Diagnosis of pancreatic neoplasia with EUS and FNA: a report of accuracy

Background EUS-guided FNA has the potential to provide diagnostic cytologic material from pancreatic lesions that are suspicious for malignancy. Objective To determine the operating characteristics of EUS-FNA in the diagnosis of pancreatic adenocarcinoma and pancreatic neuroendocrine neoplasms (PENs). Design Retrospective analysis of a prospectively maintained database. Setting Academic tertiary-care center. Patients This study involved 559 patients undergoing evaluation of pancreatic masses or diffuse pancreatic parenchymal abnormalities. Main Outcome Measurements Performance characteristics of EUS-FNA of pancreatic adenocarcinoma and PEN. Results From January 1997 to December 2005, 737 patients undergoing initial EUS-FNA evaluation for a pancreatic mass were identified. In the final analysis, 559 patients with 560 FNA-sampled lesions were included. Overall, 442 lesions were pancreatic adenocarcinoma, and 40 were PEN. The sensitivity of EUS-FNA in the diagnosis of pancreatic adenocarcinomas and PENs was 77% (95% CI, 72.8%-80.8%) and 68% (95% CI, 50.8%-80.9%), respectively, using strict cytologic criteria. Reclassification of atypical and suspicious cytologies as diagnostic of malignancy resulted in a sensitivity of 93%, (95% CI, 90.9%-99.7%) in adenocarcinoma and 80% (95% CI, 63.9%-90.4%) in PEN. Tumor size, tumor location, and number of needle passes did not significantly influence diagnosis, but immediate cytologic evaluation was influential. Limitations Retrospective analysis at a single center. Conclusions In a large, well-controlled study, EUS-FNA was found to be an accurate test (80%) for the detection of pancreatic adenocarcinoma by using aspiration cytology. The accuracy of the examination is significantly improved (94%) when atypical and suspicious samples are considered positive. Finally, only 2 to 3 FNA passes may be needed to achieve a good diagnostic yield.