Evaluating survival post-7 mo PSA response, stratified by response, disease extent, and intermittent vs. continuous ADT in mHSPC: The S9346 trial.

268 Background: A decline in prostate-specific antigen (PSA) following androgen deprivation therapy (ADT) is a well-established prognostic marker in metastatic hormone-sensitive prostate cancer (mHSPC). While S9346 trial found intermittent ADT (IAD) to be not non-inferior to continuous ADT (CAD) in mHSPC, de-intensification strategies are being considered for patients with complete PSA response. We hypothesize that IAD has different treatment effect in men with partial and complete PSA response. Methods: In the phase 3 S9346 trial men with mHSPC were randomized to IAD and CAD, if they had PSA ≤ 4.0 ng/ml after 7 months (PSA-7mo) of treatment with ADT and bicalutamide. We evaluated the association of complete PSA-7mo response (CR, PSA ≤ 0.2 ng/ml) and partial PSA -7mo response (PR, PSA 0.3-4 ng/ml) with subsequent overall survival (OS) in patients with mHSPC treated with either IAD or CAD in S9346 trial using Cox regression. We evaluated association of the disease extent and OS, using Cox regression: minimal disease defined as metastasis confined to the spine, pelvic bones, or lymph nodes; and extensive disease defined as metastasis involving ribs, long bones, or visceral organs. Results: The analysis included 1523 patients from the S9346 trial. In the IAD arm (n=763), 483 men (63%) achieved CR and 280 (37%) had PR. In the CAD arm (n=760), 473 men (62%) achieved CR and 287 (38%) had PR. A PSA-7mo CR was associated with significantly improved OS compared to PR (HR 0.57, 95% CI 0.51-0.65, p<0.0001). Extensive disease was associated with worse OS compared to minimal disease (HR 1.3, CI 1.15-1.47, p<0.0001). However, the relative treatment effect of IAD remained consistent across these subsets of patients: IAD vs CAD in subset of patients with CR HR 1.15, 95% CI 0.96-1.39; and in patients with PR HR 1.14, 95% CI 0.98-1.34. There was no statistically significant pairwise interaction between PSA response, extent of disease, and IAD vs. CAD in the multivariate OS Cox analysis (all p≥ 0.20). Conclusions: This study reaffirms the prognostic significance of a PSA-7mo CR and disease extent. IAD had the same relative treatment effect for those with a PSA CR vs. PR. These findings indicate that IAD is not an optimal therapy even in patients with good prognostic baseline clinical features and on treatment response. Clinical trial information: NCT00002651 . Multivariate Model (n=1523) Hazard Ratio 95% CI p-value CR vs. PR PSA-7mo response 0.57 (0.51, 0.65) <0.0001 Extensive vs. Minimal Disease 1.30 (1.15, 1.47) <0.0001 IAD vs. CAD 1.14 (1.02, 1.29) 0.027 IAD Arm Subset Median OS (95% CI) In months CAD Arm Subset Median OS (95% CI) In months PSA CR, ext. disease (n=228) 71 (61,83) PSA CR, ext. disease (n=193) 71 (58, 82) PSA CR, min. disease (n=255) 77 (66, 95) PSA CR, min. disease (n=280) 92 (85, 103) PSA PR, ext. disease (n=147) 43 (35, 53) PSA PR, ext. disease (n=167) 44 (36, 50) PSA PR, min disease (n=133) 43 (37, 55) PSA PR, min disease (n=120) 60 (43, 76)