Fundamentals of ERCP

Fundamentals of ERCP Image Interpretation

Read Article


Successful cannulation of the desired duct and safely carrying out any necessary interventions are the primary goals in ERCP. It is also important to obtain the best possible images, accurately interpret these images, and document the findings in the procedure report. There are common, avoidable mistakes or errors that occur while performing ERCP. These include poor technique, misinterpretation of variant anatomy and artifacts, and failure to recognize adverse events, especially perforation.1 Many of the image interpretation errors arise with less commonly encountered anatomic variants of the biliary and the pancreatic ducts.2 Most ERCP malpractice lawsuits arise after adverse events such as pancreatitis, perforations, and severe infections. These lawsuits are difficult to defend if the procedure was not indicated or if there was a failure to adequately document the justification for the risk of the procedure.3,4 Therefore, all ERCP procedures should include documentation of appropriate and acceptable indications and detailed descriptions of the findings and the interventions. Images should be correctly interpreted, and any adverse events recognized and treated early. Capturing and saving images assists in “telling the story” of the entire procedure, illustrating the interventions performed, documenting the absence of adverse events or early recognition of adverse events. The images and the report should be congruent. As will be described in this article, fluoroscopy can be utilized to assist in correctly positioning the duodenoscope, identifying the major and the minor papilla, and selectively cannulating the duct of choice. 

Proper Use of Fluoroscopy and Safety Measures 

Every endoscopist performing ERCP should have a basic understanding of the fluoroscopy equipment, proper settings, and safety measures.5 Common errors, mistakes, and pitfalls in ERCP imaging are summarized in Table 1. It is not uncommon that the endoscopist will be required to directly operate the fluoroscopy equipment and therefore should have a good working understanding of how to generate and correctly interpret adequate fluoroscopic images in a safe manner. A helpful publication from the American Society for Gastrointestinal Endoscopy (ASGE) reviews in detail radiation and fluoroscopy safety in endoscopy.6 

There are several different types of fluoroscopy units, but ERCP is performed in most centers with a portable C-arm unit, although some high-volume centers have so-called fixed table units. The principles, however, are the same. An X-ray beam is passed from the X-ray source or cathode that is usually located below the fluoroscopy table or bed upward through the patient. An image intensifier or receiver unit receives the beam and generates the images. The X-ray beams may deflect causing scatter resulting in increased radiation exposure. There is increased scatter in obese patients. The image intensifier should always be placed as close to the patient as possible without contacting the patient or interfering with the ability to move the table or the fluoroscope. Typically, placing the image intensifier 1-2 inches above the patient is ideal for good image quality without excess magnification or increased scatter. The ALARA principle (As Low as Reasonably Achievable) should always be followed to limit the exposure to the patient and the staff. 

Endoscopists and assistants should always wear thyroid shields, lead glasses, and lead aprons that fully cover their body. Adding a lead skirt or drape around the fluoroscopy unit can help prevent scatter and reduce excess exposure as can clear lead shields positioned between the fluoroscopy unit and the endoscopy staff. Increased fluoroscopy time increases cancer risk to the patient, endoscopist and all staff present at the procedure. Angulation of the beam and magnification all increase radiation exposure to patients and staff. 

Physician-controlled fluoroscopy may result in reduced radiation exposure compared to utilizing an X-ray technician.7 There are clinical determinants that may predict longer fluoroscopy times and, hence, greater overall exposure, such as patient obesity, mechanical lithotripsy, needle knife use, and malignant biliary obstruction.8 With more ERCP experience, fluoroscopy time is typically shorter.9 Radiation time should, if possible, be noted, recorded, and compared with published benchmarks to ensure excess fluoroscopy time is not being utilized and to make it a goal to reduce exposure time as much as possible. It may be helpful to turn off the auto save mode of the fluoroscopy unit to avoid excess images to review at the conclusion of a case. Selected images should be saved and sent for permanent storage in the local picture archiving and communication system (PACS). The images saved should include the scout film and final film, and those images documenting the complete biliary tree or pancreatic duct as well as images that identify distinct findings or interventions. 

Room Design and Patient Position 

The room design, type of fluoroscopy equipment, and position of the patient and the monitors are all important for a successful procedure. This also helps ensure the comfort and safety of all in the room, including proper and healthy ergonomics for the endoscopist. The position of the fluoroscope and the position of the patient also affect the anatomic appearance as well as filling and drainage of the ducts of interest. Most ERCPs are performed with the patient in the prone or semi-prone position where the papilla is better positioned for the design of the duodenoscope and for most endoscopist’s natural scoping position. Most endoscopists position the biliary tree to the left side of the image as would correspond to the patient’s right side if the patient were supine and viewed from the front and how CT and MRI images are typically saved and reviewed. However, with the patient prone, this would not be consistent with how the image was acquired. Therefore, some endoscopists prefer the right corresponding to a posterior view when the patient is prone (see Figures 1 and 2 – page 26). I.e., most ERCP images are consistent with an anterior-to-posterior view, but some prefer a posterior-to-anterior view. 

Supine positioning of the patient may be necessary in certain cases such as: extreme obesity, poor neck mobility, an open abdomen, altered anatomy, or the need to perform a laparoscopic-assisted ERCP. With a patient in the supine position, many endoscopists look at the endoscopy and fluoroscopy monitoring screens with their back towards the patient. With this positioning, the endoscopist must maintain significant rightward torque on the duodenoscope. This can be accomplished with the aid of an assistant, if needed, while the endoscopist is facing away from the patient. 

The Importance and Utility of the Scout Film 

The baseline fluoroscopic image is also termed the “scout film” and it should be reviewed and saved routinely. The scout film serves several important purposes. The scout film confirms that the fluoroscopy equipment is functioning properly, ensures a properly focused field of view and position or orientation, demonstrated the initial bowel gas pattern (prior to insufflation) and verifies that the images can be saved (see Figures 3 and 4). The fluoroscopy image should be adjusted as needed to ensure the spine is vertically oriented, or close to it, and that there is a focused field of view that encompasses the expected area of interest (either biliary tree, pancreatic duct, or both). Such a proper focused field of exam, employing alternate angles of view, and the judicious use of magnification can reduce radiation exposure and interpretation errors in ERCP (see Figure 5). The scout film also identifies any impairment of the view of the duct or ducts. 

