10081 Background: We have previously reported the translation of gene expression data into a five reagent polyclonal antibody immunohistochemistry assay targeting p53, SLC7A5, NDRG1, HTF9C, and CEACAM5 whose staining results are combined using a Cox proportional hazard model for prognostication of estrogen receptor expressing breast cancer. Monoclonal antibodies targeting these genes have now been obtained and or produced. Methods: Paraffin blocks from 229 patients seen at the University of Alabama at Birmingham Clinical Cancer Center were assembled into tissue arrays and clinical follow-up data was collected by chart review. Immunohistochemical stains were performed and cases scored as positive or negative with the five monoclonal antibody assay. In order to further test the assays predictive value for assessing outcome, the prospectively defined Cox proportional hazard model and high, moderate, and low risk cutoffs were applied to the staining results on the estrogen receptor positive (ER+) patients for whom a assessment of death due to disease (DOD) and staining results for all five biomarkers were available. Results: The Kaplan-Meier estimates of outcome confirmed that the Cox model distinguished ER+ patients with poor outcomes. The five antisera algorithm identified high risk patients with a five year estimate of DOD of 41% compared to 20% for moderate and 9% for good (p=0.006) (79 patients total). In ER+ node negative patients the model identified high risk patients with a five year estimate of DOD of 50% compared to 18% for moderate and 5% for good (p=0.01) (40 patients total). The prognosticator was independent of age, pathological stage, and tumor size. Conclusions: A panel of five antibodies can significantly improve upon traditional prognosticators in predicting outcome for estrogen receptor positive breast cancer patients. This represents the third independent validation study performed on cohorts from different institutions. Retrospective studies on clinical trial cohorts and prospective studies should be performed to further establish the utility of this test in identifying early stage breast cancer patients at high risk for recurrence. [Table: see text]
Read full abstract