Objectives: A correct preoperative definition of the hepatic duct confluence anatomy of right liver living donors is pivotal in determining their candidacy for donation and planning surgery during liver transplant. Here, we evaluated the accuracy of 3-dimensional magnetic resonance cholangiography compared with intraoperative cholangiography in assessing biliary anatomy and aimed to identify imaging characteristics that may help to predict the yield of hepatic duct orifices in the right liver graft.
Materials and Methods: All consecutive living hepatectomy donors for adult liver transplant included in this study (N = 110) were evaluated with preoperative 3-dimensional magnetic resonance cholangiography, which was performed before and after intravenous administration of gadolinium (20-40 cm3). For intraoperative cholangiography, a 4F catheter was advanced through the cystic duct, and contrast matter (5-10 mL) was injected into the biliary tree via the catheter. The number of right hepatic ducts in explanted graft was determined on the back table.
Results: Of 110 donors, 71 had type 1 (normal) biliary anatomy based on both 3-dimensional magnetic resonance and intraoperative cholangiography and 39 had abnormal biliary anatomy, with 2 having type 3 (abnormal) biliary anatomy. Normal biliary anatomy was found in back-table examination, and abnormal biliary anatomy (type 2) was found with intraoperative cholangiography. Sensitivity, specificity, and predictive values of 3-dimensional magnetic resonance cholangiography in revealing the biliary anatomy and anomalies were compared with intraoperative cholangiography findings. Observed final hepatic duct outcomes were also assessed. Use of 3-dimensional magnetic resonance cholangiography accurately predicted the biliary anatomy in 97 of 110 cases. Sensitivity was 80.4%, positive predictive value was 94.4%, specificity was 96.9%, and negative predictive value was 87.3%.
Conclusions: Three-dimensional magnetic resonance cholangiography reliably represented normal biliary anatomy; the presence of anatomic variations decreased its sensitivity, making intraoperative cholangiography or duct probing necessary tools to accurately perform right hepatic duct transection.
Key words : Biliary anatomy, Hepatic duct transection, Liver transplant
Biliary complications are among the most prevalent complications in recipients after liver transplant from living donors.1,2 The number of graft bile ducts after transplant is the most common event associated with biliary complications.3,4 Intra- and extrahepatic bile duct variations are frequently observed. Normal biliary anatomy is found in only 58% of the population5.
Hepatobiliary and liver transplant surgeries involve difficult technical procedures. Complications associated with the biliary tract can arise in many cases, and these complications can be accompanied by unexpected anatomic variations. Prevalence of these complications can range from 10% to 25%, and fatal complications have been observed in about 10% of cases.6 Preoperative and intraoperative imaging procedures for confluence of right and left ducts are greatly important. In addition, the integrity of the left hepatic duct system should be properly protected in the transection plane preferred by the surgeon.
Although anatomic variations do not constitute strict contraindications for donation, their presence could affect donor selection. The presence of multiple ducts could increase postoperative biliary morbidity chances in recipients.7,8 One of the most critical steps in donor operations is division of the right hepatic duct. Biliary complications caused by postoperative anastomotic narrowing and leaks are still major complications.4,8,9 Preoperative imaging of donor biliary anatomy is critical in selection of a suitable donor, deciding the transection plane of the right hepatic duct, and estimating the number of graft bile ducts. Presently, there are various techniques for imaging of the biliary tree. Intravenous cholangiography does not produce enough opacity in intra- and extrahepatic trees and rarely allows detailed imaging of duct bifurcation. Endoscopic retrograde cholangiopancreatography (ERCP) is an invasive method, but it provides more accurate information in imaging of biliary tree. Nevertheless, it is not a technique that can always be used in computed tomography cholangiography.
Despite its high precision, intraoperative cholangiography (IOC) is an invasive procedure. Magnetic resonance cholangiography (MRC) is an efficient method for the preoperative evaluation of potential living liver donors. In addition, it is a robust noninvasive method for imaging of bile ducts.3,10-14 Methods used for this purpose are crucial in performing transplant in only the duct area of the graft, in preventing damage to the donor’s biliary system, and in selecting right hepatic duct transection.9 In addition to preoperative findings obtained through MRC, IOC has also been an important tool in our center.
