Biliary tract cytology in specimens obtained by direct cholangiographic procedures: a study of 74 cases

A retrospective review of bile (BL) and biliary tract brushings (Br) obtained by endoscopic retrograde cholangiopancreatography (ERCP) and percutaneous transhepatic cholangiography (PTC) was undertaken to determine the sensitivity and specificity of cytology in the diagnosis of pancreaticobiliary malignancies. A total of 104 cytologic specimens (PTC-BL 15, PTC-Br 13, ERCP-BL 8, ERCP-Br 68) received between 1990 and mid-1994 from 77 patients who had undergone ERCP and/or PTC primarily for biliary stricture were reviewed. Specimens were unsatisfactory/ inadequate in 11 (10.6%), benign in 41 (39.4%), suspicious in 25 (24%), and positive for malignant cells in 27 (26%). Follow-up was available in 74/77 patients; 46 (59.7%) had tissue confirmation while 28 (32.5%) had adequate clinical follow-up based on chart review. Of those with histologic confirmation, there were 32 malignant and 14 benign cases. The overall sensitivity and specificity of PTC- and ERCP-obtained cytologic specimens were 88.9 and 95.7% respectively. There was only one false positive case (ERCP-Br). Overall positive predictive value was 96% negative predictive value 88%, and accuracy 96%. PTC had a significantly lower sensitivity rate (42.8%) and higher rate for unsatisfactory specimens (21%) compared with ERCP-obtained material (100 and 1.9%). Bile obtained by PTC or ERCP appeared less sensitive in detecting malignancies compared with endoscopic brushing using either technique (BL 50% vs. Br 100%). All three false negative cases were PTC-BL specimens. Of the 17 suspicious cases, eight were confirmed histologically as malignant, four were clinically consistent with malignancy, and five showed marked inflammatory atypia on biopsy. Positive predictive value and accuracy rate of a "suspicious cytology" diagnosis were 69 and 80.5%, respectively. Inadequate specimen, poor cellular preservation, and cells obscured by bile all interfere with proper cytologic evaluation. Experience is necessary to appreciate subtle malignant changes in well differentiated carcinomas. Communication between the cytopathologist and the clinician is critical in the accurate interpretation and proper management of the patients.