Good afternoon, everyone. Welcome to the next session with Definium. I'm Ami Fadia, biotech analyst here at Needham. It's my pleasure to be hosting Rob Barrow, who's the CEO of the company, and Dan Karlin, who's its Chief Medical Officer. Rob and Dan, thank you so much for taking the time to do this sort of fireside chat. Maybe if I can turn it over to Rob for some opening remarks, and then we can dive into Q&A.
Yeah. First of all, thanks so much for having us. It's an incredibly exciting time for us, heading into pivotal data readouts for our lead program, DT120. A lot of attention, I think, in the last few years has started to be paid to our field, and we've been very fortunate to have incredible colleagues who have executed with remarkable efficiency and really set the standard for the field and the ability to design and execute trials and a development program for DT120, which is a proprietary form of LSD that we're developing in GAD, generalized anxiety disorder, and MDD, major depressive disorder. We see this as the next wave of really promising treatments in psychiatry.
While there are several programs in development, we're really proud of the data we've been able to generate so far and heading into these readouts, starting with Major Depressive Disorder in late Q2, and then our first GAD readout in early Q3 gives us a lot of reasons to be excited about the path forward and the ability to hopefully have a huge impact for patients out in the world.
Yeah, absolutely. I think this is a very interesting year for the company with three phase III data readouts. I'm excited to see how the year evolves. Maybe if you could start with the MDD study, which is the EMERGE study. That's the first one to read out. If you could start by talking about the powering of the study, the primary endpoint, and what change are you looking to see on the primary endpoint versus placebo for it to be competitive in this space?
Yeah. Maybe I'll turn it over to Dan, and he can start us off.
Yeah. The EMERGE study is a 1-to-1 randomized, double-blind, placebo-controlled study with a true placebo that is looking at a 12-week double-blind period with a six-week primary outcome measure. The measure that's used in this is the standard one that's used in all depression studies, which is the MADRS. We'll look at MADRS in double-blind fashion out to week 12, and then beginning after week 12, we have a 40-week extension period, during which time folks who reach a moderate or worse severity on the MADRS are eligible for up to four additional open-label treatments. At that point, folks who get treatment are guaranteed to get drug. They're no longer randomized. That allows for a couple of things.
That obviously gives people an incentive to stay in the study, and the more people that stay in the study, the better your ability to measure people, because it's very hard to measure them when they're not in the study anymore. It also allows us to characterize the efficacy of the drug over a longer period of time in a controlled fashion, because unless someone, and until someone gets an open-label treatment, they remain in that initial double-blind state, right? They don't become unblinded just because they pass week 12. That gives us the ability to look at a full year from that initial treatment in that double-blind state. The other advantage of that extension phase is that because of the nature of the triggered treatment, we get to evaluate the sorts of patterns that we expect to see in the real world, right?
That real world treatment is not randomized, and it's based on usual severity. We're really excited for the richness of the data that will emerge from that second, the part Bs. The study is 80% powered to detect a 5-point difference on the MADRS, and obviously that sets a standard for ourselves. We are likely to detect a smaller change if there were to be one, but a 5-point change in the MADRS would be remarkable relative to the current existing standard. Of course, we always hope for bigger placebo-adjusted change than less. We're extraordinarily confident and hopeful in being able to get to five or greater points.
Yeah. Maybe just a quick reminder for our listeners, what is the trigger that allows a patient to get retreatment in the open-label extension portion of the study? Where, of course, the patients on drug are still blinded.
Right. What we do once folks enter, right through the double-blind period, we're doing obviously, regular MADRS scores. We not only have the primary outcome, but we get MADRS scores throughout people's participation in that first phase. We basically keep doing that into the extension phase. We alternate PHQ-9s, which are a PRO, so a patient-reported outcome, which are done biweekly. There's central raters. We always do our MADRS and HAM-As with central raters to preserve the blind of the rater. Those central rater-assessed MADRSs are done scheduled monthly, or they can be triggered by a PHQ-9 score of more than or greater than or equal to 10. 10 or higher will trigger a MADRS. A MADRS score of 20 or higher will enable someone to get an open-label treatment. We picked 20 because you got to pick something.
