Abstract
Next generation sequencing (NGS) is playing an increasingly important role in clinical research and clinical diagnostics. This expanding utilization has largely occurred outside of the U.S. Food and Drug Administration's (FDA's) regulatory process. In part, this is due to the lack of clear regulatory guidance as to how tests using NGS would undergo regulatory review; NGS-based tests do not neatly fit within the current regulatory model. The lack of clarity makes it difficult for FDA to articulate expectations, and deters companies from submitting applications. Unclear regulatory requirements and pathways make the review process more unpredictable and risky for applicants.
Another thorny issue is product change. While in vitro diagnostic (IVD) tests tend to change over time, NGS tests are far more elastic. Obtaining clearance for an NGS product that is static, e.g., cannot be updated to reflect new data or genetic insights, is not commercially viable. To FDA's credit, the agency has been working for a number of years to develop clearer guidance on NGS assays, and has solicited substantial public input. On July 8, FDA released two draft guidances. One addresses the acceptability of curated data bases, and the other data requirements for IVDs for germline mutations. Both documents are important not only for their impact on NGS tests, but also the insights they offer into FDA regulation of other kinds of IVDs.
luchschen / Getty Images
The first document, entitled “Use of Public Human Genetic Variant Databases to Support Clinical Validity for [NGS]-Based In Vitro Diagnostics” is seemingly narrow. It addresses the ability of NGS applicants to rely upon curated databases in lieu of providing clinical validity data. The use of external data to demonstrate clinical validity is critical to the viability of submitting applications to FDA for NGS IVDs. Generating its own data to prove the clinical validity of a single biomarker or genetic sequence can be challenging for a company. Trying to do that for a wide range of variants would be daunting in the extreme, and would deter or prevent applications. FDA has opened the possibility of NGS applicants referencing a database rather than generating its own data.
The implications of this approach go beyond NGS. There are other situations in which IVD manufacturers want to use external information to show the clinical validity of markers. A recurring issue for IVD companies is demonstrating clinical validity. Offering another interesting insight, FDA notes that there “are several parallels” between guidelines for variant curation and FDA review of clinical validity. To the extent that the draft guidance shows greater receptiveness to reliable third party sources, there would be a welcome spillover effect.
However, while FDA has unlocked the door to the use of curated databases, pushing the door open will not be easy. FDA is not proposing that any and every database may be cited. Rather, an applicant can only use databases that pass FDA muster, and that will not be easy.
Obtaining FDA recognition of a database will take work. For example, the database operator must not only be transparent, but must provide enough information about data sources and SOPs to allow third parties “to make fully informed medical decisions.” Variant data should be accompanied by metadata. The database operator will need to make its decision matrices publicly available. All of this information—and much more—must be submitted to FDA before the agency will recognize the database. But database recognition is not necessarily permanent. Once a database is recognized, the operator will need to remain in good standing. FDA has said that it will review the databases “on a set schedule to verify” that they continue to comply. The draft guidance does not explain what happens if FDA finds issues; how a database operator can respond; or what happens to applicants relying on the database.
Many questions abound. FDA does not say how a database operator could challenge a decision not to recognize it, nor does the document talk about the notice to database operators before their database loses its recognition. Even so, the concept of reliance on curated databases in lieu of sponsor-generated data is a step forward.
Also issued July 8, the second draft guidance addresses the use of standards for NGS-based IVDs diagnosing germline mutations. FDA stresses the narrow scope: the document excludes, for example, standalone diagnostics, screening, fetal testing, or multiple other applications. Nevertheless, the draft guidance does cover a significant clinical area.
Many mutations will not have predicate devices, i.e., there is no FDA-cleared IVD for the mutation. The agency, though, held out the strong possibility of tests undergoing review through the de novo process, rather than the more stringent pre-market approval process. The de novo is more akin to the 510(k) notification, and offers significant advantages to both FDA and sponsors, both in the initial review and once the test is on the market.
FDA did not explicitly say that it would use the de novo process. Rather, it said these tests would be good candidates for de novo. Given that FDA cannot guarantee a device will be reviewed de novo until the data are submitted, this is about the strongest favorable signal that FDA can grant.
Furthermore, FDA held out the possibility that future IVD NGS tests that fall within the classification may not need to undergo any FDA review. This could mean that a single NGS IVD germline mutation test could enable other companies to enter the market without prior FDA review.
That is, in fact, what happened to a class of tests when FDA granted de novo status to 23andMe's genetic tests. However, FDA will establish “special controls.” Later entrants will need to meet these special controls, even if they bypass FDA review.
FDA does provide a considerable amount of guidance as to the contents of these submissions. These include, among other elements, accuracy, precision, limit of detection, analytical specificity, and test run quality metrics. The draft guidance also describes the information needed for test reports, including a summary of the performance tests and the use of “clear, consistent language that can be easily understood.” While it is still unclear what will be the fate of FDA's proposal to regulate laboratory developed tests, this may give some insights into what would be expected in lab reports if FDA did regulate LDTs.
The two draft guidances collectively represent progress in understanding FDA's expectations for NGS tests. While hardly complete, they are a start. All interested parties have the opportunity to comment, with a 90-day comment period. Given the complexity of the two topics, the different perspectives of potential applicants, and the ongoing technological changes in the NGS arena, FDA will have a lot to consider before it issues these documents in final. In the meantime, NGS-based IVD manufacturers will consider how this affects their FDA strategy, and laboratories offering NGS testing will be evaluating not only these documents, but waiting to see what happens with a guidance document of even greater importance to them: FDA's guidance on laboratory-developed tests.