Endoscopic Ultrasound in Bile Duct, Gallbladder, and Ampullary Lesions

Key Points

  • In patients with low to moderate clinical probability of common bile duct (CBD) stones, endoscopic ultrasound (EUS) or magnetic resonance cholangiopancreatography (MRCP) is recommended before endoscopic retrograde cholangiopancreatography (ERCP) is performed.

  • In patients with acute pancreatitis of unknown origin or right upper quadrant pain with normal transabdominal ultrasound, EUS should be considered.

  • In patients with a CBD stricture of unknown origin, EUS should be performed and, if inconclusive, ERCP should follow with tissue sampling and consideration of cholangioscopy with or without intraductal ultrasonography (IDUS).

  • Gallbladder polyps larger than 5 mm in size may be investigated with EUS to determine the malignancy potential and the subsequent therapeutic approach.

  • Ampullary tumors can be staged with EUS and IDUS. EUS is best to differentiate between early (adenoma, T1) and advanced (T2 to T4) tumors and guide therapy.

General Examination Checklist for Biliary Stones and Tumors and Ampullary Lesions

  • Extrahepatic ducts (dilation, stones, strictures)

  • Intrahepatic ducts (dilation)

  • Left and right liver lobes (masses)

  • Gallbladder

  • Ampulla (including IDUS in the case of T1 ampullary lesions)

  • Pancreatic main and accessory ducts

  • Lymph nodes

  • Ascites

  • Portal hypertension

Endoscopic Ultrasound and Biliary Stones

Bile Duct Stones

Endoscopic retrograde cholangiopancreatography (ERCP) has long been considered the best diagnostic method for common bile duct (CBD) stones. Moreover, ERCP allows stone removal during the same endoscopic session when combined with endoscopic sphincterotomy (ES). Nevertheless, it remains an invasive procedure and carries a substantial risk of complications, although when performed by experienced endoscopists, the complications and mortality rates can be reduced to under 5% and 0.1%, respectively. Furthermore, because it can be difficult to differentiate small stones from aerobilia, a substantial proportion of ERCP procedures are completed with ES, in order to confirm the diagnosis of choledocholithiasis. ES has a complication rate of 5% to 10%, with a current mortality being less than 1%. Long-term complications, such as stenosis and nonobstructive cholangitis, are rare occurring in approximately 10% or fewer of patients.

As such, ERCP is no longer acceptable as a diagnostic tool for CBD stones albeit it remains the therapeutic modality of choice. Transabdominal ultrasonography (TUS) is a widely available, noninvasive imaging modality that should be part of the initial evaluation of any patient with clinical and/or laboratory suspicion of CBD stones. However, although TUS is very sensitive and specific for cholelithiasis, its sensitivity for the diagnosis of choledocholithiasis remains limited, even in heavily calcified CBD stones. The location and orientation of the bile duct, along with adjacent duodenal air, make imaging of the distal bile duct difficult, and abdominal fat attenuates ultrasound waves, making this technique less effective in obese patients.

Other imaging modalities such as multidetector computed tomography (CT), endoscopic ultrasonography (EUS), and magnetic resonance cholangiopancreatography (MRCP) have been effectively employed for the diagnosis of CBD stones. The sensitivity, specificity, and accuracy of helical CT range from 85% to 88%, 88% to 97%, and 86% to 94%, respectively. Nevertheless, the sensitivity of CT for detecting stones under 5 mm in size remains significantly lower than those measuring 5 mm or more. In one comparative study with MRCP and EUS, helical CT was inferior to either, although multiplanar reconstructions with multidetector CT can improve its specificity. Therefore EUS and MRCP remain the most accurate minimally invasive methods for diagnosing CBD stones.

Endoscopic Ultrasound Technique for Detecting Choledocholithiasis

EUS provides excellent sonographic visualization of the extrahepatic biliary tree. Bile duct stones are shown as echo-dense structures ( Figs. 16.1 and 16.2 ) within the ampulla or CBD, sometimes freely moving within the duct, with or without acoustic shadowing or thickening of the bile duct wall ( Videos 16.1A and B ). The accuracy of EUS was found to be higher than that of ERCP for the detection of small CBD stones ( Fig. 16.3 ), with a negative predictive value (NPV) exceeding 95% and specificity in ruling out the presence of CBD stones of 95% or higher in the majority of published studies ( Table 16.1 ). Additionally, EUS detects bile duct sludge as well as microlithiasis ( Video 16.2 ), often missed by the other imaging techniques.

Fig. 16.1

Linear endoscopic ultrasound image (7.5 MHz) of common bile duct stone (+) in a patient presenting with right upper quadrant pain and elevated transaminases.

Fig. 16.2

Linear endoscopic ultrasound image (7.5 MHz) of a shadowing 9 mm common bile duct stone without significant upstream ductal dilation.

Fig. 16.3

(A) Fluoroscopy image of a common bile duct without any filling defects at endoscopic retrograde cholangiopancreatography. (B) Small stone identified on endoscopic ultrasound (radial echoendoscope, 6 MHz). (C) Stone confirmed by biliary sphincterotomy and after balloon sweeping of the common bile duct.

TABLE 16.1

Performance of Endoscopic Ultrasound in the Diagnosis of Common Bile Duct Stones

Reference (Year) Level of Evidence a No. of Patients Included Frequency of CBD Stones (%) Endoscopic Ultrasound Performance Characteristics
Sensitivity (%) Specificity (%) PPV (%) NPV (%) Accuracy (%)
Kohut et al. (2002) 1 134 68 93 93 98 87 94
Meroni et al. (2004) 1 47 15 71 90 55 95
Liu et al. (2000) 2 139 35 98 98 100 96 99
Prat et al. (2001) 2 123 27 100 100 100 100 100
Berdah et al. (2001) 2 68 20 96 97 93 100 98
Buscarini et al. (2003) 2 463 52 98 99 99 98 97
Kohut et al. (2003) 2 55 9 75 99 100 98 98
Aube et al. (2005) 2 45 34 94 97 94 97 96
Ney et al. (2005) 2 68 32 96 99 100 97 98
Lachter et al. (2000) 3 50 66 96 75 89 93 94
Materne et al. (2000) 3 50 26 97 88 94 93 94
Scheiman et al. (2001) 3 28 18 80 95 80 96
Ainsworth et al. (2004) 3 163 33 90 99 98 94 93
Kondo et al. (2005) 3 30 86 98 50 92 100 93
Dittrick et al. (2005) 3 30 37 100 84 56 100
Jeon et al. (2016) 3 200 83 98 80 95 89 94
Netinatsunton et al. (2016) 3 141 59 98 80 98 80

