leu- J Clin Invest 2002;110:285288.
kocyte adhesion molecule, in rabbit aortic endothelium. Arterioscler 65. Aikawa M, Rabkin E, Okada Y, et al. Lipid lowering by diet reduces
Thromb 1993;13:197204. matrix metalloproteinase activity and increases collagen content of
43. Rosenfeld ME, Ross R. Macrophage and smooth muscle cell prolif- rabbit atheroma: a potential mechanism of lesion stabilization. Cir-
eration in atherosclerotic lesions of WHHL and comparably hyper- culation 1998;97:24332444.
cholesterolemic fat-fed rabbits. Arteriosclerosis 1990;10:680687. 66. Aikawa M, Voglic SJ, Sugiyama et al. Dietary lipid lowering reduces
44. Gordon D, Reidy MA, Benditt EP, Schwartz SM. Cell proliferation tissue factor expression in rabbit atheroma. Circulation 1999;100:
in human coronary arteries. Proc Natl Acad Sci USA 1990;87: 12151222.
46004604. 67. Aikawa M, Rabkin E, Voglic SJ, et al. Lipid lowering promotes
45. Aikawa M, Rabkin E, Sugiyama S, et al. An HMG-CoA reductase accumulation of mature smooth muscle cells expressing smooth
inhibitor, cerivastatin, suppresses growth of macrophages express- muscle myosin heavy chain isoforms in rabbit atheroma. Circ Res
ing matrix metalloproteinases and tissue factor in vivo and in vivo. 1998;83:10151026.
Circulation 2001;103:276283. 68. Kockx MM, De Meyer GR, Buyssens N, Knaapen MW, Bult H,
46. Sakai M, Kobori S, Miyazaki A, Horiuchi S. Macrophage prolifera- Herman AG. Cell composition, replication, and apoptosis in athero-
tion in atherosclerosis. Curr Opin Lipidol 2000;11:503509. sclerotic plaques after 6 months of cholesterol withdrawal. Circ Res
47. Collins T, Cybulsky MI. NF-kappaB: pivotal mediator or innocent 1998;83:378387.
bystander in atherogenesis? J Clin Invest 2001;107:255264. 69. Crisby M, Nordin-Fredriksson G, Shah PK, Yano J, Zhu J, Nilsson
48. Kodama T, Freeman M, Rohrer L, Zabrecky J, Matsudaira P, Krieger J. Pravastatin treatment increases collagen content and decreases
M. Type I macrophage scavenger receptor contains alpha-helical lipid content, inflammation, metalloproteinases, and cell death in
and collagen-like coiled coils. Nature 1990;343:531535. human carotid plaques: implications for plaque stabilization. Circu-
49. Tsao PS, McEvoy LM, Drexler H, Butcher EC, Cooke JP. Enhanced lation 2001;103:926933.
endothelial adhesiveness in hypercholesterolemia is attenuated by 70. Fukumoto Y, Libby P, Rabkin E, et al. Statins alter smooth muscle
L-arginine. Circulation 1994;89:21762182. cell accumulation and collagen content in established atheroma of
 CHAPTER 34 / VULNERABLE ATHEROSCLEROTIC PLAQUE PATHOGENESIS 375

Watanabe heritable hyperlipidemic rabbits. Circulation 2001;103: comparison with other human cell types. Biochim Biophys Acta
993999. 1997;1345:259268.
71. Aikawa M, Sivam PN, Kuro-o M, et al. Human smooth muscle 78. Ridker PM, Rifai N, Pfeffer MA, et al. Inflammation, pravastatin,
myosin heavy chain isoforms as molecular markers for vascular and the risk of coronary events after myocardial infarction in patients
development and atherosclerosis. Circ Res 1993;73:10001012. with average cholesterol levels. Cholesterol and Recurrent Events
72. Rollins BJ. Chemokines. Blood 1997;90:909928. (CARE) Investigators. Circulation 1998;98:839844.
73. Sakai M, Kobori S, Matsumura T, et al. HMG-CoA reductase inhibi- 79. Ridker PM, Rifai N, Lowenthal SP. Rapid reduction in C-reactive
tors suppress macrophage growth induced by oxidized low density protein with cerivastatin among 785 patients with primary hypercho-
lipoprotein. Atherosclerosis 1997;133:5159. lesterolemia. Circulation 2001;103:11911193.
74. Colli S, Eligini S, Lalli M, Camera M, Paoletti R, Tremoli E. 80. Weiss D, Sorescu D, Taylor WR. Angiotensin II and atherosclerosis.
Vastatins inhibit tissue factor in cultured human macrophages. A Am J Cardiol 2001;87:25C32C.
novel mechanism of protection against atherothrombosis. 81. Bocan TM, Krause BR, Rosebury WS, et al. The ACAT inhibitor
Arterioscler Thromb Vasc Biol 1997;17:265272. avasimibe reduces macrophages and matrix metalloproteinase
75. Bellosta S, Via D, Canavesi M, et al. HMG-CoA reductase inhibitors expression in atherosclerotic lesions of hypercholesterolemic rabbits.
reduce MMP-9 secretion by macrophages. Arterioscler Thromb Arterioscler Thromb Vasc Biol 2000;20:7079.
Vasc Biol 1998;18:16711678. 82. Perrey S, Legendre C, Matsuura A, et al. Preferential pharmacologi-
76. Guijarro C, Blanco-Colio LM, Ortego M, et al. 3-Hydroxy-3- cal inhibition of macrophage ACAT increases plaque formation in
methylglutaryl coenzyme a reductase and isoprenylation inhibitors mouse and rabbit models of atherogenesis. Atherosclerosis 2001;
induce apoptosis of vascular smooth muscle cells in culture. Circ 155:359370.
Res 1998;83:490500. 83. Ziouzenkova O, Perrey S, Marx N, Bacqueville D, Plutzky J. Peroxisome
77. Negre-Aminou P, van Vliet AK, van Erck M, van Thiel GC, van proliferator-activated receptors. Curr Atheroscler Rep 2002;4:5964.
Leeuwen RE, Cohen LH. Inhibition of proliferation of human 84. Fayad ZA, Fuster V. Clinical imaging of the high-risk or vulnerable
smooth muscle cells by various HMG-CoA reductase inhibitors; atherosclerotic plaque. Circ Res 2001;89:305316.
 CHAPTER 35 / MDCT OF CLINICAL/PRECLINICAL CORONARY ATHEROSCLEROSIS 377

35 Multidetector-Row CT Imaging of Clinical

and Preclinical Coronary Atherosclerosis

CHRISTOPH R. BECKER, MD

PREREQUISITE FOR THE ASSESSMENT ASSESSMENT OF CORONARY ATHEROSCLEROSIS

OF CORONARY ATHEROSCLEROSIS The morphology of calcifications may already give a first
Complete assessment of coronary atherosclerosis with hint for the presence or absence of significant stenoses in the
multidetector-row computed tomography (MDCT) requires coronary arteries. From electron beam CT (EBCT) studies,
motion-free, contrast-enhanced images with the highest spatial Kajinami et al. reported that the positive predictive value for
resolution available. Because of the rather long exposure time significant stenosis (=75%) was 0.04 and 0.17 in none, 0.18 and
with MDCT (approx 200 ms), patient preparation with -blocker 0.59 in spotty (Fig. 1), 0.32 and 0.87 in long, 0.40 and 0.84 in
may be necessary. Optimal results will be achieved if the heart wide, and 0.56 and 0.96 in diffuse (Fig. 2) coronary calcifica-
rate of the patient is below 60 beats per minute (bpm). The tions, respectively (5).
spatial resolution is given by the detector element size. The Besides displaying the coronary artery lumen, MDCT as a
current reasonably achievable near isotropic spatial resolution cross-sectional modality is able to display the coronary artery
is approx 0.4 mm3. wall. Coronary calcifications can easily be assessed even with-
A timely, accurate, and homogenous vascular lumen out contrast media, and represent an advanced stage of athero-
enhancement is essential for full diagnostic capability of coro- sclerosis. However, as different stages of coronary atherosclerosis
nary MDCT angiography studies and to assess coronary ath- may be present simultaneously, calcifications may also be
erosclerosis. Higher contrast enhancement is superior to associated with lesions of more early stages of the disease. The
identify small vessels in MDCT. However, coronary athero- entire extent of coronary atherosclerosis will be underestimated
sclerosis is commonly associated with calcifications that may by assessing coronary calcifications alone (6). Calcified as well
interfere with dense contrast material and hinder the assess- as noncalcified lesions can be completely assessed by contrast-
ment of the residual lumen. In a randomized control group enhanced MDCT of the coronary arteries.
study, we investigated different contrast agent densities as well In our initial experience, it seems that different histological
as peripheral venous injection rates to determine the optimal stages of atherosclerosis will present with different morpho-
contrast protocol for the assessment of coronary atherosclero- logical pattern in MDCT. Because of the limited spatial
sis. We found that a contrast medium flow rate of 1 g/s iodine resolution, it cannot be expected to assess a thinned fibrous cap
resulted in an enhancement of approx 250300 Hounsfield units (65 m) that is approx 10 times beyond the current resolution of
(HU), which allows for delineation of intermediate (91 21 MDCT (750 m). In addition, contrast uptake as an indicator of
HU) as well as highly dense plaques (419 194 HU) (1). inflammatory processes is unlikely to be detected by MDCT. In
The final vessel enhancement will also depend on physi- extremely large atheromas, a lipid core may be visible with
ological parameters such as cardiac output and central blood negative density values corresponding to fat. More commonly,
volume. The cardiac output and central blood volume (corre- atheromas may present as well-defined and homogenous humps
lates with body weight) is inversely related with the final in the coronary artery wall, without calcifications (Fig. 3). The
enhancement (2). It needs to be considered that in patients with density of these plaques may vary between 40 and 60 HU, and
heart rates above 60 bpm, -blockers are now frequently applied may reflect the ratio between lipid and fibrous tissue. These
to achieve good image quality (3,4). The consequent lower kinds of plaques are most often found in asymptomatic patients.
cardiac output may lead to a higher final enhancement in these Plaque with densities above 80 HU may be considered to con-
patients. Because MDCT imaging of the coronary arteries is tain predominantly fibrous tissue. Fibrous plaques may also
performed during the first pass of the contrast medium, the frequently be associated with calcifications (Fig. 4), indicating
central blood volume plays a minor role for the final enhancement. an advanced stage of atherosclerotic disease as well as coronary
artery disease, and may therefore be found in patients with
chronic and stable angina.
From: Contemporary Cardiology: CT of the Heart: Other types of plaques (Table 1) present with low densities
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ (20 HU) and inhomogeneous structure with irregular defined

377
378 BECKER

Fig. 1. The positive predictive value of spotty coronary calcifications

(calcified nodules) for predicting coronary artery stenosis is between Fig. 2. The positive predictive value of diffuse coronary calcifications
18% and 59%. is between 56% and 96%.

Fig. 3. Homogenous and well-defined lesion with approx 40-HU Fig. 4. Homogenous plaques with approx 100-HU density and calci-
density in asymptomatic patients most likely corresponds to fications correspond most likely to fibro-calcified plaques and may be
atheromas. found in patients with chronic stable angina.

borders (Fig. 5). We have frequently observed these kinds of whereas in chronic vessel occlusion, organizing fibrous tissue
lesions in patients with acute and unstable angina. We currently may lead to shrinkage and increased density of the vessel.
believe that these plaques may correspond to intra-coronary As mentioned above, spotty, calcified lesions may com-
thrombi. In some cases, the entire coronary artery may be filled monly be associated with minor wall changes in conventional
with low-density material with a bright rim surrounding it, in- coronary angiography (5). However, it is known that such cal-
dicating a complete thrombus occlusion. Acute thrombus oc- cified nodules may also be the source of unheralded plaque
clusion may lead to enlargement of the coronary artery vessel, rupture and consequent thrombosis, and may lead to sudden
 CHAPTER 35 / MDCT OF CLINICAL/PRECLINICAL CORONARY ATHEROSCLEROSIS 379

Table 1
Coronary Artery Plaque Entities and Morphological Appearance in Multidetector-Row CT
Plaque entity AHA type Calcification Density Shape Remodelling Symptoms
Atheroma IV No approx 40 HU Smooth Positive No
Fibroatheroma Va No approx 60 HU Smooth Positive/negative No
Fibrotic lesion Vc No approx 100 HU Smooth Positive/negative No
Fibrocalcified plaque Vb Yes approx 100 HU Smooth Negative Chronic stable angina
or absent
Thrombus VI No approx 20 HU Irregular High-grade stenosis Acute unstable angina
or occlusion

Fig. 5. Irregular, nonhomogenous, and low-density plaques in the left Fig. 6. In a patient with acute coronary syndromes, a calcified nodule
anterior descending coronary artery in a patient with acute coronary may have led to a consecutive thrombus formation (arrow).
syndrome. The lesion most likely corresponds to a thrombus in the
coronary artery.