Artifacts during ERCP are common. Overlying objects such as EKG leads, wires from the pulse oximeter, blood pressure cuff, or IV lines that need to be moved out of the examination field will be revealed during the scout film. Excess bowel gas, retained contrast, the presence of any internal hardware or prosthesis including clips, coils, or percutaneous drains (see Figures 6 and 7) that may or may not interfere with the view or need to be taken into consideration should also be noted. Spinal hardware usually does not obscure the view of the biliary tree but may require some adjustment of the fluoroscopy angle (see Figures 8–12). Calcifications should be recognized. These may include calcified hemangiomas of the liver (see Figure 13), porcelain gallbladder (see Figures 14 and 15), calcific pancreatitis, phleboliths, renal lithiasis, or calcified lymph nodes to name a few. Overlying bowel gas and retained contrast (see Figures 16 and 17) can interfere with image creation, produce shadows, and can lead to misinterpreted images. When evaluating the scout film, careful notation should be made of the appearance of the diaphragm, bowel gas pattern, stomach, and the liver edge prior to any intervention. Pre-procedure and post-procedure scout films should always be obtained to look for, and, if needed, document the presence or absence of free air. 

It should be recognized that undue pressure from the endoscope or catheters can distort ductal anatomy as well as increase the risk of adverse events. Contrast material outside the ductal systems from prior contrast imaging studies or inadvertent extraductal injection should be appreciated. Under filling of the ducts or unintentional injection of air can also result in misinterpreted images. Oblique and lateral images can be helpful in showing abnormalities, variants of anatomy, or artifacts. Analog, as opposed to digital, fluoroscopic units will, in general, produce poorer images (see Figure 18). Misinterpretation of images is, unfortunately, very common but having a routine and following basic principles will reduce such errors. 

The Role Of Contrast Media 

In the context of ERCP, there is some debate about whether to use full-strength or diluted contrast media. Commonly, the initial cannulation begins with diluted contrast (typically half-strength). Then, after cannulation of the CBD an early cholangiogram is achieved (see Figure 19) some switch to full-strength contrast if necessary for better delineation of the biliary tree or pancreatic duct. Dilute contrast, however, is best for the detection of small stones, especially in patients with dilated ducts.10 (see Figure 20). Prior to the routine use of guidewire cannulation techniques, dilute contrast was often favored due to the belief that if inadvertent pancreatic duct cannulation was achieved, there would be less risk of post-ERCP pancreatitis. This notion has been discarded. 

Full-strength contrast is favored by some endoscopists for detection of strictures in the bile duct or in the pancreatic duct and when needing to clearly visualize the biliary and pancreatic duct anatomy. Prior to cannulation, all catheters should be flushed with contrast to eliminate air bubbles. If air bubbles are inadvertently introduced they can produce artifacts that can interfere with image interpretation. 

Using Fluoroscopy to Achieve Proper Scope Postion for Finding the Papilla and Successful Cannulation 

During ERCP, the scope tip is generally positioned below and lateral to the 12th rib, at about the level of the L2 vertebral body or between L2 and L3 vertebrae. This is considered the “usual position.” The body of the pancreas typically crosses the spine at the L1-L2 vertebrae. Understanding the scope position in the duodenum relative to the spine and ribs can help reveal a hidden or difficult to identify major papilla or when locating the minor papilla. Sometimes the major papilla is partially or totally obscured by an overlying duodenal fold and is not easily identified. It may be helpful to advance the duodenoscope into the deep second portion of the duodenum and then slowly withdraw it into the usual position just below the 12th rib between the L2 and L3 vertebrae. The major papilla is then typically found in this location, though gentle lifting of folds with a catheter or sphincterotome may be needed in some situations. 

For bile duct cannulation, the scope tip will typically have a “hockey stick” configuration (see Figures 21 and 22), whereas when cannulation of the pancreatic duct is desired, the scope tip is usually in a flatter position if not advanced to the long position (see Figure 23). For minor papilla cannulation, the scope is usually initially best positioned in a longer position with the tip more proximal in the duodenum. Once the PD is cannulated via the minor papilla, the duodenoscope can be reduced though this may risk loss of scope position. Once the CBD or PD are cannulated deeply, the scope position can typically be maintained (see Figures 24 and 25). Knowing the typical and expected anatomic position of the major papilla on fluoroscopy can also aid ERCP procedures in patients with surgically altered anatomy (see Figures 26–29). 

Utilizing Fluoroscopy to Maximize the Preferred Guidewire Cannulation 

Guidewire cannulation is now commonly employed during ERCP, and the use of injection during cannulation is much less frequently performed than it was in the past. Knowing the angle of both the bile duct and the pancreatic duct can assist in selective cannulation of the duct of choice. When CBD cannulation is desired, but the PD is inadvertently cannulated with the guidewire, the so-called “double wire technique” is a very useful and often successful method to achieve selective CBD cannulation (see Figure 30). The initial guidewire is left in the PD and a second guidewire is loaded into the sphincterotome (see Figure 31). Positioning the duodenoscope in the “hockey stick” position and observing a separation of the CBD and PD with a properly bowed sphincterotome under endoscopic and fluoroscopic guidance can also aid in selective cannulation of the CBD. 

Some have suggested that the first cholangiogram should be performed with the catheter or sphincterotome in the upper bile duct just below the confluence to prevent the inadvertent flushing of debris and stones or stone fragments from the distal extrahepatic duct into the intrahepatic ducts, although others prefer to inject distally first11 (see Figure 30). 

Body Position Effect on Duct Filling and Injection Tips 

Opacification of the biliary tree is, to some extent, dependent on gravity, and contrast media is denser than bile; therefore, the dependent ducts will fill preferentially. Whether the patient is in the semi-prone position or in the supine position will therefore determine which ducts are more dependent and will fill first. In the standard ERCP prone or semi-prone position the left ducts and anterior duct of the right lobe will fill first whereas in the supine position the right posterior ducts will fill first (see Figures 33 and 34). 

