Serum Androgens as Predictors of Survival in Metastatic Prostate Cancer

Sponsor: NIH National Cancer Institute

Location(s): United States


The primary objective of this research is to demonstrate that serum androgen (SA) levels in patients with castration resistant prostate cancer (CRPC) are prognostic of overall survival (OS). A relationship of higher SA to improved survival has been observed in two phase III randomized studies, regardless of treatment arm, but never in a study in which an androgen synthesis inhibitor (ASI) such as abiraterone or ketoconazole was NOT part of the therapy. Patient serum is banked from CALGB 90401 - an NCI sponsored cooperative group randomized phase III that compared docetaxel plus prednisone (DP) to docetaxel/prednisone plus bevacizumab (DPB) in CRPC. Banked serum from this completed study will be used to measure androgen levels via CLIA certified ultrasensitive (liquid chromatography tandem mass spectroscopy) technique and these results will be associated with mature patient survival and outcome data. The underlying hypothesis of this work is that higher serum androgens are associated with improved outcomes, including survival, regardless of treatment, as it reflects a more favorable biological mileu in which the tumor remains partially dependent on androgens. We have the following specific aims:

AIM 1: To assess whether serum androgens (Androstenedione, DHEA-sulfate and Testosterone) using an ultrasensitive assay at baseline will be prognostic and predictive for clinical outcomes in mCRPC patients treated on CALGB 90401. 

AIM 2: To evaluate whether serial changes in SA from baseline to the time of disease progression in mCRPC patients treated on 90401 are prognostic for OS. 
AIM 3: To develop a prognostic model for OS for men with mCRPC that integrates SA levels. 
This proposal is the first analysis to test the hypothesis that SA is prognostic and predictive of OS in chemotherapy-naïve mCRPC patients treated with DP, standard of care chemotherapy. A positive finding will confirm the importance of androgen signaling even in the setting of very advanced disease that is being treated with first-line chemotherapy. The implications of this work are that SA could be an unappreciated determinant of outcome in patients with CRPC and should be considered in the design of targeted or risk-adapted therapies. Further, the use of this analysis could inform clinical decision-making, as our preliminary data have shown that pre-treatment SA levels may be predictive of treatment response. Since 90401 was a study of chemotherapy and did not utilize an androgen synthesis inhibitor, validating these findings would suggest that SA are prognostic and predictive of outcomes in CRPC regardless of treatment.