coronary death in very rare cases (7). Since atherosclerosis is studies, as well as from coronary calcium screening studies
a continuously ongoing process, different stages of the disease obtained with no contrast and thicker slices (8).
may be found simultaneously in the coronary artery wall. In However, because of partial-volume effects and the close
such cases, coronary calcium may be found together with relationship to the myocardium, it is much more difficult to
thrombus formation (Fig. 6). quantify noncalcified plaques and vessel wall changes in the
coronary arteries. In addition, with image reconstruction algo-
QUANTIFICATION OF CORONARY rithms that are used to visualize the soft tissue in the coronary
ATHEROSCLEROSIS artery wall, dense material such as calcifications will become
As already shown in many coronary calcium studies, exaggerated. In patients with extensive coronary calcifications,
quantification may have an impact on the risk assessment and noncalcified plaques are rarely found, most likely because the
follow-up of patients with coronary atherosclerosis. Coronary so-called blooming artifact of calcium prevents its assessment.
calcifications can easily be detected and quantified even in the Therefore, the optimal quantification algorithm for determina-
presence of contrast material if the density of calcium (>350 tion of the complete amount of noncalcified and calcified ath-
HU) is above the density of the contrast medium in the coronary erosclerosis by MDCT is still under development.
artery lumen (250300 HU). The mass of coronary calcifica-
tions can best be determined by measuring the volume and CONCLUSION
density of the plaques. Using correction factors, the mass of The newest generation of MDCT scanners now allows for
calcium can be similarly derived from MDCT angiography consistently good image quality in patients with regular sinus
380 BECKER

rhythm and heart rate in the range between 40 and 60 bpm. With 3. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama
targeted administration of contrast medium, an enhancement PM, de Feyter PJ. Reliable noninvasive coronary angiography with
fast submillimeter multislice spiral computed tomography. Circula-
can be achieved that allows for simultaneous assessment of tion 2002;106:20512054.
calcified as well as noncalcified atherosclerotic lesions in the 4. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery
coronary artery wall. Morphologic criteria allow for distinguish- stenoses with thin-slice multi-detector row spiral computed tomo-
ing between basic types of atherosclerosis such as thrombus, graphy and multiplanar reconstruction. Circulation 2003;107:
664666.
atheroma, or fibro-calcified plaques. In addition, calibrated HU 5. Kajinami K, Seki H, Takekoshi N, Mabuchi H. Coronary calcifi-
allows for estimation of the predominant component of lesions. cation and coronary atherosclerosis: site by site comparative mor-
These new features of coronary MDCT angiography may offer phologic study of electron beam computed tomography and
the chance to detect, quantify, and follow up vulnerable plaques coronary angiography. J Am Coll Cardiol 1997;29:15491556.
6. Wexler L, Brundage B, Crouse J, et al. Coronary artery calcification:
with a high fraction of lipids. However, partial-volume effects pathophysiology, epidemiology, imaging methods, and clinical
with the myocardium and coronary calcifications may interfere implications. A statement for health professionals from the Ameri-
with the detection of these plaques. Therefore, the optimal quan- can Heart Association. Circulation 1996;94:11751192.
tification algorithm is still under development. With these tools, 7. Virmani R, Kolodgie FD, Burke AP, Frab A, Schwartz SM. Lessons
from sudden coronary death. A comprehensive morphological clas-
large prospective cohort studies in an asymptomatic population
sification scheme for atherosclerotic lesions. Arterioscler Thromb
will be necessary to determine the predictive value for future Vasc Biol 2000;20:12621275.
cardiac events and the change over time under therapy of athero- 8. Hong C, Becker C, Schoepf UJ, Ohnesorge B, Bruening R, Reiser
sclerosis detected by MDCT. M. Absolute quantification of coronary artery calcium in non-
enhanced and contrast enhanced multidetector-row CT studies.
Radiology 2002;223:474480.
REFERENCES
9. Gibbons R, Chatterjee K, Daley J, et al. ACC/AHA/ACP-ASIM
1. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection guidelines for the management of patients with chronic stable angina:
and evaluation of atherosclerotic coronary plaques with multislice a report of the American College of Cardiology/American Heart
computed tomography. J Am Coll Cardiol 2001;37:14301435. Association Task Force on Practice Guidelines (Committee on
2. Fleischmann D. Use of high concentration contrast media: principles Management of Patients With Chronic Stable Angina). J Am Coll
and rationale vascular district. Eur J Radiol 2003;45:S88S93. Cardiol 1999;33:20922197.
 CHAPTER 36 / MDCT VS IVUS FOR CORONARY PLAQUE CHARACTERIZATION 381

36 Multidetector-Row CT
vs Intravascular Ultrasound
for Coronary Plaque Characterization

AXEL KUETTNER, MD

RATIONALE FOR CORONARY the fraction of interventional procedures remained constantly

PLAQUE DETECTION low at about 30% (7). Although coronary angiography has
Coronary artery disease (CAD) constitutes a major clinically become a safe procedure with only a small associated risk, the
relevant disease in the Western industrialized world, causing inconvenience for the patient as well as the economic burden
600,000 deaths annually (1). The underlying pathophysiological have fuelled the quest to find an alternative, noninvasive
mechanism seems to be coronary plaque disruption, subsequent method to visualize and assess coronary plaque burden.
thrombosis, and acute myocardial infarction. Fifty percent of
LIMITATIONS OF CONVENTIONAL CAG
afflicted patients experience this potentially life-threatening
condition without experiencing prior symptoms (2). Therefore, Although angiography is still the gold standard used for
direct visualization of epicardial coronary arteries is necessary defining coronary anatomy and to depict CAD, studies have
to assess the focal severity and clinical relevance of the vessel shown that accuracy and reproducibility of this technique can
wall alterations. The visualization of plaque has two different be challenged, and its results are not totally conclusive (8).
clinical focal points. On the one hand, obstructive coronary Visual interpretation of angiograms can exhibit significant
artery disease causing chronic ischemia to the vessel-dependent observer variability and often correlates poorly with post-
myocardial tissue needs to be assessed to determine an adequate mortem examination (9). In technical terms, angiography
revascularization strategy (condition of stable angina pectoris). depicts coronary arteries in a two-dimensional projected plane
On the other hand, precursors of the already mentioned unher- of the contrast-filled lumen. Any angiographic projection may
alded plaque rupture causing unstable angina, myocardial inf- more or less accurately depict the true extent of luminal nar-
arction, or sudden death should be assessed to take preventive rowing. Coronary interventions may increase luminal irregu-
measures to avoid these acute coronary syndromes. larity, thus further impairing the accuracy of angiography (10).
Evidence suggests that atherosclerotic plaque composition Also, assessment of lesion severity requires the measurement
and configuration are important predictors of plaque stability of the intrastenotic minimal luminal diameter as well as a nor-
(3). The risk for plaque rupture seems to depend rather on com- mal segment prior and distal of the lesion. However, post-
position rather than on plaque volume (4,5). Most ruptures mortem studies demonstrate that CAD constitutes a diffuse
occur in plaques containing a soft, lipid-rich core that is cov- disease with no truly normal segments. Remodeling represents
ered by a thin and inflamed cap of fibrous tissue (6). Small another feature of early CAD, making the true assessment of
ruptures often remain clinically silent, but more extensive the disease process difficult since true luminal narrowing may
plaque ruptures may cause the onset of unstable angina, myo- be minimal for a long time while plaque enlargement is com-
cardial infarction, or sudden death. Thus, the reliable pensated by outward growth of the vessel. Intravascular ultra-
noninvasive detection and classification of coronary lesions sound (IVUS) and multislice CT of the coronaries may be able
would constitute an important step forward in risk stratification to overcome these limitations (11,12).
of patients with known or suspected coronary artery disease. CORONARY PLAQUE IMAGING
The current gold standard to assess the morphological
USING RETROSPECTIVELY GATED MDCT
severity of CAD is invasive conventional coronary angiogra-
phy (CAG). In Germany alone, the total number of conven- Since 1999, mechanical multislice spiral computed tomog-
tional CAG has risen by 45%, from 409,000 in 1995 to more raphy (MSCT) systems with simultaneous acquisition of four
than 594,000 annual procedures in the year 2000 (7). However, slices and half-second scanner rotation have become available.
Multirow acquisition with these scanners allows for consider-
From: Contemporary Cardiology: CT of the Heart: ably improved visualization of the coronary arteries (1315).
Principles and Applications
Initial experiences have shown that coronary lesions can be
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
detected with acceptable sensitivity and specificity (1619).

381
382 KUETTNER

There is evidence that even preclinical atheroma and noncalci- (QCA) require calibration to correct for radiographic magnifi-
fied plaque tissue can be identified (2022). With the introduc- cation, a potential source of errors. IVUS uses an electronically
tion of the second generation of multislice spiral scanners with generated scale, performing direct planimetry. The tomo-
up to 16 detector rows and a gantry rotation time as low as graphic perspective of ultrasound enables an assessment of
420 ms, an improved tool to visualize coronary lesion was vessels that are difficult to image by angiography, including
introduced in 2002 (18,19). diffusely diseased segments, ostial or bifurcation stenoses, as
In the following, scanning techniques are described to well as eccentric plaques.
acquire 4- and 16-slice MSCT data sets suitable for plaque IVUS consists of a catheter incorporating a miniaturized
imaging. transducer and a console to reconstruct the images. Ultrasound
In our institution, cardiac MSCT data sets are acquired frequencies between 20 and 50 MHz are used, yielding a prac-
using 4- and 16-slice CT scanners. tical axial resolution of approx 150 m. Lateral resolution
First, a native scan without contrast medium is performed to averages 250 m. Current catheters range from 2.6 to 3.5 French
determine the total calcium burden of the coronary tree (num- (0.87 to 1.17 mm) and can be placed through a 6-French guid-
bers in parenthesis correspond to 16-slice scanning): collima- ing catheter. Mechanically rotated devices and multielement
tion 2.5 mm (1.5 mm), pitch1.5 (table feed 3.8 mm/rotation), electronic arrays are available.
tube current 133 eff. mAs (133 eff. mAs) at 120 kV (120 kV). Standard techniques for intracoronary catheter delivery are
To determine the circulation time, 20 mL of contrast medium used for intravascular examination. The operator advances or
(20 mL at 4 mL/s, 400mg iodine/mL) and a chaser bolus of retracts the IVUS device over the wire, recording the data for
20 mL saline is administered in an antecubital vein. The correct subsequent analysis. A motorized pullback device is used to
scanning delay is established by measuring CT attenuation withdraw the catheter at a constant speed (0.5 mm/s).
values in the ascending aorta, using the last slice with maximal Complication rates of an IVUS exam vary from 1% to 3%,
contrast as circulation time. By using a dual-head power injec- transient spasm being the most common. The complication rate
tor, a total of 150 mL (80 mL) intravenous contrast agent plus of major dissection or vessel closures is approx 0.5%.
a 20-mL chaser bolus is injected (50 mL at 4.0 mL/s, then IVUS is accurate in determining the thickness and
100 mL (30 mL) at 2.5 mL/s). CT imaging starts at the dia- echogenicity of vessel wall structures, but it is not consistently
phragm caudally of all cardiac structures and stops at the aortic able to provide actual histology. Validated methods do not yet
root cranial to the coronary ostia: 4 1.0-mm collimation exist for objective or automated classification of atheromatous
(16 0.75 mm), pitch 1.5 (table feed 1.5 mm/rotation), tube cur- lesions. Artifacts may adversely affect ultrasound images. The
rent 500 eff. mAs at 120 kV for both 4- and 16-slice scanning. physical size of ultrasound catheters (currently approx 1.0 mm)
To reconstruct the images, the standard built-in reconstruc- constitutes an important limitation in imaging high-grade
tion algorithm is used. The reconstruction window is set to start lesions.
at 60% RR interval for all native images as well as for the The current clinical application of IVUS is the correct
contrast-enhanced scan. If necessary, a test series is performed assessment of intermediate lesions and left main artery disease,
ranging from 35% to 75% in 2% steps. That percentage that although no absolute indications exist when the device should
displayed the least motion artifacts is chosen to reconstruct the be used. In clinical trials, however, IVUS provides an indis-
entire stack of images of the CT angiography (CTA) scan. Using pensable tool for research. Many important trials that shifted
this reconstruction strategy, sets of images are created that show treatment strategies in clinical cardiology used IVUS as refer-
no substantial motion artifacts. ence standard (23).
The calcium score is determined on an offline workstation,
based on an Agatston equivalent scoring system as well as IVUS AND CTA COMPARISON STUDY
expressed in absolute mg calcium-hydroxyapatite. As mentioned earlier, when MSCT was introduced in 1999,
On the basis of original axial slices, 3D volume-rendering soon evidence was found that even preclinical atheroma and
technique (VRT) images as well as thin-sliding maximum noncalcified plaque tissue could be identified (20). However,
intensity projections (MIP), each coronary plaque is assessed. these early results had to be confirmed to show whether MSCT
Parameters to describe plaques are density in absolute Houns- was really able to correctly assess not only the severity of the
field units (HU), presence or absence of calcifications, length lumen loss, but also plaque composition and plaque configura-
of the lesion, and the degree of luminal narrowing in %. tion. Thus, the purpose of the study was to investigate whether
plaque composition as assessed by MSCT corresponds to the
CORONARY PLAQUE DETECTION USING IVUS: gold standard, IVUS.
TECHNIQUE AND CLINICAL INDICATIONS A total of 15 patients were enrolled in the study who were
Intravascular ultrasound imaging offers several advantages scheduled for IVUS-guided percutaneous transluminal
in the evaluation of CAD. First, as a result of the imaging from angioplasty (PTCA). The angiographic criteria for enrolment
inside the vessel, IVUS provides images of the atherosclerotic in the study was a lesion with a stenosis >70% and the absence
plaque, not only the lumen. Tomographic orientation of IVUS of severe vessel angulations (>90) in the proximal vessel seg-
offers a three-dimensional visualization of the entire circum- ments. Angiographic exclusion criteria were left main artery
ference of the vessel wall and not just biplanar projections. disease, total occlusions, and vessel diameter <2 mm with
Also, correct angiographic vessel or stenosis measurements bypass lesions. According to the study protocol, all patients
 CHAPTER 36 / MDCT VS IVUS FOR CORONARY PLAQUE CHARACTERIZATION 383

Fig. 1. Intravascular ultrasound criteria for the differentiation of plaque.

