Definium Therapeutics, Inc. (DFTX) Earnings Call Transcript & Summary

All right. Awesome. Thanks so much, everybody. Always a pleasure to be moderating a discussion for Rob Barrow, CEO of Definium. I'm sure people who are dialing in here mostly know the story, but maybe Rob can give us sort of a quick snapshot of where you guys are at? Maybe talk about some of the data you generated this year with 20 the upcoming readout that we're going to dig into pretty deeply on some of the nuances there. So Rob, thanks again. Always good to see you and yes, take it away.

Yes, likewise, thanks so much for having us. It's been an incredibly exciting summer to our lead program, DT12 is an ODT form of LSC. We're developing in generalizing [indiscernible] depressive disorder. 2024, we had landmark results, first study readout for this molecule in this program, generalist where we saw an incredibly large effect size, 12 weeks after a single treatment effect at 0.81, a demonstrated dose response. And then we kicked off a Phase III program. Based on those data, we kicked off the program both in and first data, we had this summer in June. We're from a merger MDD pivotal study, where we saw an effect size of 0.83, and 8.1 on the MADRS separation. About 6 or 8 weeks later. In August, we had a second Phase III readout, our first GA readout in Phase III Voyage where we saw an effect size of 0.81. And now we're here in September and by the end of this month, we anticipate having our third Phase III read out the summer. So our team has been quite busy. We've got Breakthrough Therapy Designation for the JD program. have been working diversely over the last several years, but over especially over the course of this year that you're ready for hopefully an NDA filing if we have another positive study here.

So I'm old enough to remember when I said to you in March that you shouldn't set a bar for the MDD study. It turns out that bar was a total sand bank. So congrats. So yes, maybe let's talk about Panorama. I guess how are you thinking about the bar for success for that study? And we'll talk about this more, but maybe help frame for people the context around having the 100 dose and also the 50 dose and what different clinical outcomes for the 50 dose could or could not have in terms of implications for 120 in the program overall?

Yes. I mean at this point, of course, when we talk about bars for outcomes, I think there's sort of the fundamental truth and reality. And then there's, of course, the outside looking in perspective of what people want to see. We're talking about -- to just take a brief step back, we're talking about in market in generalizing anxiety disorder where the last Aprica Eli Lilly had [ Cymbalta ] approved initially for depression and then moved into JD. There hasn't been anything new since then. The prevalence over that time has tripled. So we have 10% of U.S. adults that is not for lack of trying it in a dozen programs that have tried and failed in late-stage development in GAD over that time. It is a very difficult indication, a very difficult scale to show any sort of separation for. And so fundamentally, at this stage, we have 2 positive studies. We need a second pivotal study in GAD and that will provide the evidence of tipi pivotal studies to support an NDA for a new product. And so if the study is positive and what we saw from the interim sample reestimation in Panorama is that the implied separation needed to demonstrate statistical significance is quite small or a little over 2 points that we would expect the study to be positive based on all those assumptions. And so fundamentally, what we need is a positive study. Now if we just had a 2.5 point separation, but it was positive, would people be excited, really not as much. We've continually set a very high bar just because of the data we've been able to generate to date. We powered the study to have 90% power to detect a 5-point difference, assuming a 10-unit standard deviation. So about a 0.5% effect size. And what we've consistently said is if we can show a separation of 4 points or greater, it's going to be among the largest effects ever seen in JD. And not only do you have a new drug, first new drug in 20 years but 1 with the most significant efficacy, the largest efficacy to be seen in technical studies. So all of those dynamics, we feel like anything that is positive, but certainly anything that's 4 points or greater separation is just remarkably strong and set us up very well. As you mentioned, that is -- the study is almost exact replica of the VOYAGE study that you read out a few weeks ago, the difference being that it also includes a lower dose of 50-microgram control. Everyone talks about functional unblinding, which is what that dose arm is there to try to address in some capacity. And really what I was trying to do is lower the information value of patients feeling something as the patients take a dose of 120 and whether it's 50 or 100 micrograms in our Phase II study, we showed that folks can reliably tell that they're on a dose of drug. Shouldn't be surprising to anyone. This is how site drugs work, but a lot of attention and a lot of inks filled around this topic in our field. And so basically we were trying to say is if you feel the effects of drug, you can't assume it's the real dose they've got. It might be a dose of drug that previously we showed had no clinical separation or statistical separation from placebo and is trying to sort of build a robustness argument. It's really not a study-specific argument. It's 1 that's programmatic. So we had a 5-arm study where we showed an FX 0.81 at a 2-arm study, where it should affect size. If we have a third study with a different design and we're consistently seeing statistically significant large magnetite efficacy, that's going to give us a ton of confidence that there is a real drug effect that's happening. And that's the point of FDA asking for these complementary designs across the program.

