Skip to main content

Study Finds Genomic Instability Scores Used in Clinical Trials Are Not Equivalent

Conference Correspondent

The genomic instability scores seen in published research and used in ongoing clinical trials are not equivalent at identifying tumors with homologous recombination deficiency (HRD), and therefore should not be considered interchangeable in predicting response to PARP inhibitors in clinical practice, according to Gordon Mills, MD, PhD, Director of Precision Oncology at the Knight Cancer Institute at Oregon Health & Science University.

Clinical trials have explored the utility of several different genomic instability scores to support PARP inhibitor use in ovarian cancer. In this trial, Dr Mills and colleagues sought to determine if these scores were consistent in their ability to identify HRD tumors, as patients with these tumors may be more likely to benefit from PARP inhibitor therapy, he explained at the Society of Gynecologic Oncology (SGO) Annual Meeting on Women’s Cancer.

In the study, whole-genome single-nucleotide polymorphism analysis was used to reconstruct tumor genomic profiles from 3278 ovarian tumors. Using these profiles, the researchers calculated 3 different genomic instability scores: myChoice HRD, loss of heterozygosity (LOH), and percentage of LOH (%LOH). Tumor mutation screening of BRCA1 and BRCA2 was also performed.

Published thresholds of 42 or 33 for myChoice HRD, 8 for LOH score, and 16% for %LOH were used as cutoffs. The correlation of positive results in BRCA-mutated tumors was determined, and in the entire study population, percentage positive agreement (PPA) between the scores was assessed.

A total of 297 BRCA-mutated tumors were identified. Among these, 92% had positive myChoice HRD results for cutoff 42, and 96% had positive myChoice HRD results for cutoff 33.

In contrast, LOH and %LOH scores were much lower: LOH scores were positive in 85% of BRCA-mutated tumors, and %LOH scores were positive in only 69% of BRCA-mutated tumors.

Correlations and PPA among all 3 scores in the entire cohort indicated a high level of concordance, but did not indicate the equivalence of the different assays, the investigators concluded.

“Compared to myChoice HRD, %LOH score identifies a much smaller percentage of BRCA-mutated tumors,” reported Dr Mills. “In addition, %LOH and myChoice HRD tests have only moderate PPA.”


  • SGO 2020 Annual Meeting on Women’s Cancer. Abstract 296.

Related Items