were examined by MSCT within 24 h prior to the coronary plaques analyzed, the test for statistical significance was highly
intervention. The examination protocol used was mentioned positive (KruskalWallis test, p < 0.0001) (Figs. 35).
earlier in the chapter. Immediately before the intracoronary Since some patients had up to five plaques, plaque density
intervention, IVUS was performed to analyze the vessel con- results had to be tested for independence of the patient itself.
figuration proximal to the target lesion and within the lesion. First it was tested whether patient groups with either one, two,
To ensure that identical plaques were assessed by the different three, four, or five plaques were significantly different from
techniques and to allow for precise correlations, landmarks each other, which was not the case (p = 0.876). Also, the inde-
such as the origin or side branches and the distance to the target pendence of patient group and plaque composition was dem-
lesion were used. onstrated (p = 0.817).
The assessed plaque configuration was classified according These data suggest that MSCT is capable of differentiation
to the following IVUS criteria published by our group and oth- among different plaque compositions. The density measure-
ers (Fig. 1): ments performed by MSCT correlate highly with the well
established IVUS criteria (soft, intermediate, and calcified).
Soft plaques: More than 80% percent of the plaque area is
There was no overlap in the mean density values among the
composed of tissue with an echogenity of lower than the
echogenity of the surrounding adventitia. If calcifica- three groups of plaques. Thus, especially soft plaques with
tions were partially present, the arc of lesion calcium presumably lipid-rich core might be identified by density val-
had to be <90% ues <50 HU. Intermediate plaques showed a density ranging
Intermediate plaque: More than 80% percent of the plaque from 50 to 119 HU. Lesions with a density >120 HU correlated
area is composed of tissue producing echoes as bright or to calcified plaques in the IVUS study (Fig. 6).
brighter than those of the surrounding adventitia but with- A more precise view by MSCT on plaque configuration by
out acoustic shadows. If calcifications were partially visualizing lipid cores, fibrous caps, or smallest calcified
present, the arc of lesion calcium had to be <90% sprinkles is at present restricted because of the limited spatial
Calcified plaque: This plaque type involves bright echoes resolution of the 4-slice MSCT used to perform the study.
with acoustic shadowing accompanying >90 of the vessel Further clinical studies are necessary to confirm the relevance
wall circumference. of a plaque score that is based on these findings. The impact of
A total of 34 plaques (right coronary artery [RCA] n = 12, 16-slice MSCT will be discussed later in this chapter.
left anterior descending artery [LAD] n = 22) were analyzed by
both methods with respect to lesion configuration (Fig. 2). On FUTURE DIRECTIONS: 16-SLICE MSCT
IVUS, 12 plaques were classified as soft, 5 plaques as interme- With the introduction of the 2nd generation of multislice
diate, and 17 plaques as calcified. When comparing these data spiral scanners with up to 16 detector rows and a gantry rotation
to MSCT data, the plaques identified by IVUS as soft had a time as low as 420 ms, an improved tool to visualize coronary
mean density of 14 26 HU, those classified as intermediate lesion was introduced in 2002 (18,19). Thus, a more reliable
had a mean density of 91 21 HU, and calcified plaques had a tool for the visualization of plaques is available. As mentioned
mean density of 419 194 HU. Calcifications were also found earlier, next to a possible prognostic factor of plaque visualiza-
in intermediate plaques; however, only small sprinkles of cal- tion, a keen interest remains whether plaque morphology also
cium deposits were found. Interestingly, even with only 34 is a predictor for the assessment of stenosis.
384 KUETTNER

Fig. 2. Comparison between intravascular ultrasound (left), multislice CT (middle), and angiography (right).

Fig. 3. Multidetector-row CT angiography-criteria soft plaque: hypodense in comparison to vessel filled with contrast, density <50 HU.

A total of 41 patients referred to our institution for conven- 1. Calcified plaque adhered to vessel wall, no vessel obstruc-
tional CAG were examined also by MDCT according to the tion
protocols given in this chapter. 2. Calcified plaque, CT morphologically complete vessel
On the basis of original axial slices, 3D VRT images as well obstruction
as thin-sliding MIPs, each coronary plaque identified by con- 3. Calcified plaque conglomerate
ventional CAG as a high-grade stenotic lesion or occlusion was 4. Noncalcified plaque
assessed. 5. Mixed plaque with calcifications predominantly present
Depending on the identified plaque composition, each 6. Mixed plaque with noncalcified tissue predominantly
plaque was classified in one of six groups as follows (Fig. 7): present
 CHAPTER 36 / MDCT VS IVUS FOR CORONARY PLAQUE CHARACTERIZATION 385

Fig 4. Multidetector-row CT angiography-criteria intermediate plaque: hypodense in comparison to vessel filled with contrast, density 50120 HU.

Fig 5. Multidetector-row CT angiography-criteria calcified plaque: hyperdense in comparison to vessel filled with contrast, density >120 HU.

In all patients, coronary angiograms were obtained using this pilot study was to principally analyze different types of plaque
4- and 5-French catheters. The angiograms were evaluated by composition and their predominance in high-grade lesions.
quantitative coronary analysis. All plaques with an underlying In 533 coronary artery segments scanned, all 69 plaques
diameter stenosis =70% or total occlusions were documented causing high-grade diameter stenosis or complete occlusion
according to a modified AHA classification (24). could be detected. These plaques were subclassified according
MSCT plaque data were analyzed unblinded to the results of to the defined criteria.
conventional CAG, and each plaque was attributed a group char- No plaque was attributed group 1 characteristics. 1 plaque
acteristic as described. Next we analyzed which of the plaque was attributed group 2 characteristics, 6 plaques had group 3
groups was predominantly present and whether there were clini- characteristics, 20 plaques had group 4 characteristics, 29
cally relevant differences between the groups. The primary aim of plaques were assigned to group 5, and 13 plaques to group 6.
386 KUETTNER

Fig. 6. Density distribution of plaques characterized by intravascular ultrasound as soft, intermediate, and calcified.

Fig. 7. Classification of different plaque composition. The first letter represents the plaque scheme, the double letter an actual CT image of
the different plaque types. A) + AA): calcified nodule; B) + BB): calcified plaque; C) +CC): soft plaque; D) + DD): mixed plaque.
 CHAPTER 36 / MDCT VS IVUS FOR CORONARY PLAQUE CHARACTERIZATION 387

Conventional coronary angiography revealed a total of 49 cation of atherosclerosis. A report from the Committee on Vascular
high-grade lesions (>70%) and 20 complete occlusions. Of all Lesions of the Council on Arteriosclerosis, American Heart Asso-
ciation. Circulation 1995;925:13551374.
these lesions, only 9/69 (13%) were caused by calcified 7. Mannebach H, Hamm C, Horstkotte D. [18th report of the statistics
plaques. Plaques of group 1 (calcified nodules) caused no of heart catheter laboratories in Germany. Results of a combined
lesion at all; two plaques of group 2 (vessel obstruction) survey by the Committee of Clinical Cardiology and the Interven-
caused 2/69 (3%) lesions; and six plaques of group 3 (calci- tional Cardiology and Angiology Working Group for ESC of the
fied conglomerate) caused a high-grade lesion (9%); so the German Society of Cardiology-Heart- and Cardiovascular Research
2001]. Z Kardiol 2002; 919:727729.
total of all calcified plaque types caused only 9/69 (13%) high- 8. Zir LM, Miller SW, Dinsmore RE, Gilbert JP, Harthorne JW.
grade lesions or occlusions. Soft plaques (group 4 plaques) Interobserver variability in coronary angiography. Circulation
caused 20/69 (29%) high grade lesions, and the majority of all 1976;534:627632.
high-grade lesions was caused by mixed plaques (42/69 9. Grodin CM, Dydra I, Pastgernac A. Discrepancies between cinean-
61%), of which predominantly calcified plaques accounted giographic and post-mortem findings in patients with coronary
artery disease and recent myocardial revascularization. Circula-
for 29/69 (42%) lesions and group 6 (predominantly tion 1974;49:703709.
noncalcified plaques) accounted for 13/69 (19%) lesions. All 10. Galbraith JE, Murphy ML, de Soyza N. Coronary angiogram
together, 60/69 (87%) of all high-grade lesions or occlusions interpretation. Interobserver variability. JAMA 1978;240(19):
contain noncalcified plaque tissue. These distribution results 20532056.
are statistically relevant (p < 0.0001, Pearsons chi-square test). 11. Kopp AF, Kuttner A, Heuschmid M, Schroder S, Ohnesorge B,
Claussen CD. Multidetector-row CT cardiac imaging with 4 and
When analyzing the data with respect to presence of calcified 16 slices for coronary CTA and imaging of atherosclerotic plaques.
or noncalcified plaque tissue alone, the differences were sta- Eur Radiol 2002;12 Suppl 2:S1724.
tistically not relevant (49/69 vs 60/69, p < 0.2921). 12. Nissen SE, Gurley JC, Grines CL, et al. Intravascular ultrasound
Intriguingly, the majority of all high-grade lesions and assessment of lumen size and wall morphology in normal subjects
occlusions are not caused by merely calcified plaques but con- and patients with coronary artery disease. Circulation 1991;843:
10871099.
tain noncalcified plaque tissue, which is characteristic for 87% 13. Becker CR, Ohnesorge BM, Schoepf UJ, Reiser MF. Current de-
of high-grade lesions in our studied population. Only 13% of velopment of cardiac imaging with multidetector-row CT. Eur J
the lesions studied were solely calcified plaques. Radiol 2000;362:97103.
These data also seem to suggest that a binary decision tree 14. Kopp AF, Ohnesorge B, Flohr T, et al. [Cardiac multidetector-row
of calcified vs noncalcified plaque tissue is not sufficient to CT: first clinical results of retrospectively ECG-gated spiral with
optimized temporal and spatial resolution]. Rofo Fortschr Geb
characterize high-grade lesions, since both entities are present Rontgenstr Neuen Bildgeb Verfahr 2000;1725:429435.
in the majority of high-grade lesions. Only the combination 15. Ohnesorge B, Flohr T, Schaller S, et al. [The technical bases and uses
yields statistically relevant results. of multi-slice CT]. Radiologe 1999;3911:923931.
If these findings can be confirmed in larger studies, this 16. Kopp AF, Schroeder S, Kuettner A, et al. Non-invasive coronary
would have a major impact on examination strategies when angiography with high resolution multidetector-row computed
tomography. Results in 102 patients. Eur Heart J 2002;2321:
looking for obstructive CAD using MSCT. These findings sug- 17141725.
gest that a combination of both examsthe detection of coro- 17. Nieman K, Oudkerk M, Rensing BJ, et al. Coronary angiography
nary calcifications in combination with a contrast enhanced with multi-slice computed tomography. Lancet 2001;3579256:
scan of the coronariesshould be performed to visualize and 599603.
detect obstructive CAD. Currently, IVUS studies are being 18. Nieman K, Cademartiri F, Lemos PA, Raaijmakers R, Pattynama
PM, de Feyter PJ. Reliable noninvasive coronary angiography with
performed to verify these data, and preliminary data confirm fast submillimeter multislice spiral computed tomography. Circula-
these findings. tion 2002;106(16):20512054.
It could be shown that MSCT allows noninvasive determi- 19. Ropers D, Baum U, Pohle K, et al. Detection of coronary artery
nation of plaque morphology, suggesting that high-grade coro- stenoses with thin-slice multi-detector row spiral computed tomog-
nary lesions are predominantly caused by mixed plaques. Solely raphy and multiplanar reconstruction. Circulation 2003;107(5):
664666.
calcified plaques, however, are a less frequent cause of severe 20. Becker CR, Knez A, Ohnesorge B, Schoepf UJ, Reiser MF. Imaging
lesions. of noncalcified coronary plaques using helical CT with retrospec-
tive ECG gating. AJR Am J Roentgenol 2000;1752:423424.
REFERENCES 21. Kopp AF, Schroeder S, Baumbach A, et al. Non-invasive
1. Wielopolski PA, van Geuns RJ, de Feyter PJ, Oudkerk M. Coronary characterisation of coronary lesion morphology and composition by
arteries. Eur Radiol 1998;86:873885. multislice CT: first results in comparison with intracoronary ultra-
2. Falk E. Coronary thrombosis: pathogenesis and clinical manifesta- sound. Eur Radiol 2001;119:16071611.
tions. Am J Cardiol 1991;687:28B35B. 22. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection
3. Falk E, Shah PK, Fuster V. Coronary plaque disruption. Circulation and evaluation of atherosclerotic coronary plaques with multislice
1995;923:657671. computed tomography. J Am Coll Cardiol 2001;375:14301435.
4. Naghavi M, Libby P, Falk E, et al. From vulnerable plaque to 23. Nissen SE, Yock P. Intravascular ultrasound: novel pathophysiologi-
vulnerable patient: a call for new definitions and risk assessment cal insights and current clinical applications. Circulation 2001;
strategies: Part II. Circulation 2003;10815:17721778. 103(4):604616.
5. Naghavi M, Libby P, Falk E, et al. From vulnerable plaque to 24. Austen WG, Edwards JE, Frye RL, et al. A reporting system on
vulnerable patient: a call for new definitions and risk assessment patients evaluated for coronary artery disease. Report of the Ad Hoc
strategies: Part I. Circulation 2003;10814:16641672. Committee for Grading of Coronary Artery Disease, Council on
6. Stary HC, Chandler AB, Dinsmore RE, et al. A definition of Cardiovascular Surgery, American Heart Association. Circulation
advanced types of atherosclerotic lesions and a histological classifi- 1975;51(4 Suppl):540.
 CHAPTER 37 / MDCT VS MRI FOR CORONARY PLAQUE CHARACTERIZATION 389

