J Hepatobiliary Pancreat Sci (2014) 21:281–287

DOI: 10.1002/jhbp.25

ORIGINAL ARTICLE

Identification of inferior pancreaticoduodenal artery during

pancreaticoduodenectomy using augmented reality-based
navigation system
Shinji Onda · Tomoyoshi Okamoto · Masaru Kanehira ·
Fumitake Suzuki · Ryusuke Ito · Shuichi Fujioka ·
Naoki Suzuki · Asaki Hattori · Katsuhiko Yanaga

Published online: 22 August 2013

© 2013 Japanese Society of Hepato-Biliary-Pancreatic Surgery

Abstract and intraoperative blood loss in group A was 415 min and
Background In pancreaticoduodenectomy (PD), early 901 ml, respectively. There was no significant difference in
ligation of the inferior pancreaticoduodenal artery (IPDA) operating time and intraoperative blood loss among the
before efferent veins has been advocated to decrease blood groups.
loss by congestion of the pancreatic head to be resected. In Conclusions The AR-based NS provided precise anatomi-
this study, we herein report the utility of early identification cal information, which allowed the surgeons to rapidly iden-
of the IPDA using an augmented reality (AR)-based navi- tify and perform early ligation of IPDA in PD.
gation system (NS).
Methods Seven nonconsecutive patients underwent PD Keywords Image-guided surgery · Superior mesenteric
using AR-based NS. After paired-point matching registra- artery · Surgical navigation system
tion, the reconstructed image obtained by preoperative com-
puted tomography (CT) was fused with a real-time operative Introduction
field image and displayed on 3D monitors. The vascular
reconstructed images, including the superior mesenteric Pancreaticoduodenectomy (PD) is one of the most complex
artery, jejunal artery, and IPDA were visualized to facilitate surgeries in gastrointestinal surgery. Although the current
image-guided surgical procedures. We compared operating mortality rate after PD is less than 5% in high-volume insti-
time and intraoperative blood loss of six patients who suc- tutions, the morbidity rate still remains relatively high.
cessfully underwent identification of IPDA using AR-based Large intraoperative blood loss and the need for blood trans-
NS (group A) with nine patients who underwent early liga- fusion are known risk factors for postoperative complica-
tion of IPDA without using AR (group B) and 18 patients tions after PD [1, 2] and a prognostic factor in patients with
who underwent a conventional PD (group C). resected pancreatic cancer [3, 4].
Results The IPDA or the jejunal artery was rapidly iden- In conventional PD, bleeding mainly arises from conges-
tified and ligated in six patients. The mean operating time tion of the pancreatic head because the efferent veins are
usually divided before most afferent arteries such as the
S. Onda (*) · K. Yanaga
inferior pancreaticoduodenal artery (IPDA). Recently, some
Division of Digestive Surgery, The Jikei University Graduate School institutions have reported a reduction in intraoperative
of Medicine, 3-25-8 Nishi-Shimbashi, Minato-ku, Tokyo 105-8461, bleeding during PD following ligation of IPDA before effer-
Japan ent veins [5–7]. However, the origin of IPDA frequently
e-mail: s-onda@jikei.ac.jp
varies, which makes it difficult to identify in some patients.
T. Okamoto · M. Kanehira · F. Suzuki · R. Ito · S. Fujioka Therefore, reconstructed images of blood vessels obtained
Department of Surgery, The Jikei University Daisan Hospital, using preoperative multi-detector computed tomography
Tokyo, Japan (MDCT) used in some institutions are extremely important.
N. Suzuki · A. Hattori
In our institution, an augmented reality (AR)-based naviga-
Institute for High Dimensional Medical Imaging, The Jikei tion system (NS), for which we have previously reported
University School of Medicine, Tokyo, Japan the feasibility and safety [8, 9], have been performed for
282 J Hepatobiliary Pancreat Sci (2014) 21:281–287

Table 1 Patients’ characteristics and results of the navigation surgery

Patient Age (years), Preoperative diagnosis Operative Operating Blood FRE Duration of
no. sex procedure time (min) loss (ml) (mm) identification
of IPDA (min)

