| Objective: As an aggressive malignant tumor,breast cancer has a high incidence,which seriously threatens the health of women all over the world and has a trend of becoming younger.The study of breast cancer biomarkers is helpful to the individualized and precise treatment of breast cancer,which is of great significance to improve the prognosis of breast cancer.INF-γ,as an important tumor immune factor,plays a key role in the occurrence and development of breast cancer.It is of profound significance to explore whether INF-γ related genes can be used as effective biomarkers for breast cancer.Methods: Breast cancer gene expression data,gene mutation data and clinical data were downloaded from METABRIC database,GEO database and TCGA database,and the gene set containing 200 INF-γ response genes was downloaded from the MSig DB official website.Univariate survival analysis and multivariate COX proportional hazard regression analysis were used to find the significant prognostic factors(P<0.1).INF-γ score was constructed based on ss GSEA algorithm,gene difference analysis was performed on the subjects grouped according to the median score,then COX proportional risk regression analysis was performed based on LASSO algorithm and Riskscore was calculated,and the cohort was divided into high and low risk groups according to the optimal cut-off value.The survival analysis was performed to determine whether the Riskscore was a breast cancer risk factor.The multivariate COX proportional hazard regression model and Nomogram model were reconstructed to evaluate whether INF-γ risk characteristics could be used as an independent prognostic factor for breast cancer.The correlation between INF-γ risk characteristics and the expression level of immune checkpoint molecules,immune cell infiltration level,gene mutations,and different treatment methods were analyzed to evaluate the prognostic value of the risk characteristics.The GSCALite website and Cellminer database were used to analyze the drug sensitivity of the model genes.Results: After univariate and multivariate COX regression analysis,five variables with independent prognostic value for breast cancer were identified,namely age(age), tumor stage(stage),breast cancer molecular subtype(PAM50),tumor mutation load(TMB)and Menopausal status(Menopausal).A C-index was 0.665(95%CI: 0.646-0.698).INF-γ score was constructed based on 176 INF-γ related genes in METABRIC breast cancer cohort,and the breast cancer cohort was divided into two groups according to the median,and yielded 251 differential genes.The risk characteristics of11 prognostic genes were constructed based on LASSO algorithm.Survival analysis results showed that the INF-γ Riskscore was a risk factor for breast cancer.After Riskscore was included,the prognosis model was reconstructed,and the C-index of the new model was 0.691(95%CI: 0.649-0.775),indicating that this Nomogram model has intermediate-level discrimination.The NRI of the new model was 14.8%(95%CI:-1.7%-28.4%,P=0.08),and the IDI was 3.7%(95%CI: 0.3%-7.6%,P=0.04).Comprehensive comparison shows that the prediction performance of the model with INF-γ risk characteristics is better.The survival analysis and ROC curve of the high and low-risk groups and clinical subgroups of the breast cancer cohort showed that the high-risk group had a worse prognosis,higher immune infiltration level and higher TMB level.Waterfall plot of mutation analysis showed that there was no significant difference in mutation landscape between high and low-risk groups.16 immune checkpoint molecules were higher in the high-risk group,and 5 immune checkpoint levels were higher in the lowrisk group.The levels of fibroblasts and angiogenesis were significantly higher in the high-risk group than in the low-risk group.The CYT score in the low-risk group was significantly higher than that in the high-risk group.High-risk characteristics were associated with worse prognosis in the breast cancer cohorts treated with chemotherapy,hormone therapy,and radiotherapy.In addition,for breast cancer patients who received hormone therapy,the risk scores were all lower than those who did not.The expression levels of the vast majority of DEG were positively correlated with drug sensitivity,suggesting that the high expression of these genes suggested that breast cancer patients may be resistant to multiple drugs,and when other genes were highly expressed,the cells had strong sensitivity to drugs.Conclusion: Based on LASSO algorithm,an INF-γ-related breast cancer risk feature for predicting the prognosis of breast cancer is constructed,which contains a total of 11 related genes.It can be used as a biomarker to guide the clinical and predict the prognosis of breast cancer.Moreover,the efficacy of the prognostic model incorporating this risk characteristic was superior to the prediction model based on traditional clinical variables only. |
PDF Full Text Request