CBD, Common bile duct; ERC, endoscopic retrograde cholangiography; ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic sphincterotomy.

a Level 1: Technique compared with ERC + systematic ES with a very short interval between the technique and ERCP; Level 2: Technique compared with ERCP + ES if positive, and clinical and biological follow-up of at least 6 months if negative; Level 3: Technique compared with ERC or with intraoperative cholangiography.

Video 16.1A

Linear Endoscopic Ultrasound Exam (7.5 MHz) of the Bile Duct Demonstrating a 15-mm Shadowing Stone in the Distal Bile Duct With Upstream Common Bile Duct Dilation

The patient presented with jaundice and intermittent right upper quadrant pain.

Video 16.1B

The Same Stone Interrogated With a Radial Echoendoscope

Video 16.2

Interrogation of the Head of the Pancreas, Distal Bile Duct, and Pancreatic Duct Using a Radial Echoendoscope in a Patient Presenting With Right Upper Quadrant Pain and Elevated Transaminases

A free-floating nonshadowing echogenic structure, likely sludge but could be a noncalcified stone, is noted in the common bile duct.

In most EUS literature, EUS radial echoendoscopes were used for assessment of choledocholithiasis. Nevertheless, the accuracy of linear EUS appeared to be comparable to that of the radial exam, as indicated in some series comparing EUS with ERCP plus ES, or choledochotomy with choledochoscopy (see Table 16.1 ). The advantages of radial echoendoscopes reside in their ability to visualize the bile duct in a long section (along its main axis) without having to torque the scope. Nevertheless, it can miss hilar stones due to the distance. A linear echoendoscope, on the other hand, provides cross-sectional or tangential views of the bile duct but requires continuous torque to visualize the entire duct: clockwise to interrogate the bile duct from the hilum to the papilla and counterclockwise to interrogate the duct in the opposite direction (see Videos 16.1 and 16.3 ).

Videos 16.3A and B

These Videos Demonstrate the Technique of Interrogating the Bile Duct From the First Part of the Duodenum Using a Linear (A) and Radial (B) Echoendoscope

The nondilated bile duct is seen coursing through the pancreatic head until it enters the duodenal wall.

The ability of either echoendoscope to visualize intrahepatic stones remains low due to the distance from the tip of the scope and the presence of several intervening structures. In addition, stones impacted at the level of the papilla can be missed unless full visualization of the duct at its insertion into the duodenal wall is possible. This ampullary view can be difficult to obtain in some cases from the duodenal bulb, where deep insertion into the second part of the duodenum followed by slow withdrawal with the tip of the scope in full upward deflection could help bring the ampulla to view and stabilize the scope within the duodenal sweep. It is also recommended to instill water in the second part of the duodenum, particularly when the radial scope is being utilized to improve acoustic coupling of the periampullary area.

The Use of Endoscopic Ultrasound, Magnetic Resonance Cholangiopancreatography, and Endoscopic Retrograde Cholangiopancreatography in the Management of Choledocholithiasis

MRCP is a noninvasive, radiation-free imaging modality and is more accurate than CT for the diagnosis of choledocholithiasis ( Table 16.2 ). The disadvantages of this technique include the limited spatial resolution, the difficulty of diagnosing CBD stones in the periampullary region, lack of availability in some areas, need for operator’s experience to interpret findings, and the high cost. Moreover, MRCP is contraindicated in patients with metallic hardware such as pacemakers or cerebral aneurysm clips and is difficult to conduct in claustrophobic patients; EUS offers higher spatial resolution than MRCP (0.1 vs. 1 to 1.5 mm), and its sensitivity for detecting choledocholithiasis does not vary with the stone size like MRCP. Thus, it is not surprising that stones missed by MRCP were always smaller than 10 mm, and that the sensitivity of MRCP decreased to approximately 65% for diagnosing stones smaller than 5 mm. Nevertheless, improvements in imaging may in the future permit the detection of even smaller stones. In a recent systematic review, the sensitivity of MRCP in detecting CBD stones was 90% with a specificity of 95%.

TABLE 16.2

Performance of Magnetic Resonance Cholangiopancreatography in the Diagnosis of Common Bile Duct Stones

Reference (Year) Level of Evidence a Number of Patients Sensitivity (%) Specificity (%) PPV (%) NPV (%) Accuracy (%)
Gautier et al. (2004) 2 99 96 99
Aube et al. (2005) 2 45 88 97 93 93
Topal et al. (2003) 2 315 95 100 100 98
Mofidi et al. (2008) 2 49 100 96
Scaffidi et al. (2009) 2 120 88 72 87 72 83
Cervi et al. (2000) 3 60 100 94
Demartines et al. (2000) 3 70 100 96 93 100
Stiris et al. (2000) 3 50 88 94 97 81
Materne et al. (2000) 3 50 91 94 88 95 92
Scheiman et al. (2001) 3 28 40 96 66 88
Kim et al. (2002) 3 121 95 95 95
Taylor et al. (2002) 3 146 98 89 84 99
Griffin et al. (2003) 3 115 84 96 91 93 92
Ainsworth et al. (2004) 3 163 87 97 95 93
Kondo et al. (2005) 3 30 88 75 96 50 86
Ausch et al. (2005) 3 773 94 98 80 99
Hallal et al. (2005) 3 29 100 91 50 100 92
Makary et al. (2005) 3 64 94 98 94 98
Moon et al. (2005) 3 32 80 83 89 71 81
De Waele et al. (2007) 3 104 83 98 91 95 94
Norero et al. (2008) 3 125 97 74 89 90 90
Richard et al. (2013) 3 70 27 83 36 77 69
Badger et al. (2016) 3 527 90 86 97 60

CBD, Common bile duct; ERC, endoscopic retrograde cholangiography; ERCP, endoscopic retrograde cholangiopancreatography; ES, endoscopic sphincterotomy.

a Level 1: Technique compared with ERC + systematic ES with a very short interval between the technique and ERCP; Level 2: Technique compared with ERCP + ES if positive, and clinical and biological follow-up of at least 6 months if negative; Level 3: Technique compared with ERC or with intraoperative cholangiography.

