Yield of endoscopic ultrasound-guided fine-needle aspiration of bile duct lesions

Background and Study Aims: It is still difficult to differentiate reliably between benign and malignant biliary tract lesions. Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) has added to the diagnostic power of EUS for other gastrointestinal tumors. A retrospective analysis of experience with FNA sampling of bile duct lesions was therefore carried out. Patients and Methods: All EUS-FNA procedures for bile duct masses or strictures were analyzed at our tertiary referral center from May 2000 through October 2002. Data for EUS findings, the results of EUS-FNA, and tissue sampling at surgery were included. EUS-FNA procedures were carried out using a 22-gauge needle. An experienced cytopathologist was present during FNA in all but three cases. Clinical follow-up details were recorded when available for patients in whom a suitable diagnostic gold standard was not available for comparison. Results: A total of 35 patients underwent EUS-FNA of bile duct lesions during the study period. There were no complications. Data for EUS-FNA of bile duct masses or strictures and tissue obtained at surgery were available for 23 patients. If positive cytology at surgical pathology is taken as the gold standard, EUS-FNA has a diagnostic yield for cancer of 100% (if atypia/inconclusive findings in the FNA sample are regarded as benign). Eleven patients had a definite malignancy on surgical pathology. Of these 11 patients, five had a finding of malignancy on EUS-FNA, giving a sensitivity of 45% (if FNA cytology reported as atypia/inconclusive is regarded as benign). Twelve patients had findings of no malignancy from tissue obtained at surgery. Of these 12 patients, nine had benign pathology and three had atypia/inconclusive findings in the EUS-FNA sample (specificity of 100% if atypia/inconclusive findings are considered benign). A further 12 patients did not have surgical specimens for comparison with EUS-FNA results. Four patients had definite findings of malignancy on EUS-FNA alone, and one patient had FNA findings suspicious for malignancy. Seven patients had negative or equivocal EUS-FNA results. These 12 patients are described but excluded from further analysis, as a gold standard was not available for comparison. However, clinical follow-up data were available for eight of these 12 patients, and in each case the follow-up findings were compatible with previous benign or malignant EUS-FNA findings. Conclusions: The practice of EUS-FNA has improved the diagnostic yield of EUS. These results suggest that it is a safe and useful procedure for investigating biliary masses or strictures that have hitherto caused considerable diagnostic confusion, especially in patients with negative brush cytology findings. The possibility of false-negative findings remains, but core biopsy needles may improve the situation. The results of further studies are awaited.