Androgen Receptor Modulation Optimized for Response—Splice Variant: A Phase 3, Randomized Trial of Galeterone Versus Enzalutamide in Androgen Receptor Splice Variant-7–expressing Metastatic Castration-resistant Prostate Cancer

Mary Ellen Taplin, Emmanuel S. Antonarakis, Karen J. Ferrante, Kerry Horgan, Brent Blumenstein, Fred Saad, Jun Luo, Johann S. de Bono

Research output: Contribution to journalArticlepeer-review

Abstract

Background:: Detection of androgen receptor (AR) splice variant-7 (AR-V7) messenger RNA (mRNA) in circulating tumor cells (CTCs) is associated with a suboptimal response to abiraterone and enzalutamide in metastatic castration-resistant prostate cancer (mCRPC). Galeterone inhibits CYP17 and AR, and induces AR protein degradation. We hypothesized that galeterone would be clinically superior to enzalutamide in AR-V7–positive (AR-V7+) mCRPC. Objective:: To screen and characterize AR-V7+ mCRPC, and evaluate galeterone compared with enzalutamide. Design, setting, and participants:: This was a multicenter randomized phase 3 trial; enzalutamide-, abiraterone-, and chemotherapy-naïve mCRPC patients had AR-V7 prescreening using a CTC-based mRNA assay. Intervention:: AR-V7+ patients were randomized (1:1) to open-label galeterone or enzalutamide; planned sample size was 148. Outcome measurements and statistical analysis:: The primary endpoint was radiographic progression-free survival (rPFS). Baseline AR-V7 status was correlated with patient characteristics. Results and limitations:: Overall, 953 men were prescreened for AR-V7; 323 (34%) had detectable CTCs, and 73/323 had AR-V7 mRNA. The AR-V7+ prevalence was 8% (73/953; 95% confidence interval [CI] 6–10%). AR-V7 was associated with indicators of advanced and high-volume disease at baseline, including higher prostate-specific antigen (PSA) level (p < 0.001), more bone metastases (p < 0.001), docetaxel for hormone-sensitive disease (p < 0.001), prior first-generation androgen deprivation therapy (p < 0.001), and shorter time from diagnosis to enrollment (p < 0.001). Of 73 eligible patients, 38 were randomized to galeterone (n = 19) or enzalutamide (n = 19); 35 dropped out before randomization. Owing to high censorship for the rPFS events, the data monitoring committee recommended early closure based on interim evidence that the primary endpoint would not be met. The PSA50 values were 2/16 (13%) and 8/19 (42%) for galeterone and enzalutamide respectively (proportion difference = −0.278, 95% CI −0.490 to 0.097). Conclusions:: The prevalence of CTC mRNA AR-V7 in first-line mCRPC was 8% (95% CI 6–10%). AR-V7+ was associated with the characteristics of aggressive and advanced disease. These men had rapid disease progression. Development of galeterone will not be pursued. Patient summary:: Of men with metastatic castration-resistant prostate cancer, 8% had the androgen receptor splice variant-7 (AR-V7) blood biomarker. The AR-V7+ patients had features of aggressive disease. Thirty-eight men were treated with either galeterone or enzalutamide; the trial was stopped early prior to determining efficacy because too many patients transitioned off the trial due to advancing cancer before having required radiographs. The prevalence of circulating tumor cells expressing androgen receptor V7 in castration-resistant prostate cancer naïve to abiraterone or enzalutamide is low and associated with aggressive disease. Treatment approaches other than available second-generation androgen axis inhibitors should be considered.

Original languageEnglish (US)
Pages (from-to)843-851
Number of pages9
JournalEuropean Urology
Volume76
Issue number6
DOIs
StatePublished - Dec 2019

Keywords

  • Androgen receptor splice variant-7
  • Castration-resistant prostate cancer
  • Galeterone

ASJC Scopus subject areas

  • Urology

Fingerprint Dive into the research topics of 'Androgen Receptor Modulation Optimized for Response—Splice Variant: A Phase 3, Randomized Trial of Galeterone Versus Enzalutamide in Androgen Receptor Splice Variant-7–expressing Metastatic Castration-resistant Prostate Cancer'. Together they form a unique fingerprint.

Cite this