From Lab to Market: Navigating Diagnostics and Regulation with Andrew L'Huilier

[00:00:01] Welcome to the Chalk Talk Gym Podcast, where we explore insights into healthcare that help uncover new opportunities for growth and success. I'm your host, Jim Jordan.

[00:00:18] Welcome back to Chalk Talk Gym. Today we have a remarkable guest, Andrew LHuilier. He's a PhD in immunology and an expert in precision medicine. Andrew's journey began in the diagnostics industry, focusing on precision medicine and companion diagnostics.

[00:00:34] Now these predict patient responses to specific diseases. And throughout his career, Andrew has been at the forefront of innovation, particularly in the regulatory landscape, navigating complex FDA and global regulations to ensure effective patient care and drug development.

[00:00:49] In this episode, Andrew shares insights into the evolution of the diagnostics industry from basic tests to advanced precision medicine. He talks about the impact of recent regulatory changes and the future of healthcare innovation.

[00:01:03] Join us as we explore how technologies and regulations are shaping our industry and what it means to patient outcomes. So Andrew, tell me a little bit more about yourself.

[00:01:13] I'm Andrew LHuilier and my background, I'm a PhD kind of immunology background. And after school, I went to Rutgers University. After school, I went into the diagnostics industry.

[00:01:27] And that really picked off, I would say, the primary focus of my career through the years. When I was starting off in the diagnostic space, I was in the CLIA environment. It's a company that has a large number of clinical labs through the Northeast.

[00:01:43] And so I really got to see how do you move from the science to actually the clinical application of these tests.

[00:01:52] And of course, the kind of really interesting area that I saw emerging at the time, this was a little more than a decade ago, was companion diagnostics.

[00:02:01] And whether something is a true companion diagnostic or not, it really falls into this bucket of precision medicine where you have patients, you want to know who's going to respond best to what particular treatment, what is really the best way to treat this patient and give them the best outcome.

[00:02:20] And I think diagnostics traditionally has aimed towards that generally, but it's been more about just defining your population. What's too high? What's too low? Are you a high risk patient? Are you not? What particular disease might you have?

[00:02:36] But not really getting into the nitty gritty of within a population, all my lung cancer patients, for example, which ones are going to respond to this drug? Which ones are going to respond to that drug instead of just carte blanche, giving everybody the same thing and crossing your fingers.

[00:02:51] Yeah. Can we talk about where the diagnostic industry fits within the continuum of health care? Because going back to the statement you just made about the beginning of the industry, it was basically saying whether you have or do not have something.

[00:03:05] But this has evolved. Can you share your insights on that?

[00:03:09] Sure. Yeah. And this is something that we'll probably touch on a little bit later in terms of CLIA and how the diagnostic industry on the clinical lab side is run.

[00:03:18] But tests used to be quite simple. You had a, in previous decades, maybe 70s and 80s, you had a much more limited number of biomarkers that could be tested for in the first place.

[00:03:32] What do you actually have a test that can reliably detect?

[00:03:35] You had a limited amount of information of those things that we can actually detect what is particularly important.

[00:03:41] And so it really fell into those pretty simple buckets of does this patient have diabetes? Does this patient have cancer of a particular type?

[00:03:52] And it was then really the bulk of the burden then to figure out what to do with this patient really fell on the clinician using the gestalt of all of their experience.

[00:04:06] How does this patient look?

[00:04:07] How does this fit into the picture of patient type A, B, and C or disease type A, B, and C?

[00:04:13] Maybe this marker is relevant to 10 diseases.

[00:04:16] And I have to use the other things that might be relevant to each one to figure out which patient I actually have, which disease this person actually has.

[00:04:25] And once you came to a diagnosis, it would then be on the clinician to figure it out.

[00:04:30] For example, I have a patient with diabetes.

[00:04:32] How do you treat?

[00:04:33] Let me just go and read the basic guidelines that are published out there to see what's the current best practice of how to deal with that.

[00:04:41] And all of the patients of a particular type would usually get the same treatment.

[00:04:45] I'm taking a little risk here, getting my statistics correct or incorrect so you can correct me.

[00:04:52] But it seems to me the diagnostic industry in the old days was like a T-test.

[00:04:56] You were either in that population or you weren't, and there was no real sense of degree.

[00:05:01] And today we have more information about that population in terms of slices of it, or what you might call it standard deviation statistic term, where you are in that.

[00:05:12] And that led to you using the term precision medicine and personalized medicine.

[00:05:17] Can you maybe define those a little bit for people?

[00:05:20] Sure.

[00:05:20] I think it's funny.

[00:05:21] Personalized medicine, the way it's been described to me is personalized medicine was a term originally coined by, I believe, FDA.

[00:05:30] And apparently some physicians got quite mad and said, hey, we're very personalized.

[00:05:35] We look at every single patient individually.

[00:05:38] Don't insinuate that we don't.

[00:05:39] And so they changed it to precision medicine.

[00:05:41] We're being more precise with our patients.

[00:05:44] And it's really this idea of saying you can have a population of patients, say diabetics, say a particular type of cancer, lung cancer, something like that.

[00:05:54] But of that total pool of patients, not everybody is going to respond the same way.

[00:06:02] People have different ages.

[00:06:04] People have different lifestyles.

[00:06:06] They might have different BMIs, different metabolism.

[00:06:08] There might be different subsets of a particular disease that are mechanistically different, even though they present the same and therefore have the same name.

[00:06:17] And based on many confounding factors, patients respond to treatments in different ways.

[00:06:24] And we've known this for a long time.

[00:06:26] And for, I would say, the majority of the history of medical treatment, it was just taken that, okay, this is the patient population.

[00:06:36] I know some patients are going to benefit more for the current standard treatment than others.

[00:06:41] And that's just how it is.

[00:06:42] And if one of my patients particularly doesn't respond well, I'll just have to figure something out at that time.

