Establishment of prognostic risk model and drug sensitivity based on prognostic related genes of esophageal cancer

Abstract

AbstractAt present, the treatment of esophageal cancer (EC) is mainly surgical and drug treatment. However, due to drug resistance, these therapies can not effectively improve the prognosis of patients with the EC. Therefore, a multigene prognostic risk scoring system was constructed by bioinformatics analysis method to provide a theoretical basis for the prognosis and treatment decision of EC. The gene expression profiles and clinical data of esophageal cancer patients were gathered from the Cancer Genome Atlas TCGA database, and the differentially expressed genes (DEGs) were screened by R software. Genes with prognostic value were screened by Kaplan Meier analysis, followed by functional enrichment analysis. A cox regression model was used to construct the prognostic risk score model of DEGs. ROC curve and survival curve were utilized to evaluate the performance of the model. Univariate and multivariate Cox regression analysis was used to evaluate whether the model has an independent prognostic value. Network tool mirdip was used to find miRNAs that may regulate risk genes, and Cytoscape software was used to construct gene miRNA regulatory network. GSCA platform is used to analyze the relationship between gene expression and drug sensitivity. 41 DEGs related to prognosis were pre-liminarily screened by survival analysis. A prognostic risk scoring model composed of 8 DEGs (APOA2, COX6A2, CLCNKB, BHLHA15, HIST1H1E, FABP3, UBE2C and ERO1B) was built by Cox regression analysis. In this model, the prognosis of the high-risk score group was poor (P < 0.001). The ROC curve showed that (AUC = 0.862) the model had a good performance in predicting prognosis. In Cox regression analysis, the comprehensive risk score can be employed as an independent prognostic factor of the EC. HIST1H1E, UBE2C and ERO1B interacted with differentially expressed miRNAs. High expression of HIST1H1E was resistant to trametinib, selumetinib, RDEA119, docetaxel and 17-AAG, High expression of UBE2C was resistant to masitinib, and Low expression of ERO1B made the EC more sensitive to FK866. We constructed an EC risk score model composed of 8 DEGs and gene resistance analysis, which can provide reference for prognosis prediction, diagnosis and treatment of the EC patients.