A novel anoikis-related gene signature predicts in lung adenocarcinoma

Abstract

Objective: To generate a prognostic model prognosis based on anoikis-related genes (ANRGs) expression for Lung adenocarcinoma (LUAD), an exploration of the prognostic value of ANRGs in LUAD was conducted. Methods: Based on the expression matrix of genes from the TCGA database, we built the co-expressed modules by weighted gene co-expression network analysis (WGCNA). Then we identified the differentially expressed ANRGs (DE-ANGs) between LUAD and normal samples by the WGCNA results, DEGs, and the 345 ANRGs. The biofunction of the DE-ANRGs was interpreted using the GO and KEGG databases. Univariate and multivariate regression models were used to verify whether the risk model could serve as an independent prognostic factor. A nomogram was utilized to predict overall survival (OS) in LUAD. Results: The expression of 56 DE-ANRGs was significantly different in tumor tissues. We established a 4-ANRG prognostic signature. In the TCGA cohort and the external GSE31210 validation cohort, the OS of the high-risk group was lower than that of the low-risk group significantly. Moreover, the prediction performance of the risk model was excellently verified by the ROC curve. In addition, both univariate COX and multivariate Cox analyses indicated that risk score could act as an independent prognostic factor for LUAD patients. The calibration curve and C-index demonstrated that the nomogram was satisfactory in predicting 1, 3- and 5-year survival in LUAD patients. Conclusions: Our study developed a novel prognostic signature of 4 ANRGs with Excellent prognostic performance for LUAD patients.