[00:06:48] And for a patient that doesn't respond to the standard, their physician might have to go through two, three, four different lines of treatment with that patient to try and find, just based on guesswork, what that patient might respond best to.

[00:07:01] And so this precision medicine idea was essentially, can we do more with our growing arsenal of scientific tools to identify all those subsets of patients ahead of time so we know which patients are going to respond to what and do best with what.

[00:07:20] So we can get this guesswork of giving the wrong patients to the treatment and get straight to the one that's going to work the best.

[00:07:27] And that has benefited, I think, both sides of the coin, both on the physician's side to pick a treatment and treat their patient with the right thing faster, but also on the drug development side.

[00:07:43] Because if you're making a clinical trial and you want to get your drug approved, traditionally clinical trials would be the same way too.

[00:07:51] I have my patient population of say, and I'm just going to get a whole bunch of diabetics and I'm going to give them my drug.

[00:07:57] And hopefully the response rate is good enough that FDA and other health authorities around the world give the thumbs up.

[00:08:05] And you had a relatively low success rate and therefore a lot of years and a lot of money wasted.

[00:08:13] And a lot of patients who were going in hopeful for these clinical trials who didn't have a good outcome.

[00:08:17] So the idea was really, well, if we can take one of these biomarkers and cut to the chase to say we're using this biomarker only to put the patients we think are going to do well in our clinical trial from the beginning, then the response rate for those patients at the end of the trial is going to be better.

[00:08:36] Therefore, I'm going to have a higher chance of getting the thumbs up from FDA and other health authorities.

[00:08:40] Therefore, I'm going to waste fewer years and fewer dollars to get my drug to market.

[00:08:47] There's a lot that you just said there.

[00:08:49] Let's unpack some.

[00:08:50] Sure, go ahead.

[00:08:51] I think there was a startup company here in Pittsburgh whose name is eluding me at the moment, but it's going back about 15 years ago to give an example of the benefit of this.

[00:09:03] So it was a situation where if I had a cancer, a particular cancer, and I got treated with the right chemotherapy the first time I had a certain result.

[00:09:15] If they gave me the wrong one the first time and I had an adjustment and I used the right one the second time, the result was not the same, meaning that I needed to pick the right chemo the first time.

[00:09:28] So I think that was step one of this process in terms of getting things right.

[00:09:32] And to your point, that's good for the physicians, good for the patient, it's good for cost, right?

[00:09:37] Just bring it down to the cost of the entire system.

[00:09:40] And then I think of the diagnostics industry moving towards helping with patient management, right?

[00:09:49] You had a scenario where we worked with a company called Ariel Therapeutics that worked on pancreatitis.

[00:09:56] And of course, if left untreated, it can evolve to pancreatic cancer.

[00:10:01] And I think that there's a whole long line of if-then statements that go in this precision medicine to manage a patient, meaning if I produce today with pancreatitis and you produce today, not only genetically are we different, but our evolution on that chronic pathway is very different.

[00:10:18] And so these have led to helping with the cost of reducing clinical trials by having better selection.

[00:10:27] But there's another piece that you hinted to that I'm curious in your comment.

[00:10:30] So the diagnostic industry, they used to say, if you don't have a drug to treat it, then why are we collecting the data?

[00:10:37] That's a 20-year-old statement.

[00:10:38] And I think one of the things that's interesting in watching some of these precision therapeutic models is they're starting to come back with data for the pharmaceutical industry that says, here's some evidence that there's something over here worth pursuing.

[00:10:52] Have you experienced or witnessed any of those?

[00:10:54] I think that's definitely the case.

[00:10:57] And you can collect a lot of interesting, just exploratory information from a clinical trial.

[00:11:06] And today you might not have a drug that blocks or enhances pathway X, but two years from now, something might be coming out of R&D into the clinical pipeline that does.

[00:11:18] And if you have past data that really helps you prove that and helps you understand that can then potentially move into the clinic a lot faster.

[00:11:28] And in terms of drug development, it's all about speed.

[00:11:32] I want to get my drug into the clinic as fast as possible.

[00:11:36] I want to get my drug to market as fast as possible, especially if I'm going to be the first to market with drug type X or indication Y.

[00:11:45] Because that represents such a huge capture of a new market that that's what really pays for the R&D in the first place.

[00:11:55] And so if you can get that kind of exploratory data up front, then that can really help with that kind of speed.

[00:12:02] So we're talking a lot of product development here and you shared with the audience at the beginning of your science background, but you play the position of regulatory person.

[00:12:10] Can you explain what that position entails for people?

[00:12:13] Sure.

[00:12:14] Sure.

[00:12:15] The diagnostics industry has one set of regulatory requirements and the drug side has a separate set of regulatory requirements.

[00:12:26] And you can further subdivide that into kind of small molecule versus biologic.

[00:12:31] So that's exemplified by the three branches of FDA.

[00:12:35] You have CDR, which regulates the small molecule and the antibody stuff.

[00:12:39] You have CBER, which regulates the biological things like cell therapies, gene therapies, things like that.

[00:12:47] And then you have CDRH, who's by the drug side, the forgotten branch of FDA.

[00:12:52] But of course, to the medical device side is everything, right?

[00:12:55] And the CBER and CDR, the biologics and the small molecules are generally regulated pretty similarly.

[00:13:04] But CDRH, because it has to deal with all medical devices, which goes from everything from Band-Aids up to a test that tells you if you have cancer, it's a huge range of types of products of how they work, of who you're using it on.

[00:13:20] So there's quite a bit of difference from the drug world to the diagnostic.

[00:13:26] And I think it's pretty safe to say that in general, drug companies, especially large pharma, are better funded than medical device companies have more money to spare.

[00:13:38] And so it is, as a pharmaceutical company who's working on a drug, goes out and says, hey, I think a CDX companion diagnostic to say what biomarker says who should get this drug.

[00:13:51] I think that's important for this particular drug.

