Improving diagnostic yield of biliary brushings cytology for pancreatic cancer and cholangiocarcinoma

Biliary brushings are currently the best accepted method to obtain a cytological diagnosis of pancreatic cancer or cholangiocarcinoma. The technique has good specificity but poor sensitivity. Two dedicated pathologists reviewed 137 consecutive biliary brushings from 127 patients between February 1997 and February 2000. The ultimate diagnosis was determined by review of radiology, operative diagnosis and patient outcome. The sensitivity, specificity, positive predictive value and negative predictive value of the original results and the review results were calculated and compared. Additional diagnostic categories 'suspicious' and 'atypical possibly benign' were included on review. After review, the sensitivity improved from 49.4% to 89.0% and the specificity remained 100%. The use of the additional diagnostic category 'suspicious' increased the sensitivity to 90.4%, at the expense of a fall of the specificity to 66.7%. We conclude that review by two dedicated pathologists and additional diagnostic categories can improve the diagnostic accuracy of biliary brushings.