Combination of ERCP-Based Modalities Increases Diagnostic Yield for Biliary Strictures.

ABSTRACT

BACKGROUND: Biliary stricture evaluation with brush cytology and intraductal forceps biopsy carries a low sensitivity, but the combination of newer modalities may improve sensitivity. AIM: To determine whether the addition of advanced modalities increases diagnostic yield of ERCP-based sampling. METHODS: This single-center retrospective study evaluates patients with biliary strictures sampled using brush cytology. Operating characteristics were calculated for individual and combinations of modalities including cholangioscopy, fluoroscopy- and cholangioscopy-directed intraductal biopsy, fluorescence in situ hybridization (FISH), and confocal laser endomicroscopy. Analyses under Standard Criteria (SC) included malignant results as "positive" and Expanded Criteria (EC) included "suspicious" and "high-grade dysplasia" results as "positive." RESULTS: A total of 614 patients were included, and 354 (57.8%) received brush cytology alone, which had a sensitivity of 38.5% (SC) to 40.3% (EC) and a specificity of 97.8% (EC) to 99.3% (SC). Combining brush cytology with fluoroscopy-guided biopsy (n = 259, 42.2%) had a sensitivity of 62.5% (SC) to 67.9% (EC) and specificity of 90.2% (EC) to 96.7% (SC). Adding FISH to brush cytology had a sensitivity of 84.2% (SC) to 87.5% (EC) and specificity of 54.1% (SC and EC), while cholangioscopy visualization addition resulted in a sensitivity of 80.4% (SC) to 92.2% (EC) and specificity of 67.3% (EC) to 89.1% (SC). There were no significant differences in sensitivity and specificity using SC and EC. CONCLUSIONS: Brush cytology has a low sensitivity, but the addition of other modalities increases sensitivity. There was no difference in specificity between the SC and the EC, supporting the inclusion of "suspicious" impressions with malignant results at our center.