Transcription Factor Profiling to Predict Recurrence-Free Survival in Breast Cancer: Development and Validation of a Nomogram to Optimize Clinical Management

Abstract

Breast cancer (BC) is the most frequently diagnosed cancer and the leading cause of cancer-related death in young women. Several prognostic and predictive transcription factor (TF) markers have been reported for BC; however, they are inconsistent due to small datasets, the heterogeneity of BC, and variation in data pre-processing approaches. This study aimed to identify an effective predictive TF signature for the prognosis of patients with BC. We analyzed the TF data of 868 patients with BC in The Cancer Genome Atlas (TCGA) database to investigate TF biomarkers relevant to recurrence-free survival (RFS). These patients were separated into training and internal validation datasets, with GSE2034 and GSE42568 used as external validation sets. A nine-TF signature was identified as crucially related to the RFS of patients with BC by univariate Cox proportional hazard analysis, least absolute shrinkage and selection operator (LASSO) Cox regression analysis, and multivariate Cox proportional hazard analysis in the training dataset. Kaplan-Meier analysis revealed that the nine-TF signature could significantly distinguish high- and low-risk patients in both the internal validation dataset and the two external validation sets. Receiver operating characteristic (ROC) analysis further verified that the nine-TF signature showed a good performance for predicting the RFS of patients with BC. In addition, we developed a nomogram based on risk score and lymph node status, with C-index, ROC, and calibration plot analysis, suggesting that it displays good performance and clinical value. In summary, we used integrated bioinformatics approaches to identify an effective predictive nine-TF signature which may be a potential biomarker for BC prognosis.