Abstract 4670: A novel companion diagnostic predicts response to the PARP inhibitor rucaparib in ovarian cancer

Abstract

Abstract Background: Genomic studies suggest that ∼50% of high-grade serous ovarian cancers (OC) have homologous recombination deficiency (HRD). Germline BRCA1/2 mutations (gBRCAmut) are expected to account for 1/3 of HRD in OC, and identification of non-gBRCAmut HRD tumors likely to respond to PARP inhibitors (PARPi) remains a challenge. Using comprehensive next generation sequencing (NGS)-based tumor genomic profiling, we developed a companion diagnostic HRD assay to predict sensitivity to the PARP inhibitor rucaparib by combining tumor BRCA1/2 status (germline and somatic) and genomic loss of heterozygosity (LOH). The HRD assay is being validated in a Phase 2 study (ARIEL2) and will be prospectively applied to the primary analysis of the ongoing Phase 3 study (ARIEL3) of rucaparib. Methods: The HRD assay uses 50-200ng of DNA from tumor FFPE specimens, which undergoes sequencing library construction and hybrid-capture of all coding exons from 100s of cancer-related genes. Libraries are sequenced to high, uniform depth (>500× unique coverage, Illumina® HiSeq) and data are processed by a customized pipeline that accurately detects all classes of genomic alterations, including BRCA1/2 base substitutions, indels, and homozygous deletions. Genomic LOH is assessed by a CGH-like analysis of sequencing coverage and >3,500 genome-wide SNPs and a tumor is classified as HRD with either BRCA1/2 alteration or high genomic LOH (LOH+). Somatic/germline status of discovered BRCA1/2 alterations is assessed by a previously-presented computational approach (“SGZ”, AACR 2014 abstract #1893), and verified against medical records where available. ARIEL2 is an ongoing single-arm (n = 180), open-label study of rucaparib in recurrent, platinum-sensitive OC patients. The primary objective is to evaluate clinical activity of rucaparib among 3 prospectively defined subgroups: tumor BRCAmut, BRCAwt/LOH+ (“BRCAness”) and BRCAwt/LOH-. Response is determined by RECIST and/or GCIG-CA125 criteria. Results: The HRD assay was performed on tumors from 121 patients, of whom 25% were found to be BRCA mutant (17 germline/12 somatic), 42% had the BRCAness signature (BRCAwt/LOH+), and 33% were biomarker negative (BRCAwt/LOH-). Efficacy data available for 61 patients revealed objective response rates (combined RECIST/CA125 criteria) at 70%, 40% and 8%, respectively. Responses were observed for all classes of genomic alterations, and in gBRCAmut and non-gBRCAmut tumors. Conclusions: Preliminary clinical data indicates that the HRD assay identifies OC patients likely to respond to rucaparib and highlights the potential for innovative companion diagnostics enabled by comprehensive genomic profiling based on NGS. Citation Format: James Sun, Iain McNeish, Robert L. Coleman, Amit Oza, Clare Scott, David M. O'Malley, Kevin K. Lin, Christine Burns, Christine Vietz, Philip J. Stephens, Murtaza Mehdi, Matthew Hawryluk, Heidi Giordano, Mitch Raponi, Lindsey Rolfe, Jeff Isaacson, Vincent A. Miller, Andrew Allen, Elizabeth Swisher, Roman Yelensky. A novel companion diagnostic predicts response to the PARP inhibitor rucaparib in ovarian cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4670. doi:10.1158/1538-7445.AM2015-4670