A Prognostic Model of Genetic Markers for Triple-Negative Breast Cancer Based on Machine Learning and Bioinformatics Analysis.

Triple negative breast cancer (TNBC) that has low survival rate and prognosis due to its heterogeneity and lack of reliable molecular targets for effective targeted therapy. Therefore, finding new biomarkers is crucial for the targeted treatment of TNBC. The experimental data from the Cancer Genome Atlas database (TCGA).First, key genes associated with TNBC prognosis were screened and used for survival analysis using a single-factor COX regression analysis combined with three algorithms: LASSO, RF and SVM-RFE. Multi-factor COX regression analysis was then used to construct a TNBC risk prognostic model. Four key genes associated with TNBC prognosis were screened as TENM2, OTOG, LEPR and HLF. Among them, OTOG is a new biomarker. Survival analysis showed a significant effect of four key genes on OS in TNBC patients (P<0.05). The experiment showed that four key genes could provide new ideas for targeting therapy for TNBC patients and improved prognosis and survival.