When using a balloon catheter, contrast should initially be injected with the balloon deflated. This will allow air bubbles or debris within the intrahepatic system to be flushed into the extrahepatic duct. The size of the balloon catheter selected should match the size of the duct to ensure appropriate occlusion and retention of contrast above the balloon and to facilitate a complete occlusion cholangiogram of the intrahepatic ducts (see Figure 35). 

In the semi-prone position, the left hepatic duct system is more dependent and fills earlier than the right ducts. In this position, the common bile duct is more posterior to the common hepatic duct and will fill earlier than the distal common bile duct. Furthermore, the cystic duct also generally courses posteriorly to the common hepatic duct with the anterior portion of the gallbladder filling first. With low-pressure contrast injections, there may be difficulty filling the right hepatic duct. If there is runoff into the gallbladder, the bile duct may be poorly opacified early. Balloon occlusion injection below the confluence will help fill the right hepatic ducts. Continuing to inject with the balloon inflated and moving distally in the duct can help delineate the distal duct. If the patient is moved into a more left lateral position, the right duct drains more preferentially. With the patient in the supine position, the right ducts are in a more dependent position and will fill preferentially as does the posterior gallbladder. Historically, if delayed biliary drainage was suspected, patients were placed supine with their head up though this is not commonly performed in the modern era. 

Balloon Sweeping and Proper Visualization of the Ducts 

Contrast injection is continued while slowly sweeping the duct with an occlusion balloon to remove biliary sludge, stones, or debris. Spot images can be captured and saved of the inflated balloon at the confluence (see Figure 36) then at the mid-bile duct, and finally in the distal CBD just above the papilla (see Figure 33). The mid-portion of the bile duct is commonly obscured by the duodenoscope and can be better visualized by gently pushing the duodenoscope into a long position with counterclockwise torque with care not to lose duct access (see Figure 38). Having a guidewire deep into the biliary tree and locked with an accessory locking device attached to the duodenoscope, use of elevator closure, or pinching the catheter or guidewire with the little finger of the endoscopist’s left hand reduces the risk of losing biliary access during this maneuver. Alternatively, the fluoroscopy C-arm can be rotated to expose the duct; but pushing the duodenoscope to the long position is typically more time efficient. Complete opacification of the biliary tree, including the intrahepatic ducts, can be performed to avoid missing intrahepatic stones, strictures, and anatomic variants. An exception to this may include cases of overt cholangitis where there is a genuine concern for the risk of precipitation of hepatic abscesses with high pressure injection. 

Discerning Bile Duct Stones Versus Air Bubbles or Pneumobilia 

Air introduced into the biliary or the pancreatic ducts can be problematic in that it can mimic stones, and both bubbles and stones are true “filling defects.” Air bubbles are commonly 2-5 mm in size, symmetrically round, and tend to cluster together or conform to the shape of the duct (see Figure 38). Tilting the patient may assist the endoscopist in distinguishing air bubbles from stones as the air bubbles usually rise in the duct with such maneuvers and bile duct stones will sink. However, floating stones do occur and can be misinterpreted as bubbles. This can result in missed identification of stones.12 Bubbles can also be long and tubular. Aside from stones and air bubbles, a filling defect may also be indicative of clots, tumor, or an intraductal parasite such as Ascaris. 

It is important to note that the use of imaging to identify stones or masses has its limitations. Dilute contrast may be best for visualization of small bile duct stones, especially in patients with dilated bile ducts.13 By paying close attention to early images on the initial contrast injection, the endoscopist can reduce the chance of missing stones. 

Larger and more dense radiopaque stones are easier to identify than smaller stones (see Figures 39–41). It is important to remember small stones can be missed in the distal bile duct at the papilla and may not be retrieved or extracted on balloon sweeps if the balloon slips past the stones without engaging them. Balloon occlusion cholangiogram and serial duct sweeps are the most effective technique for ensuring complete duct clearance (see Figure 42). Failure to perform multiple sweeps with an adequate balloon size relative to the duct size may result in inadequate clearance of the duct. The risk of inadequate duct clearance can be increased in patients with a dilated bile duct, in the presence of pneumobilia, following lithotripsy, and when a guidewire or stent is in the pancreatic duct (see Figure 43). 

Adequate Views and Sizing of Ducts 

It is important to adjust the endoscope position, and, if necessary, the fluoroscope itself to ensure that the entire duct of interest is visualized. For therapeutic interventions, it is important to know how to estimate the size of lesions, stones, and strictures. The size of stones and of the duct itself will influence the decision regarding removal techniques. These decisions include the size of sphincterotomy needed, the need for balloon sphincteroplasty, the size of any balloon or retrieval basket that may be needed. Knowing the insertion tube outer diameter of the duodenoscope is one of the easiest and quickest ways to estimate sizes. These vary by manufacturers and if pediatric or therapeutic duodenoscope from 7.5 mm to 12.1 mm.14 The most prevalent Olympus TJF190 therapeutic duodenoscope has an 11.3 mm outer insertion tube diameter while the Pentax duodenoscopes vary from 10.8 to 12.1 mm. Both the currently available single use disposable duodenoscopes also have outer insertion tube diameters of 11.3 mm.15 The endoscope diameter can serve as a “ruler” to compare any object to during the procedure. Large stones are considered those greater than 10 mm, so stones greater than the outer insertion diameter of the duodenoscope are, by definition, large stones (see Figure 41). 

Contrast streaming artifacts can occur when contrast flows along the dependent wall of a dilated duct. This may give the illusion of a normal duct caliber. However, obtaining a balloon occlusion cholangiogram will confirm the true duct size. Small periductal lymphatics sometimes fill with contrast especially during difficult cannulations in the setting of a tight stricture or if mucosal tears or false passages are created. 

