Riskscore Model Based on Oxidative Stress–Related Genes May Facilitate the Prognosis Evaluation for Patients With Colon Cancer

Abstract

INTRODUCTION: This study aims to identify the oxidative stress–related genes (OSRGs) with prognostic value and develop a riskscore model for prognosis evaluation in colon cancer. METHODS: The transcriptome data and corresponding clinical information about colon cancer were extracted from The Cancer Genome Atlas database. Differentially expressed OSRGs and transcription factors (TFs) were identified between the normal and tumor samples. A riskscore model was established with OSRGs filtered from Cox regression analysis. This riskscore model was appraised by Kaplan-Meier plot, receiver operating characteristic curve, and Cox regression analysis. The clinical relevance of riskscore and its association with immune cell infiltration were also evaluated. RESULTS: A total of 307 differentially expressed OSRGs and 64 differential TFs were identified. A TF-OSRG regulatory network was constructed in Cytoscape software. A riskscore model was established based on 17 OSRGs with independent prognostic value. This riskscore model could separate the patients into low-risk and high-risk groups. It also had good predictive ability, with an area under the curve = 0.8. In multivariate Cox regression analysis, age, T stage, and riskscore were identified as independent risk factors in colon cancer. Riskscore was significantly correlated with T stage, N stage, and immune cell infiltration. DISCUSSION: We established a useful riskscore model with 17 OSRGs for prognosticating the overall survival in patients with colon cancer. This study provides a new insight into the clinical utility of OSRG-based riskscore model, which will hopefully facilitate the prognosis evaluation of patients with colon cancer.