Construction of a prognostic score model for breast cancer based on multi-omics analysis of study on bone metastasis.

Background: Breast cancer (BRCA) is the most common type of cancer and the second leading cause of cancer-related death in women all over the world. Metastasis to bone is an indicator of poor prognosis in BRCA patients. This study aimed to develop a prognostic score model for predicting bone metastasis in patients with BRCA. Methods: BRCA-related RNA sequencing datasets and corresponding clinical information were downloaded from the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA). Differentially expressed genes (DEGs) were screened using Limma package of R software. A risk score based predictive model was constructed based on the key genes identified through univariate Cox regression and the least absolute shrinkage and selection operator (LASSO) Cox regression. The gene expression profiles in BRCA patients were analyzed by gene set variation analysis (GSVA) and gene set enrichment analysis (GSEA). Random survival forest (RSF) analysis of BRCA patients with bone metastasis was conducted to identify the key DEGs. Results: Based on DEG analysis, a total of 677 genes were identified as genes related to bone metastasis in BRCA. By univariate Cox regression and LASSO regression, 28 DEGs were identified as signature genes to develop the prognostic model. A risk score for each patient was created by incorporating the expression values of each specific gene and weighting them with the corresponding estimated regression coefficients. Patients were divided into a low-risk and a high-risk group based on the median risk score. Overall survival (OS) was significantly lower in the high-risk group. The receiver operating characteristic (ROC) curve and multi-omics analysis indicated that the model had high training/testing accuracy and a good clinical predictive value. We used extra data from GEO database to verify the robustness of the prognostic model, and the lower OS in high-risk group and area under the curve (AUC) value indicated the model had strong predictive efficacy for prognosis of BRCA. Conclusions: A prognostic prediction model was constructed based on 28 key DEGs identified through multi-omics analysis of studies on bone metastasis. The model may provide a promising method for distinguishing the high-risk BRCA patients and help on decision making in addition to prognosis prediction for BRCA patients.