Introduction

The biliary tract is anatomically divided into two sections as intra- and extrahepatic ducts. The distribution of the intrahepatic biliary ducts is in harmony with the segmental anatomy of the liver. According to the classification by Couinaud, the liver is divided into eight segments, each one having its own vascularisation, and biliary and venous drainage systems [16]. The main hepatic branches maintaining the drainage of the right and left lobes are formed by the junction of the segmental ducts [22]. The right and left hepatic ducts merge to form the common hepatic duct. More distally, the cystic duct joins the common hepatic duct to form the coledochus. The coledochus joins the pancreatic duct and leads to the sphincter of oddi at the second part of the duodenum [16, 22].

The gallbladder and the biliary tract are structures in close connection with the adjacent organs and may show various variations and anomalies. It is therefore important for surgical purposes to know their anatomy and variations in detail [16]. The biliary tract is under the threat of iatrogenic damage during open or laparoscopic cholecystectomy or liver resection surgery and liver transplants from live donors. Also, the anatomic variations of the biliary system have been reported to lay the ground for the formation of gallstones, recurrent pancreatitis, cholangitis and biliary malignancies [9]. Various methods are used in the imaging of the variations of the biliary tract and their pathologies, including ultrasonography (USG), computed tomography (CT); direct cholangiographic methods like endoscopic retrograde cholangiopancreatography (ERCP), percutaneous transhepatic cholangiography (PTC), intravenous cholangiography and T-tube cholangiography, as well as indirect methods like magnetic resonance cholangiopancreatography (MRCP) or cholescintigraphy. The uses of USG and CT for these purposes are limited and these methods bring extra advantages only if a dilatation is present in the biliary tract [6, 18, 21]. Intravenous cholangiography also has certain limitations, the leading one being the low quality of the anatomic imaging due to the inadequate opacification of the cystic duct [13]. The imaging of the biliary tract through direct contrast agent injection is performed in the PTC, ERCP, T-tube or intraoperative cholangiography techniques [18, 21]. MRCP is a magnetic resonance imaging (MRI) technique using T2-weighted sequences to evaluate the anatomy and pathologies of the pancreatobiliary system noninvasively and without contrast injection. With the help of this imaging technique, the intra- and extrahepatic biliary tracts can be evaluated rapidly, reliably and without the risk of complications [6].

3-T MRI has a higher signal-to-noise ratio than 1.5-T MRI. Until recently, MRI devices with a magnetic power of 1.5 T or less were used in the studies investigating the anatomic variations of the biliary system. In the present study, our aim is to investigate the frequency of the anatomic variations of the biliary tract using 3-T MRCP.

Materials and methods

Patient population

Patients who had undergone MRCP procedures at our hospital between November 2009 and February 2012 were retrospectively evaluated. The MRCPs included in the study were carried out in relation with stone formations, cholecystitis, cholangitis, malignancies, parasites and for various investigations on the biliary tract. All the male and female patients whose bile ducts could be viewed were included in the study without any age limit. Patients who had a history of surgery on the biliary tract except for cholecystectomy, whose images could only be inadequately evaluated due to significant respiratory artefacts during the imaging; and those with prosthetics incompatible with the MRI, pacemakers or metallic stents were excluded from the study. A total of 590 patients (between 6 and 88 years of age; mean age 51 ± 9 years) were enrolled for the purposes of the study. Approval was obtained from the Ethics Committee of the Dicle University Medical School before the initiation of the study.

Imaging

MRCP was carried out with an MRI device supplied with 3-T magnetic power (Philips, Eindhoven, Netherland) and using the standard body coil. Patients were informed about the MCRP imaging and the procedures were carried out in the radiology department of our hospital following a 6-h fasting period and after any metal items or objects on the patients which may produce artefacts were removed.

Before the imaging, T2-weighted images were obtained using the two-dimensional “single-shot fast spin echo” (SSFSE) technique as the pulse sequence. The choledochus was located in the images in the axial plane, and coronal images were obtained using the sense-magnetic resonance cholangiopancreatography-three dimensional-high resolution (S MRCP-3D-HR) sequence from the zone commencing a few centimetres above the porta hepatis and ending at the ampulla of Vater. Cross-sections of 1 mm were taken without any cross-sectional intervals. During the post processing, a radial maximal intensity projection (MIP) was performed in order to produce the three-dimensional images. Taking the point where the coledochus was located in the source images in the axial plane as the centre, slabs have been obtained from the coronal or coronal oblique planes, with 35–40 mm thick volumes and each passing through this centre. While each thick cross-section was taken, the duration of breath-hold for the patients was determined as 4.5 s.