37 Multidetector-Row CT vs Magnetic
Resonance Imaging for Coronary
Plaque Characterization

KONSTANTIN NIKOLAOU, MD, CHRISTOPH R. BECKER, MD,

AND ZAHI FAYAD, MD

INTRODUCTIONIMAGING sclerotic plaques in carotid (9,10) and aortic (11) arterial dis-
OF ATHEROSCLEROTIC DISEASE ease. In vivo imaging of the coronary artery wall is challenging
Reliable noninvasive imaging tools that can detect various because of a combination of cardiac and respiratory motion
stages of atherothrombotic disease in different vessel regions artifacts, and the tortuous course, small size, and location of the
and characterize the composition of the plaques are clinically vessels. Initial in vivo studies in human coronary arteries have
desirable (1). Such imaging tools would improve our under- used noninvasive black-blood spin-echo techniques with
standing of the pathophysiological mechanisms underlying breath-holding (12) or a real-time navigator for respiratory
atherothrombotic processes and allow us to better risk-stratify gating (13).
the disease. Future goals are optimal tailoring of treatment and By possibly combining the advantages of both techniques,
direct monitoring of the vascular response. Currently available detecting significant stenoses and describing the plaque com-
imaging techniques for the diagnosis of coronary artery disease position at the same time, information could be provided that
(CAD) are subject to several limitations. Conventional coro- may predict cardiovascular risk, facilitate further study of
nary angiography, widely accepted as the gold standard for the atherothrombosis progression and its response to therapy, and
detection of coronary artery disease, demonstrates the degree provide for assessment of subclinical disease.
of luminal narrowing, but fails to visualize the coronary artery
wall. It has been shown that plaque composition rather than the
PLAQUE IMAGINGMETHODS
severity of an actual stenosis predicts the risk of plaque rupture METHODS OF CT PLAQUE IMAGING
and acute clinical complications of coronary artery disease (24). In 2002, newly developed 16-row CT systems were clini-
Thus, new imaging techniques that can image the artery wall cally introduced, allowing for faster data acquisition with
and characterize different lesion types may allow for identifi- improved spatial resolution (14). Primary requisites for a suf-
cation and follow-up of patients at risk and for selecting appro- ficient delineation and depiction of atherosclerotic calcified
priate therapeutic strategies (5). and noncalcified plaques are similar to the requirements for a
Presently, a number of imaging modalities are employed to high-quality CT angiography (CTA) of the coronariesi.e., a
study atherosclerosis and to assess luminal diameter, wall high spatial and high temporal resolution at the same time.
thickness, and plaque volume (6). Two noninvasive imaging There are several advantages of a 16-row CT system over a
modalities, computed tomography (CT) and magnetic reso- 4-row CT system concerning the depiction of coronary artery
nance imaging (MRI), have been introduced for the study of plaques. A direct comparison of a 4-row CT and a 16-row CT
atherothrombosis. Both have been shown to be capable of image clearly demonstrates this advantage (Fig. 1). First, the
imaging vessel wall structures and differentiating various gantry rotation time in 16-row CT for cardiac investigations is
stages of atherosclerotic wall changes. The latest generation of 420 ms, allowing a temporal resolution of =210 ms. This is a
multidetector-row computed tomography (MDCT) scanners gain of about 20% over the temporal resolution of a 4-row CT
allow for sufficiently reliable detection of significant proximal system. With higher heart rates and multisegment-reconstruc-
coronary stenoses (7), quantitative measurement of atheroscle- tion algorithms, the exposure time varies between 105 and
rotic burden including calcified and noncalcified plaques (8), 210 ms, depending on the actual heart rate (15). Second, the
and characterization of the plaque components (1). MRI has slice thickness is reduced from 1.25 mm to 0.75 mm, allowing
been applied in various in-vivo human studies to image athero- for an improved spatial resolution along the z axis. This way,
using 16-row CT, almost isotropic voxels can now be acquired.
Based on the improved spatial resolution, blooming artifacts
From: Contemporary Cardiology: CT of the Heart: of calcium deposits in the vessel wall are reduced as a result of
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ
decreased partial-volume effects. This allows for improved
depiction and delineation of calcified and noncalcified plaques.

389
390 NIKOLAOU, BECKER, AND FAYAD

Fig. 1. Comparison of thin maximum intensity projection of a simulated 4-row CT (left) and 16-row CT (right) angiography, depicting a plaque
in the left main coronary artery. 4-row CT simulation is based on 16-row CT data set and uses 1.25-mm thick slices. Acquiring thinner, 0.75-
mm thick slices with a 16-row CT system (right) clearly improves depiction quality and assessability of the noncalcified plaque (arrows).
(Courtesy of T.J. Jakobs, University of Munich.)

Third, and probably most important, the complete heart can been performed utilizing a multicontrast approach with high-
now be covered in a significantly shorter breath-hold time of resolution black-blood spin-echo and fast spin echo (FSE)
less than 20 s, compared to 35- to 40-s breath-hold time on a based MR sequences. The signal from the blood flow is ren-
4-row CT. This results in a considerable reduction of motion dered black through preparatory pulses (e.g., radiofrequency
artifacts and allows for a substantial reduction in contrast vol- spatial saturation or inversion recovery pulses) to better image
ume compared to previously published protocols on 4-row CT. the adjacent vessel wall (18). However, bright blood imaging
Compared to low-pressure arterial systems such as the pulmo- (i.e., 3D fast time of flight) can be employed in assessing fibrous
nary arteries, where calcifications are absent and the injection cap thickness and morphological integrity of the carotid artery
rate can be increased to visualize the smallest arterial branches, plaques (19). This sequence enhances the signal from flowing
in coronary arteries the opacification must not exceed approx blood, and a mixture of T2 and proton density contrast weight-
300 Hounsfield units (HU) for a reliable depiction and judg- ing highlights the fibrous cap. Atherosclerotic plaque charac-
ment of calcifications. Optimization of vessel contrast-to-noise terization by MR is generally based on the signal intensities and
ratio (CNR) is also mandatory for sufficient visualization of morphological appearance of the plaque on T1-weighted, pro-
noncalcified plaques, and can be performed either by a test ton densityweighted, and T2-weighted images as previously
bolus setting (20 mL + 50 mL NaCl) or a bolus tracking. validated (see references in recent reviews by Fayad et al. [6]
Because nonenhanced blood on CT has attenuation similar and Yuan et al. [20]).
(5070 HU) to that of noncalcified plaques, this type of lesion
can be detected only after administration of contrast medium.
PLAQUE IMAGINGAPPLICATIONS
Therefore, a vessel enhancement significantly above the CT APPLICATIONS OF CT PLAQUE IMAGING
values of noncalcified lesions (150 HU) must be achieved to Noncoronary Plaque Imaging With CT
allow for reliable detection. A target attenuation of 200 HU Several precedent studies in animals and humans in other
seems best suited to fulfill this requirement. With this vessel vascular territories than the coronary arteries have demon-
enhancement, calcified coronary lesions remain detectable strated the ability of CT to differentiate calcified, fibrous, and
because their attenuation is significantly higher (16). lipid-rich plaque components based on CT attenuation (HU).
CT is described as an accurate, noninvasive means for studying
METHODS OF MR PLAQUE IMAGING detailed plaque morphology and composition in the carotid
High-resolution MR has emerged as the potential leading arteries. According to Estes et al. (21), CT accurately defined
noninvasive in-vivo imaging modality for atherosclerotic plaque features containing calcium, fibrous stroma, and lipids
plaque characterization. MR differentiates plaque components in carotid arteries. Using tissue attenuation values, CT distin-
on the basis of biophysical and biochemical parameters such as guished between lipid and fibrous stroma (means 39 12 HU
chemical composition and concentration, water content, physi- and 90 24 HU, respectively, p < 0.001). Oliver et al. (22) tried
cal state, molecular motion, or diffusion (17). MR provides to assess whether features seen at CTA might be used to predict
imaging without ionizing radiation and can be repeated over carotid plaque stability by comparing CT angiograms with his-
time. In-vivo MR plaque imaging and characterization have topathologic examinations of the carotid artery bifurcation.
 CHAPTER 37 / MDCT VS MRI FOR CORONARY PLAQUE CHARACTERIZATION 391

Fig. 2. Mixed fibro-calcified plaque in the proximal left anterior descending coronary artery (LAD), as assessed by multidetector-row CT
(MDCT) and intravascular ultrasound (IVUS). Above: MDCT image, maximum intensity projection (MIP), with a fibrous plaque area (100
HU, arrow) next to significant calcifications. Below: IVUS images in cross-sectional views along the course of the plaque, moving from
the noncalcified, fibrous plaque area (left) towards the calcifications (middle and right).

That study concluded that CTA is a promising method for arteries in vivo (27,28). Various imaging features of noncalci-
assessing the lumen and wall of the carotid artery, and that the fied and calcified plaques depicted with CT correlate well with
apparent correlation between histologic appearance and plaque histopathologic stages of atherosclerosis defined by the Ameri-
density on CT angiograms could have important implications can Heart Association (AHA) (29), as demonstrated in a recent
for the prediction of plaque stability. ex-vivo study on human hearts (30).
Coronary Plaque Imaging With CT Intravascular ultrasound (IVUS) is the reference standard
CT has become an established method for noninvasive and for invasive detecting and evaluating atherosclerotic plaques in
highly sensitive detection of coronary artery calcifications (23). vivo. For plaques with and without signs of calcification
However, calcified plaques are probably the result of repetitive detected on intracoronary ultrasound, electron beam CT
plaque rupture and healing, causing shrinkage of the vessel (EBCT) without contrast enhancement yielded a sensitivity of
lumen with subsequent stenosis (24). Earlier stages of athero- 97% and 47% and a specificity of 80% and 75%, respectively
sclerosis without calcifications might be more prone to rupture, (31). In an in vivo study on contrast-enhanced MDCT vs IVUS
resulting in acute cardiovascular events (25). Yet, this theory for the accuracy in determining coronary lesion configuration,
does not stand undisputed, as an autopsy study suggested that Schroeder et al. (16) reported a good correlation of these two
the degree of calcification was greatest for acute and healed modalities. MDCT was able to differentiate between soft,
plaque ruptures (26). Recently, it was shown that CT has the intermediate, and calcified plaques, as compared to IVUS, with
potential to identify early, noncalcified plaques in the coronary significant differences in CT attenuation values (Fig. 2). In
392 NIKOLAOU, BECKER, AND FAYAD

Table 1
Imaging Parameters for Multidetector-Row CT and Magnetic
Resonance Imaging, and Typical Appearance of Different Plaque
Types According to the Composition of the Plaque

Modality
CT (HU) MR (SI)
C/E T1w PDw T2w ToF
Thrombus 20 + to to to +
Lipids 50 + +
Fibrous 100 to + + to + to
Calcium >300

HU, Hounsfield units; SI, signal intensities; C/E, contrast-enhanced; T1w,

T2w, PDw, ToF, multicontrast MR imaging sequences; , hypointense; ,
isointense; +, hyperintense.

Fig. 3. Magnetic resonance and CT plaque characterization. Coronary artery calcified plaque. (A) Cross-sectional ex-vivo T2-weighted MR
image of a human left anterior descending artery with a small calcified lesion (arrow) and vessel wall thickening. (B) Multislice CT image of
the same lesion (arrow), showing the typical blooming effect of calcified lesions in CT images. (C) Corresponding histopathologic section;
calcium is washed out during the preparation process.

recent studies on MDCT of ex-vivo coronary arteries vs histo- calcifications caused by the high CT attenuation values of these
pathology as the gold standard, again a good correlation was lesions (type Vb lesions). Sensitivity has been reported to be
found. Lipid-rich, fibrous, and calcified plaques were differen- lower for earlier stages of atherosclerosis (type III and IVa
tiated reliably (30) (Table 1 and Figs. 35). Acute intravascular plaques), mainly owing to lack of in-plane spatial resolution
thrombi can also be detected in vivo, with a typical appearance and to partial-volume effects. Still, heavy coronary calcifica-
of the irregular thrombus with typically low attenuation num- tions may prevent adequate assessment of complex plaques
bers, in the range of 2030 HU (Fig. 6). Additionally, new with calcified and soft plaque components in direct proximity
image analysis software may enable in-vivo quantification of to each other. The reasons for this are considerable beam hard-
noncalcified atherosclerotic lesions (Fig. 7) (8). ening artifacts and partial-volume effects of calcium on CT.
MDCT possesses a high sensitivity for the detection of cal- Whereas this effect facilitates the detection of calcium, the
cified plaques, as a result of its inherent high sensitivity for same effect hampers the assessment of noncalcified compo-
 CHAPTER 37 / MDCT VS MRI FOR CORONARY PLAQUE CHARACTERIZATION 393

Fig. 4. Magnetic resonance (MR) and CT plaque characterization. Coronary artery lipid-rich plaque. (A) Cross-sectional ex vivo T2-weighted
MR image of a human left anterior descending artery with a lipid-rich lesion (arrow). (B) Multislice CT image of the same lesion (arrow),
showing the typical low density of lipid-rich tissue (44 HU). (C) Corresponding histopathologic section with a large extracellular lipid pool.

Fig. 5. (A) Type VI atherosclerotic lesion with extensive lipid accumulation (Lc = lipid core) and small intra-plaque hemorrhage (Hb) within
the atherosclerotic widened intima (I) and media (M) cell layer. Corresponding multidetector-row CT (MDCT) image shows a soft tissue lesion
with a mean density of 40 HU. (B) In contrast to lipid-rich plaques, purely fibrous (F) plaque (Type Vc) without calcium demonstrate
significantly higher attenuation (90 HU) on MDCT. (C) Calcified nodules or spotty lesions contain little pieces of calcium (Ca) that was
removed in the process of preparing the slides. On MDCT the calcification can be easily detected and partly covers the adjacent fibrous soft
tissue (110 HU). (Courtesy of C.R. Becker, Ludwig-Maximilians-University, Munich, Germany.)
394 NIKOLAOU, BECKER, AND FAYAD

Fig. 6. 41-yr-old man with atypical chest pain. (A) Maximum intensity projection of contrast-enhanced helical CT scan reveals lumen
obstruction of left anterior descending coronary artery (arrowheads) and first diagonal branch (arrow). A thrombus can be delineated with
irregular borders and the typical low CT attenuation. (B) Conventional angiogram confirms findings on helical CT scan and also reveals high-
grade stenosis of left anterior descending coronary artery (arrowhead) and first diagonal branch (arrow). Modified from ref. 28.

