Identification and Validation of an Immunological Expression-Based Prognostic Signature in Breast Cancer

Abstract

Background: Emerging evidence suggests that the immune system plays a crucial role in the regulation of the response to therapy and long-term outcomes of patients with breast cancer (BRCA). In this study, we aimed to identify a significant signature based on immune-related genes to predict the prognosis of BRCA patients. Methods: The expression data were downloaded from The Cancer Genome Atlas (TCGA). The immune-related gene list, the transcription factor (TF) gene list, and the immune infiltrate scores of samples in the TCGA database were acquired from the ImmPort database, the Cistrome Cancer database, and the TIMER database, respectively. Univariate Cox regression analysis was utilized to identify prognostic immune-related differentially expressed genes (DEGs) (PIRDEGs) in BRCA. A prognostic immune signature containing 15 PIRDEGs in BRCA was established using the least absolute shrinkage and selection operator (LASSO) model with 1,000 iterations followed by a stepwise Cox proportional hazards model with a training set of 508 samples in TCGA. An independent assessment of the prognostic prediction ability of the signature was conducted using Kaplan–Meier survival analysis with a testing set of 505 samples in TCGA. Results: We identified 466 PIRDEGs and 80 TFs among the DEGs. A gene signature containing 15 PIRDEGs was constructed. Risk scores of BRCA patients were calculated using this model, which showed a high accuracy of prognosis prediction in both the training set and testing set and could be an independent prognostic factor of BRCA patients. Conclusions: Our study revealed that a PIRDEG signature could be a candidate prognostic biomarker for predicting the overall survival (OS) of patients with BRCA.