[00:13:53] I need to go out and I need to find a diagnostic partner who I can work with to make that test for me, to find those patients for me.

[00:14:02] And of course, the drug company doesn't want to make that test themselves because their expertise is on the drug side.

[00:14:08] It's not on the diagnostic side or the medical device side.

[00:14:11] So they want to work with a partner.

[00:14:13] They go out and they find a partner.

[00:14:15] This partner might be really competent in technology and they make a really solid technology.

[00:14:21] They might be really competent in the clinical side, even, which I would say most diagnostic companies are really good at technology.

[00:14:28] A subset of those will be really good at running a clinical trial because not all diagnostics will need a full clinical trial.

[00:14:36] And then a subset of those are really good at dealing with regulatory agencies like FDA.

[00:14:44] And what pharma found as they started to go into this area of precision medicine is, hey, these diagnostic companies are really great and they make great tests and refining the right patients.

[00:14:54] But they have no idea how to get a regulatory approval from FDA, let alone across the world in multiple jurisdictions.

[00:15:03] What it was really found was that the pharma companies have to start building that expertise internally so that they can help the diagnostic partners to not only do the right steps to set themselves up for approval of the drug and the diagnostic test, but also do it globally because these are global clinical trials.

[00:15:31] And do it in the results and do it in the results that matches the pace with which pharma wants to get their drug to market.

[00:15:38] So that's interesting.

[00:15:39] I hadn't thought of it that way.

[00:15:41] We were talking earlier before we started about the recent changes in the regulatory side from the FDA perspective on this.

[00:15:49] But you would just to stick with this for a second.

[00:15:52] You were just talking about that.

[00:15:53] But it's like any major industry, if their vendor or supplier network needs help, they're going to train, they're going to educate.

[00:16:00] But it seems as I listened to that, it made me think that there's a certain alignment with the timing of the FDA also changing their regulations.

[00:16:09] So can you share what they changed and why you think they changed it?

[00:16:13] Definitely.

[00:16:13] I think at the start of our talk, I mentioned how diagnostic testing was back in the 70s and 80s.

[00:16:19] And for many years, so you have the diagnostic test, which is an actual test that's produced by a company.

[00:16:26] And perhaps that's put in a kit in a box and sold to clinical labs.

[00:16:30] The clinical labs are the companies who actually run the test.

[00:16:35] So the physician will send like a blood sample or a tumor sample to the clinical lab.

[00:16:40] The clinical lab will buy the test.

[00:16:42] They'll run the test on the sample they got from the doctor.

[00:16:45] And they'll send a report back to the doctor saying, here's the result for your patient.

[00:16:49] Your patient has disease X or they don't.

[00:16:51] And for many years, there was a gap in terms of the companies building these kitted solutions that those labs could buy versus what actually needed to be diagnosed in the population.

[00:17:06] So it became quite common for the labs to make their own what are often called homebrew tests.

[00:17:13] They would buy a bunch of reagents just from research companies, play around with them with some patient samples and see what works and what doesn't and make their own test.

[00:17:23] There's been a bit of a disconnect in the U.S.

[00:17:27] And it's similar in a number of jurisdictions around the world where you have the companies that make the kits and sell to the clinical labs.

[00:17:34] And those are generally regulated by FDA versus the clinical labs who are making their own homebrew kits are not regulated by FDA, or at least that's been a contentious issue.

[00:17:47] The labs are regulated by a separate set of legislation called CLIA.

[00:17:52] And CLIA is administered by CMS, which is Medicare and Medicaid services.

[00:17:59] And they have to basically prove to CMS on a yearly basis that they are collecting the right information, that they are validating their tests appropriately, that they have kappa systems.

[00:18:12] There's some kind of an issue.

[00:18:14] They follow up with the issue and they go to the root cause.

[00:18:16] There's a lot of requirements under CLIA that kind of say how a lab should be regulated, should be run.

[00:18:24] And the burden of proof is on the labs to show, yes, we are running our lab appropriately.

[00:18:29] And here's the documentation to back that up.

[00:18:32] And it's been assumed for many years that as long as the lab is doing what it should to run its lab properly, then that should let us assume those homebrew tests are being developed appropriately.

[00:18:47] And therefore, they can be trusted.

[00:18:50] Now, FDA has for many years said, we think that we should have jurisdiction to regulate those tests, what are known as LDTs or laboratory developed tests.

[00:19:02] Because we say that they meet the definition of a medical device.

[00:19:07] And therefore, they're regulated by CDRH.

[00:19:10] And we have the authority to review those and approve those and send out mean letters if someone hasn't done the appropriate things.

[00:19:18] And it's been a contentious issue for a long time where FDA has started to assert itself in that space and send warning letters to companies making LDTs, etc.

[00:19:30] Then because this has been such a long-running industry that does serve patients legitimately, there's been a lot of pushback against FDA.

[00:19:41] And I think it's pretty reasonable to say that both sides, the clinical lab side and the FDA side, have reasonable arguments.

[00:19:52] These LDTs are essentially the same thing as a kitted solution, except they're only being built and used in one place.

[00:20:01] Otherwise, a diagnostic is a diagnostic.

[00:20:03] If FDA has the authority for kitted diagnostics, it makes sense that they should be able to review and approve non-kitted diagnostics.

[00:20:11] However, audience, because they can't see Andrew and I looking at each other here, I typed on the chat GPT if Thernos had any influence in being the tipping point for this.

[00:20:22] And it came back.

[00:20:23] It's been one of those names.

[00:20:23] Yeah, it served as a wake-up call.

[00:20:25] So this line of whether we should or shouldn't for years, I think probably got influenced from that perspective.

[00:20:31] But I also think what you bring forward here is the diagnostic industry is starting to take a different place, right?

[00:20:39] It was a you have it or you don't have it.

[00:20:42] Now it's you have it, you don't have it.

[00:20:44] Here's the degree.

[00:20:45] Here's the probability.