Fig. 7. Quantitative volumetry of noncalcified plaques can be performed on special offline workstations (InSight, NeoImagery, San Francisco,
CA). Such volumetry is still conducted semi-quantitatively, segmenting the soft plaque borders by hand (arrow) while the software measures
plaque volume automatically according to predefined thresholds (circle).
 CHAPTER 37 / MDCT VS MRI FOR CORONARY PLAQUE CHARACTERIZATION 395

nents of lesions next to calcified components of the same ath- (Fig. 8) (33). The method was validated in swine coronary
erosclerotic plaque. Thus, the soft-tissue components in mixed lesions induced by balloon angioplasty (35). High-resolution
plaques with a high content of calcium may be unspecific and black-blood MR of both normal and atherosclerotic human
can resemble any AHA type atherosclerotic lesion on histopa- coronary arteries was performed for direct assessment of coro-
thology, although the presence of calcium within the plaque nary wall thickness and the visualization of focal atheroscle-
would suggest a classification as AHA type Vb. Yet, the ulti- rotic plaque in the wall. The difference in maximum wall
mate goal is to detect earlier stage, noncalcified plaques. Cal- thickness between the normal subjects and patients (>40%
cifications are thought to be more prevalent in advanced stages stenosis) was statistically significant. Fig. 8 shows in vivo MR
of atherosclerosis. According to Virmani et al. (2), calcified coronary plaque images from two patients. The coronary MR
plaques are likely the result of repetitive plaque rupture and plaque imaging study by Fayad et al. (33) was performed dur-
healing, causing shrinkage of the vessel lumen with consequent ing breath-holding to minimize respiratory motion, with a reso-
stenosis. Earlier stages of atherosclerosis without calcifications lution of 0.46 0.46 2.0 mm3. To alleviate the need for the
might be more prone to acute rupture, resulting in acute cardio- patient to hold his or her breath, Botnar et al. (36) have com-
vascular events (25). Yet, as described above, this presumption bined the black-blood FSE method and a real-time navigator
does not stand undisputed (26). for respiratory gating and real-time slice position correction. A
APPLICATIONS OF MR PLAQUE IMAGING near isotropic spatial resolution (0.7 0.7 0.8 mm3) was
Noncoronary Plaque Imaging With MRI achieved with the use of a 2D local inversion and black-blood
MR has been used to study atherosclerotic plaques in human preparatory pulses (36). This method provided a quick way to
carotid (17,32), aortic (18), and coronary (33) arterial disease. image a long segment of the coronary artery wall and may be
The superficial location and relative absence of motion in useful for rapid coronary plaque burden measurements. Future
carotid arteries presents less of a technical challenge for imag- studies need to address these possibilities.
ing than the aorta or coronary arteries. Some of the MR studies Monitoring of Atherosclerotic Disease With MRI
of carotid arterial plaques include the imaging and character- As shown in animal experimental studies (37,38), MR is
ization of atherosclerotic plaques (17,32), the quantification of also a powerful tool to serially and noninvasively investigate
plaque size (34), and the detection of fibrous cap integrity the progression and regression of atherosclerotic lesions in
(19). Typically, the images are acquired with a resolution of vivo. In asymptomatic, untreated, hypercholesterolemic
0.25 0.25 2.0 to 0.4 0.4 3.0 mm3 by use of a carotid patients with carotid and aortic atherosclerosis, Corti et al. (39)
phased-array coil to improve signal-to-noise ratio and image have shown that MR can be used to measure the efficacy of
resolution. In vivo, black-blood MR atherosclerotic plaque lipid-lowering therapy (statins). Atherosclerotic plaques were
characterization of the human aorta has been reported by Fayad assessed with MR at different times after lipid-lowering
et al. (18), who assessed thoracic aorta plaque composition and therapy. Significant regression of atherosclerotic lesions was
size with the use of T1-weighted (T1w), T2-weighted (T2w), observed. Despite the early and expected hypolipidemic effect
and proton densityweighted (PDw) images. The acquired of the statins, changes in the vessel wall were observed after 12
images had a resolution of 0.8 0.8 5.0 mm3 with a phased- mo of treatment. As with previous experimental studies, there
array chest coil. Matched cross-sectional aortic imaging with was a decrease in the vessel wall area and no alteration in the
MR and transesophageal echocardiography showed a strong lumen area after 12 mo (37,38). A case-controlled study dem-
correlation for plaque composition and mean maximum plaque onstrated substantially reduced carotid plaque lipid content (but
thickness. no change in overall lesion area) in patients treated for 10 y with
Coronary Plaque Imaging With MRI an aggressive lipid-lowering regimen, compared with untreated
With a combination of multicontrast MR imaging sequences, controls (40). Yet, there are no reports so far on longitudinal
differentiation of fibrocellular, lipid-rich, and calcified regions human in-vivo studies on the alteration of coronary atheroscle-
of the atherosclerotic coronary plaque is feasible, as shown in rotic plaques over time.
an ex-vivo study on human coronary arteries in correlation to
histopathology (30) (Figs. 3 and 4). Calcifications are well POSSIBLE FUTURE IMPROVEMENTS
defined by signal loss in all sequences (T1w, T2w, PDw), FUTURE POSSIBILITIES OF CT PLAQUE IMAGING
because of the low mobile proton density within the calcified Next-generation MDCT scanners will most probably allow
area. Fibrocellular regions are well distinguished by the less for even faster gantry rotation and simultaneous acquisition of
dense body of the plaque-containing regions of extracellular >16 slices. The breath-hold time may decrease to <10 s, thus
lipid deposition on T2w imaging. Fatty lesions are identified as reducing the necessary contrast medium (e.g., 60 mL) for suf-
areas of low signal intensity in T2w and intermediate signal ficient enhancement of the coronary arteries. The temporal and
intensities in T1w sequences (Table 1). spatial resolution may most likely further decrease to ideally
In vivo studies of coronary artery plaques are obviously 100 ms and 0.6-mm slice thickness for true isotropic voxel
more challenging. Preliminary studies in a pig model showed sizes. These enhancements may help in the detection, differen-
that the difficulties of coronary wall imaging are the result of tiation, and reliable quantification of calcified and noncalcified
a combination of cardiac and respiratory motion artifacts, non- coronary artery plaques. Reduction of spatial resolution and
linear course, small size, and location (35). Fayad et al. extended new image reconstruction algorithms should further reduce
the black-blood MR methods used in the human carotid artery beam hardening artifacts and partial-volume effects caused by
and aorta to the imaging of the coronary arterial lumen and wall calcifications, improving the assessment of complex mixed
396 NIKOLAOU, BECKER, AND FAYAD

Fig. 8. Upper row: X-ray angiogram from 78-yr-old female patient with mild disease on X-ray angiography in proximal left anterior descending
(LAD) artery (arrow, A). Black-blood-magnetic resonance (BB-MR) cross-sectional lumen image reveals circular lumen (B); wall shows
uniformly thickened LAD wall (B) with concentric plaque (C). Lower row: X-ray angiogram from 76-yr-old male patient shows high-grade
stenosis in proximal LAD (arrows, A). In vivo cross-sectional BB-MR images of LAD lumen (B) shows obstructed lumen (elliptical lumen
shape); wall image (C) shows large eccentric plaque with heterogeneous signal intensity. (Modified from ref. 12.)

plaques. Further optimization of multisegmental reconstruc- activity (46,47). One promising example is fibrin-specific con-
tion algorithms (41,42) yields potential to investigate patients trast agents, which allow for reliable detection of thrombotic
with higher heart rates with a constant image quality. material and for longitudinal control of thrombus progression
FUTURE POSSIBILITIES OF MR PLAQUE IMAGING or regression (48).
Thinner slices, such as those obtained with 3D MR acquisi-
tion techniques, could further improve artery wall imaging (36). SYNOPSIS
Additional MR techniques, such as water diffusion weighting MRI VS MDCT
(43), magnetization transfer weighting (20), steady-state free Both high-resolution MRI and MDCT have specific advan-
precession (SSFP) sequences (44), contrast enhancement (45), tages in the detection and differentiation of atherosclerotic
and molecular imaging (46,47) may provide complementary vessel wall changes. Using dedicated MRI techniques, a higher
structural information and allow more detailed plaque charac- in-plane resolution can be achieved, and in combination with
terization. New and improved blood suppression methods (36) multi-echo sequences, a more detailed analysis of different
are necessary for accurate plaque imaging, especially in the plaque types and plaque components is feasible. MRI does not
carotid artery bifurcation. suffer from beam-hardening artifacts and partial-volume effects
Contrast-enhanced MR angiography with the use of caused by calcifications, allowing superior differentiation of
gadolinium-based contrast agents may provide additional mixed plaques with calcified and noncalcified components.
information for plaque characterization by identifying neova- The major disadvantages of MRI are its limitations concerning
scularization in the atherosclerotic plaque, and potentially in vivo application, because the technique suffers from patient
improve the differentiation between necrotic core and fibrous movement and breathing artifacts as a result of the compara-
tissue (45). Furthermore, other nonspecific and specific contrast tively long acquisition times. This is the main reason why, in an
agents may facilitate accurate plaque constituent characteriza- in vivo setting, a detailed analysis of coronary artery plaques is
tion and the identification of specific molecular and biological hardly feasible yet, and only limited regions of the coronary
 CHAPTER 37 / MDCT VS MRI FOR CORONARY PLAQUE CHARACTERIZATION 397

artery tree can be depicted (12). Also, patients with cardiac technical requirements for optimal imaging, develop accurate
pacemakers or other specific metal parts in their body can not image analysis methods, outline criteria for interpretation,
undergo MRI. delineate the clinical indications for which CT and MR imag-
Currently, MDCT is easier and faster to perform in an in ing together should be used as an adjunct to conventional
vivo setting, and less sensitive to movement artifacts caused by imaging, and address the issue of cost-effectiveness. Finally,
patient movement or breathing artifacts, despite the relatively imaging may address the high-risk plaque, as described before,
long acquisition window of 210 to 250 ms per heartbeat. Mod- but it does not take into account the blood hypercoagulable
ern MDCT scanners with 16 detector rows allow for in-vivo state or markers of inflammation. Therefore, one of the ulti-
investigation of the entire coronary artery tree with 0.75-mm mate goals for clinicians is the identification of the high-risk
slice thickness within one short breath-hold, achieving a nearly patient through a combination of strategies such as assessment
isotropic submillimeter voxel size, with typically better out- of conventional risk factors, blood markers, and imaging.
of-plane resolution than MRI. However, MDCT applies a
considerable amount of ionizing radiation to the patient, and REFERENCES
iodine-containing contrast agents are used. 1. Fayad ZA, Fuster V, Nikolaou K, Becker C. Computed tomography
Another promising technique for noninvasive cardiac imag- and magnetic resonance imaging for noninvasive coronary angiog-
ing and possibly plaque imaging is EBCT. EBCT has mainly raphy and plaque imaging: current and potential future concepts.
Circulation 2002;106(15):20262034.
been used for accurate quantitation of coronary artery calcium,
2. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons
scanning the entire heart in a single breath-hold from rapid from sudden coronary death: a comprehensive morphological clas-
(100 ms) tomographic scans done in synchrony with the heart sification scheme for atherosclerotic lesions. Arterioscler Thromb
cycle (49). So far, no reports on the assessment of noncalcified Vasc Biol 2000;20(5):12621275.
plaque are available using this modality (50). 3. Fuster V, Badimon L, Badimon JJ, Chesebro JH. The pathogenesis
of coronary artery disease and the acute coronary syndromes (2). N
CT AND MR IMAGING COMBINATION
Engl J Med 1992;326(5):310318.
Cross-sectional modalitiesMRI and CToffer more 4. Fuster V, Badimon L, Badimon JJ, Chesebro JH. The pathogenesis
information than just displays of the patients coronary vessel of coronary artery disease and the acute coronary syndromes (1). N
lumen. The advantage of MDCT is the potentially complete Engl J Med 1992;326(4):242250.
assessment of the entire coronary artery tree within a very short 5. Pasterkamp G, Falk E, Woutman H, Borst C. Techniques character-
izing the coronary atherosclerotic plaque: influence on clinical
scan time, and MRI offers excellent soft-tissue contrast. How-
decision making? J Am Coll Cardiol 2000;36(1):1321.
ever, because of the limited scan range and long examination 6. Fayad ZA, Fuster V. Clinical imaging of the high-risk or vulnerable
time for MRI, MDCT may be used first to localize suspicious atherosclerotic plaque. Circ Res 2001;89(4):305316.
CAD lesions in the coronary arteries. With the knowledge of 7. Nieman K, Cademartiri F, Lemos P, Raaijmakers R, Pattynama P, de
the problematic site in the coronary arteries from MDCT, MRI Feyter P. Reliable noninvasive coronary aniography with fast sub-
millimeter multslice spiral computed tomography. Circulation
follow-up may be used for further lesion assessment. Improved
2002;106:20512054.
CT and MR imaging of the coronaries may in the future allow 8. Nikolaou K, Becker CR, Wintersperger BJ, Sagmeister S, Reiser
for better noninvasive evaluation of atherosclerotic plaques. A MF. Assessment of non-calcified vessel-wall changes in the coro-
comprehensive study of atherosclerosis could involve a cal- nary arteries using contrast-enhanced multirow-detector computed
cium screening measurement, contrast-enhanced CT for tomography. Radiology 2002;225:632.
9. Toussaint J-F, LaMuraglia GM, Southern JF, Fuster V, Kantor HL.
noncalcified plaque burden and stenosis assessment, and Magnetic resonance images lipid, fibrous, calcified, hemorrhagic,
multicontrast MR plaque characterization for a detailed analy- and thrombotic components of human atherosclerosis in vivo. Cir-
sis and follow-up of the composition of coronary atheroscle- culation 1996;94(5):932938.
rotic lesions. The utility of such a noninvasive combined CT 10. Yuan C, Beach KW, Smith LH, Jr., Hatsukami TS. Measurement of
and MR approach will have to be tested in future studies. atherosclerotic carotid plaque size in vivo using high resolution
magnetic resonance imaging. Circulation 1998;98(24):26662671.
CONCLUSIONS 11. Fayad ZA, Nahar T, Fallon JT, et al. In vivo magnetic resonance
evaluation of atherosclerotic plaques in the human thoracic aorta: a
CT and MR imaging are emerging as the most promising comparison with transesophageal echocardiography. Circulation
complementary imaging modalities for coronary atheroscle- 2000;101(21):25032509.
rotic disease detection. They identify flow-limiting coronary 12. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human
coronary artery lumen and wall imaging using black-blood magnetic
stenoses and calcified plaques, directly image the atheroscle- resonance imaging. Circulation 2000;102(5):506510.
rotic lesions, measure atherosclerotic burden, and characterize 13. Botnar R, Stuber M, Kissinger K, Kim W, Spuentrup E, Manning W.
the plaque components. Together, they provide unique infor- Noninvasive coronary vessel wall and plaque imaging with mag-
mation that may predict cardiovascular risk, facilitate further netic resonance imaging. Circulation 2000;102:25822587.
study of the mechanisms of atherosclerosis progression and its 14. Wintersperger BJ, Herzog P, Jakobs TF, Reiser MF, Becker CR.
Initial experience with the clinical use of a 16 detector row CT sys-
response to therapy, and allow for assessment of subclinical tem. Crit Rev Comput Tomogr 2002;43:283316.
disease. However, published studies with relatively small num- 15. Flohr T, Kuttner A, Bruder H, et al. Performance evaluation of a
bers of patients were conducted independently with either CT multi-slice CT system with 16-slice detector and increased gantry
or MR. These results should be validated in large-scale clinical rotation speed for isotropic submillimeter imaging of the heart. Herz
2003;28(1):719.
trials before CT and MR are implemented clinically, outside of 16. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection
research settings, especially for atherosclerotic disease screen- and evaluation of atherosclerotic coronary plaques with multislice
ing. This type of clinical investigation is needed to define the computed tomography. J Am Coll Cardiol 2001;37(5):14301435.
398 NIKOLAOU, BECKER, AND FAYAD

