Androgen receptor in advanced breast cancer: is it useful to predict the efficacy of anti-estrogen therapy?

Background: Androgen receptor (AR) is widely expressed in breast cancer (BC) but its role in estrogen receptor (ER)-positive tumors is still controversial. The AR/ER ratio has been reported to impact prognosis and response to antiestrogen endocrine therapy (ET). Methods: We assessed whether AR in primary tumors and/or matched metastases is a predictor of efficacy of first-line ET in advanced BC. Patients who had received first-line ET (2002-2011) were recruited, while those given concomitant chemotherapy or trastuzumab or pretreated with > 2 lines of chemotherapy were excluded. ER, progesterone receptor (PgR), Ki67 and AR expression were assessed by immunohistochemistry, and HER2 mainly by fluorescent in-situ hybridization. Cut-offs of 1 and 10% immunostained cells were used to categorize AR expression. Results: Among 102 evaluable patients, biomarkers were assessed in primary tumors in 70 cases and in metastases in 49, with 17 patients having both determinations. The overall concordance rate between primary tumors and metastases was 64.7% (95% CI 42%-87.4%) for AR status. AR status did not affect TTP significantly, whereas PgR and Ki67 status did. AR/PgR >= 0.96 was associated with a significantly shorter TTP (HR = 1.65, 95% CI 1.05-2.61, p = 0.028). AR status in primary tumors or metastases was not associated with progressive disease (PD) as best response. In contrast, Ki67 >= 20% and PgR < 10% showed a statistically significant association with PD as best response. Conclusions: AR expression does not appear to be useful to predict the efficacy of ET in advanced BC, whereas Ki67 and PgR exert a greater impact on its efficacy.