In this study, our primary aim was to evaluate the usefulness of 3-dimensional (3D)-MRC in determining the donor’s right liver biliary anatomy. In addition, we aimed to reveal whether 3D-MRC could help surgeons to make intraoperative decisions about the hepatic duct transection plane and to estimate the number of hepatic duct orifices, an important step during surgery. Finally, we aimed to contribute to selection of recipient-donor matching and to determine whether 3D-MRC could improve interpretation of findings in donors.
Materials and Methods
The present study included 110 consecutive living hepatectomy donors for adult liver transplant. All donors were evaluated using preoperative 3D-MRC. Axial breath-hold in-phase and out-of-phase images, T2 fat saturation images, and postcontrast equilibrium in-phase fat saturation axial images were used. Specific MRC sequences consisting of breath-hold thick-slab MRC images in multiple oblique coronal planes were taken in addition to thin-slab axial images and 3D coronal images. Magnetic resonance cholangiography was performed before and after intravenous administration of gadolinium (20-40 cm3). Findings of MRC were evaluated by a radiologist and a transplant surgeon and were classified according to Choi and associates.15
Intraoperative cholangiography was performed throughout donor surgery using a 4F catheter advanced through the cystic duct. Contrast matter (5-10 mL) was injected into the biliary tree via the catheter. For IOC, we used posteroanterior views via digital subtraction while donor was holding breath. At the end of the operation, the catheter was pulled back and the cystic duct was closed with 3-0 silk suture. The number of right hepatic ducts in the explanted graft was determined on the back table. In addition, sensitivity, specificity, and predictive values of 3D-MRC in revealing the biliary anatomy and anomalies were compared versus IOC findings. Observed final hepatic duct outcomes were also assessed.
For descriptive statistics, mean, standard deviation, median, minimum, maximum, frequency, and ratio values of data were used. Distribution of variables was determined using Kolmogorov-Smirnov test. Mann-Whitney U test was used for analyses of quantitative data, whereas chi-square test was used for qualitative data. Kappa fitness test was used for fitness analysis. SPSS software (version 22.0) was used for statistical analyses.
Donor demographics are shown in Table 1. As shown in Table 2, 71 donors had type 1 (normal) biliary anatomy based on both 3D-MRC and IOC. On the other hand, 39 donors had abnormal biliary anatomy, with 2 donors having type 3 (abnormal) biliary anatomy. Evaluations with 3D-MRC were consistent with IOC findings in 104 donors. In a donor who was concluded to have abnormal biliary anatomy (type 2) using 3D-MRC in IOC, normal biliary anatomy was found in graft and abnormal biliary anatomy (type 2) was found in IOC.
The use of 3D-MRC failed to correctly identify biliary anatomy in 13 cases. Of 37 donors who were concluded to have type 2 (abnormal) anatomy based on IOC, the same findings were shown using 3D-MRC. Two donors had type 3 (abnormal) biliary anatomy (Figure 1). The number of ducts in graft was consistent with bile duct anatomy determined preoperatively using 3D-MRC (P < .005). Sensitivity was 80.4%, positive predictive value was 94.4%, specificity was 96.9%, and negative predictive value was 87.3%. The number of ducts in graft was consistent with IOC findings (P < .005) (Table 3). With IOC, sensitivity was 84.8%, positive predictive value was 97.5%, specificity was 98.4%, and negative predictive value was 90.0% (Table 3).
Average age and distributions in men versus women were not different (P > .005) between patients having 3D-MRC findings consistent with the number of ducts in graft examinations and patients having findings that were inconsistent with graft examinations. Average age and distributions of men versus women in patients having IOC findings that were inconsistent with the number of ducts in graft examinations were significantly higher (P < .005) than in patients whose IOC findings were consistent (Figure 2).
