Prediction of breast cancer prognosis by gene expression profile of TP53 status

Abstract

TP53 mutations are a poor prognostic factor in breast cancers. The present study sets out to identify the gene set that determines the expression signature of the TP53 status (TP53 signature) and to correlate it with clinical outcome. Using comprehensive expression analysis and DNA sequencing of the TP53 gene in 38 Japanese breast cancer patients, a gene set from differentially expressed genes was isolated, depending on the TP53 status. As independent external datasets, two published datasets were introduced for validation of TP53 status predictions (251 Swedish samples) and survival analysis (both the Swedish and 295 Dutch samples). Thirty-three gene sets were identified from microarray analysis. Predictive accuracy of the TP53 status by gene expression profiling was 83.3% in the test set (n = 12). TP53 signature has the ability to predict recurrence-free survival (RFS) of 29 stage I and stage II Japanese breast cancers (log rank, P = 0.032), and RFS, overall survival of two independently published datasets (log rank, both P < 0.0001). Multivariate analysis has shown that the TP53 signature an independent and significant prognostic factor with a hazard ratio (HR) for recurrence and survival in two external datasets (P < 0.0001). The TP53 signature is also a strong prognostic factor in the subgroups: estrogen-receptor positive, lymph node positive and negative, intermediate/high risk in St. Gallen criteria, and high risk in National Cancer Institute (NCI) criteria (log rank, P < 0.0001). TP53 signature is a reliable and independent predictor of the outcome of disease in operated breast cancer.