Biomarker Expression Can Identify Biologically Distinct Categories of Invasive Breast Cancer
JY Shim, R Zhou, PA Thompson, AM Brewster, JL Murray, ME Edgerton, N Karadag, ML Bondy, A Sahin. MD Anderson Cancer Center, Houston; Arizona Cancer Center, Tucson
Background: Tumor size, stage, lymph node status and histologic grade are significant prognostic factors in breast cancer. We sought to investigate the usefulness of a panel of biomarkers by correlating them with clinicopathologic parameters and outcome.
Design: Tissue microarrays from 1158 patients with stage I and II invasive breast cancer were constructed and immunohistochemical staining for ER, PR, HER2, P53, Ki-67, CK5/6, CK14, CK17, EGFR, Bcl-2, COX-2 and P27 were performed.
Results: The positive rates for ER, PR, HER2, P53, CK5/6, CK14, CK17, EGFR and COX-2 were 66%, 53%, 10%, 57%, 10%, 5%, 9%, 8% and 84%, respectively. The mean percentage of Bcl-2 was 70% and that of P27 was 62%. Ki-67 positivity was low in 61%, intermediate in 26% and high in 14%. Table1 summarizes the correlation between the biomarkers and clinicopathologic parameters. Many biomarkers were associated with each other. In univariate analysis, tumor size and stage were predictive of disease-free survival, while ER, CK14, EGFR, tumor size and stage were predictive of overall survival.