Knowledge of biliary anatomy is critical, especially in hepatobiliary surgery and liver transplant surgery.16 Preoperative evaluations of potential liver donors calls for biliary anatomy knowledge. The precise determination of biliary duct anatomy in living liver donor candidates could estimate donor suitability and the number of bile duct orifices and add to surgical planning and even recipient outcomes.3,4
About 30% of individuals have anatomic variations of the biliary tract. The correct characterization of these variations is necessary to prevent iatrogenic injury in hepatobiliary surgery, especially during liver transplant from living donors. For this purpose, precise knowledge of the biliary anatomy is essential.17-21 To prevent damage to the donor’s left hepatic duct, surgeons may want to move the transection plane to the right. Development of 2 or more duct orifices could lead to undesirable outcomes in terms of high risks of ductoplasty or hepaticojejunostomy and biliary complications.
Previous studies have shown the negative effects of the number of hepatic ducts in the graft, and consequently the number of anastomoses, on postoperative complications.4,7,22 Thus, a careful postoperative description of the biliary anatomy is crucial in planning of right liver living donor transplant. In this study, we confirmed the accuracy of 3D-MRC in detection of the biliary anatomy in right lobe liver donors. Similar results have been previously reported. Specifically, Limanond and colleagues10 studied preoperative imaging modalities and reported an accuracy of close to 100% for MRC. In their study, 19 of 26 patients had type 1 anatomy, which was later confirmed by IOC. In addition, 17 of the 19 patients were shown to definitely have type 1 biliary anatomy during surgery (sensitivity of 89.5%). Five of the 7 patients were concluded not to have type 1 anatomy but had variant anatomy (sensitivity for variant anatomy of 71.4%).10 The overall accuracy of MRC was 84.6% in preoperative detection of anatomy. In a report from Kim and associates,3 for exact biliary anatomy, MRC was accurate in 90% of patients. The investigators reported that MRC could specifically describe normal biliary anatomy in 15 of 17 patients and could reveal aberrant anatomy in 12 of 13 patients.3 In our study, 3D-MRC was less successful in identification of type 1 anatomy (62 of 110 patients). However, MRC seemed correct in the estimation of anatomy for the small series. Our much larger patient population compared with those series increases the accuracy and credibility of our findings.
A recent study suggested that proliferative ERCP was superior to and more reliable than standard MRC in the precise determination of biliary variants.23 Nevertheless, this is an additional invasive procedure that increases complication risks in donors and could lead to serious outcomes, even to mortality. In addition, ERCP could not satisfy intraoperative imaging needs. On the other hand, conventional T2-weighted MRC has disadvantages in the evaluation of biliary tract anatomy. It gives some information in the absence of dilatation. Therefore, the recommendation is to use a combination of anatomic and functional methods for an exact evaluation of the biliary system.24 Volumetric 3D contrast-enhanced T1 MRC using Gd-EOB-DTPA contrast agent has recently been proposed as an effective method for obtaining a description of the biliary duct anatomy.25,26 Kinner and colleagues successfully employed Gd-EOB-DTPA-enhanced MRC in the evaluation of 30 potential living donors and could detect anatomic variations reliably.27 Compared with contrast-enhanced computed tomography cholangiography, Gd-EOB-DTPA-enhanced MRC yielded low imaging quality. Nevertheless, the quality was acceptable for the determination of biliary anatomy.28 A recent consensus statement from the European Society of Gastrointestinal and Abdominal Radiology recommended Gd-EOB-DTPA-enhanced MRC obtained at flip angles larger than 20 degrees and at 20 to 40 minutes after injection in those with normal liver functions without biliary duct obstructions.29 Finally, an advanced software (MeVis Liver Analyzer, formerly “HepaVision 2”) was developed for the management of data from MRC.30 With the use of a new algorithm, this software can provide a 3D detailed anatomy of both bile and hepatic vascularities. This development in imaging allows a more accurate preoperative mapping.31,32 However, experience with this new technology is limited, thus calling for continued use of intraoperative evaluations in the selection of precise planning of hepatic duct transection.
