NEW YORK (GenomeWeb) – A next-generation sequencing working group consisting of industry, academic, and government representatives has written a white paper addressing issues related to NGS-based companion diagnostic testing.
The group presented the paper, titled "A Blueprint for Drug/Diagnostic Co-Development: Next-Generation Sequencing (NGS) in Oncology," at a forum last week hosted by Friends of Cancer Research and Alexandria Real Estate Equities.
In the paper, which is also available online, the group laid out five main proposals that it hopes could help guide policy decisions.
First, it suggested that regulations for a cancer panel should allow flexibility in the appropriate US Food and Drug Administration medical device pathway and said that it could be based on risk classification of different panel components and use either the 510(k) or pre-market approval route depending on the specific marker.
Second, it proposed the validation studies be based on the types of alterations — SNVs versus deletions, for instance — rather than each alteration individually. Third, it suggested that the contents of a cancer panel should be determined by current utility in clinical care and clinical trials with modifications made annually. In order to address standardization, the group proposed using a common set of samples to ensure reproducibility, and finally, it suggested establishing a framework for determining the appropriate reference method against which to validate a panel.
The issues are especially pertinent because cancer panels are increasingly being used in clinical care, pharmaceutical companies and NGS developers are looking to develop multi-analyte companion diagnostics, and the FDA has said it will likely increase oversight of laboratory-developed tests; however, there have been no established standards or guidelines.
Illumina, for instance, is collaborating with AstraZeneca, Janssen Biotech, and Sanofi to leverage its FDA-cleared MiSeqDx system to develop an NGS-based panel that would serve as a universal companion diagnostic.
Thermo Fisher has a similar CDx agreement with Pfizer and GlaxoSmithKline to develop a test on its Ion Torrent PGM Dx platform, and Foundation Medicine is working with Clovis Oncology to develop a companion diagnostic.
Jeff Allen, executive director of Friends of Cancer Research, told CSN that the organization decided to tackle the issue of NGS clinical cancer panels because of its involvement with the Lung-MAP clinical trial, a multi-arm, multi-drug biomarker-driven trial that uses Foundation Medicine's FoundationOne test to screen patients upfront and determine which portion of the sub-study they qualify for.
"We thought through the larger policy implications of a trial like Lung-MAP," Allen said, adding that it became apparent that the possibility existed to shift the traditional companion diagnostic model from one test for one marker related to one drug, to an upfront NGS panel that screens for multiple biomarkers and serves as the companion diagnostic, Allen said. "We saw this as a possibility that was coming closer to reality rather than something on horizon," he said.
In addition, he said, given the FDA's recent clearance of the MiSeqDx for cystic fibrosis testing, "we thought now would be a good time to look at the landscape of cancer panels that are widely used in academic institutions to better understand and lay out a pathway for regulatory approval."
The FDA, too, has indicated the need for a paradigm shift in how companion diagnostics are developed and approved, and Allen said that his group has been in close discussions with the agency.
"We hope the blueprint will provide some sort of direction to [FDA]," he said.
Mya Thomae, Illumina's vice president of regulatory affairs, told Clinical Sequencing News that FDA "has used ideas and proposals from previous blueprint papers from Friends of Cancer Research in guidance and policy decision-making."
Illumina and other NGS technology developers have increasingly been eyeing the clinical market and Thomae said that while the guidelines in the paper have not changed the firm's clinical strategy, the paper does "support the strategy that we are currently taking."
Thomae, who was also part of the workgroup that developed the recommendations, said that the issues addressed are important because "the use of cancer panels is gradually moving into standard of care … but there has been no effort to standardize either panel performance or content."
The lack of standards has frequently been cited by experts as a major hurdle for moving next-gen sequencing into clinical use. For instance, in discussing the National Cancer Institute's MATCH trial, which will use a Thermo Fisher Ion Torrent-based panel to screen patients into different trials similar to the Lung-MAP trial, Mickey Williams, director of the molecular characterization laboratory at the Frederick National Laboratory of Cancer Research, said at a conference this year that one major question that has not been addressed is whether different platforms will give the same result.
Elizabeth Mansfield, director of personalized medicine at the FDA, has expressed similar sentiments, saying at the same conference that running different commercial tests on the same tumor currently yields different answers.
Nonetheless, because NGS panels potentially contain hundreds of targets, the workgroup in its paper acknowledged that validating each one individually would not be feasible. Rather, it suggested validating the different types of alterations — SNVs, indels, and copy number variants.
Thomae added that groups like the Genome in a Bottle Consortium are also working to develop standards. In addition, an Illumina-sponsored consortium, the Actionable Genome Consortium, aims to develop standards for using next-gen sequencing in oncology, including also defining what constitutes an actionable variant.
Allen said that Friends of Cancer Research has an ongoing discussion with the FDA about how to move forward and what the next steps should be. One potential question for future guidance, he said, is how "one could transition an [NGS panel] from research to the clinic — what types of studies and how these bridging studies should be conducted."