Androgen Receptor Burden and Poor Response to Abiraterone or Enzalutamide in TP53 Wild-Type Metastatic Castration-Resistant Prostate Cancer

Androgen receptor (AR) molecular perturbations in liquid biopsy specimens (circulating tumor cells [CTCs] and circulating tumor DNA [ctDNA]) from patients with metastatic castration-resistant prostate cancer (mCRPC) are associated with poorer outcomes on the AR signaling inhibitors (ARSis) abiraterone or enzalutamide, with contradictory claims regarding the clinical use of AR splice variant 7 (AR-V7) or AR amplification status.^1-4 By using comprehensive liquid biopsy AR profiling,⁵ we previously demonstrated how AR perturbations lost prognostic value when correcting for tumor burden estimates and clinical variables.⁶ Consistent with others’ findings,⁴ we observed that AR-driven biomarkers were outperformed by TP53 alterations, noting that 71.3% and 97.9% of patients carried relevant AR-driven biomarkers at baseline and progression, respectively.⁶ We therefore speculate that long-term chemical castration, alongside the cancer treatment trajectory, will eventually lead to perturbed AR biomarker output in all patients. In this scenario, multiple biomarker assessments of AR are necessary to identify ARSi-sensitive patients.

After obtaining approval from the Antwerp University Hospital institutional review board and written patient informed consent, we performed simultaneous profiling of AR splice variant (ARV) expression by targeted RNA sequencing of CellSearch-enriched CTCs and plasma ctDNA profiling via low-pass whole genome sequencing and targeted AR sequencing to infer genomic alterations in patients with mCRPC.⁶ Herein we report a post hoc analysis of the prognostic value of the number of ARSi outcome-associated AR perturbations at baseline with progression-free survival (PFS) in patients with TP53 wild-type mCRPC (n = 109 of 145 evaluable patients) after a median follow-up of 13 months.⁶ The summation of outcome-associated ARVs (AR45, AR-V3, AR-V4, AR-V5, and AR-V7) and AR gene alterations (amplifications and intragenic rearrangements) were correlated with PFS defined according to the Prostate Cancer Clinical Trials Working Group 3 criteria, as described previously.⁶

Comprehensive AR profiles were available for 80 (73.4%) of 109 patients (mean [SD] age at registration, 75.09 [8.2] years) (Table). We detected outcome-associated AR perturbations in 49 (61.3%) of 80 patients at baseline, with 16 (32.7%) of 49, 17 (34.7%) of 49, and 16 (32.7%) of 49 patients having 1, 2, and 3 or more significant events (ie, AR perturbations with a significant association to PFS in univariable survival analysis), respectively. Patients without any outcome-associated AR perturbation had the longest median PFS compared with patients with AR perturbation, who demonstrated a decremental PFS as the number of AR perturbations (0, 1, 2, and ≥3) accumulated (median, 13.7, 10.1, 6.1, and 2.8 months, respectively) (P < .001) (Figure, A). The decremental PFS was present in subanalyses of 57 patients treated with abiraterone who had 0, 1, 2, and ≥3 AR perturbations (median PFS, 13.7, 10.1, 5.3, and 2.3 months, respectively) (log-rank test, P < .001) and 23 patients treated with enzalutamide (median PFS, 13.8, 9.2, 9.6, and 3.4 months, respectively) (log-rank test, P = .02). In the multivariable Cox regression analysis incorporating clinical and molecular characteristics, a perturbed AR status was associated with inferior outcome and had independent prognostic value when 3 or more AR perturbations were detected (hazard ratio, 2.97; 95% CI, 1.16-7.65; P = .02) (Figure, B).

The number of AR perturbations remains independently associated with poor prognosis in TP53 wild-type disease, thereby demonstrating the prognostic value of comprehensive AR profiling. However, this study has limitations. First, this post hoc hypothesis-generating analysis warrants independent prospective validation, which will occur in our recently initiated randomized clinical trial ProBio (EudraCT number: 2018-002350-78). Second, our study was not designed to assess quantitative biomarker output such as absolute ARV expression levels or AR copy number, which likely harbors prognostic information in the context of varying tumor burden estimates. Finally, we did not include prognostic variables outside of our standard practice such as performance status, alkaline phosphatase, and lactate dehydrogenase. To our knowledge, the data represent the first demonstration of the clinical validity of “AR burden” contributed to by multiple AR gene body alterations and/or ARV expression occurring during progression of mCRPC under treatment selection pressure. Biomarker output from the AR locus appears to be a ubiquitous property of mCRPC, thereby calling into question the current single AR biomarker dogma for the prognostication of outcome in patients initiating ARSi therapy. We anticipate patient evaluation with a comprehensive AR profile; continued AR blockade could be recommended for those with a sufficiently low AR burden in the context of other prognostic factors and the specific AR alterations present.