Contrast filling of a duodenal diverticulum can sometimes cause confusion and obscure the view of the distal bile duct. The presence of an ampullary diverticulum should be noted during inspection of the papilla and subsequent filling of the diverticulum with contrast should be avoided if possible. Refluxed contrast filling of the duodenal bulb is common and may be mistaken for a partially filled gallbladder (see Figure 44). Excessive air in the stomach may make it difficult to pass the duodenoscope beyond the pylorus or difficult to maintain the proper scope position during ERCP maneuvers. A large “J shaped” stomach can also make it difficult to traverse the pylorus. Decompressing the stomach with suction and rotating the patient into a more left lateral position may help in this situation. Contrast refluxing backwards into the stomach or excessive air in the stomach may interfere with the interpretation of the pancreatogram due to overlying dye. 

Normal Bile Duct and Liver Anatomy 

Biliary anatomy can be quite variable.16 The extrahepatic bile duct is typically approximately 7-12 cm in length17 (see Figure 1). The portion of the bile duct below the cystic duct and above the papilla is, by convention, termed the common bile duct (CBD). The portion above the cystic duct and below the confluence of the right and left hepatic ducts is called the common hepatic duct (CHD) (see Figure 32). By convention, the distal CBD is that portion above the papilla and the proximal CBD is the portion nearest the liver. 

Duct diameters vary widely. In general, duct diameters are 1mm for each decade of life, until about age 60, in patients with an intact gallbladder. An easy rule of thumb is “7-11” where a CBD in a patient with an intact gallbladder of 7 mm or greater is considered dilated, and one that is 11 mm or greater post-cholecystectomy is dilated.18 Patients above age 60 may experience physiologic bile duct dilation in the absence of injury or illness. A large ultrasound (US) study found that the bile duct increased 0.4 mm/per year over age 50 and suggested a bile duct over 8.5 mm in an elderly individual would be considered abnormally dilated. However, there are some discrepancies between US and cholangiographic measures of bile duct diameter. Fluoroscopically, the bile duct also typically arises from the papilla at around the level of L2-L3 vertebrae and courses superiorly rightward into the liver (see Figure 45). 

Understanding the segments of liver and their relation to the branches of the biliary tree and anatomic variations is also important. In the Couinard classification of liver anatomy, there are 8 segments of the liver. Each segment is distinct with biliary drainage that parallels the portal drainage and can be defined by CT, MRI, and ERCP19,20 (see Figure 46). The most common hepatic ductal anatomy is a left hepatic duct (LHD) joining a confluence of the right posterior sectoral duct (RPSD) and right anterior hepatic duct (RAHD). The RPSD typically drains segments VI and VII, the RAHD segments VIII and V, and the LHD and its branches segments drain segments I, II, III and IV. 

Recognizing Biliary Tree Variants 

The most common biliary tree anatomic variants involve the RPSD.21 One of the most common is the RPSD coming off the LHD before its confluence with the RAHD (see Figures 47 and 48) and is present in about 15% of patients. In this variant the RPSD commonly passes above the portal vein creating a hump-like appearance before it crosses to its typical horizontal crossing. This is usually, but not always, posterior to the vertically coursing RAHD (see Figure 49). The next most common variant is when the RPSD does not pass the RAHD posteriorly, but it drains into the right side of the RAHD (see Figure 50). A segmental or accessory right hepatic duct that drains into the CHD or the cystic duct (CD) is also quite common. Rather than a typical bifurcation at the confluence; a trifurcation or triple confluence (triunion) of the proximal ducts is also relatively common (see Figures 51–53). Uncommon variants of the biliary tree include a CHD that may appear absent with a low union of the right and left hepatic ducts (see Figure 54). RPSD variants include an accessory LHD draining into the right anterior duct while the RPSD drains into the left accessory duct and left hepatic duct coming off the RAHD (see Figure 55). 

The cystic duct origin can be quite variable and may originate from any part of the biliary tree. Normally the CD arises from the CHD, defining the CBD below (see Figures 56 and 46). There are three common CD variants: a low CD insertion characterized by the CD fusing with the distal CBD (see Figures 57 and 58) and a CD that parallels the CHD (see Figure 59). An uncommon CD variant includes high insertion of the CD into the CHD (see Figure 60). Noting and alerting surgeons to the cystic duct variants can be quite helpful prior to cholecystectomy. The hepatic artery may cross over the bile duct and produce an indentation on the CBD that can mimic a stricture or tumor.

Biliary cysts, choledochocysts, and choledochoceles can cause confusion and misinterpretation on cholangiograms. The Todani classification (Table 2) is the commonly used system for classification of bile duct cysts. Type I cysts are the most common (90%) and have three variations. Type Ia is dilation of the entire extrahepatic bile duct (see Figures 61 and 62). 

Type Ib is focal dilation of the extrahepatic bile duct (see Figure 63). Type III is a dilation of the extrahepatic bile duct within the duodenal wall (a.k.a. a choledochocele), and is typically treated via biliary sphincterotomy. Type IV cysts are the second most common. Type V, also known as Caroli’s disease, involves multiple dilations or cysts of the intrahepatic ducts only.

Most often, the gallbladder is somewhat pear-shaped. However, it can have an hourglass shape, have septations, or have a Phrygian cap that folds over the gallbladder. The location of the gallbladder can be variable: high or intrahepatic, low, “left sided,” congenitally absent, or multiple. 

Normal Pancreatic Duct and Variants 

The pancreatic duct (PD) is formed by the fusion of the dorsal and ventral pancreatic anlagen in utero. The dorsal duct (Duct of Santorini) and the ventral duct (Duct of Wirsung) merge, with the main pancreatic duct usually emptying at the major papilla. The minor papilla is found superior and lateral to the major papilla. The CBD and the PD usually join before entering the papilla, typically within 2-3 mm of the papilla, with CBD usually superior to the PD and closer to the duodenal wall. The pancreas itself usually lies between the T12 and L2 vertebrae with the mid-body of the pancreas over L1 (see Figure 64). The PD varies in length from about 9 cm to 15 cm. It is typically 4 mm in diameter in the head, 2-3 mm in diameter in the body and 1-2 mm in diameter in the tail. Like the CBD, it may be longer and larger in diameter with increased age. The duct typically has an ascending course before crossing over the spine, but it may be more horizontal, sigmoid shaped, or descending in course and contour. A looped duct can be present as well (a.k.a. ansa pancreatica). Sometimes a pancreatic duct branch may ascend to the duct of Santorini or descend to the uncinate portion of the pancreas. 

