Novel morphological tools in predicting pancreatic neuroendocrine tumors (Pan-Net) clinical outcome

Introduction: Pancreatic Neuroendocrine Tumor (Pan-NET) clinical outcome shows a high variability among patients within the same World Health Organization (WHO) category. To date, there is a lack of markers to predict recurrence after resection, which makes it difficult to identify patients with higher risk of progression. Aim(s): To evaluate morphologic and clinicopathological features of surgically resected Pan-NET. Materials and methods: Surgical specimens of 150 Pan-NET patients were evaluated for the following diagnostic and prognostic tools: Tumor stage, Ki-67, parenchymal infiltration, neural and vascular invasion, tumor necrosis or sclerosis and tumor deposits. These data were correlated with Overall Survival (OS) and stage I-II-III Disease-Free Survival (DFS). Results: Overall, 76 (50.7%) patients showed localized (stage I-II), 22 (14.7%) locally advanced (stage III), and 52 (34.7%) metastatic disease (stage IV). Tumor size >2 cm (p<0.0001), Ki-67 index (p=0.001), parenchymal infiltration (p=0.0002), vascular and neural invasion (p<0.0001 and p=0.0008) and tumor deposits (p=0.001) were correlated with locally advanced-metastatic disease. At univariate analysis parenchymal infiltration, presence of tumor deposits and necrosis were associated with OS (p=0.03, 0.04 and 0.008) and DFS (p=0.001, 0.007 and <0.0001) while vascular and perineural invasion were correlated only with DFS (p=0.003 and 0.001). Multivariate model showed that locally advanced disease (HR=2.56, p=0.02), parenchymal infiltration (HR=3.83, p=0.008) and tumor deposits (HR=2.45, p=0.03) were correlated with lower DFS. Conclusion: Both parenchymal infiltration and presence of tumor deposits represent emerging tools in predicting tumor recurrence.