Suboptimal suppression of serum androgen levels among men treated with apalutamide and abiraterone acetate plus prednisone compared with abiraterone acetate plus prednisone alone.

Phase III studies of dual therapy with an androgen receptor (AR) antagonist, apalutamide (Apa) or enzalutamide (Enza) with abiraterone acetate (AA) plus (+) prednisone (P) in metastatic castration resistant prostate cancer (mCRPC) have not shown a survival benefit vs single agent therapy. Apa and Enza induce CYP3A4 activity, resulting in decreases in serum AA and P levels. The decrease in AA levels has not been considered clinically significant, but the potential impact of AR antagonist-mediated CYP3A4 induction on serum steroid levels, in context of clinical studies combining an AR antagonist with AA, has not been previously reported. We measured levels of AA, its metabolites and androgens using LC/MS-MS in available serum samples obtained at baseline and at 4 weeks of therapy in the PANTHER phase II study (n=86) of Apa (240 mg daily) and AA (1000mg daily) + P (10mg daily), and in the Abi Race phase II study (n=75) of AA (1000mg daily) + P (10mg daily) among men with mCRPC. Samples were batched and all sera assessed contemporaneously. Comparison of metabolite levels between studies used the Mann-Whitney test, and within study the Wilcoxon matched-pairs signed rank test. At 4 weeks, median (med) levels of AA and its primary metabolites delta-4 and keto-Abi in PANTHER vs Abi Race were 13.1 vs 39.6 ng/ml, 0.49 vs 1.93 ng/ml, and 0.75 vs 8.98 ng/ml (p=<0.0001 for all), equating to 66%, 75%, and 92% lower levels with Apa and AA + P vs AA + P. Baseline steroid levels did not differ between the studies, but steroids downstream of CYP17A were suppressed markedly less effectively in the dual therapy PANTHER study: DHEAS was detectable in 80% vs 14% of samples (med 1.23 vs 0.49 ng/ml, p=<0.0001), DHEA in 90% vs 26% (med 0.44 vs 0.01 ng/ml, p=<0.0001), AED in 70% vs 11% (med 0.017 vs 0.010 ng/ml, p=<0.0001), and testosterone in 50% vs 17% (med 0.006 vs 0.005 ng/ml, p=0.0005). Despite the decrease in AA levels, steroids upstream of CYP17A in PANTHER were markedly elevated vs treatment with AA + P alone: pregnenolone (4.1 vs 0.79ng/ml, p=<0.0001), consistent with a suboptimal prednisone-mediated suppression of ACTH. Serum AA levels at week 4 are substantially lower, and androgen levels substantially higher, among men with mCRPC treated with Apa and AA + P vs AA + P. Our data suggest that Apa decreased P levels to the point that circulating ACTH remained sufficient to mediate ongoing basal adrenal androgen synthesis. Suboptimal suppression of steroids may explain why AA and AR antagonist combination studies have not been more effective. The clinical activity when both treatment strategies are at full efficacy remains to be fully tested, and will likely require use of dexamethasone or higher than standard dosing of prednisone. This may be particularly important for treatment of prostate cancers with greater dependence on AR signaling.

Duke Scholars

Author Andrew John Armstrong Medicine, Medical Oncology

Author Susan Halabi Biostatistics & Bioinformatics, Division of Biostatistics