[00:20:46] Here's the predictability.

[00:20:47] And now we can start putting it in management systems for how we manage our populations and our patients.

[00:20:52] And I think that's the evolution.

[00:20:54] And then you're sharing that the pharmaceutical industry also needs this and is investing in this space.

[00:20:59] So there's not only governmental investment going in here, but it sounds like there's this pharmaceutical industry investment going on at the same time.

[00:21:07] Yeah.

[00:21:07] And that's very much FDA's argument is that when FDA originally decided to, even though they claim authority over those tests, not actively use that authority, what they call enforcement discretion.

[00:21:22] Their argument has been that we started doing that back in the 80s and 90s and tests were pretty simple back then.

[00:21:28] And they didn't really test for a lot.

[00:21:30] Now we have whole genome sequencing and tests predicting who's going to get what drug.

[00:21:35] And this is really becoming a much more serious part of medical management of patients.

[00:21:40] And therefore, it's time for us to step in.

[00:21:42] As we're in this transition, because I believe these new regulations were going in this summer, right?

[00:21:47] They were starting to transition in July of 2024 or somewhere in there.

[00:21:52] A big yes, but.

[00:21:54] So the final rule has come into effect.

[00:21:57] And so the clock has started for these clinical labs to really get into line and comply with what FDA has set forward in their final LTD rule.

[00:22:07] However, there are lawsuits going on right now.

[00:22:11] So I believe it's ACLA who has sued FDA.

[00:22:15] There may be more, which is a diagnostic kind of clinical lab group, industry group.

[00:22:20] There are other groups who could potentially file similar lawsuits against FDA.

[00:22:25] So that's all buzzing around.

[00:22:27] In addition to that, the Supreme Court has now knocked down the Chevron defense, which is this kind of defense that allows government agencies to courts must defer to governmental agencies for any interpretation of regulation that is reasonable.

[00:22:47] And so many governmental agencies, including FDA, including EPA, etc., have used that for many years to basically make guidance and make rules that they say we interpret things this way.

[00:23:00] It's going to be done this way.

[00:23:01] Now that's knocked down by the Supreme Court.

[00:23:03] That had been the Chevron defense came up from a Supreme Court case back in the 80s.

[00:23:08] Now it's knocked down again.

[00:23:10] So without that, FDA is going to have a harder time defending this rule in the courts.

[00:23:14] And then in addition to that, you have a very divided Congress.

[00:23:18] And that's not just this Congress.

[00:23:20] That's been Congress for 20 years.

[00:23:23] Legislation has been proposed and discussed and worked on, such as most recently the Valid Act in Congress that would formally identify and define diagnostics as a medical product that FDA specifically has jurisdiction over, including LDTs.

[00:23:39] However, for 20 years now, there's been so much argument even in Congress.

[00:23:45] Not all of the people in Congress have agreed on if that should be done or not.

[00:23:49] And those laws have never passed.

[00:23:51] So now most recently in the House Appropriations Bill for 2025, it includes a segment currently this has not passed, but it's in discussion in Congress right now.

[00:24:02] So a piece of that says that FDA will immediately halt all work and implementation of the LDT rule.

[00:24:09] If that were to go through both houses and the House and the Senate and Congress, then basically that would kill the rule.

[00:24:16] So it's very uncertain if this will actually move forward or not.

[00:24:20] Andrew, you bring up something that's giving me a deja vu here.

[00:24:24] So I taught a public policy course at CMU.

[00:24:27] So you have the code of law, right?

[00:24:29] And the code of law says, for example, thou shall not pollute.

[00:24:34] But how do you know what pollution is?

[00:24:36] And so there's a code of federal regulation and there is a clause that was negotiated.

[00:24:43] I want to say it was the 1940s, but I could totally be wrong here.

[00:24:47] I can look it up and if it's meaningful to this podcast.

[00:24:50] But there was a recognition that there needed to be experts to define the federal regulation to support the law.

[00:24:58] And I think what you're saying is, I don't know if that Chevron challenges all of it or challenges just certain components of it.

[00:25:06] But you can see, I think the audience can see the codundrum.

[00:25:10] How does the typical person in Congress know what the definition of, say, pollution is, right?

[00:25:16] What should be the carbon number and different things like that?

[00:25:19] So to your point, it's a real challenging situation.

[00:25:22] As you look at going forward in your functional role for your company with all these things changing,

[00:25:28] how can you set a direction or what sort of direction are you setting in terms of getting there over a period of time?

[00:25:37] Because I think I also recall that there was some of these regulations before being sued that were going to take 2027, 2030 before we were going to completely change the rules.

[00:25:47] So what sort of directions do you think you have to take the people you work with?

[00:25:53] Yeah, it's a very interesting question and I think it's very uncertain.

[00:25:57] I would, our initial kind of view of what this means for drug development and pharmaceutical companies is,

[00:26:06] there could be some bumps in the road, but probably it's business as usual.

[00:26:11] When we're going to, especially for a large pharmaceutical company,

[00:26:15] when we're going to try and develop and get to market a blockbuster drug,

[00:26:21] even though development of the companion diagnostic test is expensive,

[00:26:26] in the grand scheme of things of the cost of getting that drug to market, it's pretty small.

[00:26:33] And we generally do things in a pretty regulatory, robust way,

[00:26:39] where if we need what's known as an ID,

[00:26:42] the thing that FDA gives you the okay to start a medical device clinical trial,

[00:26:47] we very often get those for our tests,

[00:26:49] regardless of potential ways you could wiggle out of that currently.

[00:26:53] We go for full FDA approval through a PMA of our companion diagnostic tests.

[00:27:00] We want to make sure that there's really robust data and documentation

[00:27:06] and full approvals for everything that we need.

[00:27:10] Because even though it's a lot of money, like I said,

[00:27:13] compared to what the drug represents, it's small.

