Tumor biology, not tissue of origin, drives outcomes following resection of duodenal/ampullary and pancreatic neuroendocrine tumors

Abstract

Presenter: Ryan Schmocker MD, MS | Johns Hopkins University School of Medicine Background: Previous data has suggested that duodenal/ampullary neuroendocrine tumors (DANETs) and pancreatic neuroendocrine tumors (PNETs) have different clinical outcomes and patterns of recurrence. We report a large single institutional database to determine the pathologic characteristics, outcomes, and patterns of recurrence for these tumors. Methods: A prospectively maintained; single institution database of all pancreatic resections was queried for patients undergoing surgical resection for a neuroendocrine tumor. Patients with MEN1 syndrome and follow up less than 1 year (unless this was secondary to disease progression) were excluded. Chart review was undertaken to identify patients undergoing resection of a DANET or PNET and to obtain demographic and clinical variables. Patients undergoing surgery from 2000-2018 were included. Results: A total of 310 patients were included, 50 with DANETs and 260 with PNETs. Patients in the DANETs group were significantly older (60.0yr vs 59.6yr; p = 0.018), had higher grade tumors (22.0% G2 or G3 tumors vs. 9.3%; p = 0.019), and had more node positive resections (70.0% vs. 36.9%; p < 0.001). As expected, patients with DANETs were significantly more likely to have a local resection (18.0% vs 0.0%; p <0.001). Despite these poor pathologic features, there was no difference in overall survival (66.6 vs 60.9mo) or recurrence free survival (50.0 vs 50.4mo) when comparing the groups. The following factors on univariate analysis were associated with worse survival for all included patients: lymphovascular invasion (LVI), increasing T stage, node positivity, metastatic disease at the time of surgery, and age. DANET was not a significant predictor of overall survival. On multivariate analysis age was the strongest predictor of poor survival (HR 1.05 per year; p <0.001), followed by M1 disease (HR 3.72; p = 0.002), and LVI (HR 1.96; p = 0.050). There was no difference in the patterns of recurrence between the groups with both groups having liver recurrence as the most common site. Conclusion: In this large single institutional study, DANETs appear to have worse pathologic features at the time of resection when compared with PNETs. However, there is no difference in overall or recurrence free survival, suggesting that the underlying tumor type is a more important driver of outcomes than the specific pathologic characteristics. Increasing age, the presence of metastatic disease, and LVI were the drivers of worse outcomes in all included patients. Additional study is required to further characterize the predictors and outcomes of DANETs.