17. Toussaint JF, LaMuraglia GM, Southern JF, Fuster V, Kantor HL. 34. Yuan C, Beach KW, Smith LH, Jr., Hatsukami TS. Measurement of
Magnetic resonance images lipid, fibrous, calcified, hemorrhagic, atherosclerotic carotid plaque size in vivo using high resolution
and thrombotic components of human atherosclerosis in vivo. Cir- magnetic resonance imaging. Circulation 1998;98(24):26662671.
culation 1996;94(5):932938. 35. Worthley SG, Helft G, Fuster V, et al. Noninvasive in vivo magnetic
18. Fayad ZA, Nahar T, Fallon JT, et al. In vivo MR evaluation of ath- resonance imaging of experimental coronary artery lesions in a
erosclerotic plaques in the human thoracic aorta: a comparison with porcine model. Circulation 2000;101(25):29562961.
TEE. Circulation 2000;101:25032509. 36. Botnar RM, Kim WY, Bornert P, Stuber M, Spuentrup E, Manning
19. Hatsukami TS, Ross R, Polissar NL, Yuan C. Visualization of fibrous WJ. 3D coronary vessel wall imaging utilizing a local inversion
cap thickness and rupture in human atherosclerotic carotid plaque In technique with spiral image acquisition. Magn Reson Med 2001;
vivo with high-resolution magnetic resonance imaging. Circulation 46(5):848854.
2000;102(9):959964. 37. McConnell MV, Aikawa M, Maier SE, Ganz P, Libby P, Lee RT.
20. Yuan C, Mitsumori LM, Beach KW, Maravilla KR. Carotid athero- MRI of rabbit atherosclerosis in response to dietary cholesterol low-
sclerotic plaque: noninvasive MR characterization and identifica- ering. Arterioscler Thromb Vasc Biol 1999;19(8):19561959.
tion of vulnerable lesions. Radiology 2001;221(2):285299. 38. Helft G, Worthley SG, Fuster V, et al. Progression and regression of
21. Estes JM, Quist WC, Lo Gerfo FW, Costello P. Noninvasive char- atherosclerotic lesions: monitoring with serial noninvasive magnetic
acterization of plaque morphology using helical computed tomogra- resonance imaging. Circulation 2002;105(8):993998.
phy. J Cardiovasc Surg (Torino) 1998;39(5):527534. 39. Corti R, Fayad ZA, Fuster V, et al. Effects of lipid-lowering by
22. Oliver TB, Lammie GA, Wright AR, et al. Atherosclerotic plaque simvastatin on human atherosclerotic lesions: a longitudinal study
at the carotid bifurcation: CT angiographic appearance with histo- by high-resolution, noninvasive magnetic resonance imaging. Cir-
pathologic correlation. AJNR Am J Neuroradiol 1999;20(5): culation 2001;104(3):249252.
897901. 40. Zhao XQ, Yuan C, Hatsukami TS, et al. Effects of prolonged inten-
23. Becker CR, Knez A, Jakobs TF, et al. Detection and quantification sive lipid-lowering therapy on the characteristics of carotid athero-
of coronary artery calcification with electron-beam and conventional sclerotic plaques in vivo by MRI: a case-control study. Arterioscler
CT. Eur Radiol 1999;9(4):620624. Thromb Vasc Biol 2001;21(10):16231629.
24. Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons 41. Halliburton SS, Stillman AE, Flohr T, et al. Do segmented recon-
from sudden coronary death: a comprehensive morphological clas- struction algorithms for cardiac multi-slice computed tomography
sification scheme for atherosclerotic lesions. Arterioscler Thromb improve image quality? Herz 2003;28(1):2031.
Vasc Biol 2000;20(5):12621275. 42. Flohr T, Kuttner A, Bruder H, et al. Performance evaluation of a
25. Virmani R, Burke AP, Farb A. Sudden cardiac death. Cardiovasc multi-slice CT system with 16-slice detector and increased gantry
Pathol 2001;10(5):211218. rotation speed for isotropic submillimeter imaging of the heart. Herz
26. Burke AP, Taylor A, Farb A, Malcom GT, Virmani R. Coronary 2003;28(1):719.
calcification: insights from sudden coronary death victims. Z Kardiol 43. Toussaint JF, Southern JF, Fuster V, Kantor HL. Water diffusion
2000;89(Suppl 2):4953. properties of human atherosclerosis and thrombosis measured by
27. Becker CR, Knez A, Leber A, Treede H, Haberl R, Reiser MF. pulse field gradient nuclear magnetic resonance. Arterioscler
[Angiography with multi-slice spiral CT. Detecting plaque, before Thromb Vasc Biol 1997;17(3):542546.
it causes symptoms]. MMW Fortschr Med 2001;143(16):3032. 44. Coombs BD, Rapp JH, Ursell PC, Reilly LM, Saloner D. Structure
28. Becker CR, Knez A, Ohnesorge B, Schoepf UJ, Reiser MF. Imag- of plaque at carotid bifurcation: high-resolution MRI with histologi-
ing of noncalcified coronary plaques using helical CT with retro- cal correlation. Stroke 2001;32(11):251621.
spective ECG gating. AJR Am J Roentgenol 2000;175(2):423424. 45. Yuan C, Kerwin WS, Ferguson MS, et al. Contrast-enhanced high
29. Stary HC. Natural history and histological classification of athero- resolution MRI for atherosclerotic carotid artery tissue characteriza-
sclerotic lesions: an update. Arterioscler Thromb Vasc Biol 2000; tion. J Magn Reson Imaging 2002;15(1):6267.
20(5):11771178. 46. Ruehm SG, Corot C, Vogt P, Kolb S, Debatin JF. Magnetic reso-
30. Nikolaou K, Becker CR, Muders M. High-resolution magnetic reso- nance imaging of atherosclerotic plaque with ultrasmall
nance and multi-slice CT imaging of coronary artery plaques in superparamagnetic particles of iron oxide in hyperlipidemic rabbits.
human ex vivo coronary arteries. Radiology 2001;221:503. Circulation 2001;103(3):415422.
31. Baumgart D, Schmermund A, Goerge G, et al. Comparison of elec- 47. Flacke S, Fischer S, Scott MJ, et al. Novel MRI contrast agent for
tron beam computed tomography with intracoronary ultrasound and molecular imaging of fibrin: implications for detecting vulnerable
coronary angiography for detection of coronary atherosclerosis. plaques. Circulation 2001;104(11):12801285.
J Am Coll Cardiol 1997;30(1):5764. 48. Yu X, Song SK, Chen J, et al. High-resolution MRI characterization
32. Yuan C, Mitsumori LM, Ferguson MS, et al. In vivo accuracy of of human thrombus using a novel fibrin-targeted paramagnetic
multispectral magnetic resonance imaging for identifying lipid-rich nanoparticle contrast agent. Magn Reson Med 2000;44(6):867872.
necrotic cores and intraplaque hemorrhage in advanced human 49. Rumberger JA. Tomographic (plaque) imaging: state of the art. Am
carotid plaques. Circulation 2001;104(17):20512056. J Cardiol 2001;88(2A):66E69E.
33. Fayad ZA, Fuster V, Fallon JT, et al. Noninvasive in vivo human 50. Budoff MJ, Raggi P. Coronary artery disease progression assessed
coronary artery lumen and wall imaging using black-blood magnetic by electron-beam computed tomography. Am J Cardiol 2001;
resonance imaging. Circulation 2000;102(5):506510. 88(2A):46E50E.
 CHAPTER 38 / MDCT DETECTION OF CORONARY LESIONS 399

38 Multidetector-Row CT
for Detection of Noncalcified
and Calcified Coronary Lesions
Clinical Significance

STEPHEN SCHROEDER, MD, PhD

Immediately after the introduction of multidetector-row Due to improved spatial and temporal resolution and thin over-
computed tomography (MDCT) scanners in 1998, one became lapping slices, noncalcified coronary lesions can be visualized
aware that not only hyperdense but also hypodense areas within on contrast-enhanced axial slices noninvasively, using MDCT
contrast-enhanced axial slices could be seen (see Fig. 1). The technology (2,3,10,11). In a recently published study, we we
question to be answered was whether these areas are artifacts were able to show that noncalcified plaques might also be
or atherosclerotic plaques. present in the absence of coronary calcifications. Thus far,
As reviewed in Chapters 35 and 36, experimental as well as there is growing evidence that contrast-enhanced MDCT scans
clinical studies indicate that these lesions correspond with cal- might be useful in patients with intermediate risk profile to
cified and noncalcified atherosclerotic coronary lesions (1,2). further characterize the individuals risk for ACS. We demon-
These lesions might be characterized using MDCT by deter- strated that noncalcified atherosclerotic coronary lesions occur
mining tissue density within the plaque area (2,3). with a prevalence of 10% in this group of patients without
The detection and quantification of coronary calcifications coronary calcifications (12) (see Fig. 2).
using electron beam computed tomography (EBCT) was intro- The clinical impact of this stepwise approach needs further
duced in the early 1990s (4). There is evidence by histopatho- prospective evaluation to elucidate whether imaging of
logical studies that the calcium score correlates well with the noncalcified plaques can add additional information to conven-
entire atherosclerotic plaque burden, with calcifications tional cardiovascular risk factors for defining the individuals
accounting for approx 20% (5). risk. There are, however, limitations of methodology, especially
The evaluation of coronary calcifications is no longer lim- in differentiating atherosclerotic plaques from intracoronary
ited to EBCT, but might also be performed by new MDCT thrombi due to similar density attenuation.
technology with comparable results and precision (6,7), open- Leber et al. recently demonstrated that patients with ACS
ing this diagnostic field to more investigators. The predictive had a significantly higher number of noncalcified coronary
value of coronary calcifications on future acute coronary syn- plaques than did patients with stable angina pectoris (13). Their
dromes (ACS) is at present unclear and controversially dis- study suggests that the determination of noncalcified plaques
cussed, because results from large scale cross-sectional studies might be useful in characterizing patients at higher risk for
are still missing (8). However, there is evidence that screening ACS. However, the number of patients (n = 24) was low. Ran-
for coronary calcifications in asymptomatic patients with domized, large-scale prospective studies are needed to deter-
intermediate risk profiles might be helpful in characterizing mine whether or not the detection of noncalcified plaques is
patients at higher risk for ACS. Just recently, Rich et al. useful for risk stratification in patients with known or sus-
reported in a metanalysis on a risk ratio of 8.7 in patients with pected CAD.
positive scans (9). A major limitation of MDCT for its use as a screening test
However, even asymptomatic intermediate risk individuals is the high radiation exposure, approx 510 mSv at present
without coronary calcifications show an elevated risk ratio. (14.15), compared with conventional coronary angiography
Thus far, the exclusion of coronary calcifications is not equiva- without the use of intravascular ultrasound (IVUS) (approx 3
lent with the exclusion of atherosclerotic coronary artery dis- msv [16]). However, a combination of IVUS and conven-
ease (CAD) in this group of patients, and further diagnostic tional coronary angiography, which would be needed to
tools for noninvasive risk-stratification appear to be desirable. achieve adequate assessment of vessel lumen and vessel wall
morphology, would likely result in a much higher radiation
From: Contemporary Cardiology: CT of the Heart: exposure.
Principles and Applications
Edited by: U. Joseph Schoepf Humana Press, Inc., Totowa, NJ

