ERCP was first reported by McCune et al. in 1968 [37]. Since that time, ERCP has transformed from a mere diagnostic test to a predominantly therapeutic procedure . In the USA alone, approximately half a million ERCP procedures are performed annually. Currently, ERCP is the most widely used endoscopic procedure for evaluation of bile duct strictures [38].

On ERCP, certain cholangiographic features are suggestive of malignancy . Reported features associated with malignancy include longer length of the stricture, an abrupt transition point, irregular margins, shelf-like appearance and asymmetric narrowing of the stricture (Figs. 7.3 and 7.4) [39, 40]. Two studies have suggested that in patients with biliary stricture, intrahepatic ductal dilatation is more likely to be seen in the setting of malignancy [39, 41]. Concentric appearance and smooth transition of a stricture, on the other hand, are suggestive of a benign underlying process (Figs. 7.5 and 7.6) [31, 32]. Cholangiographic appearance of a stricture alone (without historical or clinical data) has been reported to have a sensitivity ranging from 11 to 74 % and a specificity ranging from 63 to 100 % for differentiation of benign from malignant strictures [40, 42–44].

Other studies have suggested that the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy of cholangiography (ERCP or percutaneous) in diagnosing malignancy are about 74–85, 70–75, 74–79, 70–82 and 72–80 % respectively [38]. The low accuracy rates of cholangiography in diagnosing malignancy have stimulated research in tissue acquisition and advanced imaging techniques [38].

Endoscopic retrograde brush cytology was first described by Osnes et al. at the University of Oslo in 1975 [46, 47]. Nowadays in patients with a biliary stricture, brush cytology is often performed during therapeutic ERCP. Endoscopic brush cytology during ERCP is safe, does not require special expertise and adds little to the cost of ERCP. It has therefore become the preferred initial method of pursuing a diagnosis in many patients with a biliary stricture.

The technique for endoscopic retrograde brush cytology in many institutions, including ours, is standardized. Under fluoroscopic guidance, the brush and its sheath are inserted into the duct of interest over a guidewire and positioned just distal to the stricture. The brush is then advanced from the sheath to a point proximal to the stricture and moved across the stricture in a to-and-fro manner approximately 10 times (Fig. 7.7) [48, 49]. The brush is then withdrawn into the sheath, and both are subsequently withdrawn from the endoscope as a single unit [48, 49]. The brush segment of the brushing device is cut from the supporting wire, placed in a preservative solution and transported to cytology laboratory.

Brush cytology allows easy and convenient sampling and has a low complication rate [50–52]. The diagnostic specificity of biliary brush cytology is very high and few false-positive diagnoses have been reported [50, 53]. The major limitation of this technique has been the relatively modest diagnostic sensitivity, ranging from 10 to 50 % in most series [50, 53].

The variation in reported sensitivity of brush cytology across studies is in part because of differences in patient populations. For example, brush cytology has higher sensitivity in patients with evidence of a mass on cross-sectional imaging studies [48, 54]. Another factor affecting the variation in reported sensitivity is inconsistent categorization of cytology diagnoses as positive versus negative test results. In most institutions, including ours, the brush cytology results are grouped in four categories: benign, atypical, suspicious for malignancy or malignant. Some investigators have classified equivocal (e.g. atypical or suspicious for malignancy) diagnoses as positive for the presence of malignancy , whereas others have considered equivocal diagnoses as negative for malignancy. Regardless of classification or patient population, the sensitivity of brush cytology for detection of malignancy remains disappointingly low, while specificity is excellent. In other words, a positive result on brush cytology can be trusted, while a negative test is not trustworthy.

There have been attempts to improve the sensitivity of brush cytology obtained during ERCP. Physical changes to the brushing device itself such as use of longer and stiffer brushes have not improved sensitivity [55]. Balloon dilatation of strictures, to expose underlying tissue, prior to obtaining brush samples has been tried but also not shown to be beneficial [56]. Mutation analysis of the cells obtained by brushing does not seem to improve diagnostic accuracy [57], and DNA methylation analysis of brush specimens has shown only small benefit [50].

Recently fluorescent in situ hybridization (FISH) studies on brush cytology specimens have gained interest. FISH is a technique that uses fluorescently labeled DNA probes to detect chromosomal alterations in cells [58]. FISH looks for changes in the number of chromosomes (aneuploidy), the structure of chromosomes and for losses (deletions) and gains (duplications) of genetic material [58]. Polysomy (extra chromosomes) of chromosomes 3, 7 and 17 has been associated with malignancy [53, 59]. However, only 80 % of pancreaticobiliary malignancies express these cellular alterations, thus inherently limiting the sensitivity of FISH [53, 60, 61]. In addition, some patients with benign bile duct strictures such as those with PSC, also exhibit chromosomal abnormalities. As a result, the specificity of FISH is lower than routine cytology, ranging from 67 to 88 % [60, 62]. Although FISH is not recommended as a routine screening tool for malignancy because of its low PPV, in select cases with high pre-test probability for malignancy it may improve sensitivity of brush cytology [60, 63].

