Development and validation of predictive models for distant metastasis and prognosis of gastroenteropancreatic neuroendocrine neoplasms

Imaging examinations exhibit a certain rate of missed detection for distant metastases of gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs). This study aims to develop and validate a risk prediction model for the distant metastases and prognosis of GEP-NENs. This study included patients diagnosed with gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) from the Surveillance, Epidemiology, and End Results (SEER) database between 2010 and 2015. External validation was performed with patients from the China-Japan Union Hospital of Jilin University. Univariate and multivariate logistic regression analyses were conducted on the selected data to identify independent risk factors for distant metastasis in GEP-NENs. A nomogram was subsequently developed using these variables to estimate the probability of distant metastasis in patients with GEP-NENs. Subsequently, patients with distant metastasis from GEP-NENs were selected for univariate and multivariate Cox regression analyses to identify prognostic risk factors. A nomogram was subsequently developed to predict overall survival (OS) in patients with GEP-NENs. Finally, the developed nomogram was validated using Receiver Operating Characteristic (ROC) curves, calibration curves, and Decision Curve Analysis (DCA). Kaplan-Meier analysis was employed to evaluate survival differences between high-risk and low-risk groups. A total of 11,207 patients with GEP-NENs were selected from the SEER database, and 152 patients from the China-Japan Union Hospital of Jilin University were utilized as an independent external validation cohort. Univariate and multivariate logistic regression analyses revealed that the primary tumor site, tumor grade, pathological type, tumor size, T stage, and N stage are independent predictors of distant metastasis in GEP-NENs. Additionally, among the 1732 patients with distant metastasis of GEP-NENs, univariate and multivariate Cox regression analyses identified N stage, tumor size, pathological type, primary site surgery, and tumor grade as independent prognostic factors. Based on the results of the regression analyses, a nomogram model was developed. Both internal and external validation results demonstrated that the nomogram models exhibited high predictive accuracy and significant clinical utility. In summary, we developed an effective predictive model to assess distant metastasis and prognosis in GEP-NENs. This model assists clinicians in evaluating the risk of distant metastasis and in assessing patient prognosis.