Identification of a metabolism-related gene expression prognostic model in endometrial carcinoma patients
Abstract Background: Metabolic abnormalities have recently been widely studied in various cancer types. This study aims to explore the expression profiles of metabolism-related genes (MRGs) in endometrial cancer (EC). Methods: We analyzed the expression of MRGs using The Cancer Genome Atlas (TCGA) data to screen differentially expressed MRGs (DE-MRGs) significantly correlated to EC patients’ prognosis. Functional pathway enrichment analysis of DE-MRGs were investigated. LASSO algorithm and Cox regression analysis were performed to select MRGs closely related to EC patients’ outcomes. A prognostic signature was developed and the efficacy were validated in part of and the entire TCGA EC cohort. Moreover, we developed a comprehensive nomogram including the risk model and clinical features to predict EC patients' survival probability.Results: Forty-seven differentially expressed MRGs (DE-MRGs) were significantly correlate to EC patients’ prognosis. Functional enrichment analysis showed these MRGs were highly enriched in amino acid, glycolysis, and glycerophospholipid metabolism. Nine MRGs were screened out to closely relate to EC patients’ outcomes, which are CYP4F3, CEL, GPAT3, LYPLA2, HNMT, PHGDH, CKM, UCK2 and ACACB. Based on nine DE-MRGs, we developed a prognostic signature and its efficacy in part of and the entire TCGA EC cohort was validated. The nine-MRGs signature was independent of other clinical features, and could effectively distinguish high- or low-risk EC patients and predicted patients' OS. The nomogram showed excellent consistency between prediction and actual survival observation. Conclusions: The MRG prognostic model and the comprehensive nomogram could guide for precise outcom predicting and rational therapy in clinical practice.