A whole-genome sequencing–based classifier demonstrated greater performance for predicting homologous recombination deficiency (HRD) across cancer types than standard assays, which in turn determined which patients with cancer would benefit from treatment with PARP inhibitors, according to findings published in Communications Medicine.
“A comprehensive analysis of the entire genome has advantages compared with traditional, targeted detection strategies for predicting homologous recombination deficiency,” said study senior author Juan Miguel Mosquera, MD, MSc, Professor of Pathology and Laboratory Medicine and Director of Research Pathology at the Englander Institute for Precision Medicine at Weill Cornell Medicine; Pathologist at NewYork-Presbyterian/Weill Cornell Medical Center; and Member of the Sandra and Edward Meyer Cancer Center at Weill Cornell.
Background and Study Methods
Homologous recombination deficiency assays are typically used to determine which patients with cancer are eligible for use of PARP inhibitors in approved settings, but these assays can vary significantly.
Researchers sought to study the whole genome landscape to better classify and harmonize homologous recombination deficiency across cancer types. They examined 580 tumor and normal paired samples from 453 patients with whole-genome sequencing for analyzing genome-wide signatures that were associated with a homologous recombination deficiency phenotype. The researchers used a proprietary homologous recombination deficiency classifier from Isabl Inc. for training the data on the cohort from Weill Cornell Medicine, and the algorithm was validated on a data set from the International Cancer Genome Consortium.
Key Findings
The homologous recombination deficiency phenotype was found in 21% of breast cancers, 20% of pancreaticobiliary cancers, 17% of gynecologic cancers, 9% of prostate cancers, 2% of upper gastrointestinal cancers, and in 1% of other cancer types with one case of carcinoma of unknown primary. These instances of homologous recombination deficiency were not only found among patients with BRCA1/2, as 24% of instances were found in patients with BRCA1/2 wild-type disease.
Additional genomic mechanisms of the phenotype were also identified, including biallelic mutations in FANCF, XRCC2, and FANCC.
The study found that the algorithm could better predict homologous recombination deficiency than traditional assays and even flagged several cases of false negative and false positive results where the patients' outcomes did not match expected results.
DISCLOSURE: For full disclosures of the study authors, visit nature.com.
