Nomogram individually predicts the overall survival of patients with gastroenteropancreatic neuroendocrine neoplasms

Abstract

BACKGROUND: The current study aimed to establish a novel nomogram to predict the overall survival of individual Chinese patients with gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs). Furthermore, this study sought to externally validate this nomogram using the Surveillance, Epidemiology, and End Results (SEER) database. METHODS: The records of 1183 patients with GEP-NENs treated at five high-capacity institutions in China between 2005 and 2015 were retrospectively analysed. In addition, 10 236 GEP-NEN cases from the SEER database were included as an external validation set. RESULTS: A multivariate analysis using Cox proportional hazards (PHs) regression was performed, and a nomogram was constructed. Discrimination, calibration, and external validation were performed using the SEER data set. The multivariate Cox model indicated that age, tumour size, differentiation, lymph node metastases, and distant metastases were independent covariates associated with survival. With respect to the training set, the nomogram exhibited better discrimination power than TNM classification (Harrell's concordance index (C-index): 0.837 vs 0.784, P=0.006). Discrimination was also excellent and superior to that of TNM classification for the SEER-based validation set (C-index: 0.808 vs 0.717, P<0.001). The calibrated nomogram predicted a survival rate that closely corresponded to the actual survival rate. CONCLUSIONS: We developed a nomogram that predicted the 3- and 5-year overall survival rates of patients with GEP-NENs. Validation revealed excellent discrimination and calibration for this nomogram, suggesting that it exhibits satisfactory clinical utility that might improve individualised predictions of survival risks and lead to the creation of additional clinical therapies.