Obviously, the experience of having depression and anxiety is a continuous one for people. It's not really a categorical sort of an experience. When we look at the categories on these scales once he defines the threshold between mild and moderate. When we think about moderate and worse illness with GAD and MDD, we start to think of the accumulation of functional deficits, people who are struggling to complete their ADLs, who are struggling with work and family responsibilities. We obviously try to put a lot of thought into these study plans, and when we considered who is likely in the real world to be treated, we thought, well, is someone with mild illness necessarily going to be an early candidate for our drug? The answer is, probably not, right? People probably will have to be a little sicker than that. We thought that once people start to accumulate functional effects of their illness is a good threshold to treat, and we think will probably reflect more like early real-world practice.
I'll just add one brief comment on top of that, which is that while there's, I think in recent years, been more and more attention paid on the number of past drugs that have failed patients as a sort of proxy for severity. The actual thing, when we look at HEOR data and when we really unpack what drives value in helping these patients, it's improvements in severity and function, right? It doesn't matter whether you've tried and failed one, zero, or many past antidepressants or anxiolytics. If you have severe disease and severe impairment, that has a severe consequence and shows up as a healthcare burden and a burden on the patient's life. The ability to actually intervene there not only drives value for patient, but also allows us to demonstrate the value for everyone involved in facilitating the adoption of these drugs.
Yeah. Okay. I want to sort of go back to the 5-point MADRS change, which this study is 80% powered to show, and tie it with the phase II study in GAD patients that you did, where you also measured the MADRS change, which showed up to a -6.4-point change at week 12. Maybe for our listeners, help translate why or how does that phase II study give you confidence in the repeatability of that type of data? Of course, you've given yourself some wiggle room for the MADRS change to sort of deteriorate, which we see generally in trials in larger datasets. Even though those patients were GAD patients, why is that a study that should give investors confidence in the phase III study?
Yeah. I think it really starts with understanding the diagnoses and the outcome measures we use to measure severity of these diagnoses. I don't know how many of the listeners have spent the time to go dig into the Hamilton Anxiety Scale and the MADRS. If you go dig in and you were to draw lines between the two, and the criteria on each scale, what you'd find is that virtually every item on the MADRS is captured on the Hamilton Anxiety Scale, and that the MADRS is a largely psychological scale. It's a mood disorder. We always put it into the reality that on the HAM-A, item six is depressed mood, which is effectively a single item being captured that is sort of expanded into 10 criteria on the MADRS.
The degree of overlap between both the diagnoses and the scales we use to measure severity are just so substantial that there's a large degree of read-across, right? That seeing an effect on one, there certainly could be a scenario in which only having physical symptom improvements on the HAM-A, it's not what we saw in phase II, but in that way, what we really see is that by changing the HAM-A and having a similar trajectory and magnitude of response on the MADRS, it just builds confidence that as we move fully into that population, we would expect to see some effect. The other reality is that because these patients in phase II were not in a depressive episode, we had a lower starting point.
There's more variability across the arms in phase II, but the average baseline MADRS was right at, or in some cases, just below the threshold for what it's actually required to be in our depression study, the MADRS of 26 or greater. By having a higher starting point in an MDD population where patients are in a major depressive episode, it gives us more room to separate as well. There's more room for improvement. There's more of an ability. Really what we saw in the phase II data was that on the MADRS, we virtually bottomed out the scale. We moved patients on average to a part of the MADRS that no longer has sensitivity to detect differences because everyone sort of lost that segregating power between response and not. When we unpack all of that and thought about going into depression, there's certainly historical data that would point us in the direction of there being an antidepressant effect, and what we saw in phase II just gave us even further confidence to move in that direction.
Okay. Maybe just remind us in the phase II study, how many of the GAD patients also were diagnosed as MDD patients?
Yeah. Dan, you want to cover that?
Yeah. Just over half, 60%. The critical thing to keep in mind when we think about GAD and MDD is that they do have, as Rob said, a tremendous amount of construct overlap in the disease definitions, and obviously, then therefore, the scales used to measure them have to overlap.
