Identification of a novel immune-related prognostic signature associated with tumor microenvironment for breast cancer.

BACKGROUND In the view that immune-related genes play a crucial role in breast cancer progression and long-term patient outcomes, we aimed to identify a novel gene signature based on immune-related genes to improve the prognostic prediction of breast cancer. METHODS RNA sequencing data and clinical information were obtained from The Cancer Genome Atlas (TCGA). Univariate and multivariate Cox regression analyses were conducted to establish the immune-related prognostic signature (IRPS). Then, the IRPS was validated by Kaplan-Meier analyses, time-dependent ROC curve analyses and multivariate Cox regression analyses. External validation was conducted in GSE96058. Nomogram combining IRPS with clinical factors was developed and then validated by time-dependent ROC curve analyses and calibration plots. Quantitative real-time polymerase chain reaction (qRT-PCR) was performed to validate the expression level of immune-related genes in tumor and normal tissues. RESULTS The IRPS based on 4 immune-related genes (CCL1, VGF, TSLP, FABP9) were constructed. Patients in the low-risk group had significantly better overall survival than those in the high-risk group (p = 0.0011 in the training set, p = 0.0043 in the validation set, p < 0.0001 in the entire set, p < 0.001 in the external validation set). Multivariate analyses indicated that IRPS could independently predict OS in the training set (HR, 0.48; 95% CI, 0.24-0.83; p = 0.009), validation set (HR, 0.55; 95% CI, 0.34-0.90; p = 0.018), entire set (HR, 0.52; 95% CI, 0.36-0.75; p < 0.001) and external validation set (HR: 0.74, 95% CI: 0.59-0.92, p = 0.007). Sequentially, we establish a nomogram by integrating IRPS and clinical factors, which showed satisfactory predictive performance with 3-year, 5-year, 10-year AUC of 0.701, 0.706 and 0.694. Results of qRT-PCR validated that higher expression level of FABP9, CCL1 and VGF and lower expression level of TSLP in tumor samples compared to normal tissues. CONCLUSIONS Collectively, a four-gene based IRPS was developed and validated for patients with breast cancer. As an independent and robust predictor, the IRPS was constructive to risk stratification of breast cancer.