Postoperative pancreatic fistula after resection of pancreatic neuroendocrine tumors: Incidence, predictors, and clinical impact.

4177 Background: Postoperative pancreatic fistula (POPF) remain a major source of morbidity after pancreatic surgery. Data specific to pancreatic neuroendocrine tumors (pNETs) are limited, although these neoplasms differ biologically and surgically from pancreatic adenocarcinoma, potentially influencing POPF risk and outcomes. This study evaluated the incidence, predictors, and clinical impact of POPF in a well-characterized cohort of patients undergoing resection for pNETs. Methods: All patients who underwent resection for histologically confirmed pNETs between 2010-2019 at a tertiary center were retrospectively analyzed. Clinicopathological parameters and surgical procedures were assessed for association with POPF occurrence and severity according to the International Study Group on Pancreatic Fistula (ISGPF) criteria. Statistical analyses included χ², Fisher’s exact, Spearman correlation tests, and multivariable regression analysis; significance was defined as α = 0.05. Results: Among 106 patients (44% female, mean age 61 (± 13.5) years), 85% received a formal pancreatic resection, 15% an atypical resection. More than half (53%) of pNETs were located in the pancreatic tail. The majority (64%) of tumors were G1, only 4% G3, with 72% of diseases in early stages UICC I or II. 12% of pNETs were functional and exclusively insulinomas. 53 patients (50%) developed a POPF, predominantly Grade B (60%), followed by Grade A (34%) and Grade C (6%). No surgery- or POPF-related mortality was observed. POPFs were most frequent after distal pancreatectomy for tail lesions (60% vs. 23% body vs. 17% head; p = 0.065), and in functional pNETs (77% vs. 46% in non-functional pNETs; p = 0.072). Distant metastases were inversely associated with POPF (M0 54% vs. M1 28%; p = 0.043). A nonsignificant trend toward higher POPF rates was observed after atypical resections (73% vs. 48%; p = 0.095). Higher POPF grade was associated with female sex (p = 0.016), and tumor proliferation index (p = 0.046); female sex remained independently associated with POPF grade in multivariable regression analysis (beta = 0.334, p = 0.044). Patients with POPF exhibited lower long-term mortality during follow-up (8% vs. 29%; p = 0.009) and lower mean preoperative chromogranin A levels (82.8 ng/mL vs. 167.8 ng/mL; p = 0.038). Conclusions: POPF occurred in half of all pNET resections, most commonly of Grade B severity. Tumor location, functionality, and absence of metastases were associated with higher POPF risk, and female sex with higher grade POPF. The association of POPF with lower long-term mortality likely mirrors the favorable tumor biology and localized disease profile of patients selected for curative surgery. This underscores the long-term oncological benefit of surgical resection in pNETs and provides relevant insights for postoperative risk stratification and perioperative management.