Imaging parameters in this sequence were set as follows: repetition time (TR): shortest, echo time (TE): 740 msn, band width (BW): 805 Hz/pixel, FA = 90º, FOV; FH = 30 cm, RL = 22.5 cm, AF = 8 cm, Acq matrix: 256 × 153 and NSA: 1. In order to reduce breathing artefacts, the sequences were synchronised with the patients’ breathing using the trigger function of the device. In addition, coronal images were made using the sense single-shot-magnetic resonance cholangiopancreatography (Ssh-MRCP rad) sequence (breath-hold). In the coronal plane, the settings were as follows: TR: shortest, TE: 40 msn, BW: 408.1 Hz/pixel, FA = 90º, FOV; FH = 30 cm, RL = 30 cm, Acq matrix: 320 × 256 and NSA: 1.

After the imaging was completed, unwanted sections like the intestinal contents were removed from the three-dimensional reformatted images obtained using the MIP method. Thus, the images were rendered more cognizable. No oral or intravenous contrast agents were used during the procedure. It took about 20–25 min to prepare the patients and perform the procedure. No anaesthetic support was needed for any of the patients.

Evaluation of the images

During the evaluation of the findings, the thick cross-sections, MIP images and thin collimation axial and coronal source images were evaluated in combination. The evaluation was carried out by two experienced radiologists. In the definition of the variations detected through MRCP, the following pre-defined criteria were used [18]:

Confluence of the right anterior and posterior ducts, and the common ducts of the segments 1, 2 and the segments 3, 4 at the quadrifurcation were taken into consideration [18].

The insertion of the right posterior hepatic duct to the common hepatic duct or the cystic duct at the distal aspect of the bifurcation was evaluated as an aberrant right hepatic duct variation; while its insertion into the left hepatic duct at a level close to the confluence of the right anterior hepatic duct was evaluated as a trifurcation. Drainage of the right posterior segmental duct into the left hepatic duct is another variation which may be observed at the level of the bifurcation [21, 22].

Insertion of the cystic duct into the one-third distal aspect of the extrahepatic bile duct was assessed as a long cystic duct or distal insertion; an insertion from the left was evaluated as a medial insertion, and a cystic duct length <5 mm was evaluated as a short cystic duct [1, 21, 22].

For a high localized gallbladder, the following criteria have been determined: at least half of the gallbladder located above the portal hilus, fundus located towards the superior aspect and the cystic duct directed cranially along its whole course.

The view produced by the right hepatic artery compressing the common hepatic duct frontally and causing a signal-free area within the duct similar to an intraluminal filling defect has been evaluated as vascular compression on the common hepatic duct [3].

For a pancreatobiliary junction anomaly, the criterion was the fusion of the common bile duct (choledochus) with the pancreatic duct at the proximal aspect of the duodenum to form a 15 mm or longer common duct [5, 18].

Pancreas divisum is an anatomical variation that occurs due to a lack in the fusion of the dorsal and ventral pancreatic ducts. The choledochus and the ventral pancreatic duct (Wirsung) drain into the major papilla, while the dorsal pancreatic canal (Santorini) drains into the minor papilla [12, 15]. For the purposes of the present study, the condition where the main pancreatic duct that drains the tail and body of the pancreas (Santorini) crosses the choledochus frontally and is inserted into the duodenum separately and more proximally from the choledochus was accepted as an adequate finding for the diagnosis of pancreas divisum [2].

The frequency in which the variations described in previous studies were detected in our study was expressed as percentage.

Results

Among the 590 patients included in the study, 200 were male (33.9 %), while 390 were female (66.1 %). The mean age was 54 ± 7 years in males and 50 ± 8 years in females. Variations at various levels were detected in 233 (39.5 %) of all the patients. Among these, 30 patients (12.5 %) had a history of cholecystectomy. The distribution of the variations among the patients are summarised in Table 1.