399
400 SCHROEDER

REFERENCES
1. Fayad ZA, Fuster V, Nikolaou K, Becker C. Computed tomography
and magnetic resonance imaging for noninvasive coronary angiog-
raphy and plaque imaging: current and potential future concepts.
Circulation 2002;106(15):20262034.
2. Schroeder S, Kopp AF, Baumbach A, et al. Noninvasive detection
and evaluation of atherosclerotic coronary plaques with multislice
computed tomography. J Am Coll Cardiol 2001;37(5):14301435.
3. Schroeder S, Flohr T, Kopp AF, et al. Accuracy of density mea-
surements within plaques located in artificial coronary arteries by
X-ray multislice CT: results of a phantom study. J Comput Assist
Tomogr 2001;25(6):900906.
4. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M
Jr, Detrano R. Quantification of coronary artery calcium using
ultrafast computed tomography. J Am Coll Cardiol 1990;15(4):
827832.
5. Rumberger JA, Simons DB, Fitzpatrick LA, Sheedy PF, Schwartz
RS. Coronary artery calcium area by electron-beam computed to-
mography and coronary atherosclerotic plaque area. A histopatho-
logic correlative study. Circulation 1995;92(8):21572162.
Fig 1. Axial slice image visualizing the contrast-enhanced left ante- 6. Becker CR, Schoepf UJ, Reiser MF. Methods for quantification of
rior descending artery with one hypodense and two hyperdense areas. coronary artery calcifications with electron beam and conventional
CT and pushing the spiral CT envelope: new cardiac applications.
Int J Cardiovasc Imaging 2001;17(3):203211.
7. Becker CR, Kleffel T, Crispin A, et al. Coronary artery calcium
measurement: agreement of multirow detector and electron beam
CT. AJR Am J Roentgenol 2001;176(5):12951298.
8. Rourke RA, Brundage BH, Froelicher VF, et al. American College
of Cardiology/American Heart Association Expert Consensus
Document on electron-beam computed tomography for the diagno-
sis and prognosis of coronary artery disease. J Am Coll Cardiol
2000;36:326340.
9. Rich S, McLaughlin VV. Detection of subclinical cardiovascular
disease: the emerging role of electron beam computed tomography.
Prev Med 2002;34(1):110.
10. Becker CR, Knez A, Ohnesorge B, Schoepf UJ, Reiser MF. Imaging
of noncalcified coronary plaques using helical CT with retrospec-
tive ECG gating. AJR Am J Roentgenol. 2000;175(2):423424.
11. Schroeder S, Kopp AF, Baumbach A, et al. Non-invasive
Fig. 2. Prevalence of noncalcified plaques in a study group consting characterisation of coronary lesion morphology by multi-slice com-
of 68 patients with distinct cardiovascular risk profile but without puted tomography: a promising new technology for risk stratifica-
known coronary artery disease. tion of patients with coronary artery disease. Heart 2001;85(5):
576578.
12. Schroeder S, Kuettner A, Kopp AF, et al. Noninvasive evaluation of
the prevalence of noncalcified atherosclerotic plaques by multi-slice
detector computed tomography: results of a pilot study. Int J Cardiol
2003;92(23):151155.
13. Leber AW, Knez A, White CW, et al. Composition of coronary ath-
Furthermore, improvements of MDCT technology with erosclerotic plaques in patients with acute myocardial infarction and
reduced tube current during systole, reducing radiation expo- stable angina pectoris determined by contrast-enhanced multislice
sure by as much as 50% (17), as well as the development of computed tomography. Am J Cardiol 2003;91(6):714718.
scanners with additional detector slices, allowing for shorter 14. Achenbach S, Ropers D, Regenfus M, et al. Noninvasive coronary
angiography by magnetic resonance imaging, electron-beam com-
examination times and presumably stabilized image quality, puted tomography, and multislice computed tomography. Am J
are underway. Cardiol 2001;88(2A):70E73E.
Also, the need for iodinated contrast media could be reduced 15. Achenbach S, Giesler T, Ropers D, et al. Detection of coronary
to approx 80 mL per scan, which is in the range of conventional artery stenoses by contrast-enhanced, retrospectively electrocardio-
graphically-gated, multislice spiral computed tomography. Circula-
coronary angiography (16).
tion 2001;103(21):25352538.
At present, it is open to speculation whether MDCT might 16. Becker C, Schatzl M, Feist H, et al. [Assessment of the effective dose
become a useful part of a diagnostic algorithm to further char- for routine protocols in conventional CT, electron beam CT and
acterize high risk plaques in a combined approach with other coronary angiography] Rofo Fortschr Geb Rontgenstr Neuen
imaging modalities, such as magnetic resonance, or whether Bildgeb Verfahr 1999;170(1):99104. German.
17. Jakobs TF, Becker CR, Ohnesorge B, et al. Multislice helical CT
screening for noncalcified plaques in intermediate or even high- of the heart with retrospective ECG gating: reduction of radiation
risk patients with or without coronary calcifications might exposure by ECG-controlled tube current modulation. Eur Radiol
become useful for risk prediction. 2002;12(5):10811086.
INDEX 401

Index

A sensitivity, 392
Agatston score, coronary artery calcification, 117, 130132 multislice computed tomography
Angiography, conventional 16-slice MSCT, 383385, 387
electron beam computed tomography comparison, 330 data acquisition, 381, 382
heart valve studies, 155 intravascular ultrasound comparison, 382, 383, 391
imaging requirements, 321 pathologic intimal thickening, 352
limitations, 381 rupture, 353, 356
multislice computed tomography comparison, 323 thin fibrous cap atheroma, 352, 353
Angiosarcoma, see Cardiac masses thrombosis, 353
Aortic valve, see Heart valves total occlusions, 358
Atherosclerotic plaque, see also Coronary artery vulnerable plaque
calcification endothelial cell activation, 365, 367, 368
adaptive intimal thickening, 351 features, 365
angiography limitations, 381 oxidative stress in formation, 365, 367
calcified nodule, 357 prospects for study, 371
classification, 351, 352, 377 statins
computed tomography angiography combination with pleiotropic effects, 369, 370
calcium score for plaque burden and morphology stabilization, 368, 369
determination, 333, 334
detection rationale, 381 B
erosion, 356 Bolus, see Contrast material
fibrous and fibrocalcific plaques, 358 Bypass graft assessment, see Coronary artery bypass
fibrous cap atheroma, 352 grafting
healed ruptures, 357, 358 CABG, see Coronary artery bypass grafting
hemorrhage, 358, 359 CAC, see Coronary artery calcification
intimal xanthomas, 351 Calcification, see Coronary artery calcification
magnetic resonance imaging Cardiac masses
computed tomography comparison and combination, benign cardiac tumors, 173, 174
396, 397 electron beam computed tomography
coronary plaques, 395 comparison with multidetector-row computed
monitoring, 395 tomography, 171
noncoronary plaques, 395 data acquisition, 171, 172
overview, 389 prospects, 181
prospects, 396 epidemiology, 172, 173
technique, 390 malignant tumors
morphology and degree of calcification, 112, 113 primary tumors, 177
multidetector-row computed tomography secondary tumors, 177
coronary plaques, 392, 395 multidetector-row computed tomography
data acquisition, 389, 390 comparison with electron beam computed tomography,
morphology of calcifications, 377379 171
noncalcified lesions, 399, 400 contrast material, 172
noncoronary plaques, 390, 391 data acquisition, 171, 172
prerequisites, 377 prospects, 181
prospects, 379, 380, 395, 396 thrombus imaging, 178, 179, 181
quantitative analysis, 379 Cardio-interpolation (CI), phase weighting, 55

401
402 INDEX

Autobone, 248, 249 navigator virtual endoscopy, 252

CardIQ analysis selection of optimum phase with multiphase review of
curved two-dimensional reformatting, 249, 250 reformatted images, 246
left ventricular wall motion and function analysis, 255 stenosis quantification, 250
navigator virtual endoscopy, 252 imaging requirements, 321
stenosis quantification, 250 magnetic resonance angiography comparison, 339341,
CHD, see Coronary heart disease 343346
Chest film, coronary artery calcification visualization, 113 multidetector-row computed tomography
Chest pain evaluation axial image analysis, 231233
electron beam computed tomography, 9397 bolus timing techniques, 241, 243, 244
multidetector-row computed tomography, 121, 122 image postprocessing, 230, 231
CI, see Cardio-interpolation Kawasaki disease assessment,
Cine ventriculography, left ventricular function assessment, see Kawasaki disease
183 overview, 229, 234
Computed tomography angiography (CTA) patient preparation, 229, 230
bypass graft assessment, see Coronary artery bypass scan acquisition, 230
grafting multislice computed tomography
calcification attenuation, 329 4-slice MSCT performance
calcium score combination conventional angiography comparison, 323
algorithm for complementary use, 335, 337 coronary obstruction detection, 325, 327
angiography value by calcium score image interpretability, 323325
high score, 335 16-slice MSCT performance, 327, 328
intermediate borderline score, 335 advantages, 331
no or low scores, 335 arrhythmia effects, 329
plaque burden and plaque morphology determination, clinical implementation, 331
333, 334 data acquisition, 321, 322
suspected coronary stenosis evaluation, 334, 335 motion artifacts, 328, 329
cardiac applications, overview, 10 overview, 29, 30, 3237, 259, 260, 321, 333
contrast material patient-based assessment, 328
bolus geometry, factors affecting postprocessing and data analysis, 322, 323
bolus chaser, 240 prospects, 331
demographics, 237, 238 respiration and scan time, 329
disease, 238 temporal resolution, 328
injection rate, 238 radiation dosage, 10
injection site, 241 stent and surgical material attenuation, 329
injection volume, 238 Computed tomography dose index (CTDI)
iodine concentration, 238 definition, 63
multiphasic protocols, 240, 241 measurement, 63
bolus tracking, 241 radiation dose by imaging technique, 65
fixed-delay technique, 241, 244 volume CTDI, 63
guidelines for administration, 244 weighted CTDI, 63
overview, 237 Congenital heart disease, see Multislice computed
test bolus technique, 241, 244 tomography
coronary anatomy, 219227 Contrast material
coronary artery anomalies, see Coronary artery administration, 9
anomalies computed tomography angiography
electron beam computed tomography bolus geometry, factors affecting
comparison with conventional angiography, 330 bolus chaser, 240
overview, 329, 330, 333 demographics, 237, 238
prospects, 330 disease, 238
image visualization injection rate, 238
artery identification injection site, 241
coronary vessel tree, 248 injection volume, 238
three-dimensional maximum intensity projection, iodine concentration, 238
248 multiphasic protocols, 240, 241
volume rendering, 247 bolus tracking, 241
bifurcations within different projections, 248 fixed-delay technique, 241, 244
curved two-dimensional reformatting, 249, 250 guidelines for administration, 244
multiplanar volume reformation with Autobone, 248, multidetector-row computed tomography, 172, 241,
249 243, 244
INDEX 403

overview, 237 postoperative applications, 305, 306

test bolus technique, 241, 244 preoperative applications, 303305
Coronary arteries Coronary artery calcification (CAC)
anatomy asymptomatic persons, 73
left coronary artery, 219 chest film visualization, 113
right coronary artery, 221, 222 chest pain differential diagnosis with imaging, 121, 122
angiography, see Angiography, conventional; Computed comparison of computed tomography instruments, 138,
tomography angiography; Magnetic resonance 139
angiography computed tomography angiography combination
anomalies, 225, 227 algorithm for complementary use, 335, 337
bypass graft assessment, see Coronary artery bypass angiography value by calcium score
grafting high score, 335
related veins, 224, 225 intermediate borderline score, 335
Coronary artery anomalies no or low scores, 335
aberrant sinus node artery, 269 plaque burden and plaque morphology determination,
clinical importance, 259, 260 333, 334
complex and combined anomalies, 268, 269 suspected coronary stenosis evaluation, 334, 335
ectopic origin of coronary arteries above aortic cusp, coronary artery disease correlation, 7174, 83, 84, 101,
269, 270 108, 112, 120, 121
etiology, 260 double-helical computed tomography
fistula calcification progression monitoring
anatomy, 271, 273 accuracy, 97
clinical relevance, 273 clinical significance, 97, 98
etiology, 271 transplanted hearts, 98, 99
pathophysiology, 271, 273 treatment monitoring, 98
prevalence, 271 chest pain evaluation, 9397
left anterior descending artery chronic versus acute coronary artery disease imaging,
doubled artery, 265 92
origin defects dilated cardiomyopathy differential diagnosis, 97
left circumflex, 264 prediction of angiographic obstructive disease, 93
pulmonary artery, 265 electron beam computed tomography
right sinus of valsalva, 261 clinical applications, 8688
transseptal course, 265, 268 principles, 83, 115, 129
left circumflex coronary artery, 268 reproducibility, 115
left coronary artery origin defects evolution of computed tomography techniques, 129, 130
pulmonary artery, 263 multidetector-row computed tomography
right sinus of valsalva, 261, 263 calibrated calcium mass, 107
separate ostia, 261 clinical value, 108, 109, 120124
multislice computed tomography, 259, 260, 274276 comparison with other techniques, 104, 105, 114, 115
myocardial bridges electrocardiogram gating, 106
anatomy, 273 noise, 106
clinical relevance, 273, 274 principles, 103, 104
etiology, 273 prospects, 108, 109
prevalence, 273 scanning
prevalence, 260 protocol, 106, 107
right coronary artery origin defects sequential scanning, 118
conus artery, 261 spiral scanning, 118120
distal circumflex, 261 spatial resolution improvement, 106, 115
left coronary artery branches, 261 pathophysiology, 91, 112
left sinus of valsalva, 261 phantoms for testing, 134, 135, 139, 140
posterior sinus of valsalva, 261 plaque morphology and degree of calcification,
pulmonary trunk, 261 112, 113
single coronary artery, 270 predictive value for cardiovascular events, 73, 8486,
Coronary artery bypass grafting (CABG) 101, 123, 124
internal mammary artery grafts, 301, 302 prevalence, 113
minimally invasive bypass surgery planning, progression, 86, 91, 97, 122, 359, 360
see Totally endoscopic coronary artery bypass prospects for testing, 76
grafting rationale for imaging, 79, 81
multidetector-row computed tomography scoring
overview, 302, 303 Agatston score, 117, 130132
404 INDEX