When you say -- it's super interesting. So when you say it's about influencing patient expectation bias. Like is this how -- like is this framed to patients in the way you just described to us, like I know you're speaking on a more kind of like conversational way, like, hey, you might feel something, it may or may not be efficacious. But like how operational is that kind of framing to patients on what arm could be randomized and what can.

I mean it's part of the informed consent process. You're right. I think it's quite as casually described as [ Mesahere ] for you or for those who are listening in. But it is exactly what I described is that there are 2 doses of drug in the study and to provide the context of what those doses could be. It is similar thing that we did in Phase II, where the informs Phase I describe placebo or 1 of the 4 doses of drug. And patients are typically told the probability of getting the drug and they're also told at the different dose levels, what might be the effects of those drugs so that they have the appropriate level of information to consent to the study. So yes, it is absolutely a part of the education process. It's something that we can sometimes not normally so, but sort of lean into because if it's there to try to sort of separate ties between expectancy and some biased reporting of outcomes, do you want people to be aware of it, if you just sort of hit it, it wouldn't have any purpose at all.

Right, right. Okay. And I guess the only sort of sub question that I've gotten from some investors about this is sort of what if 50 actually does a lot better than you expected? What if it's as good or what if it's better than 100 nominally, like does that could -- is there any risk that, that opens to can of worms where the FDA says, oh, now we want you to do another study looking at lower doses. And it gets back to this whole sort of lowest effective dose question, which I think it's actually been a murky question in general for neuropsych, but maybe like entertain that hypothetical scenario for us, Rob, what does that mean?

Every regulatory decision necessarily is on its benefit and risk right? It's -- for a drug that was highly toxic or we had -- if we were talking about a narrow therapeutic index and butting up against the top of that, then that argument might have a little bit more traction. Here, we're talking about intermittent sparsely intermittent dosing where patients are in a controlled clinical environment and the tolerability, the safety profile has been quite benign so far and all that we've seen. And so in that context, 1 would have to argue that you either really truly expect superior efficacy or that you'd be mitigating some risk that somehow the risk would be unacceptable at 100 micrograms and acceptable at 50%. Part of the challenge with drawing any conclusions from this is that the study was not powered around a 50-microgram dose in any way. It is not being tested inferentially in any way. And so there's really no statistical conclusion. And again, we look at things nominally and we see the scores we see the numbers are different or higher or lower. But fundamentally, to prove something, you have to establish hypothesis and power the study to be confident in the conclusion and then test it. And when you're testing multiple endpoints, you have to control for multiplicity, so you'd have to have somewhere in a statistically multiplicity-controlled way establishes the thing is happening. We're not even attempting to do that. We're not nowhere in our physical analysis, are we comparing 150 or 50 to placebo. -- it is there as an operational control. The other thing is sort of as you think about the sort of regulatory precedents, I think that's quite fascinating to look at it. I think folks can easily jump over history. One of the very few approved drugs and general thing I did this is effect for. And they're in a dose response study, they near linear dose response. And then in a Phase III study where the tested doses of 75 and 150 milligrams the 75 outperform the 150. You don't see psychiatrists concerned [indiscernible] Both of those doses were approved. They're both controlled and there was a study designed to assess both of those doses. But just because of lower or a different dose of a drug works or doesn't work, tells you nothing about the dose of interest. And so the entire point, the entire design of these studies has been to establish that is beyond a statistical degree of confidence, right, that 100 is superior to see that we're trying to establish through these studies. And the last point I'd make is that in the absence of the study we did in Phase II, this might be more appropriately elevated as a concern. But uniquely in this deal, we -- going into Phase II, we didn't know the answer. We didn't know what the right dose of drug was. So we designed a study to actually address that question and we established statistically and clinically a dose response, we established that to get clinically meaningful and durable efficacy you need the 100-microgram dose and so if there are questions about what dose is the right dose, each study can provide different point estimates of something. But when you have a body of evidence supporting the efficacy of 100 micrograms, and you statistically and clinically shown that you have a dose response and the 100 is the right dose. I'm not sure what we're really still asking more about what 50 could do or not do.

