A Novel Prognostic Four-Gene Signature of Breast Cancer Identified by Integrated Bioinformatics Analysis

Abstract

Molecular analysis facilitates the prediction of overall survival (OS) of breast cancer and decision-making of the treatment plan. The current study was designed to identify new prognostic genes for breast cancer and construct an effective prognostic signature with integrated bioinformatics analysis. Differentially expressed genes in breast cancer samples from The Cancer Genome Atlas (TCGA) dataset were filtered by univariate Cox regression analysis. The prognostic model was optimized by the Akaike information criterion and further validated using the TCGA dataset (n = 1014) and Gene Expression Omnibus (GEO) dataset (n = 307). The correlation between the risk score and clinical information was assessed by univariate and multivariate Cox regression analyses. Functional pathways in relation to high-risk and low-risk groups were analyzed using gene set enrichment analysis (GSEA). Four prognostic genes (EXOC6, GPC6, PCK2, and NFATC2) were screened and used to construct a prognostic model, which showed robust performance in classifying the high-risk and low-risk groups. The risk score was significantly related to clinical features and OS. We identified 19 functional pathways significantly associated with the risk score. This study constructed a new prognostic model with a high prediction performance for breast cancer. The four-gene prognostic signature could serve as an effective tool to predict prognosis and assist the management of breast cancer patients.