LBA21 - Maximal androgen depletion with abiraterone acetate (AA) followed by randomization of maximal androgen ablation with molecular targeted therapies d...

Date 29 September 2014
Event ESMO 2014
Session Genitourinary tumours, prostate 2
Topics Anticancer Agents
Prostate Cancer
Biological Therapy
Presenter Eleni Efstathiou
Citation Annals of Oncology (2014) 25 (5): 1-41. 10.1093/annonc/mdu438
Authors E. Efstathiou1, E. Li Ning Tapia1, A. Aparicio2, S.M. Tu1, S. Wen3, A. Hoang4, L. Pagliaro1, C. Harmon1, P. Troncoso1, J.C. Araujo5, C. Logothetis1
  • 1Oncology, University of Texas MD Anderson Cancer Center, TX77030 - HOUSTON/US
  • 2Dept Of Gu Medical Oncology, MD Anderson Cancer Center, 77030 - Houston/US
  • 3N/a, University of West Virginia, Morgantown/US
  • 4Genitourinary Medical Oncology, UT MD Anderson Cancer Center, 77030 - HOUSTON/US
  • 5Dept Of Genitourinary Medical Oncology, MD Anderson Cancer Center, 77030 - Houston/US



Background: During the course of successful mCRPC drug development the unmet need for predictive markers has become pressing. We have previously (JCO 2012, EurUr 2014, ASCO 2014) identified a candidate androgen signaling signature associated with favorable outcome and lacking in patients with primary resistance to either abiraterone acetate or enzalutamide. Here we report on a prospective testing of the signature.


This is a phase II study of AA in mCRPC patients randomized upon progression to combination with Dasatinib or Sunitinib. Patients underwent bone marrow biopsy sampling at baseline and upon progression with matched blood and bone marrow aspirate collection Endpoints included prespecified prospective assessment of tumor androgen signaling signature in pts with benefit vs primary resistance to AA (progression ≤ 4 months). Tumor Markers included, Androgen Receptor-N terminal (AR-N), AR-C terminal (AR-C), CYP17, Ki67, Glucocorticoid receptor (GR), ERG, Proliferation Index Ki67 (%), pSrc by IHC. Steroid metabolome assessed by LCMS.


Study (March 2011-April 2014) has completed accruing 170 mCRPC pts. Median Age is 67 years (range 49-88). All pts had bone metastases (100%). Diagnostic Gleason Score was ≥8 in 111/147 (76%) evaluable pts and median pretreatment PSA 21 (range 0.5-1655). Forty one pts completed study follow up and tissue and metabolome analyses at this time. The prespecified testing of the molecular signature comprising of tumor AR-N terminal overexpression the ratio of AR-C terminal expression/ AR-N terminal expression ≥ 0.8 and CYP17 expression exhibits a significant predictive value (p value <0.0001). Table includes associations with ERG, GR, Ki67.Clinical characteristics were not predictive of outcome.

Tumor Markers Primary Resistance (%) Benefit (%) P Value Fisher Test
Ki67 >20% 5 (29) 2 (13) 0.4
ERG presence 8/21 (38) 1/19 (6) 0.023
GR presence 7/21(33) 0.723
ANDROGEN SIGNATUREAR-N overexpression + CYP17 expressionAR-C/ AR-N ≥0.8 4/22 (18) 18 /19 (82) < 0.0001


Conclusion and future direction: This initial prespecified testing of previously identified andogen signaling molecular signature, is highly supportive of its predictive value in maximal androgen deprivation strategies in mCRPC. These findings warrant validation in a study selecting for treatment based on signature presence.


E. Efstathiou and C. Logothetis: Research funding, ad board, honoraria: JnJ, BMS, Pfizer, Bayer, Sanofi, Astellas, Takeda, Viamet; A. Aparicio: Research funding: JnJ, BMS, Pfizer, Bayer, Sanofi, Astellas, Takeda, Viamet. All other authors have declared no conflicts of interest.