[00:27:16] So we want to have the best possible thing we can

[00:27:19] for identifying the right patients to have our drugs.

[00:27:24] And since we're already doing things in a very robust way,

[00:27:28] we don't really see this kind of shift as that big of an impact to us.

[00:27:35] But that being said, there are many smaller biotech companies

[00:27:39] who are early in their journey.

[00:27:41] They don't have the huge capital to put to work that a pharmaceutical company does.

[00:27:47] And maybe they're more likely to try and do things with it,

[00:27:50] try and push their drug forward to a point,

[00:27:53] maybe not all the way to drug approval,

[00:27:55] but maybe they're getting to a phase three

[00:27:57] and a larger pharmaceutical company might acquire them.

[00:28:00] Or they might have enough data that they could have a successful IPO

[00:28:04] and get out into the public markets.

[00:28:06] This, I think, could probably not have a big impact on the large pharmas,

[00:28:12] but probably would have a larger impact on the little biotechs.

[00:28:16] And the diagnostic industry itself, right?

[00:28:18] And the diagnostic industry.

[00:28:20] I think what you're saying is that the pharmaceutical industry

[00:28:24] has always operated at a higher level of regulation.

[00:28:30] As a result, that behavior is burned into their infrastructure.

[00:28:35] And so it's easier to go with the higher regulation

[00:28:38] than it is to try to change a system,

[00:28:40] to tone it down only in certain circumstances.

[00:28:43] It's just easier to be steady.

[00:28:45] So for the diagnostic industry, though, to them,

[00:28:47] it might not be that.

[00:28:49] It might be having to heighten and change their systems.

[00:28:51] And during this transition period,

[00:28:54] I imagine that it's a challenge to do that.

[00:28:56] So how do you keep current on these rapid changes?

[00:29:00] Where do you go?

[00:29:01] If our audience loves to hear exactly where do you go

[00:29:04] so that they can see some of those places themselves?

[00:29:07] Before I get to that,

[00:29:08] I'd like to just comment on what you just said

[00:29:10] on the impact for the diagnostic industry.

[00:29:12] I think it'll be a bifurcation

[00:29:15] because you have the larger diagnostic companies

[00:29:18] who have had enough success

[00:29:20] that they have built a regulatory department

[00:29:23] and they have the infrastructure in-house

[00:29:26] to do the right thing to set themselves up

[00:29:28] for a regulatory approval.

[00:29:31] They're actually probably going to be happy

[00:29:33] about these rules coming into place

[00:29:35] because for them, it's a regulatory lock-in.

[00:29:38] They have now a regulatory moat

[00:29:40] protecting them from competition

[00:29:42] versus the smaller players,

[00:29:45] some of whom might be bad companies,

[00:29:47] not working very well,

[00:29:48] which is what the FDA is afraid of,

[00:29:50] but some of whom might be very good companies.

[00:29:52] They're just small and up-and-comers

[00:29:54] and they don't have the money

[00:29:55] to be able to do all of this extra work.

[00:29:58] Maybe they could grow into a really successful company

[00:30:00] and they're going to have a harder time doing that now.

[00:30:02] So it's going to be a bifurcation.

[00:30:05] That's interesting.

[00:30:06] That's an interesting perspective.

[00:30:07] When we think of startups,

[00:30:08] most people think of patents

[00:30:10] as being the unfair advantage.

[00:30:12] Talk about the unfair advantage,

[00:30:13] but an unfair advantage is infrastructure,

[00:30:16] know-how, people, talent.

[00:30:18] There's a whole lot of things that make,

[00:30:19] give your company an advantage.

[00:30:21] It seems to me after,

[00:30:23] as I listened to you say,

[00:30:24] that there's been a mom and pop aspect

[00:30:27] to the diagnostic industry

[00:30:28] that it's a word I'm looking for,

[00:30:30] I guess has been

[00:30:31] that there's nothing wrong

[00:30:32] with the mom and pop startup and scaling,

[00:30:35] but you have to be able to do it

[00:30:36] according to certain standards.

[00:30:38] And that's, I think,

[00:30:39] given the take of it,

[00:30:41] there's legitimate arguments

[00:30:42] for more and less regulation on each side, right?

[00:30:45] I don't think there's anybody who's right

[00:30:47] and I don't think there's anybody who's wrong.

[00:30:49] But you can certainly imagine for a startup,

[00:30:51] speed is essential.

[00:30:53] Speed to market and revenue is essential.

[00:30:55] And in the diagnostics world for a long time,

[00:30:58] whether or not the long-term goal

[00:31:00] was to stay as an LDT,

[00:31:02] being able to start as an LDT

[00:31:04] represented a quick path to market

[00:31:06] and generating at least some revenue.

[00:31:08] And now if this rule stays in effect,

[00:31:11] that could be hammered.

[00:31:13] Yeah, that's interesting.

[00:31:14] Going back to the other question,

[00:31:16] where do you go to find your own?

[00:31:18] Yeah.

[00:31:19] There's a lot of good places.

[00:31:21] Obviously, I'm a regulatory person

[00:31:23] and FDA themselves are a great source of information

[00:31:27] for regulatory news.

[00:31:29] All of the different branches,

[00:31:30] CEDARS, CBER, and CDRH

[00:31:32] have newsletters that they send out constantly

[00:31:34] and you can get really great information.

[00:31:37] I like a lot,

[00:31:38] there's quite a few policy,

[00:31:41] regulatory and policy news sources out there.

[00:31:44] One that I like a lot is Agency IQ,

[00:31:47] which is owned and run by Politico, I believe.

[00:31:50] So they have a lot of newsletters they put out.

[00:31:52] They have a lot of webinars that they put on.

[00:31:56] 360DX is a general diagnostic industry,

[00:32:00] what's going on in the industry kind of news forum

[00:32:02] and they're quite good as well.

[00:32:03] But there's a lot out there.

[00:32:05] Excellent.

