The association of D4, 3-keto-abi (D4A) and response to abiraterone in castration-resistant prostate cancer (CRPC).

Abstract

245 Background: Abiraterone (Abi), a potent inhibitor of 17α-hydroxylase/17,20-lyase (CYP17A1), is a standard treatment for men with metastatic CRPC. Abi is converted to D4A by 3β-hydroxysteroid dehydrogenase (3βHSD). D4A inhibits CYP17A1, 3βHSD, and steroid-5α-reductase (SRD5A) and has direct androgen receptor antagonist activity, which together make it a more potent agent than Abi in xenograft models. It is not known if conversion to D4A in patients (pts) correlates with response or resistance to Abi. Methods: Blood was collected (single time point on Abi) from CRPC pts who started Abi during 2011-2015. Abi and D4A were extracted from serum and analyzed by mass spectrometry. The purpose of this ongoing study is to assess the potential correlation between D4A and response to treatment. Results: 32 patients with CRPC had blood collected. 4 pts (12.5%) received ketoconazole and 6 (18.8%) chemotherapy prior to Abi. Median pre-Abi prostate-specific antigen (PSA) was 14.3 ng/ml (0.6-646.1). Median time from initiation of androgen deprivation therapy (ADT) to CRPC was 34.3 months (m) (6-129.6) and median time from CRPC to Abi initiation was 4.8 m (0-14.9). PSA decline > 50% (PSA50) on Abi was seen in 68% (12/31) and 73% (19/26) of pts at 3 and 6 m, respectively. Treatment with Abi was ongoing in 23/32 pts (74%), and discontinued in 8 pts (26%) due to disease progression (no follow up data on 1 pt) with median duration of Abi treatment 14.6 m (2.9-44.4 m) at last follow up. As the absolute concentration of detected levels of Abi and D4A varied significantly among pts, the percentage of the total extracted metabolites was used to assess correlation with response. Abi and D4A comprised 94.3% (73.8-97.4%) and 5.7% (2.6-26.2%), respectively, of total levels. Conclusions: Abi absorption and metabolism significantly vary among pts. Most of these pts had prolonged duration of response to Abi (74% ongoing treatment and median treatment duration 13.3 m in 8 pts that came off treatment). Longer follow up, accrual of pts with shorter duration of response, and sampling at multiple time points on Abi and at progression is ongoing to further evaluate the impact of D4A on treatment response.