Identifying candidates for prednisone to dexamethasone switch amongst patients with castration-resistant prostate cancer undergoing abiraterone therapy.

e16568 Background: Abiraterone plus prednisone (AA+P) delays symptomatic progression, palliates symptoms, and improves survival of pts with castration-resistant prostate cancer (CRPC). While inherent or acquired therapeutic resistance limits the benefit of AA+P, emerging data suggest that a switch from prednisone to dexamethasone (D) may achieve durable biochemical responses in CRPC pts progressing on AA+P. Methods: To better understand who may benefit from a steroid switch, we analyzed the outcome of pts switched to AA+D upon biochemical progression on AA+P, among CRPC pts treated at Sunnybrook Odette Cancer Centre, Toronto, Canada. Pts were included if they had undergone AA+D therapy (D 0.5 or 1.0 mg po daily) for at least 12 weeks. Results: 18 pts fit the inclusion criteria, of which 11 were chemotherapy-naïve, and 7 had received prior docetaxel. None had received P or D monotherapy previously. 8 pts (44%) achieved a > 50% PSA reduction with AA+P (PSA50), and 3 pts (17%) had a > 30% PSA response (PSA30). In 6 pts (33%) we found PSA stabilization (PSA-SD) as best response, whereas 1 pt (6%) biochemically progressed on AA+P. Mean (±SD) AA+P treatment duration was 213.44 (±151.24) days. Upon switch to AA+D, 5 pts (28%) achieved a PSA50, 1 pt (6%) a PSA30, PSA remained stable in 11 pts (61%), and 1 pt (6%) progressed biochemically. 6 pts are still continuing on AA+D after a median follow-up of 278 days. Amongst 11 AA+P responders (ie PSA50 or PSA30), 4 pts also achieved a sustained PSA50 on AA+D (202, 251, 399+, 587+ days), and 1 pt a PSA30 (274+). By contrast, only 1 of the 7 AA+P non-responders (14%) achieved a PSA response. Prior docetaxel did not affect the response to steroid switching. No unexpected adverse events were observed. Further data is being collected prospectively. Conclusions: In these mostly chemotherapy-naïve CRPC pts undergoing AA therapy, switching from P to D resulted in sustained PSA50 or PSA30 in 33% of pts. Whereas responses were seen in 5 of 11 pts (45%) with prior response to AA+P, the benefit of a P to D switch seems to be very limited in pts without prior response to AA+P.