A novel prognostic immunohistochemical biomarker panel for estrogen receptor expressing breast cancer. Poster at San Antonio Breast Cancer Symposium 2005

Ross DT¹, Ring BZ¹, Seitz RS¹, Shasteen WJ¹, Beck R¹, Tarr SM², Cheang MCU³, Yoder BJ², Budd, GT², Nielsen TO³, Hicks DG², Estopinal NC^4
1Applied Genomics Inc., Sunnyvale, CA; ²Cleveland Clinic Foundation, Cleveland, OH; ³Genetic Pathology Evaluation Centre, U. of British Columbia, Vancouver BC; ⁴Comprehensive Cancer Institute of Huntsville, Huntsville, AL

Abstract #3003

Background: Gene expression signatures have been identified that distinguish biologically and clinically significant breast tumor diversity. We have used gene expression data to target the production of hundreds of novel antibody reagents to explore whether a panel of targeted immunohistochemistry reagents can distinguish similar tumor diversity and identify patients at increased risk of poor outcome.

Methods: Eighty novel and seven commercially available antisera were used to stain tissue arrays containing a retrospective breast cancer cohort from the Comprehensive Cancer Center of Huntsville (550 patients). Cox proportional hazard analysis was used to identify a recurrence score algorithm that uses only five antisera to predict risk of recurrence in estrogen receptor-positive, lymph node-negative patients. To validate this model and antibody panel, we then tested the prognostic association of this prospectively defined algorithm by staining two independent tissue array cohorts, with linked clinical follow-up data.

Results: In both validation cohorts, the Kaplan-Meier estimates of recurrence confirmed that the Cox model distinguished estrogen receptor-positive patients with poor outcomes. In a cohort assembled at the British Columbia Cancer Agency (440 patients total) the five antisera algorithm identified ER+ patients with a high risk of recurrence that showed an overall 5 year survival rate of 56% compared to 74% for moderate and 89% for good (p=0.003). In the Cleveland Clinic Foundation Cohort (292 patients total) , the Cox model identified ER+ 'bad' patients with a five year recurrence rate of 51% compared to about 84% for patients classified as either moderate or good (p=0.0022). Although both cohorts were underpowered to test the association of the prognosticator with ER+ lymph node negative patients, using multivariable analysis the calculated risk of recurrence was independent of stage, grade and lymph node status.

Conclusion: A panel of five antibodies can significantly improve upon traditional prognosticators in predicting outcome for estrogen receptor positive breast cancer patients. Additional retrospective and prospective studies focused on early stage patients are indicated for further validation.

	(300 KB PDF)