1 61, F Pancreatic carcinoma SSPPD 480a 810a 6.82a Unsuccessful

2 79, F Bile duct carcinoma SSPPD 390 2,002 8.23 13
3 75, F Bile duct carcinoma SSPPD 407 1,088 4.54 10
4 81, M Bile duct carcinoma SSPPD 393 275 5.33 7
5 78, M Bile duct carcinoma SSPPD 470 860 5.10 9
6 75, M Bile duct carcinoma SSPPD 451 790 9.76 25
7 75, M Ampulla of Vater carcinoma SSPPD 379 393 4.01 10
Mean 415 901 6.20 12
FRE fiducial registration error, SSPPD subtotal stomach-preserving pancreaticoduodenectomy
a
The data are excluded from the analysis

hepatobiliary and pancreatic surgery. The purpose of this Registration and image overlaying
study was to evaluate the feasibility and utility of AR-based
NS for early identification of IPDA in PD. In the operating room, initial paired-point matching regis-
tration was performed by measuring the position of
anatomical landmark points of visceral organs using an
Materials and methods Optotrak pen-probe with 24 infrared emitting diode (IRED)
markers (Northern Digital, Waterloo, ON, Canada), tracked
Patients by an optical location sensor (Optotrak; Northern Digital).
The paired-point registration method computed a transfor-
From January 2012 to February 2013, seven nonconsecutive mation matrix using at least three paired points between the
patients at the Jikei University Daisan Hospital underwent patient and reconstructed images. Registration accuracy was
subtotal stomach-preserving pancreaticoduodenectomy calculated with fiducial registration error (FRE), which was
(SSPPD) using AR-based NS. The patients included in this the root mean square between corresponding fiducial points.
study were aged 61–81 years (mean, 75 years), and three of In paired-point matching registration, the fiducial points
them were female (Table 1). The protocol was approved by indicated with the pointing device were sometimes inaccu-
the institutional committee on human research. Informed rate because of organ shifting and deformities; therefore, the
consent was obtained from each patient who underwent blood vessels were used as the registration fiducial points
surgery using this system. because their shifting during surgery was considered to be
minimal. Fiducial registration points were determined as
accurately as possible and inaccurate data were excluded. In
Preoperative preparation the conventional registration method, if FRE is large, the
registration point selection process is performed according
All the patients had preoperative enhanced MDCT with to the subjective judgment of the surgeon and repeated until
1-mm thick slices and slice intervals of 0.70–0.85 mm. The the FRE is acceptable. Therefore, the registration procedure
image data were obtained using the Digital Imaging and can be time-consuming, and its accuracy can decrease. Fur-
Communications in Medicine (DICOM) standard. The 3D thermore, we developed an effective registration tool that
reconstructed images were created by segmentation using can quickly compute FRE of all fiducial point combinations
an imaging software suite, Analyze (Mayo Foundation, and display FRE in the order of lowest to highest. A screen
Rochester, MN, USA). The vascular structure and pancreas shot of the computed FRE for all the combinations of points
were manually or semi-automatically delineated and given is shown in Figure 1. However, because FRE relies only on
colors and computationally reconstructed into 3D images. the geometric alignment of the fiducial points registered, the
The segmented volume images were subsequently trans- accuracy should be checked by viewing the superimposed
formed into surface images to reduce the size of the dataset images. Consequently, the high rank FRE navigation images
and skeletonization. All data were transmitted to computers are displayed on a quad split monitor display (Fig. 2), and
in the specialized operating room where AR-based NS was the best image can be selected. This registration tool was
installed. used on the fourth and subsequent patients. The acquired
J Hepatobiliary Pancreat Sci (2014) 21:281–287 283