calcium mass, 117, 133, 134 angiography, 19

calcium volume score, 117, 132, 133 coronary calcium scoring
prospects, 139, 140 clinical applications, 8688
quality assurance tools, 134136, 138, 139 principles, 83, 115
screening reproducibility, 115
clinical trials, 124 function studies, 21
guidelines, 7476 overview, 10, 19, 21
identification of at-risk patients, 122, 123 cardiac masses
X-ray attenuation by calcium, 129 comparison with multidetector-row computed
Coronary heart disease (CHD), see also Atherosclerotic tomography, 171
plaque data acquisition, 171, 172
atherosclerotic disease classification, 112 prospects, 181
epidemiology, 111 electrocardiogram triggering, 1719
risk stratification, 111, 112 flow mode, 6, 18
Coronary veins historical perspective, 3, 5, 15
anatomy, 222, 223 instruments and specifications, 5, 6, 15, 18
great cardiac vein, 223 left ventricular function assessment, 183, 184
middle cardiac vein, 223 movie mode, 6
related arteries, 224, 225 principles, 5, 15, 16, 83
CTA, see Computed tomography angiography prospects, 21
CTDI, see Computed tomography dose index radiation dosage, 10, 19
scan modes
D continuous volume scan mode, 6, 7, 18, 21
DHCT, see Double-helical computed tomography prospective triggering, 1618
DLP, see Dose-length product step volume scan mode, 6
Dose-length product (DLP) stent patency assessment
computation, 63 criteria for patency, stenosis, and occlusion, 313
radiation dose by imaging technique, 65 historical perspective, 311, 312
Double-helical computed tomography (DHCT) prospects, 316, 317
calcification progression monitoring scan protocols, 312, 313
accuracy, 97 temporal resolution, 8, 9, 18
clinical significance, 97, 98 tuning, 16
transplanted hearts, 98, 99
treatment monitoring, 98 F
chest pain evaluation, 9397
chronic versus acute coronary artery disease imaging, Fast computed tomography, microvascular function
92 assessment
dilated cardiomyopathy differential diagnosis, 97 blood volume-to-flow relationship, 197
prediction of angiographic obstructive disease, 93 coronary microembolization, 200, 201
functional heterogeneity evaluation, 201
growth studies, 197
E
hypercholesterolemia studies, 198, 199
EBCT, see Electron beam computed tomography parameters in assessment, 195197
Echocardiography prospects, 201, 202
congenital heart disease, 161 subclinical coronary artery stenosis, 199, 200
heart valves, 155 Fibroma, see Cardiac masses
Kawasaki disease, 279, 280
left ventricular function assessment, 183 H
pericardium, 145
Electrocardiogram-correlated reconstruction, see also Image Heart valves
reconstruction angiography studies, 155
definition, 55 echocardiography, 155
Electrocardiogram gating, see Multidetector-row computed multidetector-row computed tomography
tomography; Multislice computed tomography calcification of valves, 155158
Electrocardiogram triggering, see Electron beam computed limitations, 158
tomography; Multislice computed tomography morphology imaging, 156
Electron beam computed tomography (EBCT) prospects, 158, 159
angiography, see Computed tomography angiography scanning, 156
calibration, 16 Helical/spiral computed tomography, see also Double-
cardiac applications helical computed tomography
INDEX 405

advantages, 7 single photon emission computed tomography, 183

cardiac applications, overview, 10 Lipoma, see Cardiac masses
historical perspective, 5
prospective gated scanning, 7, 8 M
radiation dosage, 10
Magnetic resonance angiography (MRA)
retrospective gated scanning, 8
computed tomography angiography comparison, 339341,
temporal resolution, 8, 9
343346
Hounsfield, Godfrey, 3, 4 performance, 330, 339341
Image reconstruction Magnetic resonance imaging (MRI)
kymogram detection for phase-correlated reconstruction atherosclerotic plaque imaging
performance computed tomography comparison and combination,
cardiac imaging, 56, 59 396, 397
thoracic imaging, 59, 61 coronary plaques, 395
principles, 55, 56 monitoring, 395
multislice computed tomography noncoronary plaques, 395
electrocardiogram-gated scans, 24, 2628, 48 overview, 389
electrocardiogram-triggered scans, 4648 prospects, 396
multisegment reconstruction, 4951 technique, 390
single-segment reconstruction, 4849, 51 cardiac masses, 181
Intravascular ultrasound (IVUS) congenital heart disease, 161
comparison with computed tomography for coronary Kawasaki disease, 280
plaque detection, 382, 383, 391 left ventricular function assessment, 183, 184, 192194
indications, 382 myocardial perfusion, 207
technique, 382 pericardium, 145
IVUS, see Intravascular ultrasound MDCT, see Multidetector-row computed tomography
Microvascular function
K fast computed tomography
blood volume-to-flow relationship, 197
Kawasaki disease coronary microembolization, 200, 201
angiography, 280 functional heterogeneity evaluation, 201
clinical features, 279 growth studies, 197
diagnostic criteria, 279 hypercholesterolemia studies, 198, 199
echocardiography, 279, 280 parameters in assessment, 195197
magnetic resonance imaging, 280 prospects, 201, 202
management, 279, 284 subclinical coronary artery stenosis, 199, 200
multidetector-row computed tomography rationale for imaging, 195
case studies, 281284 Mitral valve, see Heart valves
detectability, 281 MRA, see Magnetic resonance angiography
Kymogram detection MSCT, see Multislice computed tomography
performance Multidetector-row computed tomography (MDCT)
cardiac imaging, 56, 59 angiography, see Computed tomography angiography
thoracic imaging, 59, 61 atherosclerotic plaque imaging
principles, 55, 56 coronary plaques, 392, 395
Left ventricular function data acquisition, 389, 390
CardIQ analysis, 255 magnetic resonance imaging comparison and
cine ventriculography, 183 combination, 396, 397
echocardiography, 183 morphology of calcifications, 377379
electron beam computed tomography, 183, 184 noncalcified lesions, 399, 400
magnetic resonance imaging, 183, 184, 192194 noncoronary plaques, 390, 391
multidetector-row computed tomography prerequisites, 377
accuracy and reproducibility, 188, 191 prospects, 379, 380, 395, 396
data acquisition, 187, 188 quantitative analysis, 379
data analysis, 188 sensitivity, 392
image reconstruction, 188 cardiac masses
magnetic resonance imaging comparison, 192194 comparison with electron beam computed tomography,
principles, 184 171
prospects, 192 contrast material, 172
volumetric data, 184, 185, 187 data acquisition, 171, 172
parameters in assessment, 183 prospects, 181
406 INDEX

coronary artery bypass grafting significance, 213, 214

overview, 302, 303 two-phase contrast-enhanced computed tomography
postoperative applications, 305, 306 enhancement patterns, 211214
preoperative applications, 303305 technique, 208, 211
coronary artery calcification imaging prospective electrocardiogram gating, 103, 104, 106
calibrated calcium mass, 107 radiation exposure, 120
clinical value, 108, 109, 120124 stent patency assessment
comparison with other techniques, 104, 105, 114, 115 4-row MDCT, 313
electrocardiogram gating, 106 16-row MDCT, 313
noise, 106 criteria for patency, stenosis, and occlusion,
principles, 103, 104 313, 316
prospects, 108, 109 overview, 313
scanning prospects, 317, 318
protocol, 106, 107 Multislice computed tomography (MSCT)
sequential scanning, 118 angiography, see Computed tomography angiography;
spiral scanning, 118120 Coronary artery anomalies
scoring atherosclerotic plaque imaging
Agatston score, 117, 130132 16-slice MSCT, 383385, 387
calcium mass, 117, 133, 134 data acquisition, 381, 382
calcium volume score, 117, 132, 133 intravascular ultrasound comparison, 382, 383
prospects, 139, 140 congenital heart disease
quality assurance tools, 134136, 138, 139 anatomical assessment
spatial resolution improvement, 106, 115 anomalous venous return, 167
evolution of technique, 101, 103 aorta and collaterals, 165, 166
heart valve imaging coronary arteries, 164, 165
calcification of valves, 155158 pulmonary arteries, 164
limitations, 158 upper airways evaluation, 166
morphology imaging, 156 children and adults, 162
prospects, 158, 159 injection protocol, 163
scanning, 156 neonates and infants, 162
instrumentation, 103, 104 overview, 161
Kawasaki disease assessment, see Kawasaki disease postoperative evaluation, 167
left ventricular function assessment prospects, 169
accuracy and reproducibility, 188, 191 pulmonary atresia with ventricular septal defect, 167,
data acquisition, 187, 188 169
data analysis, 188 radiation dosage in children, 162
image reconsttruction, 188 scanning protocol, 161, 162
magnetic resonance imaging comparison, 192194 tetralogy of Fallot, 169
principles, 184 coronary calcification quantification, 3032
prospects, 192 electrocardiogram gating, 24, 2628, 46
volumetric data, 184, 185, 187 electrocardiogram triggering, 23, 24, 45, 46
minimally invasive bypass surgery planning, see Totally functional evaluation, 3739
endoscopic coronary artery bypass grafting image acquisition, 28, 29
myocardial perfusion and viability image reconstruction
early defect electrocardiogram-gated scans, 24, 2628, 48
definition, 207 electrocardiogram-triggered scans, 4648
significance, 212, 213 multisegment reconstruction, 4951
single photon emission computed tomography com- single-segment reconstruction, 4849, 51
parison, 208 phantom studies of performance, 28
wall motion relationship in chronic phase after prospects, 3942
percutaneous intervention, 207 radiation dosage, 27, 29, 50
wall thickness relationship in chronic phase after resolution, 23, 24, 26, 45
percutaneous intervention, 207 scanner numbers, 24, 45
late enhancement Myocardial bridges, see Coronary artery anomalies
definition, 207 Myocardial perfusion
significance, 213, 214 magnetic resonance imaging, 207
overview, 207 multidetector-row computed tomography
prospects, 214 early defect
residual defect definition, 207
definition, 207 significance, 212, 213
INDEX 407

single photon emission computed tomography multislice computed tomography, 27, 29, 50
comparison, 208 Reperfusion, see Myocardial perfusion
wall motion relationship in chronic phase after Rhabdomyoma, see Cardiac masses
percutaneous intervention, 207 Rotating detector array, historical perspective, 3
wall thickness relationship in chronic phase after
percutaneous intervention, 207 S
late enhancement Scan time
definition, 207 limitations, 10
significance, 213, 214 prospects, 11, 12
overview, 207 Single photon emission computed tomography (SPECT)
prospects, 214 left ventricular function assessment, 183
residual defect myocardial perfusion, 208
definition, 207 SPECT, see Single photon emission computed tomography
significance, 213, 214 Spiral computed tomography, see also Double-helical
two-phase contrast-enhanced computed tomography computed tomography and Helical/spiral computed
enhancement patterns, 211214 tomography
technique, 208, 211 Statins
Myocardium wall, morphologic imaging with two- plaque stabilization, 368, 369
dimensional reformatting, 252, 255 pleiotropic effects, 369, 370
Myxoma, see Cardiac masses Stent patency
electron beam computed tomography
N criteria for patency, stenosis, and occlusion, 313
Navigator virtual endoscopy, 252 historical perspective, 311, 312
prospects, 316, 317
O scan protocols, 312, 313
Oxidative stress, vulnerable plaque formation role, 365, 367 multidetector-row computed tomography
4-row MDCT, 313
P 16-row MDCT, 313
criteria for patency, stenosis, and occlusion, 313, 316
Papillary fibroelastoma, see Cardiac masses
overview, 313
Perfusion, see Myocardial perfusion
prospects, 317, 318
Pericardium
restenosis rates, 311
anatomy and function, 145147
computed tomography
T
constriction, 149
developmental anomalies TECABG, see Totally endoscopic coronary artery bypass
congenital absence, 147 grafting
cysts and diverticula, 147, 149 Tetralogy of Fallot, multislice computed tomography, 169
effusion, 149 Thrombi, see Cardiac masses
prospects, 151 Totally endoscopic coronary artery bypass grafting
thickening, 149 (TECABG)
tumors, 149, 150 multidetector-row computed tomography
echocardiography, 145 clinical studies
magnetic resonance imaging, 145 4-row MDCT, 291, 292, 294
oblique sinus, 147 16-row MDCT, 294, 298
recesses, 146 image reconstruction and evaluation, 291
transverse sinus, 146, 147 prospects, 298, 299
Plaque, see Atherosclerotic plaque rationale, 287
Pulmonary atresia with ventricular septal defect, multislice scanning parameters, 290, 291
computed tomography, 167, 169 operative technique, 288290
surgical techniques, 287
R Tumors, see Cardiac masses; Pericardium
Radiation dosage, see also Computed tomography dose
index; Dose-length product
U
automatic exposure control, 65, 66 Ultrasound, see Echocardiography; Intravascular
effective dose ultrasound
calculation, 66
imaging technique differences, 9, 10, 65, 66 V
electron beam tomography, 19 Valves, see Heart valves
multidetector-row computed tomography, 120, 162 Views, cardiac imaging, 9