They're distinguishable patient populations in these two studies. While in our GAD studies, a history of having a major depressive episode, which is the disease definition for Major Depressive Disorder, is allowed, patients cannot be in a major depressive episode. For the MDD studies, patients can certainly have a history of anxiety. They can have even concurrent GAD, but they must be in a major depressive episode. When you sort of acknowledge that there is this definitional overlap, what we're effectively saying is that when a patient presents with anxiety and depression, which they all have some flavor of, whether or not they are currently in a major depressive episode is the distinguishing factor. We think that, again, when we think about, if approved, translation into the real world, we think that's the sort of profile that appeals to providers because patients show up when they show up. To have something to be able to offer them, regardless of the phase of their disease is, we think, quite valuable.
Yeah. Okay. Maybe just if you could move to the GAD phase III studies. Could you provide an overview of the trial designs for the two trials there? You've talked about an interim sample size re-estimation for one of them, and we are waiting for the analyst event next week to get an update on the second one. Maybe perhaps sort of start us off there, and then I'll follow up with something more.
Yeah. Dan alluded to the MDD study design, and it's virtually identical in GAD. A single dose of drug, patients followed for 12 weeks, and then a 9-month extension period where patients can receive up to four open-label doses of drug. Of course, patients can't be in a depressive episode. They must have generalized anxiety disorder, and they must have a Hamilton Anxiety score of 20 or greater at baseline to be enrolled in the study. As you mentioned, in both the GAD studies, one of the features in phase II, we didn't include this extension period. Patients didn't have any real incentive other than the good of science, I guess, to stay in the study beyond receipt of their first dose. Unsurprisingly, particularly those who didn't respond, particularly those who were in the placebo group, left the study, especially after four weeks after the primary endpoint.
Three months later, a patient who has severe anxiety who's been taken off of background medications and just asked to stay without those medications, it's understandable that those folks would opt to return back to their standard of care. We had a not unreasonable, but a higher dropout rate than we'd like to see in the phase III program. We had designed the phase III and powered it based on some assumptions that we'd have a slight improvement there and have a variance in the data that's somewhat consistent with historical studies. What we ended up seeing for VOYAGE in the sample re-estimation is that on both of those nuisance parameters, they were better than we anticipated, which means the confidence intervals would, all else equals, be smaller and you'd need less patients to be able to detect and have the same degree of power.
In that study, we didn't need to increase the sample size, and so we are excited to get to the readout with a target N of 200 there. In PANORAMA, we likewise had the same analysis. As we've said, at the event we have next week, we'll be sharing some updates in terms of where we are with that study and the outcome of that sample re-estimation.
Thanks. Recently, and I believe that was at your earnings update, for the VOYAGE study, you provided us the update that you did the re-estimation, and you didn't need to increase the sample size there. In terms of powering, now with the number of patients that you are going to have because the dropout rate is panning out to be lower than what you had even modeled. What is the study now powered to show? In other words, what is the minimum change you need to see in the study in HAM-D for it to be statistically significant?
Yeah. I mean, for VOYAGE, which is the only one where we've done this, and there's of course a number of assumptions and forward-looking statements in any of this.
Of course.
Of course, right? Which is that you have to assume that the nuisance parameters at the interim analysis are going to be the same at the end of the study in order to rely on them.
Yes.
Those things can always change because we have more patients who've been enrolled.
Yes.
If we were to rely on what we saw, we would only need to see a little bit over a 2-point difference on the Hamilton Anxiety Scale. As a reminder, we saw 7.7 unit delta between 100 micrograms in placebo in the phase II study. If those nuisance parameters that we shared were to hold in the final analysis, it'd give us in excess of 99% power for a 5-point difference. We feel quite confident that, going into the study, we've designed and executed on the study that we set out to do, and we should get a very clear answer, and we're quite optimistic.
Yeah. I mean, I think from everything that you disclosed and talked about, from a statistical powering perspective, it seems that you've done everything that one could do to position the trial for success, for it to be a positive study. With all the caveats of we don't know what the second half of the study would look like and all of that, which I'm sure the biostatisticians can articulate much better than I can. Outside of that, right, I guess the only other question then remains is that even if the study is positive, we do want to see the magnitude of change that is meaningful and also differentiated from whatever else is there on the market today. Maybe if you could sort of comment on what investors could think about, to kind of assess that portion of it.