Outline of operative procedures and navigation surgery

The abdomen was explored through a midline incision. The

Kocher maneuver was performed to fully mobilize the head
of the pancreas and duodenum and continued to the left side
of the aorta, exposing the ventral surface of the left renal
vein. The transverse colon was subsequently lifted up, and
the middle colic artery (MCA) was identified. The MCA
was proximally dissected until the superior mesenteric
artery (SMA) was identified and taped. The superior
mesenteric vein (SMV) was exposed just behind the pan-
creas, and the gastrocolic trunk was also identified. Subse-
quently, paired-point registration was performed using the
anatomical landmark points of visceral organs as the fiducial
points. The points were selected from vascular landmarks
such as the origin of MCA, left renal vein–inferior vena cava
confluence, and the gastrocolic trunk–SMV confluence. The
real-time operative images captured by the scope were dis-
played on the monitor by superimposing them onto the
preoperative reconstructed images. The vascular recon-
structed images, including the SMA, jejunal artery, and
IPDA were visualized. The origin of IPDA was ligated, but
in cases where the IPDA originated from the first jejunal
artery, the origin of the first jejunal artery was ligated. Dis-
section of the nerve plexus around SMA is performed only
in patients with pancreatic carcinoma. In other diseases, it is
performed longitudinally for about 10 mm. Subsequent sur-
gical procedures were performed without NS, and under-
Fig. 1 A screen shot of the computed fiducial registration error (FRE) taken in a standard fashion. All surgeries were performed
for all the combinations of registration points is shown. FRE of all without additional invasive procedures for the navigation
fiducial point combinations were displayed in the order of lowest to
highest. FRE was represented by the root-mean-square error. rms err: surgery.
root-mean-square error
Comparison of operating time and intraoperative
blood loss
FRE was compared between the conventional and novel
methods. We retrospectively divided 33 consecutive cancer patients
The operative field was captured by a sterilizable rigid who underwent PD into three study groups: Group A con-
stereo-scope with a short lens tube (10 mm diameter, sisted of six patients who successfully underwent PD with
18 cm length, 0 degrees) designed for specifically open early ligation of IPDA using AR-based NS. One patient with
abdominal surgery (Shinko Optical, Tokyo, Japan), as pancreatic carcinoma in whom IPDA was unsuccessfully
previously reported [9]. This scope was mounted on identified using AR-based NS was excluded. Group B con-
two charge-coupled device (CCD) cameras, whose images sisted of nine consecutive cancer patients who underwent
were displayed on a stereo monitor display (LG, Seoul, PD with early ligation of the IPDA without using AR-based
Korea), which uses film-type patterned retarder technol- NS. Group C consisted of 18 cancer patients who underwent
ogy. Using circular polarized 3D glasses, surgeons can conventional PD between April 2011 and April 2013. All the
view the stereo images. This compact, light-weight scope navigation surgeries (Group A) were performed by three
can be managed in the operative field with one hand, or trainee surgeons. Seven of nine in Group B and 13 of 18 in
fixed to the operating table with a scope holder. The scope Group C were performed by five trainee surgeons (three of
captures the real-time operative field, and the images them are the same as above), and the others were performed
obtained are superimposed on the reconstructed images by two experienced surgeons. Among the three groups, there
displayed on the monitor display. Because the scope is were no patients in Group A with pancreatic carcinoma,
attached to the optical marker, real-time tracking is pos- while there were five in Group B and eight in Group C
sible allowing surgical navigation. (Table 2).
284 J Hepatobiliary Pancreat Sci (2014) 21:281–287

Fig. 2 The high rank fiducial

registration error (FRE) naviga-
tion images are displayed on a
quad split monitor display.
Overlaid images of the pan-
creas, arteries, veins, and tumor
were visualized

Table 2 Demographics and outcomes

Group A (n = 6) Group B (n = 9) Group C (n = 18) P-value

Group A vs. C Group A vs. B

Median age (years) 77 (75–81)* 72 (56–82)* 67 (49–80)* <0.01 0.163

Gender (M/F) 4/2 4/5 13/5 1.000 0.608
Operating surgeon (trainee/experienced) 6/0 7/2 13/5 0.280 0.486
Body mass index (kg/m2) 22.37 ± 1.73 22.36 ± 3.67 22.49 ± 2.40 0.893 0.994
Disease
Pancreatic carcinoma 0 5 8 0.066 0.044
Bile duct carcinoma 5 3 8 0.166 0.119
Gallbladder carcinoma 0 0 1 1.000 1.000
Ampulla of Vater carcinoma 1 0 1 0.446 0.400
Duodenal carcinoma 0 1 0 1.000 1.000
Combined resection of other organs
Portal vein 0 1 1 1.000 1.000
Colon 0 1 0 1.000 1.000
Intraoperative variable
Operating time (min) 415 ± 37 425 ± 53 402 ± 47 0.555 0.696
Blood loss (ml) 901 ± 618 828 ± 514 922 ± 471 0.934 0.806
* range

Statistical analysis Results

Groups were compared using the Fisher’s exact, and the Segmentation took approximately 3–6 h per patient. The
averages were compared using the Student’s t-test and images of reconstructed blood vessels obtained by preop-
Welch’s t-test. The results were considered statistically sig- erative CT showed three types of IPDA variation. IPDA
nificant when P-values were less than 0.05. originated from the first jejunal artery in patients 1, 2, and 4
J Hepatobiliary Pancreat Sci (2014) 21:281–287 285