Accepted for Publication: March 7, 2019.

Published Online: May 2, 2019. doi:10.1001/jamaoncol.2019.0869

Correction: This article was corrected on July 11, 2019, for an error in the first column of Figure, B.

Corresponding Author: Bram De Laere, PhD, Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Nobels väg 12A, Stockholm 171 65, Sweden (bramdelaere@gmail.com).

Author Contributions: Drs De Laere and Lindberg had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All CORE-ARV-CTC and ProBIO Investigators provided administrative, technical, or material support and recruited patients and collected blood samples.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: De Laere, Rajan, Dirix, Lindberg.

Drafting of the manuscript: De Laere, Rajan, Dirix, Lindberg.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: De Laere, Rajan, Dirix.

Obtained funding: Rajan, Grönberg.

Administrative, technical, or material support: De Laere, Rajan, Grönberg, Lindberg.

Study supervision: Grönberg, Dirix, Lindberg.

Conflict of Interest Disclosures: Drs De Laere and Dirix have patent WO2017207702 pending: Androgen receptor splice variants and androgen deprivation therapy. No other disclosures were reported.

Funding/Support: This study was funded by grants from The Belgian Foundation Against Cancer (grant No. C/2014/227), Kom op tegen Kanker (Stand up to Cancer), the Flemish Cancer Society (grant No. 00000000116000000206), Royal College of Surgeons/Cancer Research UK (C19198/A1533), The Cancer Research Funds of Radiumhemmet through the PCM program at KI (grant No. 163012), The Erling-Persson Family Foundation (grant No. 4-2689-2016), the Swedish Research Council (grant No. K2010-70X-20430-04-3), and the Swedish Cancer Foundation (grant No. 09-0677).

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The CORE-ARV-CTC and ProBIO investigators who participated in this study by providing administrative, technical, or material support and recruited patients and collected blood samples were Steffi Oeyen (University of Antwerp, Belgium), Markus Mayrhofer (Karolinska Institute, Sweden), Tom Whitington (Karolinska Institute, Sweden), Pieter-Jan van Dam (University of Antwerp, Belgium), Peter Van Oyen (AZ Sint-Jan Brugge, Belgium), Christophe Ghysel (AZ Sint-Jan Brugge, Belgium), Jozef Ampe (AZ Sint-Jan Brugge, Belgium), Piet Ost (University Hospital Ghent, Belgium), Wim Demey (AZ Klina, Belgium), Lucien Hoekx (University Hospital Antwerp, Belgium), Dirk Schrijvers (ZNA Middelheim, Belgium), Barbara Brouwers (AZ Sint-Jan Brugge, Belgium), Willem Lybaert (AZ Nikolaas, Belgium), Els Everaert (AZ Nikolaas, Belgium), Daan De Maeseneer (AZ Sint-Lucas Brugge, Belgium), Michiel Strijbos (AZ Klina, Belgium), Alain Bols (AZ Sint-Jan Brugge, Belgium), Karen Fransis (University Hospital Antwerp, Belgium), Nick Beije (Erasmus MC, Netherlands), Inge de Kruijff (Erasmus MC, Netherlands), Valerie van Dam (University of Antwerp, Belgium), Anja Brouwer (University of Antwerp, Belgium), Dirk Goossens (Agilent Technologies, Belgium), Lien Heyrman (Agilent Technologies, Belgium), Gert Van den Eynden (GZA Sint-Augustinus, Belgium), Annemie Rutten (GZA Sint-Augustinus, Belgium), Jurgen Del Favero (Agilent Technologies, Belgium), Mattias Rantalainen (Karolinska Institute, Sweden), Stefan Sleijfer (Erasmus MC, Netherlands), Anders Ullén (Karolinska University Hospital, Sweden), Jeffrey Yachnin (Karolinska University Hospital, Sweden) and Steven Van Laere (University of Antwerp, Belgium). We also thank Luc De Laere (AZ Sint-Jan Brugge, Belgium), Thijs Develter (AZ Sint-Jan Brugge, Belgium), Sophie Vantieghem (AZ Sint-Jan Brugge, Belgium), Sofie Herman (AZ Klina, Belgium), Gwen Colfs (AZ Klina, Belgium), Veerle Lamotte (University Hospital Antwerp, Belgium), Anita Boumans (ZNA Middelheim, Belgium), Abdelbari Baitar (ZNA Middelheim, Belgium), Roos Haeck (AZ Nikolaas, Belgium), and Goele Wallays (AZ Nikolaas, Belgium).

Additional Contributions: We thank the patients for their willingness to participate in this study and all CORE-ARV-CTC and ProBIO investigators for their assistance with patient inclusion, sampling management, and data collection. The investigators received no compensation for their contributions.