P1-06-18: Loss of E-Cadherin Expression in Clinical Breast Cancer Is Associated with an Adverse Outcome on Tamoxifen.

Abstract

Abstract Background: Despite the benefits provided by tamoxifen, not all ER+ patients respond to endocrine therapy and a substantial number of patients who initially respond ultimately relapse whilst on therapy, with endocrine resistance associating with poorer survival. The ability to identify biomarkers predictive of outcome on endocrine treatment would enable more effective stratification of patients most likely to benefit from such measures. Previously, we reported that tamoxifen is able to promote the development of a highly aggressive phenotype in ER+, HER2− endocrine-sensitive breast cancer cells in vitro that have been depleted of the intercellular adhesion molecule, E-cadherin. Here we have investigated the clinical relevance of these observations by examining the association between E-cadherin protein expression and clinical outcome in a series of primary early stage (Stage I-IIIA) invasive ductal carcinomas of the breast. Material and Methods: A series of 794 ER+ breast cancer tissues of known E-cadherin immunohistochemical status (Rakha EA et al (2005) Histopathol 46(6)) were available for this study and comprised both patients who had received tamoxifen for 5 years (n=345) or who had had no endocrine treatment (n=449). E-cadherin was further assessed in a group of ER-negative, tamoxifen-treated primary breast cancers (n=93) as an internal control. The median E-cadherin H-score staining was 120 (range 0 to 300). Results: Kaplain-Meier survival analysis revealed a significant association between reduced E-cadherin expression (< median) and overall survival (OS) in the tamoxifen-treated group at 20 years (p=0.04; Hazard Ratio [HR]=1.51, 95% CI:1.01−2.26) however this relationship was much stronger at 5 years (p=0.02; HR=2.00, 95% CI:1.10−3.65). In the ER+, tamoxifen-untreated tumours, a relationship was observed at 60 months only. Reduced E-cadherin also correlated with metastasis (regional and distant, p=0.01; HR=1.90, 95%CI:1-12-3.24) but only within the tamoxifen-treated group and only within the first 5-years, reflecting the time during which patients would have received tamoxifen. In the ER-, tamoxifen-treated cohort, no adverse impact of reduced E cadherin on these parameters was observed. Subgroup analysis of tamoxifen-treated patients revealed that the association between E-cadherin loss and adverse outcome was apparent in both HER2+ (n=39) and HER2− (n=302) cohorts. However, the association was more significant within the HER2+ group (p=0.02 for OS; HR=3.45, 95% CI:1.08−11.07) (p=0.03 for regional and distant metastasis; HR=2.78, 95%CI:1.02−7.57). These data support our in vitro observations that suppression of E-cadherin within in ER+, HER2+ breast cancer cell lines results in a significant gain in migration and invasion in response to tamoxifen. Discussion: These results are supportive of our hypothesis that low or absent E-cadherin expression in ER+, ductal breast cancer treated with tamoxifen predicts poor survival, particularly within the ER+, HER2+ phenotype. Of note, the presence of tamoxifen appears to be required for this event, suggesting that there is a patient cohort where other systemic endocrine therapy should be considered. E-cadherin represents a biomarker able to discriminate such cases. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P1-06-18.