Right, right. Okay. Fair enough. Maybe one other question on Panorama and that's just maybe talk about the population you've enrolled. There's -- you ran a really interesting experiment with Voyage where you tried to really isolate GAD symptoms. What did you learn from the voyage results? And is that read through? And is it the same expectation with how Panoramic to play out?

Yes, absolutely. There is historically been a lot of conversation and this could -- we could take up the rest of the hour probably with this is history. But the sort of pendulum for a long time, swing towards MDD as sort of central focal point of psychiatry, like all the new drugs were for MDD and even the DSM-4, JD took a sort of back seat to MDD. So that if you had MDD, you could even then arrive at the GAD diagnosis, which is, of course, not a reflection of any fundamental reality or any broader history. Sort of where the structural elements of the world when we're seeing that go in entirely the other direction now it seems. But through that process, there is this concern about pseudo specificity, particularly in the context of SRIs that all started as antidepressants, and we're trying to expand our labels into GAD. So the concern was, given the high degree of overlap between MDD and GAD, particularly between the scales between the MADRS and the AMA, the concern was that if you have someone with both GAD and MDD, it has elevated depression or mood scores that you could be having an antidepressant effect and it's showing up as an anxiolytic effect, right? You're not really treating the anxiety, you're just treating depression and making anxiety better. Has a little bit of a face validity problem...

Why does FDA even care at the end of the day, why something is working if it's working?

For regulatory purposes, we have to isolate variables like good science, right? You have to isolate a variable, try to establish that variable. So presumably, if your FDA say, well, if you've already got an MDD label, why am I going to give you a new label for a different set of symptoms if you've already proven that it works for that set of symptoms. . And so you end up with this sort of highly overlapping Venn diagram when we go for GAD approvals, we have to isolate the GAD without MDD. And so we pay a lot of attention to that in pivotal studies. And one of the most interesting kind of things that we've now learned from our development program and just looking epidemiology and disease impact is some of the most focus of clinical care are those with anxious depression or elevated anxiety and elevated depression. Drugs we have work more against the boot symptoms and less good against the anxiety symptoms. So in Voyage, I guess, if you think of it this way, while we -- in all of our GAD studies did not include patients who are in a depressive episode. And we went to greater lengths both procedurally and with minor additions to entry criteria in the Phase III studies to really isolate the GAD symptoms. So HMA scores elevated without the commensurate increase in MADRAS scores. In Phase II, we saw baseline MADRS scores in the mid- to upper 20s across the arms. That's just right sort of at the threshold for what it would require to get into an MDD study on the MADRS. And Phase III in Voyage, we saw baseline [indiscernible] around 14. Right. The end result of that is that you not only are isolating appropriately again, for regulatory burst isolating GAD population. We do 2 other things. One, you limit the dynamic range of what you can change on the HMA. So Item 6 in MA is depressed mood. Each of the items on the MA, most of them can also map onto the MADRS and the depression symptoms. So if you take away the depression symptoms, you artificially in some ways, depress the starting point -- and on an item level, you sort of take away the ability to move things, right? You can only move things as much as they're there to be moved in the first place. The other thing is you do is because the HMA has 2 subcomponents, 1 is physiological and 1 is psychological. Bias the phenotype towards a more physiological manifestation. It turns out that's a lot harder to move. It's one of the really difficult things about GAD is that you get symptoms that are captured in a HMA cardiovascular genitourinary and things that aren't just how are you feeling right now? And so the end result of all of this is we ended up with the right population for regulatory purposes. But it's a little bit different in terms of how it's going to show up from a measurement standpoint, you have a less of a dynamic range end of the scale. You have lower item level starting point, you're going to get less nominal change in scores there get less nominal variance in the data as a result. It's using a ruler instead of a yardstick, right? It's just that level of difference. And so exactly what we saw in Voyage, same study design, same dynamics that play with Panorama, we'd expect a similar sort of dynamic where what we saw in our blinded sample re-estimation was a lower-than-anticipated variance in the data reflective of all the things we've been describing. And then the nominal sort of starting point of baseline HAMA scores were a couple of points lower than in Phase II. I think we would expect something quite similar in Panera as we saw at Voyage.

Rob, people can't resist but to just hammer me with 50-microgram questions. So I'm going to throw out a few of them, if that's cool.

Okay. Yes.

One, did FDA clear the second GAD Phase III trial with the 50-microgram dose to be sufficient in addressing functional and blinding. And maybe talk about how explicitly this was discussed with FDA.