The diagnostic performance of EUS has been evaluated in two meta-analyses covering 3532 and 2673 patients. The pooled sensitivity and specificity of EUS were 89% to 94% and 94% to 95%, respectively. The evidence for the use of MRCP for the diagnosis of CBD stones has been examined in a systematic review of 10 studies and was shown to provide a high sensitivity (range 80% to 100%) and specificity (range 83% to 98%). In comparative studies of the two technologies, EUS was found to be either superior or equivalent to MRCP for the diagnosis of choledocholithiasis. One meta-analysis and two systematic reviews comparing EUS and MRCP for depicting CBD stones showed a high diagnostic performance for both modalities. Although no statistically significant differences were found between the two modalities, there was a trend towards higher sensitivity and specificity for EUS compared to MRCP. This was especially obvious in the case of small stones causing acute biliary pancreatitis. Nevertheless, the choice between these two techniques should depend on other factors such as resource availability, operator experience, and cost.

The use of noninvasive imaging modalities resulted in a considerable reduction in the number of inappropriate ERCP with bile duct cannulation. One meta-analysis comparing an EUS-guided ERCP strategy with an ERCP-only strategy found that the use of EUS significantly reduced the risk of overall complications (relative risk 0.35) by safely avoiding ERCP in 67% of patients. Whether or not an EUS or MRCP would be necessary prior to ERCP depends on the pretest probability of having a stone in the CBD. Patients suspected of having choledocholithiasis on clinical and laboratory criteria and/or ultrasound (US) can be grouped into risk groups, ranging from low to intermediate to high risk. Patients in the high-risk group include those with CBD stones on transabdominal US, clinical ascending cholangitis, and bilirubin greater than 4 mg/dL. When considering all published studies, the proportion of high-risk patients who actually have CBD stones was less than 80% (66% to 78%), whereas fewer than 40% of patients classified as being at intermediate (also called moderate) risk had choledocholithiasis (19% to 44%). Most experts agree that ERCP could be performed as a first-line approach in patients at high risk of CBD stones, although it may be impossible to avoid unnecessary ERCP procedures altogether. EUS as a first-line approach in patients in the high-risk category could still be performed if available to exclude a stone or to evaluate other causes of biliary symptoms. Moreover, EUS confirming CBD stones would justify the use of aggressive techniques, such as precut papillotomy, if needed. However, there is still no general agreement in regard to the routine utilization of EUS in such cases. Practically, the best approach is probably to perform EUS followed by ERCP (with or without ES) when a stone is found during the same endoscopic session.

Intermediate risk patients include those who present with symptoms compatible with biliary origin, along with liver test abnormalities or dilated CBD on TUS. The consensus in this group of patients is to consider EUS (or MRCP) as the first-line diagnostic approach (after TUS). This approach was evaluated in the context of laparoscopic cholecystectomy in a series of 300 patients. Choledocholithiasis was found on preoperative EUS in 19% of the intermediate risk, and on ERCP in 78% of the high risk. For low-risk patients, who typically have no biliary symptoms or liver test abnormalities, and have no CBD dilatation on TUS, no further examination is necessary in this case. We suggest an algorithm to investigate patients with suspected choledocholithiasis based on their risk stratification ( Fig. 16.4 ).

Fig. 16.4

An algorithm for the management of patients with suspected choledocholithiasis. CT, Computed tomography; CBD, common bile duct; ERCP, endoscopic retrograde cholangiopancreatography; EUS, endoscopic ultrasound; TUS, transabdominal ultrasonography; US, ultrasound.

The utilization of EUS was associated with potential financial advantages as a first-line strategy in cost-effectiveness studies. In a prospective study of 485 patients suspected of having CBD stones where EUS was always performed regardless of the risk classification, the mean cost for patients managed by the EUS-based strategy was significantly lower than that for patients who had ERCP. In another study, the EUS-guided ERCP strategy resulted in 14% reduction in ERCP procedures, and was associated with significant cost savings. Other studies have found that EUS was the most cost-effective strategy in the intermediate-risk group, whereas in patients with a probability of CBD stones above 50% (high-risk group), the most cost-effective approach was an ERCP first approach. In patients with acute biliary pancreatitis, an economic evaluation also concluded that EUS was the strategy associated with lower costs, and fewer procedures and complications. This was especially obvious in patients with severe acute pancreatitis. Finally, a randomized study comparing EUS and ERCP during the same endoscopic session versus EUS and ERCP in two separate sessions for the management of choledocholithiasis showed that the average procedure time and days of hospitalization were significantly reduced in the first group, resulting in significant reductions in total costs.

