Registry Trial to Determine pCLE Image Interpretation Criteria and Preliminary Accuracy for PSC Biliary Strictures

pCLE (probe based confocal laser endomicroscopy) is an innovative technology, which provides real- time, microscopic imaging of tissue at the cellular level via a small diameter probe. The pCLE probe has been designed to fit into standard endoscopes and has been studied extensively in the gastrointestinal tract for real time diagnosis of conditions such as Barrett's Esophagus (BE), biliary strictures and endoscopic mucosal resection of colorectal lesions. Over 300 publications have documented the safety and efficacy of pCLE in the GI tract.

Ductal pathologies are presently diagnosed, characterized and sometimes treated during Endoscopic Retrograde CholangioPancreatography (ERCP) in which a catheter is advanced through the endoscope and inserted into the biliary or pancreatic duct, where a contrast agent is injected and X-rays are taken. But multiple tissue sampling (biopsies or brushing) requires considerable time and technical expertise with the risk of losing guide wire access across the biliary or pancreatic stricture. Tissue sampling techniques have shown to lead to highly specific results for diagnosing a malignant tumor (100%), but with low sensitivity (45- 75%). Furthermore, pancreatitis is one common and serious complication of ERCP, occurring in 5-7% of cases.

The CholangioFlex confocal miniprobe was specifically designed to explore bile and pancreatic ducts (diameter ranging from 0.9 to 2.5mm). Intended to be used via an ERCP procedure, it can be passed through the operating channel of a cholangioscope or inserted through a standard catheter.

Dr. Meining examined a small series of 14 patients with biliary strictures and Mucosal imaging was performed with a miniaturized confocal miniprobe introduced via the working channel of a cholangioscope. Thereafter, targeted biopsies were taken from the same regions. All strictures could be reached, leading to a pCLE accuracy of 86%, sensitivity of 83%, and specificity of 88%. The respective numbers for standard histopathology were 79%, 50%, and 100%.

Dr. Giovannini evaluated the diagnostic accuracy of pCLE for cholangiocarcinoma detection, on 37 patients with biliary or pancreatic strictures. The CholangioFlex confocal miniprobe was introduced in the bile or pancreatic ducts using a catheter, and strictures were imaged. Tissue sampling was then performed at the same location. In this study, the overall pCLE accuracy was 86% (vs. 53% for histology), the sensitivity and specificity of pCLE were 83% and 75% respectively compared to 65% and 53% for histology.

An observational prospective, "Cellvizio ERCP registry", was conducted which enrolled 102 patients with indeterminate or suspected biliary and/or pancreatic strictures, mass or neoplasm indicated for ERCP and/or cholangioscopy. The purpose of this multicentric trial was to compare the combination of Cellvizio with ERCP imaging to ERCP alone, using the Miami Classification (a set of image interpretation criteria developed to differentiate benign from malignant strictures). Physicians could choose whether to deliver the CholangioFlex confocal miniprobe through a cholangioscope or a catheter. Patients were followed until the physicians were able to confirm malignancy through histopathology or for a year if repeat tissue sampling led to benign results.

There were no pCLE-related adverse events in the study and 89 patients were finally evaluable, of whom 40 were proven to have cancer. The sensitivity, specificity, positive-predictive value, and negative- predictive value of pCLE for detecting cancerous strictures were 98%, 67%, 71%, and 97%, respectively,compared with 45%, 100%, 100%, and 69% for index pathology. This resulted in an overall accuracy of 81% for pCLE compared with 75% for index pathology. Accuracy for combination of ERCP and pCLE was significantly higher compared with ERCP with tissue acquisition (90% vs. 73%; P .001).

Due to the relatively low specificity, a group of investigators reviewed the false positive cases of the registry and new criteria characterizing inflammatory strictures, which are known to present pCLE features very similar to malignant strictures. 60 pCLE along with final diagnosis were reviewed by 3 pCLE-experienced gastroenterologists who refined the already existing Miami classification by devising novel pCLE criteria for the characterization of inflammatory strictures. The 4 criteria devised for diagnosing dysplasia in BE were: 1) Multiple thin white bands, 2) Dark granular pattern with scales, increased space between scales, thickened reticular structures. These criteria were then reviewed and validated in consensus by 6 pCLE experts using a set of 40 pCLE sequences. The overall accuracy was 82.5% vs. 81% for the prospective registry (n=98), resulting in a sensitivity of 81.2% (vs. 98% for the prospective study) and a specificity of 83.3% (vs. 67% for the prospective study). The corresponding interobserver agreement was fair (k=0,37). This new criteria are currently tested in a prospective multicentric trial, aiming at evaluating the impact of pCLE on the management of patient with indeterminate biliary stricture.

The purpose of the present pCLE Registry will be to study a condition with a high unmet medical need and preliminary differentiating criteria: Primary sclerosing cholangitis.

This condition offers challenges for the diagnostic pathway and could benefit from direct visualization of the tissue via a minimally invasive ductal approach.