Comprehensive multi-omics analysis identifies an immune related predicting model in early stage breast cancer

Abstract

Background Breast cancer is known the highest incidence cancer in women. Tumor-infiltrating immune cells were reported closely related to cancers’ fate but it’s function still not very clear in breast cancer. The aim of our study is to build a infiltrating immune cells based nomogram model to better predict patients survival and explore its relationship with immune features. Methods We first use CIBERSORT to analyze 22 immune cells’ status in two unrelated breast cancer cohorts (TCGA and METRABRIC). The univariate and multivariate Cox analyses were used to analyze the prognostic ability of immune cells and other clinicopathological factors. Nomogram model were built by using independent prognostic factors. Different immune signatures and gene enrichment analysis were performed in different nomogram risk groups. Results Multivariate cox analysis showed that Macrophages M2 with HR of 1.733 (95% CI: 1.013-2.966) in TCGA cohort and 1.334 (95% CI: 1.125-1.581) in METABRIC cohort is the only independent prognostic factor among the 22 immune cells in early stage breast cancer. Macrophages M2, age, TNM stage and molecular types were used to build the nomogram model. The AUC of the ROC of nomogram reached 0.732 in TCGA cohort and 0.702 in METABRIC cohort. Using the nomogram score can better classified patients to low and high risk group. High risk group showed higher malignant signatures and predicted immunotherapy response rates than low risk group which consistent with the gene entrenchment analysis. Conclusion In this study, we revealed that M2 macrophages could predict the OS of breast cancer patients. Based on M2 and other clinical features we established a nomogram model which were significantly different in immune features that can better assess the OS risk or be a predictor for the immunotherapy response in breast cancer for further research.