Tissue samples for histological investigation can be obtained from biliary strictures by using a biopsy forceps that is directed to the site of the stricture using fluoroscopy. Fluoroscopy-guided forceps biopsy of biliary strictures is technically more difficult and time consuming than brushing and has a higher risk profile with rare reports of bleeding and biliary perforation . It is therefore less widely used. However, forceps biopsy can provide a sample of subepithelial stroma that is usually not sampled by brush cytology . As a result, at least theoretically, it can diagnose a subset of cholangiocarcinomas that do not project into the biliary lumen and only affect the subepithelial bile duct wall.

Forceps biopsy of biliary strictures is usually carried out after placement of a guidewire in the bile duct [64]. The guidewire keeps the biliary sphincter open, thereby allowing easier passage of the forceps through the sphincter. It also delineates the course of the bile duct on fluoroscopy, which is of help in navigating the biopsy forceps in the appropriate direction through the bile duct (Fig. 7.8). Although in most cases the biopsy forceps can be passed through the biliary sphincter even without a sphincterotomy , a prior sphincterotomy will ease the passage of the forceps and facilitate the process. Higher number of biopsies will likely increase the yield at the cost of higher complication rates. Specialized wire-guided biliary forceps are available and easier to use [65, 66].

In older literature, the overall cancer detection rate of forceps biopsy is higher than brush cytology , ranging from 43 to 81 % [67–69]. More recent studies, have continued to confirm higher sensitivity for forceps biopsy with comparable specificity to brush cytology [70]. It has been suggested that three or more biopsy samples are required to maximize sensitivity [71].

It seems that the sensitivity of tissue sampling techniques for detection of malignancy improves when different modalities are combined. For example, in a study of 58 patients, the sensitivity of transpapillary brush cytology was 41.4 % and the sensitivity of forceps biopsy was 53.4 %. When combined, the diagnostic sensitivity increased to 60.3 % [70]. In another study involving 133 patients with a biliary stricture the sensitivity of brushing alone, FNA alone and biopsy alone were 30, 30 and 43 % respectively. The combination of brushing and biopsy increased the sensitivity to 55 % and when all three modalities were combined the sensitivity further increased to 62 % [72].

Multiple other studies have confirmed that sampling of a biliary stricture with two or more techniques is the most effective method for diagnosis of malignant strictures [73, 74]. Consequently, some endoscopists prefer multimodality tissue sampling during ERCP in patients with biliary strictures when malignancy is highly suspected. In our practice, when there is suspicion of malignancy, brushing and biopsy of the strictures are often obtained during initial ERCP and if negative, with repeated ERCP procedures.

As opposed to the two-dimensional image offered by cholangiography, cholangioscopy offers a three-dimensional image of the bile duct lumen. In recent years, cholangioscopy has gained significant interest as a complementary procedure to ERCP for diagnosis and treatment of various biliary disorders, particularly indeterminate biliary strictures .

Available dedicated cholangioscopes in the USA are typically fiberoptic and reusable or semidisposable. Video cholangioscopes have limited availability and typically offer higher quality imaging. Cholangioscopy can be performed in one of three ways: two operator, single operator or direct peroral. In both the single and dual operator systems, the cholangioscope is advanced down the working channel of the therapeutic duodenoscope, while the newer direct peroral cholangioscopy (DPOC) technique involves passing an ultralsim upper endoscope through the mouth and directly into the bile duct. DPOC will be discussed further, later in this chapter.

The two operator system uses a reusable cholangioscope with single plane tip deflection (up-down), a working channel for accessories, air/water and suction buttons. Biliary sphincterotomy and stricture dilatation are performed as needed to facilitate passage of the cholangioscope. Although biliary cannulation can be achieved directly with the tip of the cholangioscope, most endoscopists prefer cannulation over a guidewire (Fig. 7.10). The guidewire is advanced down the cholangioscope and used to cannulate the duct, or if backloading the wire, a catheter should be advanced down the working channel of the cholangioscope to capture the wire and avoid damaging the channel. Care must be taken to keep the elevator maximally open to avoid damaging the cholangioscope. The duodenoscope tip is usually positioned close to and underneath the ampulla as the cholangioscope is advanced into the duct. Back tension on the guidewire may help. Once in position, the guidewire is removed to allow use of the working channel. The bile duct is irrigated with sterile saline solution through this channel to enable adequate visualization, followed by slow withdrawal of the cholangioscope, allowing systematic inspection of the biliary mucosa. The cholangioscope position can be adjusted by moving it or the duodenoscope with the assistant operating the up-down knob on the cholangioscope. When advancing accessories down the channel, the elevator should be open with the angle of the duodenoscope and cholangioscope reduced, or the accessories may need to be preloaded into the cholangioscope. A specially designed breastplate to which the cholangioscope is attached can allow single operator use [75].

