Momentum from the 2011 launch of two highly anticipated personalized medicine cancer treatments — Pfizer's Xalkori and Roche's Zelboraf — continued to spur pharmaceutical companies and test makers to engage in drug and diagnostic codevelopment collaborations throughout 2012.
However, during the year, advances in high-throughput technologies like next-gen sequencing also made it undeniably clear that the single-marker/single-drug model of personalized medicine will likely be obsolete before it even fully emerges. As such, industry players are wondering what this impending data challenge means in terms of the regulations they will have to meet to launch innovative products.
Although the US Food and Drug Administration didn't release any guidelines last year focused specifically on molecular diagnostics, agency officials from the Center for Devices and Radiological Health said that they are working to develop processes to advance personalized medicine and make this information more accessible to the public.
During 2012, the agency issued a draft guidance on a new effort called the Pre-Submission Program, through which test manufacturers can seek FDA's input on an investigational device before formally applying for regulatory approval (PGx Reporter 7/18/2012). Additionally, the FDA issued draft guidelines that industry can use to enrich clinical trials, by, for example, enrolling patients with certain genomic markers, to study the safety and efficacy of a drug (see related story, in this issue.)
According to the agency's priorities for 2012, the final guidance on companion diagnostics was slated to be cleared internally within CDRH by June 30. By Dec. 31, CDRH had planned to clear a draft guidance on the co-development of therapeutics and devices.
For 2013, CDRH has prioritized releasing the final companion diagnostics guidance by March 31 and issuing the codevelopment draft guidelines by Sept. 30. Also by Sept. 30, CDRH, in collaboration with the National Institute of Standards and Technology, plans to develop genomics standards — specifically to "select and sequence the human DNA and microbial DNA reference materials to support analytical validation of genetic sequencing technologies."
In a recent interview with PGx Reporter, Alberto Gutierrez, director of CDRH's Office of In Vitro Diagnostics, and Elizabeth Mansfield, director of personalized medicine, discussed what's been going on at the agency over the past year in this space, and what's coming in 2013. Below is an edited transcript of the interview.
2012 was a quiet year in terms of regulatory guidance on molecular diagnostics from the FDA. Can you give us any idea of what has been going on within the agency on this front throughout the year?
Alberto Gutierrez: Although it has been a quiet year perhaps on [the molecular diagnostics guidance front], the center really has had a lot of guidances this year. The center has worked [to release guidelines on issues] that affect the entire center and that requires quite a bit of work on our part. So when you said this was a quiet year, my first reaction was, 'Oh my god, I hate to see things any more active than this.'
If you narrow it down to molecular diagnostics, we didn't have as many [guidances] coming out this year. But the [center did put out] other guidances, in terms of determining things like benefit/risk, and guidances that have to do with the new [Medical Device User Fee and Modernization Act] negotiations. We continue to work on a lot of guidances that affect our programs but are not directly related to molecular diagnostics.
This year the FDA simultaneously approved two new Rx/Dx products: a new indication for Erbitux and a companion KRAS test (PGx Reporter 7/11/12); as well as Perjeta and two HER2 companion tests (PGx Reporter 6/13/12). Were there any unique lessons learned from these approvals?
Elizabeth Mansfield: I think the splashiest one this year was the KRAS [companion diagnostic] approval that went along with cetuximab. That was interesting because the drugs had already been relabeled for safety reasons, both Erbitux and Vectibix. And this represented a new approval of Erbitux, for first-line therapy for which we were able to get enough testing data to actually establish that the test was clinically validated for selecting patients for therapy, rather than as a safety marker. So, you want patients who don't have a KRAS mutation to treat. So, that was pretty important because it had turned the corner in terms of the amount of data that we had seen.
The Perjeta approval, in terms of the companion diagnostic, was simply an extension of the diagnostic label, because it was used essentially in the same way that the diagnostic had always been used. So, there wasn't much exciting there.
For the next KRAS diagnostic that comes alongside a drug, will the test developer have an easier pathway in gaining approval of follow-on companion tests?
AG: So, each one of these has unique challenges, and what data they end up having to develop or not, to a certain extent, depends on the challenges and compared to what we know.
So, it's case by case. Even if you're looking at another KRAS test for another drug, for example, the submission requirements for the test will depend on the unique situation for the drug.
EM: Right, the drug will dictate how we view the clinical significance of the … test or the companion diagnostic. Also, if it comes from a different manufacturer, which it could in some cases, or the same manufacturer could have a different platform, then we would be looking at new analytical validation data.
If it was a companion KRAS test or HER2 test but for a brand new drug, could the diagnostic manufacturer use retrospective data? Or is that also case by case?
