Abstract
Background. The utility of repeat EUS in patients with suspicion for pancreatic cancer after non-diagnostic EUS-FNA study is not well established. Aim. Determine the accuracy of repeat EUS-FNA in patients with suspected pancreatic cancer and prior non-diagnostic EUS-FNA. Methods. Retrospective cohort study. Results. From 2002 to 2008 in our institution 28 patients underwent repeat EUS-FNA for suspected pancreatic cancer. Initial EUS showed a pancreatic mass in 24 (85.71%), no mass in 3 (10.71%) and possible mass in 1 (3.58%). FNA was performed and was negative for malignancy in all patients. Repeat EUS showed pancreatic mass in 27 patients (96.42%) and no mass in 1 (3.58%). FNA was performed in all patients and cytology was positive for malignancy in 6 (21.43%). Out of the 28 patients, 17 (60.71%) were eventually confirmed to have cancer. Overall repeat EUS-FNA correctly determined the true final status in 17 out of 28 patients providing sensitivity for the diagnosis of cancer of 35% (95% CI 14%-62%), specificity 100% (95% CI 72%-100%), and overall accuracy of 61%, (95% CI 28%-72%). Conclusion. Repeat EUS-FNA provides reasonable accuracy and may be worthwhile in patients with suspected pancreatic cancer who had had prior negative EUS-FNA.