Abstract
Researchers have developed a new tool, HRDetect, to pinpoint tumors that display BRCA deficiency but don't harbor BRCA1/2 mutations. Evaluating their method in breast, ovarian, and pancreatic cancers, they identified patients whose tumors were potentially vulnerable to PARP inhibition but who didn't carry these mutations.
Researchers have developed an algorithm to identify patients whose tumors might respond to PARP inhibition even though they don't carry BRCA1/2 mutations.
BRCA1/2 mutations impair cells' ability to repair double-strand DNA breaks through homologous recombination (HR), leaving tumors susceptible to PARP inhibitors such as olaparib (Lynparza; AstraZeneca). Even if tumors don't harbor germline or somatic mutations in the BRCA genes, they can display functional BRCA1/2 deficiency through, for instance, inactivation of related HR pathway genes. Patients with these tumors may also benefit from PARP inhibitors, but the best way to identify them remains unclear. Two companies, Myriad Genetics and Foundation Medicine, have produced tests for “BRCAness” that have been used in clinical trials but aren't yet FDA approved.
In this study, Serena Nik-Zainal, PhD, of the Wellcome Trust Sanger Institute in London, UK, and colleagues devised a new computational method for measuring BRCA deficiency. In a previous paper, they reported whole-genome sequencing of 560 breast cancers. For the new work, they identified 77 tumors from patients with BRCA1/2 deficiency, to use as a training set. These tumors, they found, often carry six specific mutational signatures. One signature, for instance, features multiple small duplications throughout the genome. The chief characteristic of another signature is numerous large tandem duplications.
Using these six parameters, the researchers created a weighted algorithm that gauges a tumor's probability of BRCA deficiency. They then tested the effectiveness of their tool, called HRDetect, in the original cohort of 560 patients. HRDetect recognized 47 patients in this group who lacked mutations in either gene but had a high BRCA1/2 deficiency score.
Nik-Zainal and colleagues then evaluated sequencing data from patients with pancreatic or ovarian cancer, who also often carry BRCA mutations. In both cohorts, HRDetect once again identified patients who didn't have BRCA1/2 mutations but whose tumors still displayed BRCAness.
Additionally, using breast tumor biopsies, the researchers showed that HRDetect could predict the likelihood of response to anthracyclines, which are also effective against tumors with BRCA1/2 deficiency. HRDetect scores were high in four patients who had complete responses to anthracyclines, but low in five patients who had residual disease.
“Our algorithm helps identify, with much greater precision, tumors that have the features of BRCA1 or BRCA2 deficiency,” says Nik-Zainal. The study findings suggest that a larger fraction of patients than previously thought carry such tumors, she says: Only 1% to 5% of patients with breast cancer have germline BRCA mutations, but HRDetect classified about 20% in all as BRCA-deficient. Nik-Zainal notes that these patients are not just candidates for treatment with PARP inhibitors. They might also respond to other therapies that exploit defective DNA repair, such as platinum-based chemotherapy.
“I think it's a great study from the perspective that they were able to pull out signatures that were predictive of BRCA status,” says Fergus Couch, PhD, of the Mayo Clinic in Rochester, MN, who wasn't involved in the research. Whether HRDetect outperforms other techniques for identifying BRCA-deficient tumors isn't clear without a head-to-head comparison, he says. Couch also cautions that researchers haven't established that PARP inhibitors are effective against tumors that display BRCA-like defects but don't harbor BRCA1/2 mutations. –Mitch Leslie