Table 1 Distribution of the variations in the patients, in whom single variations were detected using the MRCP
Full size table

In this study, a total of 275 variations were observed. Among these, a single variation was found in 191 patients (32.3 %), while 2 variations were detected in 42 patients (7.2 %). The most commonly observed variation at the level of the bifurcation was the insertion of the right posterior segmental branch into the left hepatic duct in 71 patients (12.1 %) (Fig. 1). On the other hand, the most commonly observed variation at the level of the cystic duct was the medial cystic duct insertion in 58 patients (9.8 %) (Fig. 2).

Fig. 1
figure 1

Right posterior segmental branch insertion into the left hepatic duct in MRCP (a) and the schematic view (b)

Full size image
Fig. 2
figure 2

Medial insertion of the cystic duct in MRCP (a) and the schematic view (b)

Full size image

Details of the 42 patients, in which duplicate variations were detected, are summarised in Table 2. According to these results, the most commonly observed duplicate variations were the medial insertion of the cystic duct into the main hepatic duct due to the fusion of the right posterior segmental branch with the left hepatic duct (11 patients) (Fig. 3), distal medial insertion of the cystic duct into the main hepatic duct due to the fusion of the right posterior segmental branch with the left hepatic duct (7 patients), and the medial insertion of the cystic duct into the main hepatic duct with trifurcation (5 patients).

Table 2 Distribution of the variations in the patients, in whom duplicate variations were detected using the MRCP
Full size table
Fig. 3
figure 3

Fusion of the right posterior segmental branch with the left hepatic duct and medial insertion of the cystic duct into the main hepatic duct in MRCP (a), schematic view (b) and MIP images (c) (duplicate variations)

Full size image

Discussion

Anatomical variations of the biliary tract were observed in 233 (39.5 %) of the 590 patients included in the study. When the variations are divided into groups as those at the bifurcation level, cystic duct level, pancreatobiliary junction level and those outside these classifications, it has been found out that the most frequently observed variation in general and at the bifurcation level was the drainage of the right posterior hepatic duct into the left hepatic duct before the bifurcation, which was observed in 71 patients (12.2 %). The most frequently observed variation at the level of the cystic duct was medial insertion observed in 58 patients (9.8 %). At the pancreatobiliary junction level, a pancreatobiliary junction anomaly was detected in 1 patient (0.2 %), while a pancreas divisum was viewed in 1 patient (0.2 %). Among the other anomalies, vascular compression on the common hepatic duct was observed in 17 patients (2.9 %) and a high located gallbladder was observed in 2 patients (0.3 %). With duplicate variations in 42 cases, the total number of the detected variations was 275.

MRCP is a non-invasive imaging method with a low risk of complication and an accuracy rate matching ERCP in pancreatobiliary diseases. The method does not involve any ionising radiation, contrast agent injection or premedication [6]. The fact that it can be applied during acute attacks of pancreatitis and cholangitis is able to reveal the bile ducts both in the proximal and distal aspects of the stricture, gives the chance to view the extraductal structures as well as the conventional T1-T2-weighted images, and can detect the stricture in bilioenteric anastomoses constitute the strengths of this method in comparison to ERCP. On the other hand, the limitations of MRCP are its low resolution, inability to show minor ductal pathologies and the hindrance to perform therapeutic interventions during the procedure [23]. In the literature, there are studies reporting a lower [6] or similar [4] number of variations compared to our study.

Although the anatomic variations of the biliary system are usually of no clinical importance, they may lead to confusions during diagnostic examinations. Also, these variations may complicate endoscopic or percutaneous interventions, open or laparoscopic cholecystectomy operations and liver transplantations from live donors, leading to iatrogenic traumas [5, 16, 21, 22].

The right hepatic artery may compress on the common hepatic duct from outside and leads to an image similar to an intraluminal filling defect caused by a stone or stricture present in the duct [19, 20]. In a study performed on 475 patients using a 1-T MRI device, 12 patients (2.5 %) had this variation [4]. In our study, 17 patients (2.9 %) were observed to have this variation and the ratio was in line with the literature. This variation may be mistaken for a cystic artery during the operation and thus surgically closed, leading to serious complications like the disturbance of liver functions [16]. The typical view and location help the viewer to distinguish the condition from an actual filling defect [16]. However, if doubts related to the defect are still continuing, the hepatic artery crossing the common hepatic artery may be revealed with an MRI angiography performed during the same session [19, 20].

