NEW YORK (GenomeWeb) – The US Food and Drug Administration was very busy last year updating its policies, regulations and expertise so that it could better facilitate access to personalized medicine products.
It was also a banner year at the FDA in terms of newly approved personalized drugs and molecular diagnostics. The agency continued to enable direct-to-consumer access to genetic tests, by authorizing 23andMe's pharmacogenetic tests, but it cautioned patients and doctors against using tests without the agency's clearance or approval. And much to the surprise of many in the lab industry, the FDA provided extensive feedback to lawmakers on a new regulatory framework for diagnostics, including lab-developed tests (LDTs). The agency was particularly active in the area of genetic testing and molecular diagnostics, which are integral to the delivery of personalized medicine. For example, the agency released a draft guidance on class labeling for companion diagnostics; a final guidance on how diagnostic sponsors can use information from public variant repositories to support the claimed relationships between disease and the variants their tests gauge; and recognized the first expert-classified variants from the NIH-supported Clinical Genome Resource consortium for this purpose, which can be accessed in ClinVar database.
The FDA's Office of In Vitro Diagnostics and Radiological Health (OIR) spearheaded a lot of these efforts under the leadership of new director Timothy Stenzel. As the former COO at Invivoscribe, a company developing companion diagnostics and next-generation sequencing-based diagnostics in oncology, his presence at OIR is an indication that the FDA anticipates the field of personalized medicine and advanced molecular diagnostics to be far more integrated in mainstream healthcare.
Stenzel took us through the precision medicine-focused efforts he took part in during his first year at the agency. Below is an edited transcript of the interview.
The FDA has made several announcements regarding the use of real-world data. Can you provide information about what OIR is thinking about real-world data and how that information might be used in the context of reviewing IVDs?
We're a member of the Medical Device Innovation Consortium, and we interact with them on a regular basis on this and many topics. These are active discussions that are going on about how to use real-world evidence for IVDs. In general, our office and the FDA [are] willing to consider the use of real-world evidence in support of regulatory decisions as long as we can determine that the evidence meets the standard for a particular need. We're willing to do that now.
We're very open to this and we believe it will advance regulatory science. It's not just [for the] premarket [setting]. It can be for post-market surveillance and recalls to inform what's going on out there and how we best deal with a particular situation. The standard for evidence is pretty much the same as we evaluate any evidence. Is the data secure? Is it easily verified as being accurate and reliable?
It seems there are a lot of efforts at the agency to modernize the regulatory process for medical devices and diagnostics. The agency announced several changes focused on predicate device modernization, dual FDA clearance and CLIA waivers, and the de novo pathway. Was there a specific reason for all the activity? Was this the result of feedback FDA received from industry or the agency's internal decisions?
Most of these changes apply to all devices and not just to those in our office, IVDs. However, we have been involved in these discussions and implementation of these policies. We're pursuing these changes to keep pace with the ever-increasing complexity and rapidly evolving technologies in our space, and other spaces, that the FDA has authority over. It's a matter of … evaluating where we stand now and are there ways we can advance regulatory science to launch innovative, lifesaving technologies as quickly as possible to help patients.
FDA regulation of LDTs was once again a topic of conversation at the end of 2018 with the agency providing input on draft legislation that is now known as VALID. The centerpiece of the FDA's recommendation is precertification. The FDA has a precertification pilot in the digital health realm. Would you consider a pilot for diagnostics as well?
We did not propose a pilot for this. We don't think it's necessary. We're learning a lot through the digital health precertification program. There are differences there as well, but we learn in real time from these efforts engaging sponsors and the public in innovative ways to advance regulatory science …
It's safe to say … the FDA would support a legislative solution for advancing public health. We're willing to give our opinion to Congress when asked.
Going a little further on this topic, FDA's input to Congress was very detailed, perhaps more detailed than some in the lab community expected. A lot of the folks I've spoken to in the lab and pathology community were nervous about this, especially those who have argued historically that the agency doesn't have the statutory authority to regulate this space. Do you have a response?