Every one of our studies has been discussed explicitly starting for Phase III in a Phase II meeting, including many rounds of detailed protocol review and advice letters and information requests throughout the development program. So we've had a -- I mean FDA has been incredibly collaborative throughout the development program and has been very close to the study designs and the conduct of these studies with us over the course of the last several years. . The thing that's so interesting is that I want to -- as a point maybe worth clarifying, and maybe picking on the wording, but functional binding happens with psych drugs. Period. It just does, right? I mean [indiscernible] is a great example of this other anxiety drug. People who take 2 milligrams at Xenex, know they take their drug regardless if they're a randomized study or not? The thing we're trying to do is not eliminate functional mining because that's not possible. What we're trying to do is add an element of the study design that increases confidence in the reliability of the outcomes we're drawing from the development program. That's why 50 is there. And so that's the purpose it serves, and that's what we've been discussing in that context with FDA.

Yes. Yes. Okay. Makes sense. And then I guess -- do you expect 50 to perform the same way as it did in Phase II? Like I asked another way, do you think that's a real signal in Phase II where there's like this modest effect, but it's clearly less than $100

I think Yes. Based on what we've been able to model from a dose response standpoint, we think that there is probably a sort of threshold effect where the activity of the drug is more reliably achieved at 100 micrograms, and is it 50? And that while some people will take 50 and we tell us these sort of PDMS-endpoints trying to measure how an effect the drug is having put a lot of credence in those endpoints because they're a little bit fuzzy. But some people are going to take 50 micrograms and have the most profound experience they've ever had. But we're not worried about individual patients, we're looking at averages here. So when we look at the dose response model A, we assumed it was going to be monotonic I would assume that higher doses are going to have higher effect or at least not less than clear dose. And as what we saw really is that sort of hit this plateau at 100 micrograms, it doesn't increase further at 200. But at 50, we see a much lower effect. So regardless of what we see from the ballparking here, but the roughly 50 patients in the study who will have received 50 micrograms. That 1 study is point [indiscernible] of what the effect is. We feel highly confident in the dose response that we've chosen the right 1 in 100 micrograms and that it could show up sort of anywhere between separation for placebo and very similar to 100, and it will have any impact on the actual results of this even?

Yes. Okay. All right. I'm going to hit you with one more 50 question that I got, and I promise we can move on. Did FDA suggest you include the 50 dose? Or did you suggest it and why? Well, you kind of already said sort of the context.

Yes. FDA has suggested that all sponsors include effectively intermediate dose control our argumentation and I think the sound logic here is that if you're going to use a control for this purpose, it needs to be a dose that is reliably functional, right, that it is reliably detectable, otherwise, it's unclear what purpose it would serve. . So we chose the dose based on data we had generated in Phase II that we don't believe is clinically efficacious based on the evidence we have, but that is reliably detectable sort of the best dose 1 I could choose, but as sort of codified FDA's guidance, they're asking sponsors to do this. And really, again, what they've said is use complementary designs between your Phase II and III program to support the robustness of the conclusions. And that's exactly what we've done. So at this stage, we've got 2 of 3 studies that are positive and serve that purpose. We feel that there's a quite low relevance to what happens in one single study for one very small arm to construct the ultimate argument, which is that 100 micrograms is with an effective dose to treat GAD symptoms.

Yes. Okay. Fair enough. One other -- maybe a couple of other regulatory questions. What did you make of this FDA guidelines that just came out? And how do you make sense of the seemingly confusing comment around 1-year blinded data as like not an explicit requirement, but like a sort of like would be the best way to -- I mean, look, obviously, like a study of 100,000 people would be the best way to characterize efficacy. But look, the FDA is contentional to some degree of what they put in these things. So maybe speak about that?

