Clinical and molecular determinants of survival in pancreatic neuroendocrine tumors: A comprehensive analysis from SEER and TCGA databases.

Abstract

e16355 Background: Neuroendocrine tumors (NETs) are the 3rd most common endocrine tumor, with pancreatic NETs (pNETs) accounting for 6% of all NETs. Next-generation sequencing has helped identify mutations and provided insights into prognosis and potential treatment strategies. In this study, we aim to investigate the prognostic factors and survival outcomes associated with clinical, pathological, and molecular characteristics through a combined analysis of the SEER (Surveillance, Epidemiology, and End Results) and TCGA (The Cancer Genome Atlas) databases. Methods: Using the SEER 2000-2021 database, we identified 10,992 patients based on a combination of ICD-O-3 codes. Clinical variables including age, gender, ethnicity, tumor grade, location of metastasis, and type of treatment, were analyzed. Similarly, data from the TCGA (The Cancer Genome Atlas) accessed via cBioPortal was used to assess survival outcomes and identify correlations with specific genes. Univariate and multivariate analyses were conducted using the Log-rank test and Cox regression, respectively, to evaluate median survival and variable interactions. Results: The median age at diagnosis was 61 yrs, and 45% of patients were women. The cohort was predominantly White (66%). The median tumor size was 25 mm, and 87% of tumors were functional. The median Overall survival (OS) was 97 months. In univariate analysis sex (p < 0.001), age (p < 0.001), tumor size, localized disease (p < 0.001), grade of differentiation (p < 0.001), liver metastasis status (p < 0.001), and surgical treatment (p < 0.001) were identified as prognostic factors. The results of the multivariate analysis are summarized in Table 1. Data from the TCGA database included 98 cases, and mutations in MEN1 (37%), DAXX (17%), ATRX (9%), and PTEN (7%) did not correlate with survival. Correlation between functional tumors, gene mutations, and metastatic disease was assessed using Fisher’s exact test, but no statistical significance was found. Conclusions: In our large population database analysis, female sex, younger age, smaller tumor size, absence of liver metastasis, localized disease, grade I tumors, and surgical treatment were identified as strong predictors of survival. The most common mutations, including MEN1, DAXX, ATRX, and PTEN, were not found to be significant prognostic factors for survival or associated metastases. Multivariate analysis of factors associated with poor survival prognosis. Factor Reference level Hazard ratio p-Value Male Female 1.09 (1.03 to 1.58 0.002 Age > 61 Age <61 1.82 (1.71 to 1.93) <0.001 Tumor size > 25 mm <25 mm 2.15 (1.74 to 2.59) Presence of liver mets No liver mets 2.19 (1.81 to 2.67) Distant mets Localized disease 2.59 (2.38 to 2.89) Regional mets 2.66 (1.84 to 2.31) Grade II Grade I 1.33 (1.18 to 1.49) Grade III 3.15 (2.79 to 3.55) Grade IV 3.51 (2.90 to 4.21) No surgery Surgery performed 2.94 (2.72 to 3.18)