The present study showed that intraoperative imaging of biliary anatomy using IOC could be an important modality that can help surgeons during operative procedures. However, because intraoperative cholangiography images are taken in coronal plane, there is a risk for overlapping of structures. Therefore, images should be taken in rotational perpendicular angles.
The number of hepatic ducts in the graft is important in prediction of postoperative morbidity. Lim and associates showed the importance of estimation of number of hepatic duct orifices.33 They found that a combination of 2 or 3 imaging techniques (3D-MRC, MRC, and/or contrast-enhanced 3D-MRC) increased the accuracy and reliability in estimation of hepatic duct number compared with use of 2-dimensional (2D) MRC only. However, our present study showed both normal and variant biliary anatomies. Here, the chance of single hepatic ducts in the graft area was high. In our study, 3D-MRC showed that 62 donors had normal anatomy (type 1), which was confirmed by IOC. In 3 cases, duct anatomy observed in graft was abnormal (type 2). Two factors, that is, confluence angle and distance between secondary ducts and confluence, were evaluated in these cases. Measurement of biliary confluence angle in MRC images may often be unreliable due to large variations. However, biliary confluence angle could yield clues for liver parenchymal atrophy and lesions and masses in periportal area. Hence, its use is crucial in monitoring of these patients.5
Confluence angle can be measured using a blunt tool advanced through the cystic duct or a small choledochotomy. If insertion is made through the cystic duct, its chance of yielding an accurate result is low due to perpendicular angle incompatibility. In addition, it could result in iatrogenic injury to the donor’s biliary duct. Insertion through choledochotomy is an additional invasive procedure and may not be suitable for living liver donors. In cases where the distance between the secondary duct and confluence is 1 mm or less, the chance of getting more than 1 duct orifice is high. Length of the right hepatic duct measured during surgery was 1 mm or less in 13 cases in our present study. In a recent study, Kim and associates stressed that length of right hepatic duct was a factor determining the number of ducts.3 The investigators used axial and coronal 2D-MRC images to determine duct length and evaluated their usefulness for this purpose. However, their study had some limitations. Direct measurement was not made for the images used in the study; therefore, no comparison was made versus IOC images. In additions, all cases were not subjected to right lobe procedures. A quality 3D-MRC could provide enough knowledge for right hepatic duct, and need for axial 2D-MRC is limited. However, measurement of right hepatic duct was not the focus of the present study.
Our present study had a limitation. Artifacts in 3D-MRC due to movements were not completely eliminated. In more advanced studies, Gd-EOB-DTPA-enhanced MRC measurements could improve the evaluation of the biliary system in living donors using hepatocyte-specific agents. We believe that 3D-MRC is reliable in preoperative evaluation of biliary anatomy in right liver donors. Nevertheless, along with new developments in this technique, new software that can allow improved and optimal results that are correlated with intraoperative finding is necessary. In cases where hepatic duct confluence anatomy is normal, if right main hepatic duct is short, a single duct may not be obtained. In division and dissection of biliary duct of right living liver donors, combined use with IOC should be continued.
DOI : 10.6002/ect.2020.0220
From the 1Department of Hepatopancreatobiliary Surgery, Şişli Etfal Hamidiye Training and Research Hospital; and the 2Departmant of Transplant Surgery, Faculty of Medicine, Acıbadem University, İstanbul, Turkey
Acknowledgements: The authors have not received any funding or grants in support of the presented research or for the preparation of this work and have no declarations of potential conflicts of interest. The data used to support the findings of this study are included within the article.
Corresponding author: Tüysüz Umut, Department of Hepatopancreatobiliary Surgery, Şişli Etfal Hamidiye Training and Research Hospital, Istanbul, Turkey
Figure 1. Compliance With Number of Ducts
Table 1. Donor Demographics
Table 2. Correlation Between Preoperative Imaging and Operative Findings
Table 3. Sensitivity, Specificity, Positive Predictive, and Negative Predictive Results of Magnetic Resonance Cholangiography
Figure 2. Comparison of Intraoperative Cholangiography and Magnetic Resonance Cholangiography Images