[00:32:05] What's the biggest lesson you've learned thus far

[00:32:07] in your healthcare journey?

[00:32:10] Biggest lesson in healthcare journey?

[00:32:12] I would say the biggest lesson is to really be open

[00:32:17] to things that are unexpected.

[00:32:20] In terms of, for example,

[00:32:22] I fell into diagnostics out of school accidentally

[00:32:26] and it seemed like an interesting enough place

[00:32:30] to start my journey in a career.

[00:32:32] And that just happened to be the time

[00:32:35] that we had this big kind of struggle

[00:32:40] between Optivo and Keytruda

[00:32:44] coming to market for the first time.

[00:32:46] And of course, Keytruda is from Merck,

[00:32:48] Optivo from BMS.

[00:32:50] There are these PD-1 inhibitors

[00:32:53] and there was a lot of talk at the time

[00:32:56] of number one,

[00:32:58] the benefit that can come to these patients.

[00:33:00] But number two,

[00:33:01] whether or not you really need to select your population

[00:33:04] based on one of these companion diagnostic tests.

[00:33:08] And until then,

[00:33:10] there had been talk of companion diagnostics before that.

[00:33:13] And there have been diagnostic tests approved by FDA

[00:33:17] and put on the market before that.

[00:33:20] But there was always this kind of argument

[00:33:22] in the pharma industry.

[00:33:24] Is this really worth it?

[00:33:26] You tell me that I can have a higher chance

[00:33:28] of getting my drug approved,

[00:33:29] but what you're really telling me is

[00:33:31] I have to now go pay 20, 30 million dollars

[00:33:35] to limit the number of patients

[00:33:38] that are eligible for my drug.

[00:33:40] That sounds like a terrible business decision.

[00:33:43] And that was the argument

[00:33:45] from the pharma industry for a very long time.

[00:33:47] And it really took that struggle

[00:33:51] of who's going to get a PD-1 inhibitor

[00:33:54] to market first,

[00:33:55] which obviously Merck with Keytruda won

[00:33:58] with their PD-L1 companion diagnostic

[00:34:01] that went along with it.

[00:34:03] That was really the first indication to the industry,

[00:34:06] I think,

[00:34:06] that not only is this good for patients,

[00:34:09] but this is actually something

[00:34:10] that makes business sense.

[00:34:12] And remind me,

[00:34:13] that was immunotherapy.

[00:34:15] So that was really the first kind of

[00:34:16] big immunotherapy breakout.

[00:34:18] Yeah.

[00:34:18] If you had PD-1,

[00:34:20] if PD-1 is the target,

[00:34:22] you couldn't help you?

[00:34:23] Was that the story with that one?

[00:34:25] PD-1 is the target of the drug.

[00:34:27] And PD-L1 is activator of PD-1.

[00:34:31] These patients and their tumors,

[00:34:33] they have a lot of PD-1.

[00:34:34] PD-L1 comes in

[00:34:35] and shuts off the immune system

[00:34:37] by interacting with PD-1.

[00:34:39] These anti-PD-1 drugs,

[00:34:41] what they do

[00:34:42] is they just block that interaction

[00:34:44] and it kicks up the immune system

[00:34:46] and helps the immune system fight the tumor

[00:34:47] because you don't have that shutdown

[00:34:50] of the immune system anymore.

[00:34:52] And a lot of people

[00:34:54] at a lot of companies

[00:34:55] were hoping,

[00:34:56] hey, I think,

[00:34:57] I just want to get this drug approved

[00:34:58] for everybody with lung cancer,

[00:35:01] everybody with liver cancer,

[00:35:02] whatever indication you're going into.

[00:35:04] I don't want to worry about this test

[00:35:06] limiting who I can give it to.

[00:35:08] And the other side of the coin was,

[00:35:10] if this shutdown of the immune system

[00:35:12] isn't happening,

[00:35:12] if you don't have the PD-L1

[00:35:15] to activate the PD-1,

[00:35:17] then those patients

[00:35:18] probably aren't going to benefit

[00:35:20] from the drug.

[00:35:21] And therefore,

[00:35:21] we should be just picking the patients

[00:35:23] who have really high PD-L1.

[00:35:25] And that's the attack

[00:35:27] that they took with Keytruda

[00:35:28] and it panned out

[00:35:30] and it really proved the business case.

[00:35:32] So what do you see

[00:35:33] as the biggest opportunity

[00:35:34] or threat

[00:35:35] in the next five or 10 years

[00:35:37] in the market?

[00:35:39] I would say,

[00:35:40] and this kind of ties back

[00:35:42] to all of the policy discussions

[00:35:44] we've been having

[00:35:44] with the LVT rule

[00:35:45] and all of that.

[00:35:46] On the one hand,

[00:35:47] you have FDA

[00:35:49] introducing this LVT rule.

[00:35:51] On the other hand,

[00:35:53] you have FDA saying,

[00:35:54] don't worry about it too much

[00:35:55] because we're going to down classify

[00:35:57] the risk category

[00:35:59] of the companion diagnostic tests

[00:36:02] and other tests too.

[00:36:03] Companion diagnostics,

[00:36:04] they're a class three,

[00:36:05] which means it's the highest level

[00:36:07] of burden

[00:36:08] for regulatory approval.

[00:36:10] You need a lot of really good data

[00:36:11] to prove it out.

[00:36:12] And they're saying,

[00:36:13] oh, we're going to knock it down

[00:36:14] to a class two.

[00:36:15] We're pretty comfortable

[00:36:16] with companion diagnostics now

[00:36:17] because they've been on the market

[00:36:18] for long enough.

[00:36:19] We know how to deal with them.

[00:36:21] And so that kind of,

[00:36:23] on the one hand,

[00:36:23] you're saying,

[00:36:24] oh, maybe there's,

[00:36:25] it's going to be harder

[00:36:26] because I have new regulatory

[00:36:28] requirements from FDA.