In a study conducted on 475 patients using a 1-T MRI device, a total of 115 patients (24.2 %) were observed to have various anatomic anomalies at different levels. Another study carried out on 122 patients with a 1.5-T MRI device revealed various anatomic anomalies at different levels of the intra- and extrahepatic bile ducts of 31 patients (25.2 %) [6, 18].The most commonly observed variations in both studies were the drainage of the right posterior hepatic duct into the left hepatic duct in 27 (5.7 %) and 8 patients (6.4 %), respectively. Again, the most commonly observed variation in two separate studies other than the above-mentioned ones was the drainage of the right posterior hepatic duct into the left hepatic duct. However, the ratios of the variations observed in these were higher than the previous ones with 13 and 19 % and these results were also in line with our findings [8, 17]. The most commonly viewed variation in our study was the drainage of the right posterior hepatic duct into the left hepatic duct and it was detected in 71 patients, with a frequency of 12.5 %. Also, the frequency of duplicate variations observed in 42 patients (7.2 %) in our study was significantly higher compared to the 11 patients (2.3 %) in the previous study conducted on 475 patients [6]. We are of the opinion that the use of the 3-T MRI with a higher resolution plays an important role in these different results.

In liver transplants from live adult donors, usually the right lobe of the liver is transplanted and the right bile duct is also included in the graft. In the choledococholedocostomy and choledochojejunostomy methods, the graft and the bile ducts of the receiver are anastomosed [11]. In normal biliary anatomy (bifurcation), it is relatively easier to perform a single biliary-enteric anastomose. However, in case of trifurcations or the variations where the right posterior hepatic canal is inserted into the left hepatic canal or the main hepatic canal, two anastomoses are needed in order to prevent bile leaks or segmental atrophy which may develop subsequent to the surgery.

The technique requiring two anastomoses is more complicated and these individuals are not regarded as donors and usually screened out. The standard surgical techniques in use provide that the bifurcations where the right posterior hepatic duct is inserted at a point closer than 1 cm to the bifurcation are accepted as trifurcations [10, 11]. To the best of our knowledge, a variation we have come across the patient in our study (0.8 %), which involves the combined insertion of the cystic duct and the main hepatic duct into the ampulla is the first one observed in MRI studies, although it has been reported in surgical studies in the literature [14] (Fig. 4).

Fig. 4
figure 4

The variation in MRCP where the cystic duct and the main hepatic duct are inserted to the ampulla in combination (a–c) and the schematic view (d)

Full size image

Intraoperative cholangiography is a traditional method used to view the biliary anatomy before right hepatic lobe resections [7]. Performed under optimum conditions, intraoperative cholangiography provides a high-quality view of the biliary tract. However, due to the inconvenient conditions during the surgery, this method may not always be performed successfully. A study revealed that intraoperative cholangiography performed on 20 patients has provided high-quality results suitable for an appropriate evaluation of the left and right central hepatic ducts in nine patients (45 %). On the other hand, during an MRI cholangiography, the central ducts could be fully revealed in 25 out of 28 patients (89 %). Investigators claim that the MRI technology is still developing and it may replace intraoperative cholangiography during right lobe resections and shorten the duration of the surgery in the future [21]. In our study, the MRCPs of 85 patients (12.6 %) were in an inferior condition preventing assessment and were therefore excluded from the study. Since intraoperative cholangiography is not performed in our hospital, we did not have the chance to make a comparison with MRCP.

The limitations of the study resulted from the relative difficulty in the evaluation of the biliary tracts of the study patients with pathological findings. Also, since we could not compare 3-T MRCP and intraoperative cholangiography in this study, we could not evaluate if any of the patients who underwent MRCP needed intraoperative cholangiographies.

In conclusion, the 3-T MRCP we performed has revealed similar or higher numbers of variations compared to the studies with 1- and 1.5-T devices. Accordingly, since MRCP procedures performed with 3-T devices reveal more variations, the method may reduce iatrogenic traumas like the closure or resection of the wrong duct and thus significantly reduce the mortality and morbidity rates in open or especially laparoscopic cholecystectomy, hepatic segmentectomy or lobectomy operations and the ever-increasing transplantation surgeries. Also, donors with variations rendering them unsuitable for liver transplantations may be determined through MRCP and screened out to prevent operations with high morbidity risk.