CHARLOTTE, North Carolina (GenomeWeb) – At the Association for Molecular Pathology here this week, Thermo Fisher Scientific announced a new research panel for profiling gene expression in tumors and their surrounding microenvironment, which it hopes has the potential to predict response to cancer immunotherapies with more sensitivity than current assays.
The new offering, called the Oncomine Immune Response Research Assay, enables targeted sequencing of 396 gene transcripts using Ion Torrent next-generation sequencing platforms, with low sample input requirements.
Although the assay is currently for research use only, partners who have been evaluating the tool in an early-access context are hoping that it can eventually be used clinically to predict patients' responses to the rapidly growing contingent of immune-oncology drugs.
Jeffrey Conroy, senior vice president of technology development for OmniSeq — a spinout of the Roswell Park Cancer Institute — reported on his company's early-access work with the panel at a workshop hosted by Thermo Fisher at the AMP meeting yesterday.
"We were very excited to get early access to this assay," Conroy said, adding that the panel was attractive because of its applicability to FFPE samples, its cost-effectiveness, and its targeted nature, which helps improve signal-to-noise significantly over whole-transcriptome sequencing.
"We were hoping this could be the assay that met the unmet need for checkpoint inhibitor therapy prediction," he added.
OmniSeq has been using another Thermo Fisher Oncomine assay — the company's comprehensive 144-gene panel — in clinical testing since June.
"We have been lucky in the four months since we got approval [for that test] to have assayed 400 patients," matching a significant proportion of patients to either US Food and Drug Administration-approved drugs, or precision medicine clinical trials, Conroy said.
However, he added, a large number of patients are now also being directed toward immunotherapies, something that the comprehensive assay doesn't speak to at all.
"This is going to be a burgeoning field that we need to get some sort of treatment decision support around," he said.
OmniSeq and Roswell aren't alone in this. As immune checkpoint inhibitors have advanced, interest in assays that can sensitively predict patient response to these drugs has followed. At the same time, the field has increasingly recognized that currently available immunohistochemistry assays that measure PD-L1 expression leave a lot of room for improvement.
Existing tests only have about 40 percent positive predictive value for those who test positive, Conroy estimated at the AMP meeting.
Hoping to do better, OmniSeq has been working on the first steps of analytical validation and assay robustness testing for the new Thermo Fisher Oncomine assay.
The OIRRA panel was designed to comprehensively cover 395-genes involved in the immune system and its response to cancer, Conroy explained during the workshop. The assay uses just 10 ng of FFPE RNA, and users can multiplex four, eight, or 32 samples per run, according to Thermo Fisher.
In development and validation studies, OIRRA demonstrated more than 20-fold improvement in detection over transcriptome analysis, high reproducibility for detecting low expression cytokine-encoding genes, and high sensitivity and specificity for detecting differential expression as low as two-fold, Thermo Fisher reported.
At the AMP meeting, Conroy shared some of OmniSeq's early validation data for the new assay, and said that the company is now moving forward with retrospective studies to try to show that classifying patients by their tumor micro-environmental gene expression can predict their response to checkpoint inhibition.
Based on the firm's analytical validation work (in about 350 specimens from patients with lung cancer, renal cell carcinoma, melanoma, ovarian cancer, or bladder cancer), Conroy said that the panel met OmniSeq's main expectations for robustness as well as correlation with other measures of gene expression.
Specifically, the results indicated that the panel was repeatable, applicable to FFPE samples with high sensitivity compared to frozen tissue, and that various aspects of specimen variability, like age, architecture, stromal content, and mounting, didn't seem to introduce problematic bias into gene expression results.
When the group compared gene expression results to IHC or PCR-based analysis of specific targets like CD8, they saw high concordance. Looking at PD-L1, it appeared that the RNA-seq enabled by the panel actually might provide better information on the spectrum of PD-L1 expression across different patients than do existing assays.
OmniSeq is now working on a retrospective study aimed at showing that it can algorithmically categorize patients into low-, medium-, or high-expression groups using OIRRA, and then predict their response to immunotherapies based on which subset they fall into.
If that goes well, the company hopes it will be able to connect with collaborators that could support a more extensive prospective follow-up.
"We hope that as we continue to grow the training set we will refine the algorithm and get prediction rates above [current] industry standards," he said.
Speaking with GenomeWeb yesterday, Andy Felton, VP of marketing and product management for the clinical sequencing division at Thermo Fisher, stressed that the company has produced the panel as an RUO product.
However, this does not necessarily preclude labs from integrating technologies or reagents into their clinical testing of patient samples or using them to guide clinical therapeutic decision-making by gaining approval by CLIA and/or local authorities like the New York State Laboratory Evaluation Program.
For example, OmniSeq secured NY State approval to test patients with Thermo Fisher's 144-gene pan-cancer Oncomine Comprehensive Assay, which is also classified as RUO.
Thermo Fisher has its own internal programs for sequencing-based diagnostic development, notably an AmpliSeq-based universal companion diagnostic test that it is developing with Pfizer and Novartis.
Felton said that the company also recognizes CDx potential in immune gene expression sequencing, but it doesn't have any plans for advancing a clinical diagnostic similar to OIRRA that it can discuss publicly.
Along with the new OIRRA panel, Thermo Fisher also recently launched one other tissue-based targeted sequencing research kit — the Oncomine BRCA Research Assay for comprehensive analysis of BRCA 1 and 2 in FFPE samples.
At the AMP meeting, the company also announced that it is planning to expand its menu in the next few weeks with two new liquid biopsy panels — the Oncomine Breast cfDNA Assay and the Oncomine Colon cfDNA Assay— which will join its already available NSCLC cfDNA assay.
Both of the new cell-free DNA panels, like the existing lung cancer assay, will enable detection of primary driver and resistance mutations from cell-free DNA down to a level of 0.1 percent, the company said in its workshop at the AMP meeting yesterday.
Unlike companies that sell NGS-based liquid biopsy test services as LDTs, Thermo has focused on releasing RUO kits that any lab can implement.
Though the future may hold a path for these assays through the FDA, becoming diagnostic, rather than RUO kits, Felton told GenomeWeb that this possibility is far enough off that Thermo doesn't have any definitive plans it could share.
Startup Resolution Biosciences, which launched last year, has also said it hopes to develop targeted sequencing-based liquid biopsy IVDs.