Stratification of prognosis of triple-negative breast cancer patients using combinatorial biomarkers

Abstract

Background Triple-negative breast cancer (TNBC) is highly diverse group of cancers, and generally considered an aggreive disease aociated with poor survival. Stratification of TNBC is highly desired for both prognosis and treatment decisions to identify patients who may benefit from le aggreive therapy. Methods This study retrieved 192 consecutive non-metastasis TNBC patients who had undergone a resection of a primary tumor from 2008 to 2012. All samples were negative for ER, PR, and HER2/neu. Disease-free-survival (DFS) and overall-survival (OS) were evaluated for expreion of immunohistochemical biomarkers (P53, Ki-67, CK5/6 and EGFR), as well as clinicopathological variables including age, tumor size, grade, lymph node status, pathologic tumor and nodal stages. The cutoff values of the basal biomarkers, EGFR and CK5/6, were estimated by time-dependent ROC curves. The prognostic values of combinatorial variables were identified by univariate and multivariate Cox analysis. Patients were stratified into different risk groups based on expreion status of identified prognostic variables. Results Median age was 57 years (range, 28-92 years). Patients' tumor stage and nodal stage were significantly aociated with OS and DFS. EGFR and CK5/6 were significant prognostic variables at cutoff points of 15% (p = 0.001, AUC = 0.723), and 50% (p = 0.006, AUC = 0.675), respectively. Multivariate Cox analysis identified five significant variables: EGFR (p = 0.016), CK5/6 (p = 0.018), Ki-67 (p = 0.048), tumor stage (p = 0.010), and nodal stage (p = 0.003). Patients were stratified into low basal (EGFR≤15% and CK5/6≤50%) and high basal (EGFR>15% and/or CK5/6>50%) expreion groups. In the low basal expreion group, patients with low expreions of Ki-67, low tumor and nodal stage had significantly better survival than those with high expreions/stages of three variables, log-rank p = 0.015 (100% vs 68% at 50 months). In the high basal expreion group, patient with high basal expreion of both biomarkers (EGFR 15% and CK5/6 50%) had worse survival (mean DFS = 25 months, 41.7%event rate) than those patient with high expreion of either one marker (mean DFS = 34 months, 25.5% event rate). Conclusions Immunoexpreion of basal biomarkers, EGFR and CK5/6, is useful in predicting survival of TNBC patients. Integrated with Ki-67, tumor and nodal stages, combinatorial biomarker analysis provides a feasible clinical solution to stratify patient risks and help clinical decision- making with respect to selecting the appropriate therapies for individual patients.