Alkaline phosphatase (ALP) decline and pain response as markers for overall survival (OS) in patients (Pts) with metastatic castration-resistant prostate cancer (mCRPC) treated with radium-223 (223Ra) in the REASSURE study

5041 Background: Ra-223 extends OS in pts with mCRPC. Predictive markers of response to Ra-223 are needed to help select pts who would benefit most from Ra-223 therapy. The ALSYMPCA study suggested a correlation between ALP decline and longer OS. Here, we evaluated whether ALP decline is associated with OS in the global real-world REASSURE study. Methods: Pts treated with Ra-223 were grouped by baseline ALP ≤147 U/L vs >147 U/L and any ALP decline vs no decline at week 12 from the first Ra-223 dose. The 147 U/L cut-off was selected based on the highest upper limit of normal for ALP from literature. Pts with a trend of decreasing ALP at closest value to week 12 between weeks 8 and 16 were included in the any decline group. Association of ALP decline with OS was assessed in the ≤147 U/L and >147 U/L groups separately. Median OS is provided with an unadjusted hazard ratio (HR) (95% confidence interval [CI]). Multivariate Cox models provided adjusted HRs (95% CI) for the association of ALP decline with OS. Some baseline covariates were not included in the models due to missing data. Results: 785 of 1465 pts had baseline ALP measurements, of whom 779 had week 12 ALP measurements: 443 had ≤147 U/L and 336 >147 U/L. In the ≤147 U/L group, median OS was 23.0 months (m) (95% CI 20.9–25.7) in pts with ALP decline (n=329) and 16.4 m (95% CI 14.1–20.4) in pts with no decline (n=114). In the >147 U/L group, median OS was 12.9 m (95% CI 11.7–14.3) in pts with ALP decline (n=295) and 8.1 m (95% CI 5.6–10.3) in pts with no decline (n=41). Comparison of unadjusted and adjusted HRs is shown in the Table. Conclusions: Pts with an ALP decline in first 12 weeks of Ra-223 treatment had longer OS. The effect of other baseline variables, including age, PSA, Hgb and prior treatments, provided adjustment but did not change this outcome. [Table: see text]