Two adjunct ultrasound-based technologies have been studied in the last 2 decades: extraductal catheter probe EUS (EDUS) and intraductal ultrasonography (IDUS). The use of EDUS has been evaluated in two studies and appears to be accurate in detecting CBD stones and is nearly as accurate as linear array EUS for that purpose. In the earlier published prospective study, EDUS with a radial scanning catheter probe was performed before ERCP and ES in patients with suspected CBD stones. EDUS detected 33 of 34 bile duct stones. In eight patients, the stones were missed on ERCP but seen after ES. The same authors conducted a subsequent prospective trial to compare the diagnostic potential of EDUS with that of conventional EUS, where EDUS was found to be nearly as accurate as linear array EUS. IDUS has also been proposed to assess for CBD stones ( Fig. 16.5 ). In a prospective study of patients with suspected CBD stones who underwent ERCP, 20 MHz IDUS exam was performed in those with equivocal cholangiograms or cholangiographic evidence of stones. Interestingly, no stones were found in 36% of patients with a positive finding on ERCP, which is likely due to the presence of aerobilia. In 35% of patients with a negative ERCP, sludge or stones were found on IDUS and were confirmed on ES. Another study demonstrated that the addition of IDUS to confirm complete stone clearance after ES decreased the recurrence rate of CBD stones (13% in the non-IDUS group compared to 3% in the IDUS group). In one prospective trial, the sensitivity of MRCP, ERCP, and IDUS for the diagnosis of choledocholithiasis was 80%, 90%, and 95%, respectively. The accuracy of IDUS plus ERCP was superior to that of ERCP alone in this study. IDUS could be particularly helpful in nonopaque stones, as recently demonstrated in a study including patients with various calcium density CBD stones, where IDUS identified all 148 patients with duct stones (100%) as opposed to ERCP which missed three stones in the same group of patients. However, IDUS cannot be proposed as a routine procedure because of the morbidity associated with ERCP. It can be utilized prior to ES in patients in whom CBD stones have been found on EUS or MRCP, but not on ERCP, or following ES to confirm complete stone clearance.

Fig. 16.5

Two-dimensional intraductal ultrasonography showing a shadowing common bile duct stone.


EUS is the ideal alternative to cholangiography for the evaluation of choledocholithiasis, selecting only those patients with confirmed CBD stones for ERCP. MRCP can be utilized as an alternative when there are contraindications to sedation, or when EUS is not available. ERCP should be avoided, if biliary EUS proved normal, unless symptoms persist or recur during follow-up. Ideally, EUS and ERCP should be combined in a single endoscopic session whenever possible to reduce risks of repeated sedation and minimize cost. When this approach is not feasible, high-risk patients could be managed with ERCP at first.


Transabdominal ultrasound (TUS) is an excellent modality for the diagnosis of cholelithiasis, with exceedingly high sensitivity and specificity, but its performance is limited in smaller size stones and large body habitus. Because of its value in diagnosing small CBD stones, EUS has also been evaluated for detecting cholelithiasis ( Figs. 16.6 and 16.7 ; Video 16.4 ). EUS can influence the management of patients with biliary pain and normal initial imaging with TUS or CT. For example, Thorboll et al. studied patients with a normal TUS, but with suspected gallstones based on clinical grounds and detected cholelithiasis in 18/35 patients (52%).

Fig. 16.6

(A) A shadowing gallbladder stone incidentally noted during linear (7.5 MHz) endoscopic ultrasound (EUS) exam. (B) A solitary shadowing gallbladder stone (linear EUS image; 7.5 MHz) with significant acoustic void beyond the stone due to its echodensity. This was incidentally noted during assessment of an esophageal subepithelial lesion.

[A] Figure courtesy of Dr. John DeWitt.

Fig. 16.7

Linear endoscopic ultrasound image (7.5 MHz) of multiple small calcified gallstones.

Figure courtesy of Dr. John DeWitt.

Idiopathic acute pancreatitis (IAP) can be the result of biliary sludge or microlithiasis undetected by other imaging techniques ( Fig. 16.8 ). Although the reported incidence of occult gallstones in IAP varies (ranging from 10% to 73%), this remains the most common cause of pancreatitis in patients with intact gallbladder. In one study, gallstones were found by EUS in 14 of 18 patients with negative findings on TUS. In a larger series, 168 patients with IAP were referred for EUS which identified gallbladder sludge or very small stones in 40% of patients, with or without associated CBD stones, that had been missed by other examinations. Overall, EUS was able to detect a cause for the acute pancreatitis in 80% of patients. Yusoff et al. reported that EUS established a presumptive diagnosis in 31% of 201 patients with a single episode of IAP, the most frequent causes in those with in situ gallbladders being chronic pancreatitis and biliary sludge. A systematic review evaluating the role of EUS in idiopathic pancreatitis showed a high diagnostic yield, especially in patients with a single idiopathic episode, and in patients with recurrent episodes and gallbladder in situ. Moreover, a cost analysis identified EUS as the most cost-effective initial test in the evaluation of IAP when compared with other strategies including ERCP with manometry and bile aspiration, and laparoscopic cholecystectomy. Therefore an EUS-based strategy appears to be the best approach to evaluate patients with IAP because of the high diagnostic accuracy of EUS, not only for gallbladder sludge and stones, but also for pancreatic diseases, and its minimally invasive nature. In patients with multiple unexplained attacks, particularly in those postcholecystectomy, ERCP and sphincter of Oddi manometry should be considered after negative EUS results.

Fig. 16.8

Linear endoscopic ultrasound image (7.5 MHz) of extensive gallbladder sludge.

Video 16.4

Linear Echoendoscope Examination of the Gallbladder

Several irregular calcified gallstones are seen in the gallbladder lumen. The patient was asymptomatic from biliary standpoint.


EUS is the most effective method for confirming the presence or absence of CBD stones. Its utility in avoiding unnecessary ERCP has been validated in patients at low or moderate risk of CBD stones. MRCP can probably be used as an alternative to EUS if available and there are no contraindications. EUS remains the preferred diagnostic test in the setting of acute pancreatitis where biliary stones can be very small and can be missed on MRCP. For patients at high risk of CBD stones, ERCP ± ES (in case CBD stones were found on cholangiography) can be considered as the first-line approach; however, if EUS is available and can be performed during the same endoscopic session with ERCP, then that would be optimal. EUS is now a well-established modality after TUS for the diagnosis of gallbladder stones and sludge in patients with unexplained right-upper-quadrant pain, and also in those with an acute pancreatitis of unknown origin.

