Prospective study of biliary cytology in suspected perihilar cholangiocarcinoma

Background: The diagnostic value of biliary cytology for hilar bile duct stricture is uncertain. This study prospectively examined three methods for the evaluation of biliary cytology in a consecutive group of patients. Methods: Preoperative bile sampling by aspiration through a drainage catheter (aspiration samples), saline flush through a drainage catheter (saline samples) or direct sampling from a drainage bag (bag samples) was performed in consecutive patients with suspected perihilar cholangiocarcinoma who underwent resection after endoscopic nasobiliary drainage or percutaneous transhepatic biliary drainage. All bile sampling was performed three times on separate days. The accuracy of cytology in the diagnosis of carcinoma was determined. Results: Of 100 consecutive patients with hilar strictures, 97 had histologically proven cholangiocarcinoma. The proportion of these 97 patients who had a positive finding on cytology in at least one of three sampling sessions was 55 per cent for aspiration samples, 48 per cent for bag samples and 38 per cent for saline samples (P = 0.021, aspiration versus saline). Tumour length correlated significantly with overall positivity. For aspiration samples, sensitivity was 55 per cent, specificity was 100 per cent and accuracy 56.0 per cent. Conclusion: For biliary cytology, sampling by catheter aspiration is more effective than catheter flushing or sampling from a drainage bag. Repeated sampling increases sensitivity. Biliary cytology has modest diagnostic yield, but is easy to perform, highly specific, and can provide a definitive diagnosis.