Congenital variants of the pancreaticobiliary tree include anomalous pancreaticobiliary duct union or junction; pancreas divisum, annular pancreas, an Ansa variant (Ansa pancreatica) (see Figure 65). Very rarely the pancreatic duct can be bifid or trifid 22 (see Figure 66). 

Pancreas divisum, where the dorsal and ventral pancreatic ducts fail to fuse in utero, is relatively common. Pancreas divisum can be complete or incomplete (see Figure 67). It is estimated to be present in 5-12% of all adults. It is usually easily recognized when the ventral duct is completely opacified and there is failure to opacify the dorsal or main pancreatic duct. Pancreas divisum can be confirmed by opacifying the dorsal duct at the minor papilla, though this is often unnecessary and may increase the risk of pancreatitis or adverse events. On the other hand, over injection of the ventral duct in divisum can cause acinar filling especially if divisum was not anticipated. A Santorinicele is a small cystic dilation of the dorsal pancreatic duct at the minor papilla in pancreas divisum. When present the duct of Santorini may be dilated23,24 (see Figure 68). 

Annular pancreas is a rare congenital anomaly. When present, the pancreas partially or completely encircles the duodenum. It will have a characteristic appearance of the pancreatic duct encircling the duodenum and often the duodenoscope before heading out to the body and tail of the pancreas (see Figures 69 and 70). Pancreas divisum coexists in as many as 45% of adults with annular pancreas. It is estimated that at least 1/3 of these individuals may suffer chronic pancreatitis25,26 (See Figure 71). 

Bile Duct Strictures And Tumors 

A contracted biliary sphincter may mimic a stricture or distal CBD stone. Transient narrowing or tapering of the distal duct in the absence of upstream ductal dilation argues against the presence of a true stricture. However, both benign and malignant strictures are commonly encountered during ERCP. Distinguishing between the two can be difficult at times and such strictures are commonly termed indeterminate. There are several possible causes of biliary strictures.27 Intrahepatic duct or CHD dilation is more common in malignant strictures but whether the stricture is smooth in contour or irregular does not adequately distinguish etiology.28 Malignant strictures tend to be much longer and irregular with shelf-like edges (see Figure 72). Strictures from ampullary cancers tend to be short and smooth (see Figure 73). Extrinsic compression from tumors or adenopathy can result in marked biliary ductal dilation (see Figure 75). Post-liver transplant strictures are usually short and often smooth but may be asymmetric or have a shelf-like margin as well (See Figure 76). Pancreatic cancer in the head of the pancreas causes the most common malignant biliary stricture, usually a severe distal stenosis 2-4cm long (see Figures 77–84). Iatrogenic strictures commonly include post anastomotic after liver transplantation or bile duct resection and post cholecystectomy bile duct injuries (BDI). Stones can occur above anastomotic strictures (see Figures 87–96). 

Whenever a biliary stricture is noted and a clear, benign underlying cause is not suspected, sampling should be performed that should include brushings for cytology and/or biopsy. Digital cholangioscopy with direct visualization and directed biopsies have higher yields than brushing alone. 

Primary sclerosing cholangitis is characterized by the presence of focal or multifocal strictures, pruning or rarefaction of the biliary tree, ductal irregularities, beading or saccular dilations of the intrahepatic ducts commonly alternating with segmental strictures29 (see Figures 97–101). Secondary sclerosing cholangitis also can be seen (see Figures 102–107). 

The Bismuth-Corlette classification has been the most widely used system for bile duct tumors and benign strictures with a modification used for main hepatic duct injury. However, its prognostic value has been called into question in more recent years30 (see Table 3). 

Bismuth Type IV lesions extend to and involves both the right and left hepatic ducts to the second order hepatic ducts; and is commonly referred to as a Klatskin tumor (see Figure 108). Treatment of Klatskin tumor is rarely surgical and requires bilateral stenting (see Figure 109). Bismuth Type V lesions produce a stricture involving both the common bile duct and the cystic duct.31 Cholangiocarcinoma is more accurately classified as intraductal iCCA, perihilar pCCA or distal dCCA.32 A complete occlusion cholangiogram of the biliary tree should be obtained and documented in cases of known or suspected biliary malignancies or PSC (see Figures 110 and 111). Careful delineation and documentation of the extent of ductal involvement with high quality cholangiograms is important for treatment decisions especially surgical candidacy. Not only is digital cholangioscopy often warranted to obtain directed biopsies for confirmatory diagnosis but is often requested by the surgeon preoperatively (see Figure 112). 

Bile Duct Injuries (BDI) and Leaks 

At the time of ERCP, it is critical to recognize any biliary or pancreatic duct injury, extravasation, leaks, and perforations. It is important to clearly visualize and document the sites of any leaks and/or bile duct injuries (BDI) prior to making decisions on treatment options. Many leaks and BDI are readily apparent on cholangiogram. When a post-surgical bile leak or BDI is suspected to have occurred but is not clearly seen, then a balloon occlusion cholangiogram can be performed. Common areas for bile leaks to occur in the biliary tree are in the cystic duct remnant following cholecystectomy (see Figures 113–117), the CHD region (see Figure 117), the duct of Luschka (see Figure 118) or in low lying, right sided, proximal intrahepatic ducts that overlie the gallbladder fossa. 

Careful attention should be paid to avoid confusing the cystic duct stump with the hepatic duct in cases of post-cholecystectomy stenosis. Stenosis in the biliary tree, consistent with BDI, following cholecystectomy is usually at the level of the CHD. This is often complete and difficult to pass even with a small diameter hydrophilic guidewire. A common mistake is confusing the CD stump and an occluded CHD and then repeatedly pushing a guidewire into the cystic duct stump. The two ducts are commonly superimposed on fluoroscopy, so changing the angle or axis of fluoroscopy may be necessary to distinguish the two.

The Strasberg classification of bile duct injuries is a widely used system to define the injuries by the location33 (see Table 4). Type E is an injury to the main hepatic duct that is further classified according to the Bismuth system34 (See Figures 119 –123). 

