New Model Predicts Recurrence Risk in Pancreatic Tumors

A predictive model effectively identifies patients with non-functioning pancreatic neuroendocrine tumors (NF-PanNETs) at high risk for early recurrence (ER). Developed from a cohort of 496 patients, the model leverages preoperative clinical, pathological, and imaging variables, achieving an AUC of 0.91 in training and 0.84 in validation. Neoplastic venous thrombosis emerged as the strongest recurrence predictor, with a 71% ER probability, while specific tumor characteristics provided additional risk stratification.