Diagnostic Checklist

CBD or Gallstone

  • Hyperechoic mobile structure with or without acoustic shadowing

Associated Signs

  • Dilation of extrahepatic ducts and/or cystic duct

  • Thickening of the gallbladder and/or bile duct wall

  • Thickening of the ampulla

  • Pericholecystic fluid

Endoscopic Ultrasound in Bile Duct Strictures

Benign and malignant bile duct strictures remain a diagnostic challenge for the gastroenterologist. TUS and CT imaging can reliably demonstrate dilated bile ducts, but are often inadequate in assessing underlying cause of dilation. ERCP is highly accurate for the confirmation of obstructive jaundice, but little additional information could be obtained for tumor staging, as only indirect tumor signs, such as stenosis, supra-stenotic dilation, or both, can be visualized, and the tumor itself is generally not well seen. MRCP, on the other hand, can demonstrate a small tumor or focal narrowing in the bile duct, but appears to add little to ERCP for the diagnosis of malignancy except for detecting contiguous tumor invasion or metastasis. However, MRCP could be superior in investigating the anatomic extent of the lesion compared to ERCP because it displays the biliary tree proximal to the obstruction.

Intraductal tissue sampling is commonly used at the time of ERCP. Brushing has a low sensitivity for the diagnosis of bile duct tumors of 27% to 56%, owing to their desmoplastic nature or submucosal spread of the tumor, and is often negative for extrinsic tumors (pancreatic cancer, gallbladder cancer, metastatic lymph nodes). New ancillary cytology-based techniques have been developed over the last decade to improve the sensitivity of routine cytology. Such techniques include fluorescence in situ hybridization (FISH) analysis, which detects chromosomal polysomy using fluorescent probes, and digital image analysis (DIA) technique to assess for the presence of aneuploidy. FISH plays an increasing role in the diagnosis of malignancy from ERCP-obtained samples, whereby in one study it was shown to increase the sensitivity of brush cytology from 21% to 58%, and increased sensitivity to 72% when combined with brush cytology results. Recently, PCR-based DNA mutation profiling has been studied in biliary brushings and was shown to augment the sensitivity of cytology and FISH from 32% for cytology alone to 73% when all three modalities were combined.

Additional means of tissue sampling include direct forceps biopsies during ERCP, which has a higher sensitivity than brush cytology alone, with sensitivity ranging from 44% to 89% in cholangiocarcinoma and 33% to 71% in pancreatic cancer. Nevertheless, this technique is limited by its low NPV. This has led to the development of new methods to abrade the tumor surface in order to improve cytologic yield including a combination of stricture dilation, endoscopic needle aspiration, and biliary brush cytology. Currently, bile duct biopsy under direct cholangioscopy presents the most promising method of biliary sampling using the single operator cholangioscopy system (SOC), with a sensitivity up to 90% based on direct biopsies as will be discussed later in this section.

Endoscopic Ultrasound-Fine-Needle Aspiration Considerations in Biliary Strictures and Tumors

EUS has proved to be a useful tool in assessing biliary obstruction, as it readily visualizes the entire CBD. Therefore it can be helpful in the differential diagnosis of bile duct strictures and neoplasia, and local tumor staging ( Fig. 16.9 ). The ability to acquire tissue by EUS-guided fine-needle aspiration (FNA) significantly improves the diagnostic yield when evaluating biliary strictures, and is associated with a minimal risk of complications ( Fig. 16.10 ). In a recent meta-analysis of 20 studies involving 957 patients, EUS FNA had a pooled sensitivity and specificity for diagnosis of malignant biliary stricture of 80% and 97%, respectively. As the distal CBD is located immediately under the echoendoscope transducer when examined from the duodenal bulb, EUS performs extremely well for evaluating distal biliary strictures ( Video 16.5 ). EUS FNA is highly accurate in diagnosing malignancy in distal biliary strictures, particularly in patients with masses within the pancreatic head. In this setting, the overall EUS FNA sensitivity and specificity rates range from 81% to 91% and from 71% to 100%, respectively. However, its sensitivity for proximal biliary strictures drops and has ranged between 25% and 89% ( Tables 16.3 and 16.4 ). Nevertheless, the reported accuracies are lower for cholangiocarcinomas, mainly due to the inclusion of the hilar cholangiocarcinomas (Klastkin tumors) and the difficulty visualizing and sampling such tumors via EUS due to distance from the probe. Moreover, proximal biliary lesions tend to often be small and diffusely infiltrating, unlike the distal biliary ones that frequently present as focal solid masses ( Fig. 16.11 ). Cytologic diagnosis is an important adjunct to EUS and helps direct patient management and avoid unnecessary surgery (see Fig. 16.10 ). Limited experience with a forward-viewing linear echoendoscope (Olympus Medical Center Valley, Pennsylvania) suggests improved imaging of hilar strictures and easier EUS FNA technique.

Fig. 16.9

Linear and radial endoscopic ultrasound images (7.5 MHz) of four distinct cholangiocarcinoma lesions in patients presenting with jaundice and bile duct stricture without a definite mass on computed tomography scan or magnetic resonance cholangiopancreatography. The masses are outlined in various colors. CBD, Common bile duct; PV, portal vein; SMA, superior mesenteric artery; SMV, superior mesenteric vein; ST, stent; TU, tumor.

Fig. 16.10

Cytology photomicrograph from a cholangiocarcinoma sampled by endoscopic ultrasound fine-needle aspiration after negative biliary brushings on endoscopic retrograde cholangiopancreatography. Glandular epithelium with nuclear enlargement, pleomorphism, and scant cytoplasm was diagnostic of malignancy (Diff Quick, ×40).