Fig. 4 Augmented reality (AR)-based navigation image during

pancreaticoduodenectomy of patient 7; inferior pancreaticoduodenal
Fig. 3 Augmented reality (AR)-based navigation image during artery (IPDA) directly originated from superior mesenteric artery
pancreaticoduodenectomy of patient 4; inferior pancreaticoduodenal (SMA). MCA middle colic artery, RCA right colic artery
artery (IPDA) originated from jejunal artery. MCA middle colic artery,
RCA right colic artery, SMA superior mesenteric artery

and from SMA in patients 3, 6 and 7. In patient 5, the

posterior IPDA originated from the replaced hepatic artery
that arose from SMA and the anterior IPDA originated from
the jejunal artery.
In all seven cases, AR navigation images were success-
fully displayed on the 3D monitor. In our first patient
(patient 1), IPDA was not identified with image guidance
because of bleeding from around SMA due to severe
pancreatitis. However, in the other patients, the jejunal
artery or IPDA was identified (within about 10 mm limited
dissection around SMA) and ligated following image
guidance. The mean duration of identification of IPDA
was 12 min (ranging 7–25 min). The overlaid images on
the display monitors are shown in Figures 3 and 4. The
intraoperative snapshot of the displayed image without infu-
sion image is shown in Figure 5. The registration accuracy
(FRE; mm) of six successful cases was 6.20 mm (Table 1).
In cases 2 and 3, in which the fiducial points were manually Fig. 5 Intraoperative snap shot without infusion image of patient 7;
selected, the mean FRE was 6.39 mm, and in case 4 to 7, in inferior pancreaticoduodenal artery (IPDA), which directly originates
which the fiducial points were selected with the computer- from superior mesenteric artery (SMA), is identified and taped. RCA
right colic artery
assisted tool, the mean FRE was 6.05 mm. There was no
statistically significant difference between the two types of
calculation. The time required for registration was 1–2 min
for each procedure. blood loss. The operating time and intraoperative blood loss
No complications related to NS were encountered during for early ligation of IPDA without using AR (group B) was
or after these procedures. The mean operating time of suc- 425 min and 828 ml, respectively, and those of conventional
cessful cases (group A) was 415 (379–470) min, and the PD (group C) was 402 min and 922 ml, respectively. There
intraoperative estimated blood loss was 901 (275–2,002) ml was no significant difference in operating time and
(Table 1). A learning curve was observed with regard to intraoperative blood loss among the groups (Table 2).
286 J Hepatobiliary Pancreat Sci (2014) 21:281–287

