Cholangioscopic imaging: distinguishing good from bad

It can be assumed that 30 years ago, clinicians thought that a barium enema was the diagnostic state of the art and represented the ultimate advance in evaluating pathology in the colon. Today, in the era of high definition imaging with colonoscopes, that view seems strange and completely antiquated. But we are possibly now on the verge of a similar technological shift in the evaluation of biliary diseases. Endoscopic retrograde cholangiopancreatography (ERCP) is still the standard procedure for endoscopic evaluation and treatment of diseases in the biliary tract. Like the barium enema, however, black-and-white radiographic imaging has limited sensitivity and specificity, and it is virtually impossible to distinguish between malignant and benign strictures by relying on quite unspecific features of the duct contours at cholangiography. A large study on patients with indeterminate pancreaticobiliary pathology showed sensitivity and specificity rates of 51 % and 54 % for the cholangiographic appearance on its own [1].

“There can be no doubt that direct cholangioscopic visualization is a valuable tool for differentiating between malignant and benign strictures – particularly combined with directed intraductal biopsies.”

This is the current situation: the conquest of the small bowel with new endoscopy equipment and balloons has left the biliary and pancreatic ducts as the last blind spot for endoscopists. Major efforts are being made to close the gap, but this could still be the greatest technical challenge yet in the era of endoscopy. Various cholangioscopy devices have been proposed during the past 20 years, and quite a few of them have already been withdrawn from the market. There are two major techniques that are now available and in clinical use [2]: direct peroral cholangioscopy (POC) and mother – baby cholangioscopy.

Direct POC does not use a dedicated cholangioscope to access the common bile duct, but rather an ultraslim gastroscope. The technique provides high resolution imaging and a working channel that allows for most standard forms of endoscopic therapy. However, inserting the device into the common bile duct requires additional auxiliary accessories and considerable maneuvering skills, and the depth of insertion into the biliary tree is limited in most cases. Moreover, there are still some safety concerns with the technique.

Mother – baby cholangioscopes with dedicated single-operator cholangioscopy (SOC) devices are easier to handle, but the devices are fragile, costly, and mostly have poor, fiberoptic-based imaging that significantly limits their clinical value. This was illustrated by a recent study with a simple design: investigators blinded to any clinical information evaluated 30 SOC videos. The accuracy for distinguishing between benign and malignant SOC findings was a disillusioning 50 % [3], and the level of interobserver agreement was extremely low. There is obviously a strong need for improvement of the optical capabilities of these systems.

This issue of Endoscopy includes two interesting studies on mother – baby cholangioscopy [4] and direct POC [5], both of which focus on the imaging quality. The article by Shah et al. [4] presents a multicenter retrospective study from the United States that reports on clinical experience with a new single-use, digital, single-operator, intraductal cholangiopancreatoscopy system (IDCP). The device is a further development of the SOC system known as SpyGlass (Boston Scientific, Inc., Marlborough, Massachusetts, USA), which provided limited-quality fiberoptic imaging. The case series included 108 patients with a variety of pancreaticobiliary diseases. A total of 74 patients had indeterminate strictures or dilations, and nearly 40 % of these patients had neoplasia. The authors report impressive results: the technical success rate was 100 %, and the visual appearance had a sensitivity of 97 %, a specificity of 93 %, a positive predictive value of 90 %, and a negative predictive value of 98 %. Targeted biopsy sampling yielded a sensitivity of 86 % and a specificity of 100 %. The authors conclude that the new chip technology in the IDCP system provides enhanced image resolution that makes it possible to differentiate reliably between benign and malignant stenoses by allowing recognition of mucosal details such as tumor vessels and mucosal irregularities.

These results are comparable with the findings of another recently published case series investigating the same cholangioscopic device [6], which found that the sensitivity and specificity of the visual appearance with SOC for diagnosing malignancy were 90 % and 95.8 %, respectively. In comparison with the new prototype with digital imaging accuracy, its fiberoptic predecessor showed poorer results, with sensitivity and specificity levels of 78 % and 82 % in the largest prospective multicenter study on single-operator cholangiopancreatoscopy [1], including 96 patients with indeterminate pancreaticobiliary strictures. However, the extraordinarily good results for differentiation of neoplasia need to be interpreted with some caution and should not be attributed to improved visualization alone. The endoscopists were not blinded to the clinical data and impression, and the visual impression might therefore have been biased by knowledge of the clinical history and radiographic imaging.

POC is a single-operator cholangioscopy procedure that provides high resolution imaging with the standard ultraslim endoscopes that are routinely used for transnasal gastroscopy. The Korean group led by Jong Ho Moon has previously published pioneering work on this exciting cholangioscopy technique, and in the current issue of Endoscopy they report on their initial experience using computed virtual chromoendoscopy imaging with I-SCAN (Pentax Medical, Tokyo, Japan) to differentiate between benign and malignant disease in 20 patients [5]. Virtual chromoendoscopy methods such as narrow-band imaging enhance the vascularization of superficial mucosal structures and mucosal irregularities in comparison with conventional white-light images [7]. I-SCAN may therefore also be useful for improving observation. Investigators found that the quality of visualization using I-SCAN was significantly better than that of white-light imaging. The reason for this was that I-SCAN highlighted the surface microvascular architecture by increasing the contrast. Although the outcome parameters in the study are rather subjective with regard to the quality of visualization, this pilot study provides important arguments for a specific focus on tumor vessels – this is a lesson that was learned in the last century from assessment of other types of neoplasia in the gastrointestinal tract. However, in the present pilot study, I-SCAN failed to provide better diagnostic accuracy than white-light imaging alone. This might be due to the small size of the study, with no prior statistical calculation of the power needed to confirm the findings. Our personal experience in a large cohort of patients [8] suggests that adding computed virtual imaging to white-light imaging at cholangioscopy may highlight some features of a lesion, but it rarely changes the diagnosis.