Penetrating trauma to biliary tree is rare, but can be seen in patients with hepatic gunshot wounds (see Figure 124). 

There are rare reports of portal vein (PV) cannulation as the PV runs parallel to the CBD.35 Resistance to guidewire or catheter placement, rapid disappearance of contrast, and opacification of the PV branches may be clues to PV cannulation.36 PV cannulation most commonly occurs in patients with cholangiocarcinoma or other malignancy that weakens the biliary wall and allows communication between the biliary tree and the portal venous system. 

Pancreatic Duct Strictures, Injuries, and Leaks 

Pancreatic duct strictures are common in patients with chronic pancreatitis (see Figure 125), pancreatic cancer, and patients with prior pancreatic surgery who have ductal anastomoses (see Figures 126 and 127). PD leaks, disruption, or a disconnected PD may occur from severe acute pancreatitis with necrosis or after pancreatic surgery (see Figures 128 and 129). PD injury can also occur because of ERCP. Traumatic injuries to pancreas can occur, most commonly following blunt force trauma such as in motor vehicle accidents. PD injuries from trauma are relatively uncommon, occurring in only about 2% blunt trauma cases. Most blunt trauma injuries to the pancreas occur in the junction of the body and tail, where the gland is compressed against the spine posteriorly, causing a crush injury and resulting in partial or complete pancreatic duct transection.37 Penetrating injuries to the pancreas from gunshots or stab wounds are more common than blunt trauma injuries. A classification of PD injuries associated with ERCP that is commonly utilized by endoscopists is described by Takishima38 (see Table 5 and Figure 130). 

Duodenal and Ductal Injuries and Perforations 

Duodenal injuries are a risk of ERCP. The endoscopist should be familiar with the Stapfer classification of duodenal perforations39 (see Table 6). Type 1 is endoscopy related and invariably requires surgery carrying a substantial risk of morbidity and mortality. The best outcome 

in a type I perforation is when this is identified and intervened upon early.40 Type II is the most common and is a periampullary perforation often related to sphincterotomy and may be treated with endoscopic clips or by placing a covered metal biliary stent. Type III is a ductal or duodenal perforation caused by endoscopic instruments but not a guidewire. Type III is less common but can require surgery. Type IV is the presence of retroperitoneal air due to a guidewire puncture. Types II and IV duodenal perforations rarely require surgery. Being aware of and avoiding each of these perforations as well as recognizing their appearance both endoscopically and fluoroscopically are key to early endoscopic or surgical intervention and better patient outcomes.41 

All endoscopists performing ERCP and advanced endoscopic therapeutic interventions must be familiar with and comfortable recognizing free air on fluoroscopy (see Figures 131 and 132). If a perforation is suspected, an upright chest X-ray should be done immediately as this is more sensitive than an abdominal X-ray for the detection of free air. Free air under the diaphragm is easily seen on an upright chest X-ray (see Figures 133 and 134). Rigler’s sign is present when both sides of bowel wall are well defined due to free intra-abdominal air adjacent to gas-filled loops of bowel.42 Sometimes only a subtle triangle of free air can be visible outside the bowel wall as a sign of pneumoperitoneum. 


The keys to avoiding imaging pitfalls and mistakes during ERCP are to routinely perform a pre-procedure scout film, always obtain high-quality balloon occlusion cholangiograms and pancreatograms with adequate filling and visualizations of the entire biliary tree (or pancreatic duct if indicated). Strictures should be recognized, well defined, and sampled. All images should be reviewed in real time as well as after the procedure, with selected images saved that “tell the story” of the exam, findings, and interventions. Detailed descriptions of the findings including of variant anatomy, and interventions should also be documented in the ERCP report. Adverse events should be recognized early and treated appropriately. 


  1. Fink AS, Perez de Ayala V, Chapman M, Cotton PB. Radiologic pitfalls in endoscopic retrograde pancreatogra­phy. Pancreas. 1986;1(2):180-7. doi: 10.1097/00006676- 198603000-00013. PMID: 3575303. 
  2. Gulliver DJ, Cotton PB, Baillie J. Anatomic vari­ants and artifacts in ERCP interpretation. AJR Am J Roentgenol. 1991 May;156(5):975-80. doi: 10.2214/ ajr.156.5.2017963. PMID: 2017963. 
  3. Cotton PB. Analysis of 59 ERCP lawsuits; mainly about indications. Gastrointest Endosc. 2006 Mar;63(3):378- 82; quiz 464. doi: 10.1016/j.gie.2005.06.046. PMID: 16500382. 
  4. Cotton PB. Twenty more ERCP lawsuits: why? Poor indications and communications. Gastrointest Endosc. 2010 Oct;72(4):904. doi: 10.1016/j.gie.2010.01.058. PMID: 20883872. 
  5. Adler DG, Lanke G. Training in the Use of Fluoroscopy for Gastrointestinal Endoscopy. Successful Training in Gastrointestinal Endoscopy. Second Edition. Edited by Jonathan Cohen. 2022 John Wiley & Sons Ltd. 
  6. Kwok K, Hasan N, Duloy A, Murad F, Nieto J, Day LW. American Society for Gastrointestinal Endoscopy radia­tion and fluoroscopy safety in GI endoscopy. Gastrointest Endosc. 2021 Oct;94(4):685-697.e4. doi: 10.1016/j. gie.2021.05.042. Epub 2021 Aug 13. PMID: 34399965. 
  7. Kakodkar S, Haider A, Hoff RT, Zamfirova I, Chi K. Reduced Fluoroscopy Time with Physician- Controlled Fluoroscopy During Endoscopic Retrograde Cholangiopancreatography: A Community Hospital Experience. Cureus. 2021 Mar 8;13(3):e13771. doi: 10.7759/cureus.13771. PMID: 33842147; PMCID: PMC8026404. 
  8. Oh CH, Dong SH, Kim JW, Kim GA, Lee JM. Radiation exposure during endoscopic retrograde cholangiopancre­atography according to clinical determinants. Medicine (Baltimore). 2020 Mar;99(13):e19498. doi: 10.1097/ MD.0000000000019498. PMID: 32221072; PMCID: PMC7220418. 
  9. Jorgensen JE, Rubenstein JH, Goodsitt MM, Elta GH. Radiation doses to ERCP patients are significantly lower with experienced endoscopists. Gastrointest Endosc. 2010 Jul;72(1):58-65. doi: 10.1016/j.gie.2009.12.060. Epub 2010 Apr 24. PMID: 20421102; PMCID: PMC2906113.
  10. Mishkin D, Carpenter S, Croffie J, Chuttani R, DiSario J, Hussain N, Liu J, Somogyi L, Tierney W, Petersen BT; Technology Assessment Committee, American Society for Gastrointestinal Endoscopy. ASGE Technology Status Evaluation Report: radiographic contrast media used in ERCP. Gastrointest Endosc. 2005 Oct;62(4):480- 4. doi: 10.1016/j.gie.2005.07.006. PMID: 16185956. 
  11. Dietrich CF, Bekkali NL, Burmeister S, Dong Y, Everett SM, Hocke M, Ignee A, On W, Hebbar S, Oppong K, Sun S, Jenssen C, Braden B. Controversies in ERCP: Technical aspects. Endosc Ultrasound. 2022 Jan- Feb;11(1):27-37. doi: 10.4103/EUS-D-21-00102. PMID: 34677144; PMCID: PMC8887038. 