Discussions replaced right hepatic artery from SMA. In these cases, the
navigation system may be useful. Kawai identified IPDA by
With the recent development of medical imaging and com- measuring the distance between the IPDA and SMA roots
puter technology, image-guided surgery has been widely using preoperative MDCT images. In addition, Horiguchi
applied [10]. Previously, we developed an AR-based NS for identified the IPDA by measuring the distance between IPDA
hepatobiliary and pancreatic surgery [8, 9]. Pancreatic and MCA roots [5, 6, 18]. Their methods were applied only
surgery seems suitable for AR-based NS because the pan- when IPDA directly originated from SMA. The location of
creas is a retroperitoneal organ, and intraoperative organ the IPDA root was identified, but neither the angle nor
shifting and deformity are minimal. Moreover, in pancreatic direction of the IPDA was measured. However, using
surgery, precise registration can be performed because of the AR-based NS, the surgeon can intuitively view the location
anatomical proximity of major blood vessels (SMA and of the vessel as well as angle and direction from the preop-
SMV) as fiducial points. erative reconstructed images displayed on the navigation
Recently, PD with SMA-first dissection has been monitor display. In addition, in cases in which IPDA origi-
reported [11–13]. This technique resulted in better outcomes nates from the first jejunal artery, IPDA can also be identified.
than conventional PD in terms of operating time, blood loss, Dissection of the nerve plexus around SMA is a standard
number of lymph nodes retrieved, margin positivity and procedure for pancreatic carcinoma but is not necessary for
operative morbidity. While introducing early identification cholangiocarcinoma, ampullary carcinoma, and duodenal
and ligation of the IPDA in PD in our institute, we focused carcinoma [20]. AR-based NS provides accurate position
on this technique using AR-based NS. In the past, two types information of the root of the target arteries and allows the
of surgical procedures for SMA-first dissection have been surgeons to perform minimal dissection of the target region
described: the posterior and mesenteric approaches [5–7, and prevent unnecessary dissection of the nerve plexuses
11–16]. The posterior approach technique was described in around SMA.
detail by Pessaux et al. [11]. In brief, SMA is dissected first This study showed the feasibility and safety of AR-based
and the posterior pancreatic capsule is dissected early NS for identification of IPDA in PD, but failed to demon-
during the surgery. The posterior part of the pancreas head is strate a statistically significant advantage in terms of oper-
dissected off the vessels first without dividing the pancreatic ating time and intraoperative blood loss as compared
neck. with conventional PD. Of course, operating time and
In contrast, the mesenteric approach, a modified conven- intraoperative blood loss depend on many factors, including
tional technique, was used in this study [5–7, 16]. The body mass index, disease, inflammation, experience of
periarterial connective tissue and nerve plexuses around the surgeon, and so forth, which can introduce bias. Moreover,
SMA were dissected, and IPDA was subsequently detected since AR-based NS is not yet available for clinical routine
and ligated. Finally, the pancreas was transected and con- use, it is difficult to perform a controlled study. However, a
nective tissues including the efferent veins were dissected. learning curve has been observed in terms of blood loss
The aim of this technique is to identify SMA from the during PD even with the limited number of patients evalu-
anterior aspect. In the SMA-first approach, our AR-based ated in this study.
NS seems useful for understanding the anatomy of SMA. In addition, AR-based NS seems to contribute to surgical
Anatomical variations of IPDA were classified into four training and education of inexperienced surgeons. The navi-
types based on origination site and modified from earlier gation images, which can be shared with the surgical team,
reports as follows [17–19]: (1) the IPDA originating from are intuitive and easy to understand.
the first jejunal artery; (2) originating directly from SMA; In the present study, we also developed an effective reg-
(3) the anterior IPDA and posterior IPDA independently istration tool, which can quickly compute the best combi-
originating from the SMA; (4) the posterior IPDA originat- nation of the fiducial registration points. With regard to
ing from the replaced hepatic artery from the SMA and FRE, although statistical significance was not observed, the
anterior IPDA originating from the jejunal artery. efficiency may be greatly improved. In recent years, lapa-
In the present study, three types of IPDA variations were roscopic pancreatic surgery has been rapidly developing.
observed: (1) originating from the first jejunal artery, (2) However, limited views and a lack of tactile sensation make
directly originating from SMA, (3) posterior IPDA originat- it difficult for the surgeons to precisely and efficiently iden-
ing from the replaced hepatic artery from SMA and anterior tify the target vessel. In future research, we plan to extend
IPDA separately originating from the jejunal artery. In our NS to laparoscopic pancreatic surgery, which is more
patients where IPDA directly originated from SMA, identi- important than laparotomy.
fication of IPDA seemed difficult because the IPDA usually In conclusion, AR-based NS can provide precise ana-
originated from the dorsal aspect of SMA. Moreover, it was tomical information, which can allow rapid identification
difficult to identify the target vessel in the presence of the and ligation of IPDA in PD.
J Hepatobiliary Pancreat Sci (2014) 21:281–287 287