Congress is in charge of this. We cannot comment on our interactions with Congress, especially on VALID. However, we've noted publicly that we'd support a legislative solution. We look forward to working with Congress, patient groups, etc. to find that solution. And any other public comments that the Commissioner [Scott Gottlieb] has previously said about legislative proposals, we fully support.
Turning the DTC testing space, last year there was some confusion with FDA's authorization of 23andMe's PGx tests, which was followed by a cautionary statement on such tests that are not approved by the agency. This has been challenging for stakeholders to figure out because on the one hand, 23andMe's test isn't authorized as a diagnostic or to make medical decisions. But then the agency seems to be saying broadly that there are issues with unapproved PGx tests, as most tests on the market aren't. A lot of stakeholders have interpreted this to mean that although the FDA spent years updating drug labeling with PGx information, that the agency is now backing away from the entire field of pharmacogenetics. Can you clarify this for the community?
We did things two things fairly closely together. We authorized 23andMe. They came in with a submission, we reviewed it. We had very specific authorizations there with limited indications. That reflects two things. One is that it is a direct-to-consumer test and we handle those differently because they're defined as not having a health professional involved in the process. That would add some potential risks to providing information directly back to patients without a clinician being involved.
And the safety communication was really around the fact that we have concerns about some providers who are making predictions around drug response where there may not be sufficient scientific evidence to support making those recommendations or claims.
If I can be a little more specific: genetics experts had issues with specific portions of that cautionary statement. For example, there's a portion where FDA said that "the relationship between DNA variations and the effectiveness of antidepressant medication has never been established." The Clinical Pharmacogenetics Implementation Consortium, which the agency takes part in, stated outright that this was an inaccurate statement, and former FDA officials [involved in PGx testing] have also said this is possibly not accurate. Do you have a response?
We stand by the discussion and scientific discussion that we provided in the safety notice. We worked very closely with the Center for Drug Evaluation and Research on that. The press release was a joint press release between Center for Devices and Radiological Health and CDER.
At the same time, we are open to reviewing evidence as presented to us in a pre-submission or in an actual submission. There have actually been discussions around that. We anticipate seeing an uptick in applications in non-DTC situations, where clinicians are involved. So, we anticipate seeing a lot of activity in this area.
One area that’s been unclear is FDA's regulatory thinking on third-party data interpretation companies that take raw data from firms like 23andMe and analyze it for novel, health-related associations. The agency previously has said it is looking into these companies, but there's never been any enforcement action that's been publicized. Does the FDA believe it has authority to regulate such companies?
This is a very interesting new area where there has been a lot of new development and we're very interested in having further discussions with the community on this.
Another new strategy in consumer genetics is the use of third-party physician networks to approve test orders. Some people in the field say this is a conflict of interest, but it is unclear what FDA's position is on this type of physician ordering arrangement. Can you provide any insights on this? Does the FDA have statutory authority to regulate such practices? I've heard from some regulatory experts that maybe the FDA doesn't.
There are different commercial models out there clearly. One is the DTC model where a health professional is not involved. Quite a different model is where a clinician is involved. We see clear differences between those two [models.] We're keeping an eye on all these practices. At this point, the absence of a patient-physician relationship adds additional risk.
Last year was the first time the agency recognized a public genetic variant database, and, through the recognition of ClinVar and ClinGen, outlined a pathway for how other groups can gain similar recognition for variant databases. How should drug and diagnostic sponsors use this resource that you've already authorized? And have you seen an uptick in other sponsors wanting to go through this process to gain this type of recognition?
We have seen a real uptick in the interest in this. We were super excited about the first recognition. The way that these databases can be used is that the individual test developers, or sponsors, they don't have to independently reproduce the information or interpretations that are in an FDA recognized database. They can use that as a source they can refer to in their applications. This is really a way to reduce the burden on test developers and speed the process for getting innovative technologies to the patient.
Lastly, what can we look forward to in 2019 from OIR?
Our guidance A and B lists are published. What we really expect in 2019 is to have another record year for very novel and innovative devices [being] approved. We're seeing an ever-increasing uptick in this very exciting area of molecular diagnostics and personalized medicine. Our focus is how can we get safe and effective devices to the market as quickly as possible.