The core message of the two studies is that improved imaging with digital optics or computed virtual chromoendoscopy helps the examiner to appreciate mucosal details and distinguish between different types of bile duct pathology. The normal bile duct mucosa is observed as a flat surface, often with shallow pseudodiverticula and fine networks of thin microvessels. A benign inflammatory mucosal lesion may be observed as a fairly homogeneous papillogranular surface, or it may have a scale-like appearance suggesting hyperplasia. Cholangioscopic findings suggesting a malignant lesion show thick, irregular, and tortuous vessels corresponding to neovascularization, an irregular granular surface, an appearance resembling fish eggs, and mucosal friability. However, some features of neoplastic tissue in the bile ducts may not be as specific as those in the gastrointestinal tract. In a case series of 111 patients with biliary strictures, Seo et al. [9] attempted to correlate the cholangioscopic findings with the histopathology of cholangioscopy-guided biopsies and surgically resected specimens. They classified the cholangioscopic findings associated with cholangiocarcinoma into three categories: a nodular type, with intensive neovascularization; a superficially spreading papillary type, with no neovascularization; and an infiltrative type, with smoothly tapered narrowing and the mucosa showing subtle white elevations. Shah et al. [4] report that the most common cholangioscopic findings in patients with neoplasia include tumor vessels and infiltrative strictures; however, these two signs were present in fewer than 50 % of neoplastic lesions. This is because cholangiocarcinomas appear with widely differing morphologies, and they are often desmoplastic, with very firm tissue due to abundant fibrosis that may also lead to higher rates of false-negative biopsies. In addition, many cases are difficult to evaluate because the majority of patients with stenoses have already been treated with stent placement or dilation, leading to mucosal irritation, inflammation, and vulnerability, independently of the pathology. In clinical reality, this can make the assessment of subtle mucosal findings quite difficult.

At present, ERCP-based cholangiographic evaluation with conventional cross-sectional images is still the primary modality for diagnosing biliary tract diseases. However, cholangioscopic technology and imaging are continuously being improved in order to enhance diagnostic accuracy and open up new horizons. Even today, there can be no doubt that direct cholangioscopic visualization is a valuable tool for differentiating between malignant and benign strictures – particularly in combination with directed intraductal biopsies. What we need is a robust endoscopic technology that can provide high resolution imaging with a simplified set-up. Cholangioscopy would then be suitable for extended clinical use beyond the level of tertiary hospitals. We are not quite there yet, but we are getting closer step by step!

• References

• 1 Chen YK, Parsi MA, Binmoeller KF. et al. Single-operator cholangioscopy in patients requiring evaluation of bile duct disease or therapy of biliary stones (with videos). Gastrointest Endosc 2011; 74: 805-814
• 2 Pohl J, Meves VC, Mayer G. et al. Prospective randomized comparison of short-access mother-baby cholangioscopy versus direct cholangioscopy with ultraslim gastroscopes. Gastrointest Endosc 2013; 78: 609-616
• 3 Sethi A, Widmer J, Shah NL. et al. Interobserver agreement for evaluation of imaging with single operator choledochoscopy: what are we looking at?. Dig Liver Dis 2014; 46: 518-522
• 4 Shah R, Raijman I, Brauer B. et al. Performance of a fully disposable, digital single-operator cholangiopancreatoscope. Endoscopy 2017; 49: 651-658
• 5 Lee Y, Moon J, Choi H. et al. Direct peroral cholangioscopy for diagnosis of bile duct lesions using a high-definition ultraslim endoscopy combined with I-SCAN: a pilot study. Endoscopy 2017; 49: 675-681
• 6 Navaneethan U, Hasan MK, Kommaraju K. et al. Digital, single-operator cholangiopancreatoscopy in the diagnosis and management of pancreatobiliary disorders: a multicenter clinical experience (with video). Gastrointest Endosc 2016; 84: 649-655
• 7 Itoi T, Sofuni A, Itokawa F. et al. Peroral cholangioscopic diagnosis of biliary-tract diseases by using narrow-band imaging (with videos). Gastrointest Endosc 2007; 66: 730-736
• 8 Meves V, Ell C, Pohl J. Efficacy and safety of direct transnasal cholangioscopy with standard ultraslim endoscopes: results of a large cohort study. Gastrointest Endosc 2014; 79: 88-94
• 9 Seo DW, Lee SK, Yoo KS. et al. Cholangioscopic findings in bile duct tumors. Gastrointest Endosc 2000; 52: 630-634