730PD - Prediction of PARP inhibitor response and resistance utilizing a CTC phenotypic classifer in patients (pts) with metastatic castration-resistant pr...

Date 09 October 2016
Event ESMO 2016 Congress
Session Genitourinary tumours, prostate
Topics Prostate Cancer
Presenter Felix Feng
Citation Annals of Oncology (2016) 27 (6): 243-265. 10.1093/annonc/mdw372
Authors F.Y. Feng1, S. Daignault-Newton2, A. Jendrisak3, Y. Wang3, S. Greene3, A. Rodriguez3, J. Lee3, L. Dugan3, J. Siddiqui4, J. Louw3, C. Johnson3, P. Twardowski5, C. Albany6, M. Stein7, W.M. Stadler8, L. Kunju4, A.M. Chinnaiyan4, M. Landers3, R. Dittamore9, M. Hussain10
  • 1Radiation Oncology, UCSF Helen Diller Family Comprehensive Cancer Center, 94143-0511 - San Francisco/US
  • 2Biostatistics, University of Michigan, 48109-5010 - Ann Arbor/US
  • 3Translational Research, Epic Sciences, Inc, 92121 - San Diego/US
  • 4Michigan Center For Translational Pathology, University of Michigan, 48109-5010 - Ann Arbor/US
  • 5Medical Oncology And Experimental Therapuetics, City of Hope, 91010 - Duarte/US
  • 6Medical Oncology, Indiana University Simon Cancer Center, 46202 - Indianapolis/US
  • 7Division Of Medical Oncology, The Cancer Institute of New Jersey, New Brunswick/US
  • 8Medical Oncology, The University of Chicago Medical Centre, 60637 - Chicago/US
  • 9Translational Research, Epic Sciences Inc., 92121 - San Diego/US
  • 10Medical Oncology, University of Michigan, Ann Arbor/US



PARP inhibitors (PARPi) have efficacy in mCRPC harboring homologous recombination DNA repair deficiencies (HRD), but there is no non-invasive assay to predict PARPi response. Previous work characterizing single CTCs from mCRPC pts has identified subclonal populations of CTCs with unique phenotypes and somatic genomic instability consistent with HRD and resulting in worse outcomes following abiraterone (A) treatment. NCI 9012 evaluated A alone with or without the PARPi veliparib (V) in mCRPC pts. We now determine if the addition of V to A improves pts response and its association with phenotypic HRD CTC signature.


84 baseline & 58 on-therapy blood samples from unique pts randomized to V + A vs A arms on NCI 9012 (baseline n = 44 vs 40, follow-up n = 34 vs 24, respectively) were analyzed with the Epic Sciences CTC platform. CTC analysis included digital pathology of 20 discrete phenotypic cell features inclusive of AR, CK, size and shape. 3323 single CTCs were characterized and analyzed for HRD+ signature utilizing the previously developed classifier. A subset of CTCs (n= 309) from 40 samples (30 patients) were individually sequenced and analyzed for genomic instability and resistant mechanism.


The HRD+ CTC phenotype was confirmed to have high cell level concordance (79%) vs. genomic scar (NGS). In baseline samples, HRD+ CTC phenotype occurred in 39% (V + A) and 30% (A), and had a higher ORR (>50% PSA drop) in the V + A vs the A cohort (88% vs 42%, p = 0.01). No significant difference observed in V + A vs the A cohort (62% vs 78%, p = 0.24) for HRD- patients. For patients with both baseline and short term follow up (


Prevalence of a unique morphologically derived CTC subtype predicts genomic HRD and sensitivity to V + A but not A alone. Novel markers of resistance have identified patients not suitable for PARPi despite HRD+. Further validation is ongoing.

Clinical trial identification

NCI 9012 trial

Legal entity responsible for the study

Felix Feng




F.Y. Feng: Advisory boards: Medivation, Astellas. President and co-Founder of PFS Genomics, a molecular diagnostics company focused on personalizing radiotherapy for breast cancer patients.

A. Jendrisak, Y. Wang, S. Greene, A. Rodriguez, J. Lee, L. Dugan, J. Louw, C. Johnson, M. Landers, R. Dittamore: Epic Sciences employee. All other authors have declared no conflicts of interest.