[00:36:29] But on the other hand,

[00:36:30] it's going to be less hard

[00:36:32] because everything's down classified.

[00:36:34] But then you have Europe

[00:36:37] and Europe has this new

[00:36:39] IVDR legislation

[00:36:40] that went into effect

[00:36:41] a couple of years ago

[00:36:42] that has wreaked havoc

[00:36:44] on clinical trials.

[00:36:45] It has made it really hard

[00:36:48] to start biomarker selected

[00:36:52] pop clinical trials

[00:36:53] in Europe

[00:36:54] with European patients.

[00:36:55] I'm not familiar with it.

[00:36:56] Tell us a little bit about that.

[00:36:59] Sure.

[00:36:59] It's the swing of the pendulum, right?

[00:37:01] In the old days,

[00:37:02] there was a legislation,

[00:37:03] the diagnostics legislation

[00:37:04] in Europe was called IVDD.

[00:37:06] It was very lackadaisical.

[00:37:09] Most tests,

[00:37:10] you could just self-certify.

[00:37:11] So you could say,

[00:37:12] oh yeah,

[00:37:13] you just submit something

[00:37:14] to the notified body

[00:37:15] in Europe and you say,

[00:37:16] oh yeah,

[00:37:16] I've done my due diligence.

[00:37:18] I looked at my test really hard.

[00:37:20] It's good.

[00:37:21] Thumbs up.

[00:37:21] And they'd give you a certificate

[00:37:22] and you were good to go

[00:37:23] on the European market.

[00:37:25] And actually,

[00:37:26] a lot of diagnostics companies

[00:37:27] would go to Europe first

[00:37:29] to launch their products

[00:37:30] because it was so much easier

[00:37:31] to get to market.

[00:37:33] Now,

[00:37:33] they saw

[00:37:34] a number of tests

[00:37:35] had big problems,

[00:37:37] newsworthy problems

[00:37:38] where people were diagnosed

[00:37:39] incorrectly

[00:37:40] and it led to a lot of big deals.

[00:37:42] And they went

[00:37:44] and they revamped

[00:37:45] the legislation

[00:37:46] called,

[00:37:47] it's now called IVDR.

[00:37:48] So IVDR

[00:37:49] is the much more strict version

[00:37:52] of diagnostics regulation

[00:37:53] in Europe.

[00:37:54] And so in the past

[00:37:55] where you could just go

[00:37:56] into Europe

[00:37:56] and say,

[00:37:57] hey,

[00:37:57] I'm starting up

[00:37:58] my clinical trial

[00:37:58] for my drug.

[00:37:59] I need a test

[00:38:00] to test for these patients

[00:38:01] and pick who I'm going

[00:38:02] to put in my trial.

[00:38:04] You could just have

[00:38:05] the diagnostic company

[00:38:06] self-certify that test.

[00:38:08] You start the clinical trial

[00:38:09] and then you could go

[00:38:10] for a higher level claim

[00:38:11] later if needed

[00:38:12] at the end of the trial.

[00:38:15] Now,

[00:38:16] everything is reworked.

[00:38:19] Whereas,

[00:38:19] I believe it was

[00:38:20] 80% of tests

[00:38:21] could be self-certified

[00:38:22] in the old days,

[00:38:23] now it's 10%.

[00:38:24] So now,

[00:38:26] 90% of the tests

[00:38:27] need pre-approval

[00:38:29] to go into

[00:38:30] the clinical trial.

[00:38:31] They need

[00:38:32] a very heavy review

[00:38:33] by the notified body

[00:38:35] of all of the data

[00:38:36] to get a CE mark

[00:38:38] to get certified

[00:38:38] in Europe

[00:38:39] to go on the market there.

[00:38:40] And the biggest problem

[00:38:42] is really time.

[00:38:43] We need to have

[00:38:44] our drug clinical trials

[00:38:45] starting yesterday.

[00:38:47] But now,

[00:38:48] if we want to have

[00:38:49] a novel test

[00:38:50] to select the patients,

[00:38:52] then we need

[00:38:53] to do a performance

[00:38:54] study application

[00:38:55] in Europe first.

[00:38:56] And that can take

[00:38:58] six months,

[00:38:59] nine months

[00:38:59] to get approved.

[00:39:00] So now,

[00:39:01] all of a sudden,

[00:39:01] we have this

[00:39:02] six to nine month delay

[00:39:03] to start our clinical trial

[00:39:05] of the drug

[00:39:06] just because we have

[00:39:07] to wait on the test.

[00:39:09] And

[00:39:11] especially in

[00:39:12] cancer,

[00:39:13] which is where

[00:39:13] most companion

[00:39:14] diagnostics are used,

[00:39:16] we're coming up

[00:39:17] with novel biomarkers,

[00:39:18] new biomarkers

[00:39:19] all the time.

[00:39:20] Things we didn't realize

[00:39:21] could be used

[00:39:22] to predict

[00:39:22] who might

[00:39:23] respond to a drug.

[00:39:24] So there's no test

[00:39:26] out there already

[00:39:26] that can already

[00:39:27] be approved

[00:39:28] and used in the

[00:39:29] clinical trial.

[00:39:30] There's no choice

[00:39:30] but to make a new test.

[00:39:31] And it's just

[00:39:32] made it really hard

[00:39:33] in Europe.

[00:39:34] But then similarly

[00:39:35] in other major markets,

[00:39:37] Japan has its own

[00:39:38] kind of

[00:39:40] requirements

[00:39:40] where

[00:39:41] in many countries,

[00:39:42] including the US,

[00:39:43] including in Europe,

[00:39:44] if you don't have

[00:39:45] the test completely ready

[00:39:46] by the time

[00:39:46] the drug is approved,

[00:39:47] they'll let you finish

[00:39:48] it on the back end.

[00:39:49] It's not ideal,

[00:39:50] but you can do it.

[00:39:51] In Japan,

[00:39:52] they won't let you do that.

