Background and study aims: Cytological methods may be used to differentiate benign from malignant biliary strictures. We evaluated the diagnostic accuracy of an endobiliary cytotechnique which can easily be performed during endoscopic retrograde cholangiopancreatography.
Patients and methods: Cytological samples were obtained by brushing biliary strictures via a guide wire in 86 patients with strictures of unknown status. Samples were classified by an expert cytologist as normal (including reactive cells), severely dysplastic (atypical cells suspicious of malignancy) and clearly malignant. A final diagnosis was achieved in 78 patients based on intraoperative findings and histological investigation, autopsy or prolonged follow-up. Strictures were malignant in 57 cases (31 pancreatic carcinoma, 20 cholangiocarcinoma, 6 others) and benign in 21 cases (11 chronic pancreatitis, 5 chronic nonspecific inflammation, 5 others).
Results: The overall results for brush cytological investigation were sensitivity 56.1%, specificity 90.5%, positive predictive value 94.1%, negative predictive value 43.2 %, and accuracy 65.4 %. Sensitivity was significantly higher (P<0.005) in cholangiocarcinoma (80%) compared with pancreatic carcinoma (35.5%). The overall specificity of less than 100 % resulted from dysplasia in two patients with chronic pancreatitis and inflammatory mass. No procedure-related complications occurred.
Conclusions: Brush cytology is helpful for differentiating between benign and malignant biliary strictures, especially in suspected cholangiocarcinoma. Dysplastic cells may occur in the absence of malignancy and their presence should therefore be interpreted cautiously.