NEW YORK – Thermo Fisher Scientific is throwing its hat into the ring of homologous recombination deficiency cancer testing with a new HRD algorithm that is based on its comprehensive cancer profiling assay.
The new tool, called genomic instability metric (GIM), is a complex endpoint generated from the firm's Oncomine Comprehensive Assay (OCA) Plus, a targeted sequencing panel covering more than 500 genes. It studies genome segmentation to determine copy number changes, allele-specific copy number calls, and unbalanced copy number events to produce a score ranging from 1 to 99. Using cutoffs, GIM can call breast cancer samples as HRD positive or negative and can help stratify ovarian cancer samples, Jon Sherlock, senior director of product management for oncology at Thermo Fisher, said during a corporate workshop at last week's Association for Molecular Pathology annual meeting.
Sherlock also said that Thermo Fisher will port OCA Plus, which is currently available on the Ion Torrent GeneStudio sequencer, to the Ion Torrent Genexus platform in the first half of 2023. GIM scores will be available for cancer profiling on the GeneStudio platform in the coming weeks. Sherlock presented a figure suggesting good correlation between GIM scores generated on the firm's different NGS instruments.
The announcements underscore the growing importance of HRD testing in cancer profiling, especially for recommending patients for therapy with PARP inhibitors. Eric Vail, a molecular pathologist at Cedars-Sinai Medical Center whose lab runs OCA Plus and who has been an adviser to Thermo Fisher, said he was "absolutely" interested in using GIM. "The more info the better," he said. However, he cautioned that HRD testing will be more complicated than measuring some of the other complex biomarkers that preceded it, such as tumor mutational burden (TMB) or microsatellite instability.
"There's even more heterogeneity in HRD than there is in TMB," he said. "The actual definition of TMB is the same," given in number of mutations per Mb. "In HRD tests, the definition of what it is differs in each of these different tests. That's what makes it so difficult."
GIM joins a cluttered field of HRD scores, led by Myriad Genetics' MyChoice test and several academic methods for scoring HRD, including HRDetect from the Wellcome Sanger Institute. Illumina has partnered with Merck to develop an HRD assay based on Myriad's technology, launched in June, that is available outside the US and Japan. Amoy Diagnostics has partnered with AstraZeneca to develop an HRD companion diagnostic assay, and Foundation Medicine's FoundationOne test also determines HRD status. Sherlock, in his talk, compared GIM to scores from genomic scar analysis (GSA), an algorithm published in 2021 by researchers at BGI.
Criteria used to determine HRD vary from method to method. Variants in the BRCA1 and BRCA2 genes, genome-wide loss of heterozygosity (LoH), and other NGS-based data can all be included. The overarching goal is to identify tumors without obviously pathogenic BRCA gene variants but where homology-directed genome repair is still impaired, so that PARP inhibitors — which disrupt a separate DNA repair mechanism — contrive a form of synthetic lethality in the cancerous cells.
During the Q&A portion of the workshop, an attendee noted that HRD score development seems to be following the path already taken by tumor mutational burden analysis, which necessitated studies to compare and harmonize TMB scores.
"That's something that will definitely happen," Sherlock said in an interview after the workshop. "There need to be comparisons and harmonization work so the community can understand the relative pros and cons of performance of each. Now that we have the GIM score, we're fully supportive of those efforts."
He noted that Thermo Fisher has given four separate groups early access to the GIM score "to put GIM through its paces," but declined to name them. "So far things are looking good," he said.
In general, Vail said, clinicians interested in HRD scores need to make sure that they're validated against clinical populations, which is something Thermo Fisher still needs to demonstrate.
"They need to show they ran it against 100 [samples analyzed by Myriad's test] and say, 'This is our concordance. These are our discrepancies,'" he said.
He had no doubt Thermo Fisher will perform such a validation study. "I'm sure we'll do a study, too, when this becomes available," he added.