Disparities in overall survival (OS) outcomes based on treating facility volume in pancreatic neuroendocrine tumors.

Abstract

150 Background: Pancreatic neuroendocrine tumors (pNETs) comprise less than 5% of all pancreatic tumors and 7% of all neuroendocrine tumors (NETs). There has been considerable progress in understanding the biology of pNETs and numerous therapeutic options emerged, especially with the collaboration of various specialties in the multidisciplinary setting. In this this study, we sought to analyze the association of facility volume, treatment modalities offered, and different risk adjusted outcomes in pNETs. Methods: We extracted data from National Cancer Data Base (NCDB) that covers 70% of all newly diagnosed cancer cases in the U.S and Puerto Rico. Patients with pNETs diagnosed between 2004 & 2017 were included and classified into tertiles based on hospital volume. Volume–outcome relationship was determined by using Cox regression adjusting for patient demographics, comorbidities, tumor characteristics, insurance type and therapy received. Kaplan Meier estimates of OS were compared with log-rank test. The primary predictor of interest was the facility volume defined as number of pNETs treated/year. A total of 7202 patients with pathologically confirmed pNETS were treated at 840 facilities. The median annual facility volume was 5patients/year. Facilities were classified into (T:mean cases/year) T1: < 3; T2:4-8; T3:≥9 cases/year. Results: A total of 7202 pNET patients were treated at 840 facilities. The median annual facility volume was 5 patients/year. Facilities were classified into (T: mean cases/year) T1:<3; T2:4-8; T3:≥9 cases/year. The unadjusted median OS by facility volume was: T1: 71 months (m), T2: 136 m, and T3: not-reached (p < 0.001). On multivariable analysis, compared with patients treated at T3 facilities, patients treated at lower-tertile facilities had higher risk of death [T2 hazard ratio (HR), 1.17 (95% CI, 1.03-1.33); T1 HR, 1.45 (CI, 1.30-1.67), p < 0.0001] and such a difference in OS was more pronounced in stage IV disease (Table). Patients at T3 facilities (vs T1) were more likely to receive surgical resection of the primary tumor (75 vs 49%), lymph node dissection performed at the time of surgery (70 vs 42%), and a better R0 resection (72 vs 46%) (p < 0.01). Conclusions: Patients who were treated for pNETs at high-volume centers (> 9cases/year) had significantly higher OS and were more likely to receive surgical resection along with lymph node dissection and R0 resection. There was a 45% increased risk of death in patients treated at low volume centers (< 3/year) as compared to high-volume centers (> 9/year). While OS difference was noticed in all stages, the difference was more pronounced in stage IV disease demonstrating the importance of multi-disciplinary approach in pNETs management and incorporating such quality measures in low-volume facilities.[Table: see text]