TABLE 16.3

Operating Characteristics of Endoscopic Ultrasound-Fine-Needle Aspiration in Biliary Strictures

Reference (Year) Number of Strictures Number of Strictures Confirmed Malignant a Hilar Biliary Strictures Overall Performance in All Strictures Sensitivity in Hilar Strictures (%)
Sensitivity (%) Specificity (%) PPV (%) NPV (%) Accuracy (%)
Fritscher-Ravens et al. (2000) 10 10 10 80 100 80
Rosch et al. (2002) 43 26 3 79 62 76 66
Lee et al. (2004) 42 24 1 47 100 100 50
Eloubeidi et al. (2004) 28 21 15 86 100 100 57 88 67
Fritscher-Ravens et al. (2004) 44 32 44 89 100 100 67 91 89
Rosch et al. (2004) 50 28 11 75 100 100 58 70 25
Byrne et al. (2004) 35 11 3 45 100 100
Meara et al. (2006) 46 30 87 100
DeWitt et al. (2006) 24 23 24 77 100 100 29 79 77
Saifuku et al. (2010) 34 17 0 94 82 84 93 88
Mohamadnejad et al. (2011) 81 81 30 73 100 59
Ohshima et al. (2011) 22 18 2 100 100 100 100 100
Nayar et al. (2011) 32 24 32 52 100 100 54 68 52
Tummala et al. (2013) 342 248 92 81 92
Weilert et al. (2014) 68 65 94 100 100 50 94
Tellez-Ávila et al. (2014) 39 28 39 79 100 100 42 82

a Based on surgical pathology, unequivocal cytology, or prolonged clinical follow-up.

TABLE 16.4

Performance Characteristics of Fine-Needle Aspiration in Biliary Strictures Based on Location of Stricture

Study Design Overall Sample Size Primary Stricture/Tumor Site and Number Per Site Diagnostic Sensitivity of Malignancy on FNA (Including Suspicious Interpretation) Stent Presence at Time of EUS
Rosch et al. (2004) Prospective 50 Hilar
3/11 (27%)
Eloubeidi et al. (2004) Prospective 28 Proximal
18/21 (86%) 27/28 (96%)
Lee et al. (2004) Retrospective 42 CHD
11/24 (47%) 40/42 (95%)
Byrne et al. (2004) Retrospective 35 CHD
9/14 (64%)
Fritscher-Ravens et al. (2004) Prospective 44 Hilar 44 32/36 (89%) 44/44 (100%)
DeWitt et al. (2006) Retrospective 24 Proximal 24 17/24 (71%)
Saifuku et al. (2010) Retrospective 34 Distal 34 16/17 (94%)
Mohamadnejad et al. (2011) Retrospective 81 Proximal
54/74 (73%) 64/74 (86%) a
Nayar et al. (2011) Retrospective 32 Proximal 32 24/32 (75%)
Weilert et al. (2014) Prospective 15 Proximal
11/15 (73%) b 8/51 (16%) c

Proximal tumors are those designated as hilar or occurring in the common hepatic duct. Distal tumors are those designated as distal or occurring in the common bile duct.

CBD, Common bile duct; CHD, common hepatic duct; EUS, endoscopic ultrasound; FNA, fine-needle aspiration.

a The diagnostic sensitivity of EUS FNA was 45 of 64 (70%) vs. 9 of 10 (90%) for patients with and without a stent, respectively.

b This increased to 13/15 (87%) when additional were from FNA of nonprimary sites (lymph node, liver lesion).

c Patients with pancreatic masses included in the general cohort ( n = 34) in addition to the patients with biliary strictures.

Fig. 16.11

(A) ERC image of a patient with mid bile duct stricture presenting with jaundice and proximal common bile duct dilation on magnetic resonance cholangiopancreatography. (B) Corresponding linear endoscopic ultrasound images of the patient in 11a confirming an irregular 19 mm bile duct mass without encasement of the portal vein. Fine-needle aspiration confirmed adenocarcinoma. PV, Portal vein; TU, tumor.

Figures courtesy of Dr. John DeWitt.

Video 16.5

Interrogation of the Bile Duct Using a Linear Echoendoscope in a Patient Presenting With Obstructive Jaundice

A solid distal bile duct mass is noted, in association with loss of the characteristic layers of the bile duct wall in that location. A bile duct stent appears as a hyperechoic structure in the lumen of the bile duct. The bile duct wall upstream from the stricture appears thickened.

Technique: Visualizing and sampling biliary strictures during EUS can pose a challenge to the endosonographer (see Fig. 16.9 ). Bile duct lesions are best visualized and sampled from the duodenum. Distal CBD lesions can sometimes be missed if the exam is only performed from the duodenal bulb. Therefore scope advancement to the second part of the duodenum followed by slow withdrawal through the duodenal sweep while maintaining full upward deflection is recommended to localize small distal and periampullary lesions. During FNA, maintaining a close apposition of the echoendoscope to the duodenal wall is essential to help stabilize the scope and minimize the amount of tissue the needle has to traverse. This position creates an angulation in the tip of the scope, and therefore 25-gauge needles are recommended due to ease of advancement through the scope. Additionally, 25-gauge needles have been shown to be equivalent to 22-gauge needles in their diagnostic accuracy, and may result in less bleeding during FNA, particularly when close to vascular structures like the portal vein and hepatic artery. If the clinical suspicion of a malignancy is high but no mass can be seen on EUS, targeting the stricture area with FNA under fluoroscopic guidance after CBD stenting might help improve tissue yield ( Video 16.6 ).

Video 16.6

Examination of the Head of the Pancreas Along With the Common Bile Duct Showed a Dilated Bile Duct Upstream From a Mass

Endoscopic retrograde cholangiopancreatography was completed prior to endoscopic ultrasound (EUS) with plastic stent placement across a distal bile duct stricture. On EUS, a segment of loss of interface between the mass and portal vein indicates tumor invasion. The site of fine-needle aspiration is confirmed to be within the stricture on fluoroscopy. Final cytology review confirmed adenocarcinoma.

Presence of biliary stents: A biliary stent was in place at the time of EUS in a significant proportion of patients undergoing assessment of biliary strictures (see Table 16.4 ). The presence of an indwelling biliary stent can impair EUS imaging due to stent-induced artifacts and sludge build up in its lumen ( Video 16.7 ). However, impaired imaging is probably of little impact on mass visualization and FNA as demonstrated by two studies. It is conceivable, though, that a stent may hinder the detection of a very small bile duct mass or prevent the detection of a very distal lesion in the bile duct in the setting of a metallic stent. Imaging can be optimized by examination of various locations, by limiting the amount of air insufflated that may pass through the stent into the duct, or by removing the stent prior to EUS whenever possible. The presence of a stent can facilitate detection of a biliary mass as the stent often courses through the lesion aiding detection, particularly with plastic stents ( Fig. 16.12 ).

