Conditional survival after surgical resection of well-differentiated pancreatic neuroendocrine tumors: A NDCB analysis.
Abstract
631 Background: Surgical resection remains the standard treatment for low-grade pancreatic neuroendocrine tumors (PanNETs) and offers the best chance for durable survival. While prognostic factors such as tumor grade, nodal status, and margin status are well described using traditional survival endpoints, conditional survival estimates remain less defined. We aimed to evaluate conditional survival (CS) after resection of low-grade PanNETs and identify clinical and pathologic factors influencing long-term outcomes. Methods: Patients diagnosed with PanNET who underwent surgery, had a well-differentiated tumor grade, and had recorded dates of surgery and last follow-up date were included. Demographic, clinicopathologic, and treatment variables were collected, including age, sex, tumor grade, T and M stage, resection margin, CA-19-9 status, and Mitotic Index. Kaplan-Meier analysis was used to estimate OS, and CS at 1–3 years post-resection was calculated as the probability of surviving an additional 3 years given survival to the specified time point. Subgroup analyses stratified CS by key clinicopathologic factors and subsequent nomograms were then constructed. Internal validation was performed with k-fold cross validation. Results: A total of 7,681 patients with low-grade (G1–G2) PanNETs underwent curative-intent resection. Median age was 58 years, 52.1% were male, and 75.1% had node-negative disease. Most tumors were stage I–II (84.9%) and resected with negative margins (89.3%). Three-year overall survival (OS) for the entire cohort was 93%, with conditional survival (CS) remaining stable across time: 94% after 1 year, 93% after 2 years, and 92% after 3 years. Age >65 years (3-year CS 87% vs 97% in <50), nodal positivity (89% vs 94%), distant metastases (83% vs 94%), and R1/2 resections (≤90% vs 93%) were associated with persistently lower CS. Three multivariable nomograms incorporating age, tumor stage, nodal status, mitotic count, CA 19-9, and margin status demonstrated fair discrimination for 1-, 2-, and 3-year conditional survival respectively (AUC > 0.72) and provides individualized survival estimates. Conclusions: Conditional survival after resection of low-grade PanNETs remains excellent and stable over time, with increased risk seen in older, node-positive, and margin-positive patients. Incorporating CS and the derived nomogram into postoperative counseling allows for dynamic, individualized risk assessment and may optimize surveillance intensity for low-risk patients. Conditional survival estimates (%) for low-grade pancreatic neuroendocrine tumor. If patient has survived (year) Total survival time (year) 0 1 2 3 4 5 1 0.97 2 0.95 0.98 3 0.93 0.96 0.98 4 0.90 0.94 0.95 0.97 5 0.88 0.91 0.93 0.95 0.97 6 0.86 0.88 0.9 0.92 0.94 0.97
Article Details
Journal Info
Journal of Clinical Oncology
Lippincott Williams & Wilkins
Authors (2)
Aman Singla
University of Southern California, Los Angeles, CA
Gautam Kumar Malhotra
University of Southern California, Los Angeles, CA