Introduction Genomic instability (GI) scores and gene panels that assess deficiencies in the homologous recombination (HR) DNA repair pathway to support PARP inhibitor use in ovarian cancer may not be equivalent. We compared the proportion of patients identified as candidates for PARP inhibitor use by three measures of HR deficiency: a 3-biomarker HR deficiency (HRD) score, percent loss of heterozygosity (%LOH), and an 11-gene panel.

Methods Whole-genome SNP analysis was used to reconstruct ovarian tumor genomic profiles for two cohorts (clinical laboratory cohort, N = 3,689; SCOTROC4 trial, N = 176). Mutation screening was performed for 11 HR pathway genes (ATM, BARD1, BRCA1, BRCA2, BRIP1, CHEK2, MRE11A, NBN, PALB2, RAD51C, RAD51D) for a subset of tumors from SCOTROC4 (N = 153). Samples were considered positive if they had HRD score>threshold (thresholds at 33 and 42 evaluated),%LOH>16%, or a pathogenic variant in one of the 11 HR genes. Correlation between results from%LOH and the 11-gene panel were compared to HRD score.

Results Among patients with positive HRD scores, up to 51% were negative by%LOH and up to 63% were negative by the 11-gene panel. Only 3% of patients identified as positive by%LOH and 7% positive by the 11-gene panel were negative by HRD score.

Conclusions/Implications These data show that HR deficiency tests used in clinical trials are not equivalent and should not be considered interchangeable in predicting PARP inhibitor response in clinical practice.