Development and verification of a personalized immune prognostic feature in breast cancer

Abstract

Immune-related genes have great potential as prognostic markers in many types of cancer. Therefore, we have attempted to develop immune-related gene markers to enhance the prognosis of breast cancer; 1159 samples of breast cancer gene expression data and clinical follow-up messages were downloaded from TCGA and GEO, which were classified into training set, test set, and validation set. In the training set, the gene pairs are established according to the relative expression levels between 320 immune genes, in which the prognosis-related gene pairs are screened, and Lasso is used for feature selection to screen the robust biomarkers. A prognostic model of immune gene correlation was set up and verified. Sixty-six IRGPs were obtained, and 17-IRGPs signature was established. 17-IRGPs signature is an independent prognostic indicator for BC patients, which can stratify the risk in the training set and testing series, and AUC of five years survival was greater than 0.7; 17-IRGPs signature had better classification performance in patients with advanced BC. In addition, we compared the prognostic characteristics of 17-IRGPs with four reported breast cancers and clinical stages; 17-IRGPs achieved the highest average C index (0.7, P < 0.05), and functional analysis found that the dysregulated immune environment may be the cause of the observed difference in survival between patient groups defined by our characteristics. 17-IRGPs signature was constructed as a newly developed prognostic indicator to calculate the survival of BC patients. Impact statement Breast cancer is among the highest prevalent malignant tumors worldwide with a low survival ratio. Immune-related genes have great potential as prognostic indicator in many types of tumors. Therefore, we have attempted to develop immune-related gene markers to enhance the prognosis of breast cancer. 17-IRGPs signature was constructed as a newly developed prognostic indicator to predict the survival of BC patients.