NEW YORK (GenomeWeb) – Following the recent release of two draft guidances related to next-generation sequencing tests, the US Food and Drug Administration yesterday hosted two webinars to explain the recommendations and to encourage stakeholders to provide feedback on the documents by Oct. 6.
"We hope that you will submit comments regarding the substance of each of these guidances, what we got right, what we didn't, what you think we should change," Laura Koontz, a member of the personalized medicine staff within FDA's device division, said during one of the webinars, which discussed the technical and regulatory aspects of the guidelines. The second webinar discussed what the draft guidances might mean for patients and providers, and also encouraged their feedback.
Earlier this month, the FDA released a draft guidance outlining how test developers could demonstrate the analytical validity of NGS tests for hereditary diseases. The agency simultaneously released another draft guidance outlining how administrators of publicly accessible genetic variant databases might garner the agency's recognition, so such repositories could provide clinical validity support for genetic tests.
"Both of these guidances rest on the foundation of transparency, that NGS test performance and the evidence supporting the clinical interpretation of genetic variants should be available to users of these tests," David Litwack, a policy advisor within the Office of In Vitro Diagnostics and Radiological Health (OIVD) in FDA's device division, said during one of the webinars.
Since the agency released the recommendations, some stakeholders have raised a number of objections, questioning the agency's statutory authority and scientific expertise to oversee NGS tests. The most vociferous opposition has come from members of the lab community, who suspect that the draft NGS guidances might be an underhanded way for the agency to subject lab-developed tests to oversight, which they have held off for more than two decades.
The FDA has emphasized that the draft guidelines present a "conceptual framework," not requirements, and have nothing to do with the agency's plans to regulate lab-developed tests. In a recent interview, Elizabeth Mansfield, director for personalized medicine and molecular genetics at FDA's OIVD, said the recommendations could help test developers improve the quality of NGS tests even if they don't bring them in for agency review. She also invited stakeholders to communicate their objections during the 90-day comment period.
"These are proposals put out for comment," Mansfield told GenomeWeb. "They are not final decisions. So, reacting really negatively to what we've proposed is exactly what we have a comment period for. It's really important for people to comment on what they do or don't like about the guidances, what they think they say, and so on."
Yesterday's webinars appear to be part of the agency's efforts to engage its detractors into more constructive dialogue.
Litwack noted during the first webinar that the agency decided to address analytical validity standards for germline tests because the agency believes most of the uses are moderate risk. "We would rather start with the relatively lower risk uses for NGS and extend to the higher risk uses, say, oncology and somatic variant testing," he said. "Technically, we viewed this [type of germline NGS testing] as the simplest cases."
In the draft guidance, the FDA outlined a standards-based approach that test developers could use to establish the analytical validity of NGS tests as Class II, moderate risk devices. Importantly, the FDA believes germline NGS tests could even be exempt from premarket review through the de novo process, if developers follow special requirements. A de novo 510(k) submission allows "FDA to review and down-classify a new technology, a new type test," Litwack said.
Koontz added that the draft guidance on public variant databases, in combination with the analytical validity guidance, "discuss the possible future down-classification and exemption from premarket review of NGS-based tests that demonstrate conformity with the standards outlined in the analytical standards guidance and use assertions from FDA-recognized databases."
During the second webinar, she explained to patients and providers that it would be overly burdensome for any test developer to have to generate all the clinical validity data for the potentially millions of variants that might be detected by NGS tests, especially for the exceedingly rare variants that only show up in one or two people in the world.
Public variant databases would allow for the aggregation of genetic variants seen in labs and in patients around the world. "Crowdsourcing evidence in this manner could mean that better tests get to the market faster and benefit patients sooner," Koontz said. "These databases can be incredibly useful to patients seeking to learn more information about their own disease or condition."
FDA's draft document lays out a path for administrators to voluntarily apply for recognition of a variant database and demonstrate that it meets the agency's guidelines. "Because transparency is a key component of these recommendations, the guidance document only applies to publicly accessible databases, not proprietary databases," she added.
In its effort to encourage feedback on the draft guidances, the FDA has also released federal register notices, in which it puts forth specific questions for consideration.