12. Kamel PL, Vanagunas A. Floating common bile duct stones. A cause of a false negative endoscopic ret­rograde cholangiogram. J Clin Gastroenterol. 1991 Aug;13(4):457-9. doi: 10.1097/00004836-199108000- 00019. PMID: 1918854. 

13. Mortelé KJ, Ros PR. Anatomic variants of the biliary tree: MR cholangiographic findings and clinical applica­tions. AJR Am J Roentgenol. 2001 Aug;177(2):389-94. doi: 10.2214/ajr.177.2.1770389. PMID: 11461869. 

14. ASGE Technology Committee, Varadarajulu S, Banerjee S, Barth BA, Desilets DJ, Kaul V, Kethu SR, Pedrosa MC, Pfau PR, Tokar JL, Wang A, Wong Kee Song LM, Rodriguez SA. GI endoscopes. Gastrointest Endosc. 2011 Jul;74(1):1-6.e6. doi: 10.1016/j.gie.2011.01.061. PMID: 21704803 

15. Peter S, Bang JY, Varadarajulu S. Single-use duodeno­scopes: where are we and where are we going? Curr Opin Gastroenterol. 2021 Sep 1;37(5):416-420. doi: 10.1097/ MOG.0000000000000755. PMID: 34010243; PMCID: PMC8373446. 

16. Hribernik M, Gadzijev EM, Mlakar B, Ravnik D. Variations of intrahepatic and proximal extrahe­patic bile ducts. Hepatogastroenterology. 2003 Mar- Apr;50(50):342-8. PMID: 12749217. 

17. Chuah KB, Yap CK, Ng HS. Extrahepatic bile duct length in the Singapore population. Singapore Med J. 2001 Apr;42(4):165-9. PMID: 11465316. 

18. Bachar GN, Cohen M, Belenky A, Atar E, Gideon S. Effect of aging on the adult extrahepatic bile duct: a sono­graphic study. J Ultrasound Med. 2003 Sep;22(9):879- 82; quiz 883-5. doi: 10.7863/jum.2003.22.9.879. PMID: 14510259. 

19. Gazelle GS, Lee MJ, Mueller PR. Cholangiographic segmental anatomy of the liver. Radiographics. 1994 Sep;14(5):1005-13. doi: 10.1148/radiograph­ics.14.5.7991810. PMID: 7991810. 

20. Puente SG, Bannura GC. Radiological anatomy of the biliary tract: variations and congenital abnormalities. World J Surg. 1983 Mar;7(2):271-6. doi: 10.1007/ BF01656159. PMID: 6868640. 

21. Alghamdi HM, Almuhanna AF, Aldhafery BF, AlSulaiman RM, Almarhabi A, AlQurain A. The Prevalence of Hjortsjo Crook Sign of Right Posterior Sectional Bile Duct and Bile Duct Anatomy in ERCP. Can J Gastroenterol Hepatol. 2017; 2017:2532610. doi: 10.1155/2017/2532610. Epub 2017 Jul 12. PMID: 28785551; PMCID: PMC5529631. 

22. Borghei P, Sokhandon F, Shirkhoda A, Morgan DE. Anomalies, anatomic variants, and sources of diag­nostic pitfalls in pancreatic imaging. Radiology. 2013 Jan;266(1):28-36. doi: 10.1148/radiol.12112469. PMID: 23264525. 

23. Chalazonitis NA, Lachanis BS, Laspas F, Ptohis N, Tsimitselis G, Tzovara J. Pancreas divisum: mag­netic resonance cholangiopancreatography findings. Singapore Med J. 2008 Nov;49(11):951-4; quiz 955. PMID: 19037565. 

24. Manfredi R, Costamagna G, Brizi MG, Spina S, Maresca G, Vecchioli A, Mutignani M, Marano P. Pancreas divisum and “santorinicele”: diagnosis with dynamic MR cholangiopancreatography with secretin stimulation. Radiology. 2000 Nov;217(2):403-8. doi: 10.1148/radiol­ogy.217.I00nv29403. PMID: 11058635. 

25. Lehman GA, O’Connor KW. Coexistence of annu­lar pancreas and pancreas divisum–ERCP diagnosis. Gastrointest Endosc. 1985 Feb;31(1):25-8. doi: 10.1016/ s0016-5107(85)71960-8. PMID: 3979764. 

26. Gromski MA, Lehman GA, Zyromski NJ, Watkins JL, El Hajj II, Tan D, McHenry L, Easler JJ, Tirkes T, Sherman S, Fogel EL. Annular pancreas: endoscopic and pancreatographic findings from a tertiary referral ERCP center. Gastrointest Endosc. 2019 Feb;89(2):322- 328. doi: 10.1016/j.gie.2018.09.008. Epub 2018 Sep 18. PMID: 30240880; PMCID: PMC6679934.