Fig. 16.12

A linear exam of a hilar cholangiocarcinoma being aspirated. A plastic biliary stent is seen in the lumen of the duct which could help localize the lesions on endoscopic ultrasound. PV, Portal vein

Figure courtesy of Dr. John DeWitt.

Video 16.7

Linear Endoscopic Ultrasound (EUS) Examination of a Bile Duct Mass in a Patient Who Presented With Painless Jaundice and Underwent Endoscopic Retrograde Cholangiopancreatography Where a Fully Covered Metallic Stent Was Placed Ahead of the EUS Procedure

Despite artifacts from the stent, the bile duct mass was clearly visualized on EUS and staging confirmed no vascular invasion. Fine-needle aspiration cytology confirmed adenocarcinoma.

Cytopathologic considerations: Cytologic specimens from FNA of biliary strictures (manifesting as biliary thickening on EUS) or small solid lesions tend to be of low cellularity and highly contaminated by duodenal and pancreatic epithelium. Even when cellularity is adequate, other factors can limit the interpretation even by highly experienced cytopathologists. Dysplasia can be very difficult to discern or be misinterpreted, especially in the setting of reactive inflammation or due to the presence of a biliary stent. Using the same argument, well-differentiated carcinomas can be difficult to characterize from reactive atypia. Such limitations often lead to many FNA samples being interpreted as atypical or suspicious without positively confirming malignancy. A multimodality tissue sampling approach, as described earlier in this section, is essential to provide a solid tissue diagnosis in this group of patients.

Tumor seeding: The perceived risk and implications of tumor seeding following EUS FNA for cholangiocarcinoma continues to be a source of debate. Tumor seeding by disseminating tumor cells across the needle track has been reported following EUS and imaging-guided FNA of various sites. The risk of clinically apparent tumor seeding following FNA is estimated to be approximately 1/10,000 or less. However, the reported rates likely underestimate the true occurrence of seeding due to the high mortality and short survival of patients with pancreatobiliary malignancies. Additionally, deposits along the small needle track are practically indistinguishable from local tumor recurrence that occurs with time and therefore it can be difficult to prove the origin of the deposits with certainty. This hypothesis is supported in a study that evaluated the rates of peritoneal carcinomatosis in matched pancreatic adenocarcinoma cohorts diagnosed by either EUS or percutaneously guided FNA. Peritoneal carcinomatosis developed in one and seven patients (2% vs. 16%; P < .025) in the EUS FNA versus percutaneous FNA group, respectively. The findings indicate a potential difference in tumor seeding risk between biopsy approaches and potentially greater frequency of tumor seeding than historically appreciated. Similarly, a meta-analysis of eight studies identified tumor seeding in 3% of patients following hepatocellular carcinoma biopsy.

The clinical ramifications of needle track seeding were demonstrated in another study including 191 patients with hilar cholangiocarcinoma who underwent primary tumor FNA as part of liver transplant evaluation. A total of 16 patients underwent transperitoneal FNA (13 percutaneous, 3 EUS). During intraoperative staging, peritoneal deposits were discovered in only 14/175 (8%) patients who did not undergo FNA versus 5/6 (83%), with a positive preoperative FNA ( P = .01). On the other hand, El Chafic et al. reported no difference in progression-free or overall survival between patients with cholangiocarcinoma who underwent preoperative FNA and those who did not in patients undergoing curative intent resections. Although the presence of peritoneal metastasis can be explained by the disease stage and other factors not well assessed by the studies published so far, FNA of biliary masses has been since adopted as a contraindication to liver transplantation by many centers.

The Performance of Intraductal Ultrasonography in Biliary Strictures

With the advent of high-frequency (20 MHz) mini-probes over a guidewire, IDUS has emerged as a feasible and promising imaging technique in the diagnosis of biliary strictures. Mini-probes can be easily inserted through the papilla without the need for sphincterotomy in most cases and are capable of performing linear and radial imaging simultaneously in two or three dimensions in one scanning operation. Despite the limited penetration depth, a precise image of an intraductal lesion is often possible, allowing assessment of invasion or compression of adjacent structures. IDUS should be performed prior to drainage in order to avoid inflammatory artifacts, and therefore is best performed by ERCP experts during the same procedure. Literature demonstrates that IDUS can be advanced through biliary strictures in 86% to 100% of cases allowing a complete exam of the stricture, mostly without previous dilation. Most failures were due to tight strictures of the hilum or intrahepatic ducts that the guidewire could not traverse. As with EUS, three layers are seen in the bile duct wall with IDUS. The first hyperechoic layer corresponds to the mucosa in addition to a border echo; the second hypoechoic layer is the smooth muscle fiber with fibroelastic tissue; and the third hyperechoic layer is the thin and loose connective tissue with a border echo. The criteria for malignancy in a stricture were described as disruption of the normal three-layer sonographic pattern of the bile duct wall (outer echogenic, middle hypoechoic, inner echogenic), a hypoechoic infiltrating lesion with irregular margins, heterogeneous echo-poor areas invading surrounding tissue, and continuation of the main hypoechoic mass into adjacent structures ( Fig. 16.13 ). Findings considered diagnostic of a benign stricture include preservation of the normal three-layer sonographic wall pattern, homogeneous echo patterns, smooth margins, hyperechogenic lesions, and the absence of a mass lesion. Additional criteria proposed include interruption of the bile duct wall, presence of any sessile morphology of a tumor, and tumor size greater than 10 mm. The accuracy of IDUS in differentiating benign from malignant strictures ranges from 76% to 92% in series of patients with various causes of biliary strictures. The vast majority of patients without the above-mentioned criteria and with negative samplings do not have a malignant lesion with 95% accuracy and 100% NPV. The presence of two of the criteria, even with negative biopsies, is highly suspicious of malignancy. An additional negative predictive criterion added by Krishna et al. was wall thickness of ≤7 mm, which was associated with NPV of 100% in the absence of extrinsic compression. When recently compared to CT and EUS in a large surgical cohort, the combination of ERCP and IDUS provided a statistically superior accuracy in diagnosing malignant biliary strictures. Additionally, IDUS is very effective in confirming an extrinsic compression by a vascular structure or by a stone impacted in the cystic duct and compressing the CBD (Mirizzi’s syndrome). Biliary papillomatosis, frequently misdiagnosed with other imaging such as EUS, ERCP, and magnetic resonance imaging (MRI), appears on IDUS as normal biliary ducts with alternating areas covered by polypoid lesions protruding into the lumen ( Fig. 16.14 ). In 30 patients with cholangiocarcinoma studied by IDUS, biliary papillomatosis was shown in three (10%) and confirmed by biopsy or surgery. The clinical impact of this diagnosis can be important, as young patients with biliary papillomatosis without advanced cholangiocarcinoma should be treated with pancreaticoduodenectomy in combination with partial hepatectomy or liver transplantation. In primary sclerosing cholangitis (PSC) presenting with dominant strictures, IDUS has been traditionally considered no more accurate than other imaging modalities in the diagnosis of cholangiocarcinoma. However, recent studies show encouraging results. In a prospective study, 40 patients with PSC underwent ERCP with IDUS which was associated with a sensitivity, specificity, accuracy, positive predictive value (PPV), and NPV of 88%, 91%, 90%, 70%, and 97%, respectively, for predicting malignancy.

