Abstract P4-11-05: Does androgen receptor (AR) expression impact on residual risk? A TEAM pathology study

Abstract

Abstract Introduction: Several studies have suggested that AR expression, particularly in luminal cancers following endocrine therapy, may be associated with improved outcome in early breast cancer. We performed an analysis of AR expression in the TEAM pathology cohort to test the hypothesis that AR would represent an independent predictor of residual risk following adjuvant endocrine therapy. Methods: Triplicate 0.6mm2 TMA cores from the TEAM pathology cohort (N=4,598) were stained using AR clone ER179(2) on a Ventana automated staining platform and analysed by image analysis using the Ariol Image analysis platform. Continuous histoscores were generated as previously described (Bartlett et al, JCO 2011). Results: AR histoscores were generated from 3866/4598 (84%) of available cases. Median AR histoscores were 227 (interquartile range 195-262). In a univariate Cox proportional hazard model with AR histoscore as a continuous variable, increased AR histoscores were significantly associated with a reduced hazard of distant disease relapse or death from breast cancer. The hazard ratio (HR) associated with a 50 unit increase in the histoscore was 0.88, 95% confidence interval (95%CI) 0.83-0.93, P<0.0001. There was evidence that a log transformation of the histoscore resulted in a better fitting model (P<0.0001) resulting in the following model estimates HR= 0.93, 95% CI 0.89-0.98, P=0.006. However, a multivariate model of AR histoscore including other known prognostic factors such as age, grade, tumour size, number of positive nodes, HER2 status, ER and PgR histoscores found AR histoscore was not independently prognostic for distant relapse or death (HR=1.00, 95% CI 0.94-1.06, P=0.96). There was no significant interaction between AR expression and type of endocrine treatment (Tamoxifen →exemestane versus exemestane alone) in either univariate (HR 1.003 95%CI 0.91-1.11, p=0.96) or multivariate (HR 0.92, 95%CI 0.81-1.04, p=0.18) analysis. Logistic regression analysis was performed to investigate the association between AR histoscore (2 groups above and below the median) and the known prognostic factors mentioned above. Increased AR histoscore is associated with good risk factors; young age, low grade, small tumours, decreasing ki67 and increasing ER and PgR histoscores. Conclusion: AR expression is common in luminal breast cancers. However, in this study AR histoscore does not add residual risk information beyond what can already be assessed using conventional prognostic factors. High AR expression is associated with good prognostic factors, including young age, low tumour grade, small tumour size, lower Ki67 and higher ER/PgR expression. Citation Format: John MS Bartlett, Cassandra L Brookes, Fu J Yan, Mary Anne Quintayo, Jane Bayani, Jane Starczynski, Cornelis JH van de Velde, Annette Hasenburg, Dirk G Kieback, Christos Markopoulos, Luc Dirix, Caroline Seynaeve, Daniel W Rea. Does androgen receptor (AR) expression impact on residual risk? A TEAM pathology study [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-11-05.