Establishing a Diagnosis in Indeterminate Pancreaticobiliary Strictures: Is Confocal Laser Endomicroscopy the Answer?

Abstract

Despite the advances that have been made in diagnostic modalities during recent years, the differentiation between benign and malignant pancreaticobiliary strictures remains a significant diagnostic challenge. Conventional cancer markers are of limited diagnostic utility and radiologic assessment with cross-sectional imaging may not provide a definitive diagnosis in the absence of a mass lesion [1–3]. Endoscopic evaluation has, for many years, relied heavily on cholangiographic appearance and biliary brush cytology, which, despite a specificity approaching 100 %, shows a low sensitivity of 15–57 % for malignancy [4]. Confocal laser endomicroscopy (CLE) provides microscopic information on mucosal tissue in real time and aims to enhance the diagnostic accuracy of endoscopic examination. Employing technology designed to prevent light scattering, CLE generates highly focused images of very small mucosal areas [5, 6]. It has been used in detecting dysplasia/cancer in various settings, such as colorectal polyps, inflammatory bowel disease, and Barrett’s esophagus [7–9]. A CLE probe has been devised fitting in the working channel of a cholangioscope or in a catheter that can be inserted through the working channel of a standard duodenoscope under fluoroscopic or cholangioscopic guidance allowing examination of indeterminate biliary strictures. Fluorescein is given intravenously 2–3 min prior to mucosal imaging allowing also for visualization of blood vessels [10]. In recent years, a number of studies have attempted to determine the performance characteristics of this modality for the differentiation of benign and malignant pancreaticobiliary strictures [10–13]. Also, data from a multicenter prospective study were used to reach a consensus definition of biliary and pancreatic probe-based CLE [13]. Characteristics most suggestive of malignancy included thick dark bands ([40 lm), thick white bands ([20 lm), dark clumps, or epithelial structures, whereas thin dark or white bands were suggestive of benign lesions (Miami classification criteria). It has been shown that combining two or more criteria for malignancy significantly increases the sensitivity and predictive values of the method [13]. In the largest study investigating the performance characteristics of CLE in indeterminate biliary lesions, a total of 89 patients were included [12]. Utilizing the Miami classification criteria, CLE was found to have a sensitivity of 98 %, a specificity of 67 %, and a negative and positive predictive value of 97 and 71 %, respectively. It compared favorably with conventional histopathology that had a sensitivity, specificity, negative and positive predictive values of 45, 100, 69, and 100 %, respectively [12]. The excellent sensitivity and negative predictive value of CLE in indeterminate pancreaticobiliary strictures suggest that it could be an important adjunct to ERCP for distinguishing benign from malignant disease and that it may be used to guide biopsy sampling. Integrating this very promising new technology into every-day clinical practice involves defining not only its performance characteristics but also interobserver agreement for the perceived final diagnosis and for the different diagnostic criteria applied. In the initial validation of the Miami classification criteria, three criteria indicative of malignancy and one indicative of benign strictures showed moderate interobserver agreement as suggested by significant Fleiss kappa values between 0.40 and 0.60 [13]. Apart from these findings, however, data on interobserver E. Kalaitzakis (&) Department of Gastroenterology, Skane University Hospital, 221 85 Lund, Sweden e-mail: evangelos.kalaitzakis@medicine.gu.se; kalvag@hotmail.com