Identifying and Validating of an Autophagy-Related Gene Signature for the Prediction of Early Relapse in Breast Cancer

BACKGROUND: Compelling evidence has demonstrated the pivotal role of autophagy in the prognosis of breast cancer. Breast cancer (BC) patients with early relapse consistently exhibited worse survival. METHODS: The autophagy-related genes were derived from the Human Autophagy Database (HADb) and high-sequencing data were obtained from The Cancer Genome Atlas (TCGA). Discrepantly expressed autophagy genes (DEAGs) between early relapse and long-term survival groups were performed using the Linear Models for Microarray data (LIMMA) method. Lasso Cox regression analysis was conducted for the selection of the 4-gene autophagy-related gene signature. GSE42568 and GSE21653 databases were enrolled in this study for the external validation of the signature. Then patients were divided into high and low-risk groups based on the specific score formula. GSEA was used to discover the related signaling pathway. The Kaplan-Meier curves and the receiver operating characteristic (ROC) curves were used to evaluate the discrimination and accuracy of the 4-gene signature. RESULTS: A signature composed of four autophagy-related mRNA including APOL1, HSPA8, SIRT1, and TP73, was identified as significantly associated with the early relapse in BC patients. Time-dependent receiver-operating characteristic at 1 year suggested remarkable accuracy of the signature [area under the curve (AUC = 0.748)]. The risk score model based on the autophagy-related signature showed favorable predicting value in 1-, 2-, and 3-year relapse-free survival (RFS) in training and two validating cohorts. The GSEA displayed gene sets were remarkably enriched in carcinogenic activation pathways and autophagy-related pathways. The nomogram involving three variables (progesterone receptor status, T stage, and 4-gene signature) exhibited relatively good discrimination with a C-index of 0.766. CONCLUSIONS: Our study establishes an autophagy-related 4-gene signature that can effectively stratify the high-risk and low-risk BC patients for early relapse. Combined with the clinicopathological variables, the signature could significantly help oncologists tailor more efficient treatment strategies for BC patients.