Considerable heterogeneity exists in the risk of recurrence after curative resection of pancreatic neuroendocrine tumors (PanNETs). Current clinicopathological parameters (e.g., grade, TNM stage) have limited capacity for individualized risk stratification. There is an urgent need for a clinically accessible and operationally simple predictive tools to identify high-risk PanNETs for optimizing personalized postoperative treatment and follow-up strategies. This study aims to identify risk factors associated with recurrence of PanNETs and to develop a clinically applicable risk stratification tool based on multi-center cohorts and multi-omics validation for identifying high-risk patients.