Observation Versus Resection for Small, Localized, and Nonfunctional Pancreatic Neuroendocrine Tumors A Propensity Score Matching Study

Objectives: The best treatment modalities for small (1-2 cm), localized, and nonfunctional pancreatic neuroendocrine tumors remain controversial. Therefore, we aimed to evaluate whether surgical resection provides survival benefit over observation in those patients. Methods: From 1973 to 2015, all eligible patients were retrieved from the Surveillance, Epidemiology, and End Results database. Propensity score matching (1:2) method was performed. The primary endpoints evaluated were overall survival (OS) and cancer-specific survival (CSS). Results: We identified 681 patients, of which 122 and 559 patients received observation and resection, respectively. Propensity score-matched patients who underwent surgery (n = 183) had significantly improved OS (P = 0.008) compared with matched patients who underwent observation (n = 106), but there was no difference in CSS (P = 0.310). On multivariate analysis, resection could improve OS but not CSS. Besides, poorly differentiated/undifferentiated tumor had a worse OS and CSS. Subgroup analysis showed that patients 60 years and older who underwent resection could achieve a longer OS and CSS. Conclusions: This disease exhibits a very good prognosis. Patients undergoing resection were associated with comparable 5-year CSS but longer 5-year OS compared with those receiving observation. Elderly patients (>= 60) may obtain benefit from surgery, whereas the treatment of younger patients should be individualized.