For the analytical validity draft guidance, the agency wants stakeholders to point out design, development, and validation activities important for germline NGS tests not currently addressed by the proposal. The agency also has questions about whether the current recommendations adequately differentiate between targeted panels and whole-exome sequencing; what additional design, development, and validation activities need to be addressed for the analytical validation of whole-genome sequencing tests for germline conditions; and in which scenarios minimum performance thresholds are not useful. h
For the draft guidance on variant databases, the agency wants stakeholders to weigh in on whether the same type of recommendations could be used for recognizing repositories containing both somatic and germline variants. The FDA also wants feedback on how such databases might support the clinical validity of tests using technologies other than NGS; how often the agency should review FDA-recognized databases as evidence on variants evolves and what the regulatory implications would be for tests that rely on these assertions; and the types of information FDA should collect to mitigate conflicts of interests that database personnel might have.
Importantly, the FDA also wants advice on how it should deal with discordant variant interpretations between databases. Labs that have not yet submitted to NIH's ClinVar, an archive of genotype/phenotype associations, have noted the discrepancies in variant interpretation between labs as a reason. However, those contributing to ClinVar are trying to reduce variant interpretation differences by working together to resolve discrepancies and with the interpretation help from expert groups.
The FDA is also thinking about how to manage variant interpretation differences from a regulatory standpoint. During the second webinar, one participant wondered what the agency planned to do if a lab reported a variant differently or a variant not in a recognized database. "We realize that laboratories often will see different variants and new variants they've never seen before," Mansfield responded. "We believe in that case you would be doing your own expert interpretation and you would report [that]."
"One of our goals is to allow the clinical test to keep up with the science," she explained. "In a sense, this could be considered a type of off-label use. It would be an expert interpretation by the laboratory professional."
By submitting to and using FDA-recognized public databases to establish the clinical validity of variants, Mansfield noted, test developers should be able to update their tests as the evidence evolves without having to come back to the agency for review every time. The new interpretation "would be something that you hopefully would put into the database, it would be in the database, and you could point to the database and say it was there," she said.
No existing publicly-accessible variant database currently has FDA's recognition, and some stakeholders have expressed concern that the agency may be placing a high bar for achieving and maintaining recognition for repositories that historically haven't been adequately funded or staffed.
In the Federal Register notice, the FDA estimated that it will take 80 hours annually to put together and submit an application for recognition, 20 hours to perform activities related to maintaining recognition, and one hour to post the information on a website. The agency also estimated there would be five applications per year from "respondents" — or administrators of genetic databases — "that may meet FDA recommendations for recognition and seek such recognition."
The FDA has said it is promulgating these guidances on NGS testing as part of the Precision Medicine Initiative, and plans to do more work in this area. According to Litwack, it is a high priority within the agency to issue recommendations for other NGS technologies and applications, such as "liquid biopsy" tests and diagnostics intended for use in oncology. The public will have an opportunity to also weigh in on those recommendations.
"To fully realize the vision outlined in these draft guidances and expand it to other indications in the future, we need your help," Koontz said. "We need community input on the technical recommendations of potential standards for other uses, like oncology, and for standard-developing organizations to issue consensus standards for NGS tests."
She noted that similar to FDA's involvement in creating the Genome in a Bottle reference genome, the agency will continue to work on the creation of reference materials in partnership with private and public organizations.
"FDA encourages community aggregation of genetic variant information to support data sharing efforts necessary for the creation and continuation of sustainable, high quality databases, and finally, curation," she added. "As we've heard at all of our public workshops, databases are dependent on external experts to assist in interpretation and curation activities and there is always more work to do. So, please consider getting involved there, too."
In putting together the draft NGS guidelines, the FDA held several public workshops to gather input from NGS test developers. Later this fall, sometime during the comment period for the draft guidances, the agency said it plans to hold another public workshop, this time to discuss the proposed recommendations.
Once the comment period is closed after Oct. 6, the agency will weigh the input it receives, incorporate changes to the guidances, and "hopefully move toward finalization and implementation," Koontz said.