Ability of serum alkaline phosphatase (ALP) changes to complement PSA changes and predict survival in men with metastatic castration-resistant prostate cancer (mCRPC) receiving docetaxel and prednisone (DP).

e15019 Background: PSA decline ≥30% and ≥50% within 90 days of starting chemotherapy are suboptimal surrogates for overall survival (OS) and only capture a proportion of total benefit. Changes in ALP were shown to complement PSA changes as prognostic factors for overall survival [OS] in men treated with chemotherapy in the TAX327 trial dataset. This analysis was performed to independently validate these findings in a different dataset. METHODS: 221 men recruited to a randomized phase II trial of DP+AT-101 (bcl-2 inhibitor) versus DP+placebo were studied. No significant differences in outcomes were observed; hence, both arms were combined for analysis. Changes in ALP and PSA prior to day 90 were measured and evaluated for prognostic ability to predict OS from day 90 using the Kaplan-Meir method and Cox proportional hazards regression. For inclusion, patients were required to have bony metastases at baseline, ALP ≥120 u/L and survival beyond day 90. ALP normalization was defined as an ALP of <120 and PSA response was defined as decline ≥30% from baseline within 90 days of randomization. RESULTS: 100 men were eligible for analysis h ALP≥120 and 45 had normalization of ALP by day 90. OS from day 90 was significantly improved (HR: 0.38, 95% CI: 0.21-0.68) among those with normalization (1-year OS: 71.5%, 95% CI: 55.3-82.7) compared with those not having normalization (47.3%, 95% CI: 33.1-60.2). ALP normalization was independent of PSA response as a prognostic factor. 1-year OS for men with ALP normalization and with/without PSA-response was 76.9% / 62.5%, and for men without ALP normalization and with/without PSA-response was 56.6% / 30.0%. 29 men had increased ALP at day 90 from baseline and these men had significantly worse survival (9.7 months vs. 17.4 months) even after adjusting for PSA progression (hazard ratio:2.43; 95% CI: 1.48-4.00). CONCLUSIONS: ALP changes on DP based therapy complement PSA changes as prognostic factors associated with OS. Although not a surrogate for OS, ALP changes may be useful in estimating prognosis and facilitating treatment decisions in the setting of PSA flare or lack of PSA response to docetaxel.

Duke Scholars

Author Andrew John Armstrong Medicine, Medical Oncology