The single operator reusable system (Spyglass, Boston Scientific, Marlborough, MA) consists of several parts: reusable optical fiber; disposable 10Fr catheter with 4-way tip deflection and three ports (optical probe port, accessory channel, and irrigation port) that is attached to the duodenoscope with a silastic band; and disposable 3Fr biopsy forceps. There is no suction port on the catheter, and a syringe can be attached to the working channel to provide manual suction. The optical fiber is preloaded into the catheter and the system is advanced through the working channel of the duodenoscope in a similar fashion as the reusable cholangioscopes. Once positioned inside the bile duct, the optical fiber is gently advanced beyond the tip of the catheter to enable visualization; the two dials can be adjusted and locked to adjust the tip of the catheter. In the near future, introduction of a new digital system with a chip at the tip of the catheter to provide images will obviate the need for an optical fiber.

In one study, ERCP with cholangioscopy was associated with higher rates of cholangitis thought to result from saline infusion during cholangioscopy [76]. In our centre we avoid cholangioscopy procedures in the setting of acute ascending cholangitis . Saline infusion should be limited to the lowest rate that allows adequate quality of the image. An adequate sphincterotomy , allowing the excess saline to exit through the sphincter, likely decreases the risk of cholangitis. Saline can also be suctioned through the working channel of the cholangioscope. Prophylactic antibiotics should be used.

It is well-known that the presence of irregularly dilated and tortuous blood vessels (so-called tumour vessels) due to neovascularization at the site of pancreatic or biliary strictures is indicative of malignancy [79–81]. Tumour vessels can be detected by direct visualization using a cholangioscope (Fig. 7.11, Video 7.1) [77, 79]. Narrow band imaging (NBI) is an imaging technique especially suited for visualization and characterization of mucosal vascular pattern. Use of cholangioscopes with NBI capability facilitates detection of neovascularization at the site of biliary strictures and thereby diagnosis of malignancy (Fig. 7.12) [77, 82]. Intraductal nodules or masses can also be indicative of malignancy and be easily detected by cholangioscopy (Fig. 7.13) [80]. Intense vascularization is associated with the nodular type of cholangiocarcinoma and less so with the infiltrative type. The infiltrative type may involve only the subepithelial layers of the bile duct wall and cannot be detected by cholangioscopy , which visualizes the superficial layers. An infiltrative mass may only be visible as tapering of the lumen causing a stricture. The papillary type of cholangiocarcinoma is characterized by numerous papillary projections [72]. Biliary strictures caused by extraluminal compression, such as those associated with pancreatic cancer , cannot be detected by cholangioscopy, unless at later stages when the tumour has infiltrated and penetrated the bile duct wall [78].

In theory, peroral cholangioscopy can improve diagnosis of indeterminate biliary strictures by directly visualizing the mucosa at the stricture, and allowing targeted biopsy [78]. Studies to assess the value of stricture visualization by cholangioscopy have reported high sensitivity for detection of malignant lesions . In one of the largest cholangioscopic studies to date, diagnostic fiberoptic cholangioscopy using the Spyglass system was performed in 226 patients with various biliary disorders. In patients with a biliary stricture, the ssensitivity for the diagnosis of malignancy was 51 % for ERCP impression, 78 % for cholangioscopic impression and 49 % for targeted biopsy [83]. Smaller studies using video cholangioscopes with better imaging capability have reported even higher sensitivity for detection of malignancy by visualization of the stricture site alone [82, 84, 85]. Overall, the findings of these studies suggest that addition of cholangioscopy enhances the diagnostic performance of ERCP, especially its capability to diagnose indeterminate biliary strictures [77].

Cholangioscopy-guided targeted biopsy is defined as biopsy of the sites that are affected by disease under direct cholangioscopic visualization (Fig. 7.14) [78]. On a practical note, when using the Spyglass system, there may be resistance to the passage of the biopsy forceps through the cholangioscope. The site of resistance is usually at the bend of the cholangioscope where it exits the tip of the duodenoscope and enters the bile duct. Moving the cholangioscope back and forth while continuously advancing the biopsy forceps usually allows passage of the forceps. A larger accessory channel in the new digital Spyglass system is expected to solve this problem.

Theoretically, targeted biopsy should improve cancer detection rate in malignant biliary strictures by allowing sampling of the sites that appear suspicious. In a large multicentre study, the sensitivity of fiberoptic cholangioscopy-guided targeted biopsy for diagnosis of indeterminate biliary strictures was only 49 %, far below the sensitivity of cholangioscopic visualization (78 %) [83]. However, the specificity of targeted biopsy was higher than visualization alone (98 vs. 82 %) [83]. Another study compared the diagnostic accuracy of peroral video cholangioscopic visual findings with that of video cholangioscopy-guided forceps biopsy for diagnosis of indeterminate biliary lesions. The sensitivity and specificity for visual findings were 100 and 91.7 % and for biopsy were 38.1 and 100 %, respectively [86].