Diagnostic performance and factors influencing the accuracy of EUS-FNA of pancreatic neuroendocrine neoplasms

Background Multiple studies have investigated sampling adequacy of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) for pancreatic neuroendocrine neoplasms (pNENs). However, none have described the diagnostic performance of EUS-FNA for pNENs, or the influencing factors. The aim of this study was to evaluate the diagnostic accuracy of EUS-FNA, with post-operative pathological diagnosis as the gold standard, and factors predictive of inadequate EUS sampling. Methods From 1998 to 2014, a total of 698 patients underwent pancreatic resection and 1455 patients underwent EUS-FNA sampling for pancreatic lesions. A total of 410 cases underwent both surgical resection and preceding EUS-FNA. Of these, 60 cases (49 true pNEN, nine non-diagnostic, two misdiagnoses) were included. We studied diagnostic performance of EUS-FNA and factors that were associated with failed diagnosis. Results Of the 60 cases, EUS-FNA yield was 49 true-positive cases, two misdiagnoses, and nine non-diagnostic cases (including six suggestive cases). Sensitivity, specificity, and accuracy were 84.5, 99.4, and 97.3 %, respectively; including the six suggestive cases, diagnostic values were 94.8 % sensitivity (55/58), 99.4 % specificity (350/352), and 98.7 % accuracy (405/410). In multivariate analysis, sampling adequacy rates were significantly lower when lesions were located in the pancreatic head [odds ratio (OR) = 10.0] and in tumor-rich stromal fibrosis (OR = 10.45). Tumor size, needle type, tumor grading, presence of cystic component, and time period were not significant factors. Conclusions EUS-FNA offers high accuracy for pNEN. However, location of the tumor in the pancreatic head and presence of rich stromal fibrosis negatively impacts sampling adequacy.