Introduction:

Androgen deprivation therapy with GnRH analogs may suppress serum testosterone levels, but adrenal androgen levels are not typically suppressed. In the CALGB 9583 randomized trial of ketoconazole (keto), adrenal androgen levels were associated with response and survival. We prospectively assessed these results to validate prior findings in a randomized study investigating keto and hydrocortisone (HC) with anti-PSMA radionuclides (previously reported to delay time to metastatic disease).

Methods:

In a randomized phase 2 trial in patients with high-risk (PSA doubling time <8 mo and/or absolute PSA >20) non-metastatic (M0) castration-resistant prostate cancer, we assessed AR synthesis inhibition with radiolabeled anti-PSMA radioimmunotherapy. All patients received keto (400 mg TID) with HC and were randomized 2:1 to either ¹⁷⁷Lu-J591 or ¹¹¹In-J591 after 1 month of keto/HC. To investigate effect of ketoconazole on adrenal androgen levels and patient outcomes, serum levels of androstenedione, DHEA, DHEA-S, and testosterone) was measured at baseline, after 1 month prior to PSMA radionuclide therapy, and at the time of PSA progression. We assessed changes in hormonal levels and examined associations with PSA response, metastasis-free survival, and overall survival.

Results:

Following keto/HC and PSMA-radionuclides, PSA declined by >50% in 87.5% and >90 in 45.8%. After 1 month of keto/HC, androstenedione, DHEA, DHEA-S and testosterone levels dropped. DHEA declined from median of 2.02 ng/mL (95% CI 1.36-3.53) to 0.76 ng/mL on C2D1 (0.44-0.99) (linear mixed model β -0.775, p<0.001), DHEAS median 0.49 µg/mL (0.32-0.82) to 0.08 µg/mL (0.06-0.20) (p=0.001), androstenedione median 0.58 ng/mL (0.44-0.99) to 0.14 ng/mL (0.10-0.26) (p<0.001), testosterone median 26 ng/dL (20-39) to undetectable (<20 ng/dL) (<20-29) (p=0.08). Using Cox regression analysis, it was determined that higher levels of pre-treatment testosterone were associated with PSA decline: PSA decline >50% was more common in those with higher baseline testosterone (59%) than those with testosterone below the limit of detection (<20 ng/dL, 41%) and PSA decline >90% was 67% vs 33%. Baseline serum testosterone was significantly related to metastasis-free survival (HR= 0.94, 95 % CI: 0.876-0.999, P=0.047), and overall survival HR = 0.44, 95% CI: 0.20-1.00, p=0.050. Baseline androstenedione levels were significantly associated with worse overall survival (HR=1.29, 95% CI: 1.03-1.61, p=0.026), though when controlling for type of radionuclide (¹⁷⁷Lu vs ¹¹¹In), statistical significance decreased.

Conclusion:

Ketoconazole with hydrocortisone therapy in patients with high risk M0 CRPC results in androgen suppression, with pre-treatment testosterone and androstenedione associated with outcomes.

Funding: NIH, DOD, Weill Cornell Medicine, Prostate Cancer Foundation