P5-13-19: Borderline Estrogen and Progesterone Receptor Expression and Efficacy of Anti-Estrogen Therapy Analyzed by Subpopulation Treatment Effect Pattern Plot Analysis.

Abstract

Abstract Background: The American Society of Clinical Oncology (ASCO) and the College of American Pathologists (CAP) recently recommended that estrogen receptor (ER) and progesterone receptor (PR) values be considered positive in the evaluation of breast cancer specimens if there are at least 1% positive tumor nuclei by immunohistochemisty (IHC) scoring. This change constitutes a lower cutoff value than the 5% or 10% widely used in the past and increases the number of patients that can now be treated with anti-estrogen therapy, a therapy which carries significant treatment efficacy and a favorable toxicity profile. Despite the new recommendation however, it remains unclear if patients whose tumors display borderline ER and/or PR expression truly benefit from anti-estrogen therapy. Methods: We examined a total of 608 patient samples from the OHSU Knight Cancer Institute Breast Cancer Tissue Repository for which we have quantitative ER and PR information and their clinical outcome. Of those, 282 patients did not receive any systemic treatment and 165 received anti-estrogen therapy (the majority tamoxifen). Cytosolic steroid hormone receptors were quantitated by conventional steroid binding assay and Scatchard plot analysis or by enzyme-linked immunoassay system. For immunoassays, human estrogen receptor (ER) and progesterone receptor (PR) kits were obtained from Abbott Laboratories and assays were conducted according to manufacturer's specification. We performed subpopulation treatment effect pattern plot (STEPP) analysis to explore the presence of interaction between treatment effects and quantitative ER and PR levels. Clinical outcome used in this study includes 5-year overall survival and 5-year breast cancer free interval. Results: Multi-variate analysis finds neither ER nor PR expression adjusted by treatment status is associated with either overall survival (P= 0.20, median follow up of 63 months) or breast cancer free interval (P= 0.82, median follow up of 60 months). For both ER and PR, there is no interaction between treatment outcome (5 year survival probability) and levels of hormone receptor expression based on Kaplan-Meier estimate (P=0.36 and P= 0.65, respectively). Patients with borderline ER expression appear to benefit from anti-estrogen therapy while patients with borderline PR expression appear to have worse outcome with anti-estrogen therapy. Conclusions: Though descriptive in nature, our STEPP analysis provides support for the recommendation by ASCO and CAP to lower the cutoff for ER positivity in breast cancer specimens. There is overall no interaction between the level of hormone receptor and treatment benefit for all patients from low ER to high ER. Therefore, treatment will benefit all patients including those with low ER levels. Extending hormone therapies to the group of patients whose tumors express borderline PR values, however does not seem to confer therapeutic benefit. More study is needed to determine the benefits of offering anti-estrogen therapy to patients whose tumors express low PR. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P5-13-19.