[00:39:53] The drug is actually

[00:39:54] delayed

[00:39:54] for its approval

[00:39:56] until you have

[00:39:57] the test ready

[00:39:58] and fully approved.

[00:39:59] In China,

[00:40:00] the regulations

[00:40:01] are shifting

[00:40:01] all the time.

[00:40:02] And so there's

[00:40:04] the general theme,

[00:40:05] I would say,

[00:40:06] is just uncertainty.

[00:40:07] There's a lot

[00:40:08] of uncertainty

[00:40:09] because things are

[00:40:09] changing so rapidly

[00:40:11] in terms of how

[00:40:12] things are regulated.

[00:40:13] But you not only

[00:40:14] have the new technology,

[00:40:15] you have the new

[00:40:16] regulations to go along

[00:40:17] with the new technologies.

[00:40:19] They often don't

[00:40:20] mesh up

[00:40:20] and it just creates

[00:40:22] a lot of confusion.

[00:40:23] And now you have

[00:40:25] AI going into

[00:40:26] everything.

[00:40:27] You have things like

[00:40:28] the AI Act

[00:40:28] in Europe

[00:40:29] that's regulating

[00:40:30] AI in addition

[00:40:31] to medical devices

[00:40:33] and diagnostics

[00:40:34] and it's just like

[00:40:35] layer upon layer

[00:40:36] of regulation.

[00:40:37] And you have to

[00:40:38] square it all

[00:40:39] within the context

[00:40:40] of a global clinical

[00:40:41] trial

[00:40:41] and making a global

[00:40:43] strategy.

[00:40:43] It all becomes

[00:40:44] pretty complex.

[00:40:45] And these things

[00:40:46] we're learning

[00:40:47] as we go,

[00:40:48] I think of

[00:40:48] some of the discussions

[00:40:49] on the balance

[00:40:51] between legislation,

[00:40:52] technology,

[00:40:53] and advancement

[00:40:54] in things like

[00:40:55] maybe the Tesla car

[00:40:56] or the Uber car

[00:40:57] that's an auto drive

[00:40:59] and everyone's far

[00:41:00] until there's an accident

[00:41:01] that happens

[00:41:02] and the industry

[00:41:02] stops and the insurance

[00:41:04] industry said,

[00:41:04] who's responsible?

[00:41:05] And you saw it

[00:41:06] through that.

[00:41:06] And so sadly,

[00:41:07] I think that's just

[00:41:07] the evolution

[00:41:08] of every market

[00:41:09] development

[00:41:09] as it relates

[00:41:10] to any sort

[00:41:10] of technology.

[00:41:11] So I guess

[00:41:11] we're in the middle

[00:41:12] of that

[00:41:13] in the diagnostic

[00:41:14] area at this point.

[00:41:15] I would say so.

[00:41:16] And there's always

[00:41:17] the kind of two tracks

[00:41:18] that you can go with.

[00:41:19] You can keep

[00:41:19] the status quo

[00:41:20] and wait for there

[00:41:21] to be a problem

[00:41:22] and react

[00:41:23] to that problem

[00:41:24] with new regulation.

[00:41:25] Or you can try

[00:41:27] and preempt

[00:41:28] the issues

[00:41:29] down the road

[00:41:30] with legislation

[00:41:31] now

[00:41:31] in the hopes

[00:41:32] that you can

[00:41:33] avoid problems

[00:41:34] in the future.

[00:41:35] And obviously

[00:41:36] both have

[00:41:37] their pros and cons.

[00:41:39] If you're reacting,

[00:41:40] that means someone

[00:41:40] got hurt.

[00:41:41] That's not a good thing.

[00:41:42] If you're trying

[00:41:42] to preempt it,

[00:41:44] very often

[00:41:45] you might see

[00:41:46] a huge damper

[00:41:48] on an entire industry

[00:41:49] so that,

[00:41:50] again,

[00:41:50] the AI Act

[00:41:51] in Europe,

[00:41:51] they're worried

[00:41:52] that it's going

[00:41:52] to almost kill

[00:41:54] competitiveness

[00:41:55] for AI companies

[00:41:56] coming out of Europe

[00:41:57] compared to other countries.

[00:42:00] And if that means

[00:42:01] we're getting

[00:42:02] things to market slower,

[00:42:03] does that mean

[00:42:04] we're helping patients

[00:42:05] less because

[00:42:06] they're not getting

[00:42:06] the best?

[00:42:07] That's another risk.

[00:42:08] And so there's

[00:42:09] the risk of what happens

[00:42:10] and the risk

[00:42:10] of what doesn't happen.

[00:42:12] And how do you

[00:42:12] do what's best?

[00:42:13] It's not easy.

[00:42:14] No, it's a complex

[00:42:15] topic for sure.

[00:42:16] Is there anything else

[00:42:17] you'd like to share

[00:42:18] with the audience?

[00:42:19] I would like to share

[00:42:20] that I think the field

[00:42:22] is evolving

[00:42:22] really rapidly.

[00:42:23] Even though

[00:42:24] there's all

[00:42:25] these complexities,

[00:42:26] I'm pretty hopeful

[00:42:27] personally

[00:42:27] that is going

[00:42:29] to be resulting

[00:42:30] in more and more

[00:42:31] better outcomes

[00:42:32] for patients.

[00:42:33] Excellent.

[00:42:34] Ended on a positive note,

[00:42:35] right?

[00:42:35] Yeah, I appreciate

[00:42:36] your time.

[00:42:37] This has been great.

[00:42:38] Thank you.

[00:42:39] I've enjoyed it too.

[00:42:40] Thank you so much.

[00:42:43] Thanks for tuning

[00:42:43] into the Chalk Talk

[00:42:45] Gym podcast.

[00:42:46] For resources,

[00:42:48] show notes,

[00:42:48] and ways to get in touch,

[00:42:50] visit us at

[00:42:53] chalktalkgym.com.