Fig. 16.13

Two-dimensional intraductal ultrasonography image of a small solid mass (white crosses) protruding in the lumen of the common bile duct consistent with early stage cholangiocarcinoma.

Fig. 16.14

Two-dimensional intraductal ultrasonography image showing biliary papillomatosis with intrahepatic polypoid spread (arrows).

Cholangioscopy in Biliary Strictures

As the design, maneuverability, and optical resolution of cholangioscopes continue to improve, peroral cholangioscopy is emerging as an important adjunct to ERCP in the assessment of biliary strictures, particularly the proximal ones. Direct visualization and targeted biopsy of bile duct lesions should be performed whenever possible during cholangioscopy. When compared with ERCP brush cytology, peroral cholangioscopy was 100% sensitive and 89% specific for biliary strictures and increased the diagnostic accuracy to more than 90% in earlier literature. In a Japanese multicenter trial, the accuracy of endoscopic retrograde cholangiography (ERC) alone, ERC with cholangioscopy, and ERC with cholangioscopy + biopsy for the diagnosis of bile duct malignancy were 74%, 84%, and 93%, respectively. More recently, Nguyen et al. reported on the utilization of EUS FNA before considering cholangioscopy in brushing-negative biliary strictures. The need for cholangioscopy was avoided in 60% of patients where EUS FNA provided tissue diagnosis, resulting in reduction of complications by 2.5% and in cost savings. However, in patients with proximal biliary strictures, the performance of EUS-FNA remains suboptimal. Siddiqui et al. demonstrated that cholangioscopy provided a definitive diagnosis in 77% of patients where ERCP-guided cytology brushing and EUS FNA were both inconclusive. Finally, a recent cost utility analysis demonstrated that SOC with targeted biopsies provided a more cost-effective strategy compared to ERCP with brushings and FISH in patients with PSC strictures. Based on the above, we recommend an EUS FNA approach first in distal biliary strictures, reserving cholangioscopy to mid and proximal biliary strictures with complementary EUS FNA.

A new digital cholangioscope is increasingly available and provides high-definition endoscopic images and allows direct tissue sampling. Navaneethan et al. reported on the use of this system in 98 patients in a multicenter trial. Superior views of the ductal lumen and mucosa were obtained in all 44 patients with indeterminate biliary strictures. Among the 44 patients who underwent SOC-guided biopsies, the specimen was adequate for histologic evaluation in 43 (98%). The sensitivity and specificity of SOC visual impression for diagnosis of malignancy was 90% and 96%, respectively. The sensitivity and specificity of SOC-guided biopsies for diagnosis of malignancy was 85% and 100%. Similar results were recently reported by Tanaka et al. from Japan. Cholangioscopy-related complications were reported to be as high as 7% ; however, more recent data from the digital SOC indicate an overall adverse event rate of 3% that were mainly postprocedural pancreatitis and cholangitis.

Multimodality Approach to Bile Duct Strictures

Because the performance of different diagnostic tests remains suboptimal, the decision concerning the use of the various imaging and tissue sampling modalities remains critical and often involves a combination of studies. In a prospective study of 142 patients with cholestasis and common hepatic duct dilatation of unclear etiology showed that a diagnostic algorithm with MRCP followed by EUS was highly sensitive and specific (90% and 98%, respectively) for the early diagnosis of extrahepatic bile duct carcinoma. The respective limitations and risks of EUS (+/−FNA) and ERCP + IDUS should be considered as well prior to the assessment of any biliary stricture. If the stricture is localized at the level of the CBD, EUS should be proposed after noninvasive imaging modalities, due to its excellent performance in distal biliary lesions and its ability to sample tissue ( Fig. 16.15 ). Needle track seeding should not be a concern while sampling distal biliary masses in surgically fit patients with highly suspected malignancy because the resection field encompasses the needle track. In patients with more proximal strictures, EUS and EUS FNA have several limitations and digital SOC, IDUS, and other ERCP-based tissue sampling techniques should be considered instead. In view of its low NPV in proximal strictures, EUS FNA should be reserved for negative or nondiagnostic ERCP brush cytology or cholangioscopy results and only if high probability of malignancy exists. Nevertheless, some authors propose the systematic addition of EUS FNA to ERCP brushings to optimize the diagnostic yield.

Feb 19, 2020 | Posted by in GASTROENTEROLOGY | Comments Off on Endoscopic Ultrasound in Bile Duct, Gallbladder, and Ampullary Lesions

Full access? Get Clinical Tree

Get Clinical Tree app for offline access