Abstract
Background: Biliary strictures are diagnosed using endoscopic retrograde cholangiopancreatography (ERCP) with brush cytology and biopsy. However, brush cytology shows a sensitivity of 9-56.1% and a diagnostic accuracy of 43-65.4%, while biopsy demonstrates a sensitivity of 48%. Both methods exhibit high specificity but limited sensitivity. While rapid on-site evaluation (ROSE) is effective in endoscopic ultrasound-guided fine needle aspiration (EUS-FNA), its application in ERCP-obtained samples remains underexplored. Methods: This prospective pilot study was conducted at Okayama University Hospital from April 2019 to July 2024. Patients requiring ERCP-guided sampling for bile duct strictures were included. ROSE was applied to brush cytology with up to three additional attempts and to imprint cytology from biopsy samples with up to two attempts. Diagnostic accuracy was assessed based on pathology and clinical course. Results: Among 37 patients (median age: 73 years, add range, and male-female ratio: 27:10), 18 had hilar and 19 had distal bile duct strictures. Brush cytology required one, two, or three attempts in twenty-six, six, and five cases, respectively, whereas biopsy required one or two attempts in thirty-five and two cases, respectively. Among the thirty-seven cases, thirty-five were malignant and two were benign. The B-ROSE group showed a sensitivity, specificity, and accuracy of 71.4%, 100.0%, and 73.0%, respectively, compared to lower accuracy in the conventional group, where single brush cytology attempts yielded a sensitivity of 48.6% and an accuracy of 48.6%, and single biopsy attempts showed a sensitivity of 68.6% and an accuracy of 70.3%. Conclusions: B-ROSE improves diagnostic accuracy, reduces repeat sampling, and minimizes patient burden in ERCP-based diagnosis of bile duct strictures, making it a valuable addition to current diagnostic protocols.