I think when we start talking about some of the bars and what good, bad, and great looks like in these studies, of course, the minimum bar is we need a positive study. When you have a positive study, that means almost certainly there are some patients who are getting a substantial benefit from the drug. We think particularly in these indications where there's tens of millions of patients and where there have been very few, the last approved GAD drug was Cymbalta in 2007. In a world where you haven't had new drugs in almost 20 years, and you have a growing prevalence and a huge burden, new drugs, we think, have a clear path to adoption.
Now, given what we've seen so far and what we aspire to achieve, both in development and out in the real world, a 2.5-point delta that's statistically significant wouldn't be all that exciting, right? It just wouldn't in the context of what we think we can do. Where we end up getting to, particularly in GAD, where there's far fewer drugs, is that a four-point or better delta would be the best outcome one could expect from any drug that is approved today, right? If you did a study with any of the approved anxiolytics, you wouldn't expect to do better than four. If we're able to exceed that bar, we feel like we could position as perhaps one of the best drugs to have been developed for GAD.
A similar sort of logic holds with MDD, where we see 3.5 points be the sort of delta you get for most approved therapies, even some of the newer drugs that are getting quite a bit of adoption today. If we're in excess of that, of course, we powered the study for five. If we're in excess of four, that would be good. If we're in excess of five and we start getting up higher than that or anywhere close to what we've seen in the past, that would be unprecedented numbers that we just haven't seen in psychiatry and certainly in depression. We've seen a lot of studies fail over the years in depression. Again, we have high hopes, not only for the drug but for what it can achieve for these patients, and we're hopeful that the data will back that up.
Yeah. An important feature of the drug is really dependent upon the durability, or at least sort of it's an important component of the clinical profile of the drug. Of course, this year, we're going to have three very important data readouts of the three studies. I think that the elucidation of kind of what the durability of the treatment can look like will probably come with more mature data over time. While we are waiting for that data, is there something that we can look to from either the readouts of the phase III studies or going back to literature that can help us sort of think about what could be the frequency or the number of treatments in a year that one can expect with an LSD-based drug?
We've seen some academic papers where there's a multi-year longer effect from a couple of doses of drug. A lot of the academic studies historically have used a sort of two dose in a month or so regimen. Now, we've seen our phase II, a single dose give us at least 12 weeks of durability in GAD and on the MADRS symptoms. I think you're exactly right. It's what we're trying to characterize in the phase III program to really have a definitive answer around what that's going to look like. As we aggregate data, of course, we don't want to be putting out data that would be not representative, either too good or not good enough, right? That would be sort of a mischaracterization of what the real full data set could enable. We expect to be able to share some early findings when we get to the top-line data from the phase III studies later this year.
Okay, that's great. This is maybe a little bit of a tangential question, but help us understand maybe whether it's anecdotal or in the work that you guys have done around the experience of patients that have a prior history of using LSD. Has that been studied during some of the work that you have done or in literature that we can look to to understand, over time, how do patients that have a prior experience, how do they respond to treatment?
Yeah. It's something that we capture in our data in our clinical trials, and generally, what we look to is the data in our study generally representative of the population. There's only between, like, 15%-25% of U.S. adults that have used psychedelics at one point. Of course, that's grown substantially in the last decade or so as some of this research has emerged. We capture that. I think there have been attempts and certainly in public settings, discussions about prior use of drugs as some sort of variable that could be considered. We don't really see the logic of how that would systematically bias anything or sort of change expectations. I mean, one could construct arguments for why it would make the drug look better or worse again.
In the absence of knowing that there's a singular direction that something could bias the data, really what you have is just probably more noise. It's not something that we think is a sort of substantial thing, but it's something, like everything, any of these variables, particularly ones that have been sort of subject to public discussion, we pay attention to, and we capture the data, and we really try to make sure we understand what's going on.
