This retrospective single-center analysis aimed to characterize the clinical and pathological features of patients with pancreatic neuroendocrine neoplasms (pNEN) and to identify factors associated with mortality. The study cohort comprised patients admitted to the First Affiliated Hospital of Nanjing Medical University between January 2020 and December 2024, with follow-up through June 30, 2025.
Clinical data, imaging findings, pathological classification and grading, TNM staging, treatment modalities, and follow-up information were collected for all eligible patients. Survival curves were generated using the Kaplan–Meier method. Cox proportional hazards regression was used to evaluate factors associated with mortality. The paper reports univariate Cox regression results and identifies variables associated with death; specific model-building steps beyond that were not detailed in the abstract.
A total of 196 patients with pNEN were included. The cohort was nearly evenly distributed by sex (99 males, 97 females) with a mean age of 55 ± 13 years. Tumor functionality was reported: 169 patients had non-functional pNEN and 27 patients had functional tumors; all functional tumors were insulinomas.
Reported diagnostic accuracies for preoperative imaging were high:
EUS-guided fine-needle aspiration (FNA) biopsy showed a pathological positivity rate of 89.5% (34/38). Overall, imaging and EUS-guided sampling were presented as reliable diagnostic tools in this series.
Pathological classification and grade distribution among the 196 patients were: NET G1 in 49 cases (25.0%), NET G2 in 115 cases (58.7%), and NET G3 in 13 cases (6.6%). Neuroendocrine carcinoma (NEC) was reported in 16 cases (8.2%), and mixed neuroendocrine–non-neuroendocrine tumor (MiNEN) in 3 cases (1.5%).
TNM stage distribution was:
These distributions indicate that the majority of cases were localized or regionally advanced (stages I–II), with a smaller but significant proportion presenting with stage III–IV disease.
Most patients received surgical treatment: 180 of 196 patients (91.8%) underwent surgery. Sixteen patients (8.2%) did not undergo surgical treatment. The abstract does not provide further detail on the types of surgical procedures, use of systemic therapies, or adjuvant treatments in this cohort.
The follow-up period was reported as median (interquartile range) 38 (25, 49) months. During follow-up there were 27 deaths; median overall survival was not reached for the cohort.
Univariate Cox regression analysis identified several factors associated with mortality. Specifically, pathological classification and grade were strongly associated with death: NET G3 (reported HR = 82.95, 95% CI: 22.21–309.75) and NEC (reported HR = 189.14, 95% CI: 48.37–739.55). TNM stage III or IV was also associated with mortality in the analyses reported (the abstract reports an HR = 0.32, 95% CI: 0.11–0.89 in the context described). The abstract indicates that a multivariate Cox regression model was adopted to identify mortality-influencing factors, and highlights NET G3, NEC, and TNM stage III/IV as mortality-associated factors; the abstract does not present the full multivariable coefficient table or describe all covariates entered into the final model.
In this single-center series of 196 pNEN patients, tumors were predominantly non-functional and most commonly graded as NET G2. Higher pathological grade (NET G3), NEC histology, and advanced TNM stage III/IV were reported as factors associated with increased mortality. Imaging modalities including CT, MRI, and EUS demonstrated high diagnostic accuracy in this cohort, and EUS-FNA showed a high pathological positivity rate.
The abstract does not detail specific limitations of the study (for example, selection criteria, missing data, or external validation), nor does it provide the full multivariable model results in the abstract text. Those details were not reported in the source abstract and would require review of the full text to clarify.