Conflict of interest None declared. 9. Onda S, Okamoto T, Kanehira M, Fujioka S, Suzuki N, Hattori A,
et al. Short rigid scope and stereo-scope designed specifically for
open abdominal navigation surgery: clinical application for
Author contribution Study design: Shinji Onda, Tomoyoshi hepatobiliary and pancreatic surgery. J Hepatobiliary Pancreat Sci.
Okamoto, Shuichi Fujioka. Acquisition of data: Shinji Onda, 2013;20:448–53.
Tomoyoshi Okamoto, Masaru Kanehira, Fumitake Suzuki, Ryusuke 10. Shuhaiber JH. Augmented reality in surgery. Arch Surg. 2004;
Ito, Shuichi Fujioka. Analysis and interpretation: Shinji Onda, Asaki 139:170–4.
Hattori. Manuscript drafted by: Shinji Onda. Revision: Tomoyoshi 11. Pessaux P, Varma D, Arnaud JP. Pancreaticoduodenectomy: supe-
Okamoto, Naoki Suzuki, Asaki Hattori, Katsuhiko Yanaga. Computer rior mesenteric artery first approach. J Gastrointest Surg. 2006;
operator: Naoki Suzuki, Asaki Hattori. 10:607–11.
12. Weitz J, Rahbari N, Koch M, Büchler MW. The “artery first”
approach for resection of pancreatic head cancer. J Am Coll Surg.
References 2010;210:e1–4.
13. Shrikhande SV, Barreto SG, Bodhankar YD, Suradkar K, Shetty
1. Braga M, Capretti G, Pecorelli N, Balzano G, Doglioni C, Ariotti G, Hawaldar R, et al. Superior mesenteric artery first combined
R, et al. A prognostic score to predict major complications after with uncinate process approach versus uncinate process first
pancreaticoduodenectomy. Ann Surg. 2011;254:702–7. approach in pancreatoduodenectomy: a comparative study evalu-
2. Gouma DJ, van Geenen RC, van Gulik TM, de Haan RJ, de Wit ating perioperative outcomes. Langenbecks Arch Surg. 2011;396:
LT, Busch OR, et al. Rates of complications and death after 1205–12.
pancreaticoduodenectomy: risk factors and the impact of hospital 14. Dumitrascu T, David L, Popescu I. Posterior versus standard
volume. Ann Surg. 2000;232:786–95. approach in pancreatoduodenectomy: a case-match study.
3. Kazanjian KK, Hines OJ, Duffy JP, Yoon DY, Cortina G, Reber Langenbecks Arch Surg. 2010;395:677–84.
HA. Improved survival following pancreaticoduodenectomy to 15. Varty PP, Yamamoto H, Farges O, Belghiti J, Sauvanet A. Early
treat adenocarcinoma of the pancreas: the influence of operative retropancreatic dissection during pancreaticoduodenectomy. Am J
blood loss. Arch Surg. 2008;143:1166–71. Surg. 2005;189:488–91.
4. Sohn TA, Yeo CJ, Cameron JL, Koniaris L, Kaushal S, Abrams 16. Nakamura M, Nakashima H, Tsutsumi K, Matsumoto H, Muta Y,
RA, et al. Resected adenocarcinoma of the pancreas-616 patients: Ueno D, et al. First jejunal vein oriented mesenteric excision
results, outcomes, and prognostic indicators. J Gastrointest Surg. for pancreatoduodenectomy. J Gastroenterol. 2013;48:989–95.
2000;4:567–79. 17. Bertelli E, Gregorio FD, Bertelli L, Civeli L, Mosca S. The arterial
5. Horiguchi A, Ishihara S, Ito M, Nagata H, Shimizu T, Furusawa blood supply of the pancreas: a review III. The inferior
K, et al. Pancreatoduodenectomy in which dissection of the effer- pancreaticoduodenal artery: an anatomical review and a radiologi-
ent arteries of the head of the pancreas is performed first. J cal study. Surg Radiol Anat. 1996;18:67–74.
Hepatobiliary Pancreat Surg. 2007;14:575–8. 18. Horiguchi A, Ishihara S, Ito M, Asano Y, Yamamoto T, Miyakawa
6. Kawai M, Tani M, Ina S, Hirono S, Nishioka R, Miyazawa M, S. Three-dimensional models of arteries constructed using
et al. CLIP method (preoperative CT image-assessed ligation of multidetector-row CT images to perform pancreatoduodenectomy
inferior pancreaticoduodenal artery) reduces intraoperative bleed- safely following dissection of the inferior pancreaticoduodenal
ing during pancreaticoduodenectomy. World J Surg. 2008;32: artery. J Hepatobiliary Pancreat Sci. 2010;17:523–6.
82–7. 19. Murakami G, Hirata K, Takamuro T, Mukaiya M, Hata F,
7. Ishizaki Y, Sugo H, Yoshimoto J, Imamura H, Kawasaki S. Kitagawa S. Vascular anatomy of the pancreaticoduodenal region:
Pancreatoduodenectomy with or without early ligation of the infe- a review. J Hepatobiliary Pancreat Surg. 1999;1:55–68.
rior pancreatoduodenal artery: comparison of intraoperative blood 20. Farnell MB, Pearson RK, Sarr MG, DiMagno EP, Burgart
loss and short-term outcome. World J Surg. 2010;34:2939–44. LJ, Dahl TR, et al. A prospective randomized trial comparing
8. Okamoto T, Onda S, Matsumoto M, Gocho T, Futagawa Y, standard pancreatoduodenectomy with pancreatoduodenectomy
Fujioka S, et al. Utility of augmented reality system in hepato- with extended lymphadenectomy in resectable pancreatic head
biliary surgery. J Hepatobiliary Pancreat Sci. 2013;20:249–53. adenocarcinoma. Surgery. 2005;138:618–28.