Yeah, I don't have a viewpoint on it one way or the other as to whether it would make a patient respond better or worse. It's just sort of hard to know. I guess, are you controlling for that in order to maintain balance between the drug and the placebo arm, or do you think there's just not that much representation to matter?
No. We don't balance randomization or anything on that. It's just simply a matter of there are more important variables that need to be balanced than something that has a relatively low incidence and that has an uncertain impact on anything. We just think that as larger randomized studies go, there should be a reasonable balance between the arms by the nature of randomization, right? That's just sort of how it works.
Fair. Maybe if we could move to some of the commercial side of the things. Can you talk a little bit about the overlap between the GAD and the MDD patient population? I think people sort of understand that there is a reasonable amount of overlap. But tell us about sort of what's the growth profile of the two patient populations, and how are the GAD patients being treated today, and does having a drug that potentially will have both the labels allow you to tap into the GAD market deeper than what you probably would've if it was just MDD? Maybe just if you could understand that better.
Yeah. I'll say high level, and I'll turn it over to Dan. One of the things that's so striking is that GAD diagnoses and treatments have been growing at a much faster rate, in large part because it's a disorder that was so severely overlooked for a long time. As the attention, the sort of spotlight focused on MDD and SRIs for 20-30 years up to this point, both screening and the availability of drugs was much more heavily skewed towards depression.
In psychiatry, the reality is, if a patient shows up and they need a drug that could help with anxiety or depressive symptoms, and all of the drugs are labeled for depression, it doesn't make a whole lot of sense to not give them a depression label and give them an anxiety label, because it's just going to create reimbursement challenges and difficulties getting access to care. What we've seen is sort of this tip of the iceberg phenomenon, where we still see a high incidence, high prevalence of GAD, but no one cared to look for about 20 years. The estimates we had from 20 years ago that GAD had about a 3% U.S. adult prevalence. Newest research we have, both from claims data and from high-quality epidemiological studies, is that rate is in excess of 10%.
That growth, even over the last six or seven years, that growth has just been staggering. There's no shortage of events out in the world that can call people's attention to the anxiety that probably was already there in the first place, but that is becoming more pronounced and sort of exacerbated. Anxiety's been the sort of sleeper indication, perhaps, for quite a while now. Having both a label, of course, as Dan was describing before, means that regardless of when a patient shows up, whether they're in a depressive episode or not, we'd hope to have an option for them. Dan, I don't know if you want to add any kind of color.
Probably the only thing I'd add is that the existing chronic treatments, mainly SRIs, are not very effective for anxiety cluster symptoms. Often, even if people's anhedonic symptoms are treated with an SRI, and it works, they can be left with a really high anxiety burden. Benzodiazepines, obviously, are useful for acute anxiety, but increasingly, people are very cautious about using them for chronic anxiety. In a world where, as Rob said, it's been almost 20 years since the last GAD approval because it's hard to treat, having something that treats GAD well would potentially really be a game changer for that patient population and give psychiatrists a tool that they really desperately want and need.
I want to next talk about the commercial model from the perspective of a clinic, and that seems to be really front and center for a lot of discussions that we're having with investors. It seems that the clinicians are trying to understand what that revenue model would look like. If you could sort of break it down for us, what are the components of that revenue model from a clinic's perspective? What is clear, and what is not yet clear about each of those components in terms of how the differentiation between products might impact their motivation to select one drug over the other?
Yeah. Dan, who has built out networks of clinics in the past, maybe I'll turn it to him to comment on that one.
Yeah. I think something that is often lost in the conversation about clinical models and the clinical reimbursement is how much that is in place for most of the activities related to a novel drug, even a novel sort of treatment modality, which is that evaluation and management, right? When a provider who can prescribe, an advanced practice nurse or a physician, sees a patient, they bill an E&M code. The complexity of that evaluation and management decision allows for higher billing. Certainly, coming to the decision to use a more complicated medicine allows for some increased billing in the E&M department. There's a session delivery, right, with our drug. There certainly will be a monitored session. That session gets to be billed under time-based codes that already exist in psychiatry for things like psychotherapy, which are time-based.