Yes. FDA has been very collaborative with us and I'm sure with other sponsors, and nothing in the guidance was a surprise I think, to any of us in the field, right? And I think those who are sort of direct interactions with FDA's viewpoints is when guidelines like that are published, obviously, the draft guidance was 3 years before that. And folks didn't get an updated view 3 years later. But over the course of those 3 years, we've had dozens of interactions with FDA. And I'd say that the final guidance that was issued is aligned with our program and is aligned with the arguments that we and others have been making and the sort of dialogue we've had with FDA since the inception of our development program. So there's nothing particularly surprising I think that there were some refinements between the draft guidance and the final guidance, one of which was an explicit statement that for an initial application, you need 12 weeks of data. So like -- of course, I mean, studies in a post-approval setting, if you look at most antidepressants that have been approved, they're approved on 2 acute studies and then post marketing, people do randomize the withdrawal study. Now there's some complexity with this category of drugs and how you might do that here a randomized withdrawal study. But yes, the longer and the larger the study is more informative value had ideal studies would be huge and forever. It's just sort of impractical. I think the most important thing here is that what is clearly stated there and again a reflection of the need and the opportunity here, is that you establish acute efficacy and that's what you need, what not to you say in the guidance what you need to establish for additional application. Subsequent research can be done that could be informative -- but even what's in that guidance, our -- you've heard me say this, but people talk about double-blinded studies. They're quadruply blinded studies. And in our studies, at all points throughout the duration of the study, they're at least single-blinded. So the guidance says, most of the format design would be blinded on your followed with effectively triggered retreatment. Exactly what we're doing. The outcome assessors are blinded no matter what, no matter when we are in the study, whether we've unblinded, whether patients take an open-label drug, they don't know that no matter what -- now a different -- once a patient takes open-label drug from that point forward, of course, they are unblinded and they're treating physicians unblinded. Once we unblind Part A, we, of course, are unblinded to the treatment assignment. But throughout the duration, at least 1 of those 4 legs of the stool remains blinded. And so our interpretation is that even interpreting that as some sort of direct relevant to the studies that are ongoing. So it is very much in line with what we've done. I guess the last reflection there would to would be FDA is really thoughtful and I think they've been really again, I can only speculate for other programs. But seemingly, there's been the same degree of constructive dialogue for this field, which is why they came out with guidance. It would be pretty shocking if they had given out several breakthrough therapy designations had meetings with us to align on the program and with others in the field, and I'm thinking of the facility programs and [indiscernible] programs. And each time there is an agreement on the study design for the Phase IIIs, which included exactly the study design we're doing and it's somehow guidance would be issued that is directly in contrast of that. That will be a little bit odd. So I think even in that context, I understand that -- these things have a lot of nuance to them, but we just don't see that as anything other than aligned with what we're doing.

Yes Okay. Makes sense. When you guys -- can you -- again, more questions coming on the regulatory side. Can you talk about just like when you firmed up your regulatory alignment on the designs of the MDD and GAD studies -- was this before like this guidance was kind of crystallized and what does the timing of that matter at all?

I think the guidance as issued a couple of months ago. So it was quite a bit before.

I know. Obviously, I know that this was before a couple of months ago, but there was like a draft guidance and stuff like that. Like I guess maybe just like as this guidance has evolved, like you firmed up your program like a number of years ago. Like is there anything that's kind of changed with the FDA and the guidance side that isn't aligned with your program, Rob?

No. I think what we saw is that our conversations with FDA or crystallized in the finalized guidance and that FDA is thinking continued to evolve. Again, I'm sure based on dialogue with us and others in the field and engaging with the field broadly -- and so while that guidance was sort of out there for the world to see a few months ago, it is a reflection of the iterative and very granular discussion we've been having with them over the past 5 years, definitely from the point of having our Phase II data in early 2024 and the design of the Phase III program. So I just there's just not thing there that we were surprised by or that is misaligned with the discussions we've been having with FDA.

Right. Makes sense. So assuming the Panorama study is positive, I guess a 2-part question, but when do you plan to hold a pre-NDA meeting? And two, what's your degree of confidence going into a meeting like that, that you have enough data not just to file GAD but also even MDD with it?

Absolutely going to have pre-NDA meeting. That's exactly we're going to have that discussion and come away with, hopefully, clarity on filing pathway for both of these indications. All the things we were describing before about the overlapping symptomatology and epidemiology in these indications is exactly why the law was updated in 1996 to allow for 1 study approvals and highly adjacent indications with really compelling evidence. So there's a sound rationale for at least having that discussion. Now practically, either the areas we feel like we're in great shape, right? I think we were to file GAD first, go through a review cycle and immediately file an sNDA. We don't think it's going to have any sort of adverse downstream effect in terms of launching the drug, updating the label and so forth. But there's a huge degree of urgency in both of these populations. And given the overlap, the comorbidity, there's a reasonably sound argument to say prescribers, patients would benefit from being informed about the utility in both the indications. That's the argument that we think we'll be constructing and as we come away from a pre-NDA meeting, we'll be able to offer more clarity on exactly what the plan is there.

This whole set of pseudo-specificity thing that you had to do in ad. Could they make you do the inverse of that?