Effect of Prior Local Therapy on Response to First-line Androgen Receptor Axis Targeted Therapy in Metastatic Castrate-resistant Prostate Cancer: A Secondary Analysis of the COU-AA-302 Trial

Men with localized prostate cancer are often treated with local therapy (LT). However, a proportion of these patients will eventually develop recurrence and progression requiring systemic therapy. Whether primary LT affects the response to this subsequent systemic treatment is unclear. We investigated whether the receipt of prior prostate-directed LT influenced the response to first-line systemic therapy and survival in docetaxel-naïve metastatic castrate-resistant prostate cancer (mCRPC) patients. This is an exploratory analysis of the COU-AA-302 trial, a multicentric double-blinded phase 3 randomized controlled trial in which mCRPC patients with no to mild symptoms were randomized to receive abiraterone plus prednisone or placebo plus prednisone. We compared the time-varying effects of first-line abiraterone in patients with and without prior LT using a Cox proportional hazard model. The cut points were chosen using grid search, and were 6 and 36 mo for radiographic progression-free survival (rPFS) and overall survival (OS), respectively. We also investigated whether there was any difference in treatment effect on score change (relative to baseline) in various patient-reported outcomes (measured by Functional Assessment of Cancer Therapy—Prostate [FACT-P]) over time depending on the receipt of prior LT. The adjusted association of prior LT with survival was determined using weighted Cox regression models. Among 1053 eligible patients, 64% (n = 669) received prior LT. We did not find any statistically significant heterogeneity of time-dependent treatment effect from abiraterone on rPFS in patients with (hazard ratio [HR]: 0.36 [95% confidence interval: 0.27–0.49] at ≤6 mo; 0.64 [0.49–0.83] at >6 mo) or without (HR: 0.37 [0.26–0.55] at ≤6 mo; 0.72 [0.50–1.03] at >6 mo) prior LT. Similarly, there was no significant heterogeneity in time-dependent treatment effect on OS with (HR: 0.88 [0.71–1.10] at ≤36 mo; 0.76 [0.52–1.11] at >36 mo) or without (0.78 [0.60–1.01] at ≤36 mo; 0.55 [0.30–0.99] at >36 mo) prior LT. We did not find sufficient evidence of a difference in treatment effect from abiraterone on score change over time in prostate cancer subscale (interaction p = 0.4), trial outcome index (interaction p = 0.8), and FACT-P total score (interaction p = 0.6) depending on the receipt of prior LT. Receipt of prior LT was associated with a significant improvement in OS (average HR: 0.72 [0.59–0.89]). This study demonstrates that the efficacy of first-line abiraterone and prednisone in docetaxel-naïve mCRPC do not vary significantly based on the receipt of prior prostate-directed LT. Further studies are needed to explore the plausible mechanisms of the association of prior LT with superior OS. This secondary analysis of the COU-AA-302 trial suggests that survival benefits and temporal changes in quality of life with first-line abiraterone in docetaxel-naïve mCRPC do not differ significantly among patients who received versus those who did not receive prior prostate-directed local therapy.