UC San Diego

The increased complexity of percutaneous procedures continues to demand new imaging techniques to assist these interventions and improve clinical planning and outcomes. Recent developments in 4D X-ray computed tomography (4DCT) technology, including dose reduction protocols, iterative reconstruction methods, and wide-range detectors, allow for full 3D volumes to be acquired across the entire cardiac cycle within a single heartbeat with very low radiation dose. In addition, with the use of a contrast bolus, 4DCT provides highly reproducible and high-resolution metrics of left ventricular (LV) endocardial function. This dissertation thoroughly investigates the development and application of novel methods for evaluating both global and regional LV function using 4DCT data. These novel methods present highly translational research which can be implemented immediately on many existing CT scanners as well as on existing datasets, retrospectively. First, we evaluate the accuracy and reproducibility of previously developed and validated metrics, like CT SQUEEZ, and newly proposed algorithms for measuring LV rotation, torsion, longitudinal strain, and circumferential strain through phantom experiments and cross-modality comparison studies. Then, we demonstrate the feasibility of these methods in patients with various cardiovascular disease states, such as mitral regurgitation (MR) and heart failure. In transcatheter mitral valve (TMVI) patients we show that analysis of 4DCT-derived metrics of LV function can provide a more precise understanding of the effect of MR and TMVI on regional LV function and remodeling than current imaging methods. In heart failure patients, we demonstrate that 4DCT-derived metrics of LV shape, global and regional LV function, and dyssynchrony have the potential to predict response to cardiac resynchronization therapy (CRT) through the development of the lead placement score (LPS). Overall, these results demonstrate that 4DCT imaging, combined with the novel methods we have developed, is a remarkably promising candidate for transforming clinical cardiac imaging, especially for patients undergoing transcatheter-based procedures.