EM: Yeah that's really case by case. It depends on how the retrospective data were collected — how many samples they have available from the clinical trial and so on. So we really look at all of that when we look at retrospective trials.
You recently approved a new leukemia drug, Iclusig (ponatinib), and the interesting situation with that drug was that it was initially intended to be developed with a companion diagnostic but then you determined that a test was not necessary (PGx Reporter 12/19/12). Can you provide some insights into how that came about?
EM: The way the clinical trial was set up, when we saw it, it appeared that they would need a companion diagnostic to pick out the population of patients that would respond to the drug. But the way they did their trial, it gave them the opportunity to look at everyone, and not just the marker-defined population. It turned out that the drug worked acceptably on everyone, so there was no need to pick out that population.
Has there been any forward movement with regard to regulation of laboratory-developed tests at the FDA? The Office of Inspector General's 2013 workplan mentions looking at HHS regulation of LDTs. Does this mean things might go back to the drawing board?
EM: OIG operates under their own rules and they decide what they want to do investigations in … We know they chose it, but it wasn't at anyone's direction other than their internal. So, we're not sure where it's going ourselves.
Some industry observers are doubtful that the agency is going to finalize the companion diagnostic draft guidance anytime soon, particularly because the field is moving so quickly. Is FDA taking a wait-and-see approach or are there plans to finalize the companion diagnostic draft in the near term?
AG: I believe that's part of the CDRH [2013] priorities and … if I decide not to follow through here, I'm going to be in trouble. We plan to put the companion diagnostic guidance out as final.
Drugs that the agency has approved with a companion test are already being studied in new patient populations with different response markers. For example, drug developers are studying whether patients with ROS1-positive lung tumors may respond to ALK inhibitors similarly to patients with ALK-positive tumors. In such situations, would the drug developer have to develop a separate test and submit a new pre-market approval for each new PGx marker associated with a drug? Has the FDA given any thought to how to quickly advance multimarker companion tests?
EM: The easiest rapid way is for somebody to actually show that [these new markers] mean something. We have multiple ways of intercepting data. For example, [in the case of] the drugs that work in KRAS wild-types, the lack of effect of these drugs in patients with KRAS mutations was discovered after the drug was approved. We used essentially the published literature to put the safety warning on the drug label. So it depends a lot on how much information is available, where it comes from, as to how we can relabel the drug.
But at this point, would the test developer need to get a PMA for separate test? Or are you in discussions with industry about approving multigene companion tests?
EM: There's nothing to stop us from doing multigene tests, it's just that we have to have the evidence that they're all important and all clinically valid for that drug.
In terms of direct-to-consumer genetic testing, the US market is basically whittled down to one company at this point, and that company — 23andMe — is seeking FDA approval for its Personal Genome Service (PGx Reporter 8/1/2012). Are you planning on issuing formal guidance on DTC tests in the near future?
AG: I believe that one was also in our priorities list of the guidances we will be putting out this year. So, yes, we are putting out a guidance on direct-to-consumer [genetic testing].
Currently, test developers using next-generation sequencing technologies are primarily working with pharmaceutical companies on biomarker discovery projects, but there are also long-term collaborations underway to possibly use these technologies for companion diagnostics. Many industry players believe that in the future, NGS technologies, operated out of an LDT model, will be the best way to advance personalized treatment strategies for complex diseases such as cancer. Where does FDA's work on regulating NGS technologies stand? If such technologies are used as a companion test, does the FDA have a view on how these tests will be best implemented, as LDTs or as kits?
EM: We are ready to clear or approve them as either a kit or an LDT. We have no particular bias one way or the other. Our interest is to make sure that the tests are accurate and reliable. We're not blind to the fact that next-gen sequencing will be hugely influential at least over the next decade. Who knows what the next technology is? But, of course, we are working in that area.
What can we look forward to from CDRH in 2013 in terms of FDA regulation in the personalized medicine space?
EM: We just have a lot of things going on right now. There are new technologies popping up and the whole companion diagnostic thing still needs a lot of polishing. That's a sea change in the way we've done business, so we're still trying to make that process as smooth as it can be.
There's a pretty good chance, I guess, that we would have some more products approved in the companion diagnostic model next year. It depends on what comes in. But we think we have it running pretty well. We're spending a lot of time on investigational issues and drug clinical trials that use tests that are like companion diagnostics, because there's a lot of confusion there. So, we're hoping to straighten that out to some degree.
We're working on making some of our information more public. We've just published a list of the genetic tests that we've cleared or approved, except for the viral tests … because we never tracked those before. Last year, we pulled all the companion diagnostics out on the page. So, we're trying to make things more accessible and easier to navigate.