Surgical Outcomes and Diagnostic Concordance in Patients With Pancreatic Neuroendocrine Tumors Undergoing Pancreatectomy: A Retrospective Cohort Study.

Background/Aim: Pancreatic neuroendocrine tumors (PanNETs) are heterogeneous neoplasms for which surgical resection remains the only potentially curative therapy. However, preoperative diagnostic accuracy - particularly tumor grading - often varies, complicating treatment decisions. This study evaluated diagnostic concordance between preoperative assessments and postoperative pathology, as well as surgical outcomes and prognostic factors in patients undergoing pancreatectomy for PanNETs. Patients and Methods: We retrospectively reviewed the clinical records of 32 patients who underwent surgical resection for PanNETs. Patient demographics, tumor characteristics, surgical procedures, and postoperative outcomes were analyzed. Preoperative imaging and cytology-based diagnoses were compared with final pathological findings to evaluate diagnostic concordance. Prognostic factors were assessed using Kaplan-Meier survival analysis. Results: Histological grading of resected specimens showed G1 in 53%, G2 in 41%, and G3/NEC in 6%, with a 38% discordance rate from preoperative biopsy. Lymphatic, venous, and perineural invasions were identified in 16%, 44%, and 13% of cases. Lymph node metastasis occurred in 22%. Among 30 patients with follow-up >12 months, eight developed recurrence, most commonly in the liver. The 3-year and 5-year disease-free survival (DFS) rates were 78% and 69%, while overall survival (OS) rates were 96% and 91%. Tumor number >2 and histologic grade G3 were significantly associated with reduced DFS (p<0.05). Lymphatic invasion and metachronous liver metastasis were significantly associated with reduced OS (p<0.01 and p<0.05, respectively). Histological grading mismatch was not associated with survival outcomes. Conclusion: Pancreatectomy for PanNETs is a safe and effective treatment with favorable long-term outcomes. Histological factors such as tumor grade, lymphatic invasion, and tumor number significantly predict recurrence and survival.