27. Singh A, Gelrud A, Agarwal B. Biliary strictures: diagnostic considerations and approach. Gastroenterol Rep (Oxf). 2015 Feb;3(1):22-31. doi: 10.1093/gastro/ gou072. Epub 2014 Oct 28. PMID: 25355800; PMCID: PMC4324869. 

28. Bain VG, Abraham N, Jhangri GS, Alexander TW, Henning RC, Hoskinson ME, Maguire CG, Lalor EA, Sadowski DC. Prospective study of biliary stric­tures to determine the predictors of malignancy. Can J Gastroenterol. 2000 May;14(5):397-402. doi: 10.1155/2000/467567. PMID: 10851279. 

29. Khoshpouri P, Habibabadi RR, Hazhirkarzar B, Ameli S, Ghadimi M, Ghasabeh MA, Menias CO, Kim A, Li Z, Kamel IR. Imaging Features of Primary Sclerosing Cholangitis: From Diagnosis to Liver Transplant Follow-up. Radiographics. 2019 Nov- Dec;39(7):1938-1964. doi: 10.1148/rg.2019180213. Epub 2019 Oct 18. PMID: 31626561. 

30. Paul A, Kaiser GM, Molmenti EP, Schroeder T, Vernadakis S, Oezcelik A, Baba HA, Cicinnati VR, Sotiropoulos GC. Klatskin tumors and the accuracy of the Bismuth-Corlette classification. Am Surg. 2011 Dec;77(12):1695-9. PMID: 22273233. 

31. Sainani NI, Catalano OA, Holalkere NS, Zhu AX, Hahn PF, Sahani DV. Cholangiocarcinoma: current and novel imaging techniques. Radiographics. 2008 Sep- Oct;28(5):1263-87. doi: 10.1148/rg.285075183. PMID: 18794305. 

32. Banales JM, Marin JJG, Lamarca A, Rodrigues PM, Khan SA, Roberts LR, Cardinale V, Carpino G, Andersen JB, Braconi C, Calvisi DF, Perugorria MJ, Fabris L, Boulter L, Macias RIR, Gaudio E, Alvaro D, Gradilone SA, Strazzabosco M, Marzioni M, Coulouarn C, Fouassier L, Raggi C, Invernizzi P, Mertens JC, Moncsek A, Rizvi S, Heimbach J, Koerkamp BG, Bruix J, Forner A, Bridgewater J, Valle JW, Gores GJ. Cholangiocarcinoma 2020: the next horizon in mechanisms and management. Nat Rev Gastroenterol Hepatol. 2020 Sep;17(9):557- 588. doi: 10.1038/s41575-020-0310-z. Epub 2020 Jun 30. PMID: 32606456; PMCID: PMC7447603. 

33. Todani T, Watanabe Y, Narusue M, Tabuchi K, Okajima K. Congenital bile duct cysts: Classification, operative procedures, and review of thirty-seven cases including cancer arising from choledochal cyst. Am J Surg. 1977 Aug;134(2):263-9. doi: 10.1016/0002-9610(77)90359-2. PMID: 889044. 

34. Strasberg SM, Hertl M, Soper NJ. An analysis of the problem of biliary injury during laparoscopic chole­cystectomy. J Am Coll Surg. 1995 Jan;180(1):101-25. PMID: 8000648. 

35. Kalaitzakis E, Stern N, Sturgess R. Portal vein cannula­tion: an uncommon complication of endoscopic retro­grade cholangiopancreatography. World J Gastroenterol. 2011 Dec 14;17(46):5131-2. doi: 10.3748/I46.5131. PMID: 22171149; PMCID: PMC3235598.

36. Fukuda K, Sasaki T, Sasahira N. Portal vein guidewire insertion: rare complication of wire-guided cannulation during single-balloon enteroscopy-assisted endoscopic retrograde cholangiopancreatography. Dig Endosc. 2021 Jan;33(1I3-e4. doi: 10.1111/den.13843. Epub 2020 Oct 28. PMID: 33118241

36.    Debi U, Kaur R, Prasad KK, Sinha SK, Sinha A, Singh K. Pancreatic trauma: a concise review. World J Gastroenterol. 2013 Dec 21;19(47):9003-11. doi: 10.3748/wjg.v19.i47.9003. PMID: 24379625; PMCID: PMC3870553. 

37.    Takishima T, Hirata M, Kataoka Y, Asari Y, Sato K, Ohwada T, Kakita A. Pancreatographic classification of pancreatic ductal injuries caused by blunt injury to the pancreas. J Trauma. 2000 Apr;48(4):745-51; discus­sion 751-2. doi: 10.1097/00005373-200004000-00026. PMID: 10780612. 

38.    Stapfer M, Selby RR, Stain SC, Katkhouda N, Parekh D, Jabbour N, Garry D. Management of duodenal perforation after endoscopic retrograde cholangio­pancreatography and sphincterotomy. Ann Surg. 2000 Aug;232(2):191-8. doi: 10.1097/00000658-200008000- 00007. PMID: 10903596; PMCID: PMC1421129. 

39.    Cirocchi R, Kelly MD, Griffiths EA, Tabola R, Sartelli M, Carlini L, Ghersi S, Di Saverio S. A systematic review of the management and outcome of ERCP related duodenal perforations using a standardized classification system. Surgeon. 2017 Dec;15(6):379-387. doi: 

40.    Bill JG, Smith Z, Brancheck J, Elsner J, Hobbs P, Lang GD, Early DS, Das K, Hollander T, Doyle MBM, Fields RC, Hawkins WG, Strasberg SM, Hammill C, Chapman WC, Edmundowicz S, Mullady DK, Kushnir VM. The importance of early recognition in manage­ment of ERCP-related perforations. Surg Endosc. 2018 Dec;32(12):4841-4849. doi: 10.1007/s00464-018-6235- 8. Epub 2018 May 16. PMID: 29770887. 

41.    Lewicki AM. The Rigler sign and Leo G. Rigler. Radiology. 2004 Oct;233(1):7-12. doi: 10.1148/ radiol.2331031985. Epub 2004 Aug 27. PMID: 15333763. 

Download Tables, Images & References