There's already the ability, and folks have already negotiated reimbursement when they're in network to get paid for initial hours of providing a service and then add on hours of providing a service. Obviously a service that might be split over 6-8 weeks in the psychotherapy world being delivered in a single day will be a different set of codes that will have to be negotiated. We'll obviously front-run that with payers to be sure that this isn't catching anybody by surprise. If you can generate many weeks of efficacy out of a 6-8-hour session in one day versus in multiple weeks, that actually seems to come out in the wash pretty well.
Then for clinics that opt to, there's the buy and bill model where effectively the clinic upfront buys drug and then is able to dispense it based on prescriptions that they're writing and there's a margin there for the clinic. For clinics that are doing this a lot, they get this additional benefit of doing the buy and bill margin there. The way we see it, particularly for session-based monitoring, is that in medicine in general, but certainly in psychiatry, you get paid for work you do for the patient, with the patient. Work that's done without the patient is generally uncompensated. Consolidating, and this all speaks to patient burden as well, of course, because as a patient, you're not getting benefit from the time you spend driving to and from the clinic, time you spend in the waiting room, right?
All of those things are time without benefit. We see our consolidating the treatment into a treatment day as actually being maximally efficient for patients and providers. That there's less in-between patient time spent turning over rooms and things like this because it's basically a patient per room per day. For patients, that's less trips to the clinic, and for providers, that's less uncompensated work. Fewer trips to the clinic.
Yes, that makes sense. That certainly makes a lot of sense. Is there at this point clarity with regards to what the FDA would require in terms of the qualification of the person who's monitoring a patient? And whether there's going to be need for a second person that monitors the patient, and whether that can be somebody who does that over maybe a central monitor as opposed to also being in the room?
Yeah.
We've had a lot of discussions with FDA and certainly have seen an evolution of thinking and data swaying that conversation, right? At the end of the day, requirements that get put in place via REMS stand as barriers to access, make it difficult for people, particularly in less urban settings, to actually get access to drugs. We have a vested interest and a deep desire to do this responsibly and well and safely, but not at the cost of overburdening patients and providers and not letting the folks who could benefit from this actually get access to it. We recognize there's a delicate balance there. What we've done through development and the design and execution of these studies is take, at every turn, a sort of high-grained, very detailed look at even things like when and how and if and why there's an intervention needed.
What is the session monitor doing in the room? If they're just there to be there and not actually serving any purpose, why are we requiring it in the first place? We can have certainly philosophical arguments and sort of policy arguments, but really what should win the day are data. If we can show that, in fact, no one really needed to do much, that makes a pretty compelling case, especially when tied to the policy and access argument. Until any sponsor is in late review cycle and having label and REMS discussions, it's of course not going to be known with certainty. This is going to be true for any drug. We are optimistic, and we've been taking a very sort of data-driven approach to every argument we've made and everything we've done in development. We're optimistic that as we get to those later stages, we'll have a rational thought partner to help us navigate that balance between safety and access.
Okay. I know we are coming up on time, but if I could squeeze in one last question. I wanted to maybe see if you could give us any more color around the white glove type service that you sort of talked about, at least at a high level. Maybe if you can provide some color on what does that really mean in terms of how it makes things easier either for the patients or the clinics.
I think the easiest to point to is how we operationalize our trials and development as the sort of proof point that we have that we're really proud to have excelled at, we think, over the last several years. Which is that when PIs sign up to work on one of our programs, we try to give them the best support they've ever had from a sponsor and the best support to get set up, to operationalize things, to navigate any sort of complexities that could come up. We of course have a ton of institutional knowledge gained from doing that as well, and can then share that with other providers or research sites. That same model is going to transition over into a commercial setting if we get there, right?
Which is as prescribers, as the providers are saying, "I want to be involved in the delivery of care here," are raising their hands, we intend to maximize, of course, compliantly, the degree of support to make that easy for them, to make it such that a desire to be involved can turn into actually being involved in the provision of care. The same is true for patients, making sure that they have what they need and that the care team that they have and that they will work with throughout a treatment cycle is going to support them and enable them to have probably some of the best psychiatric care that they've ever had.
There's often this overlook, despite best intentions and despite thousands of psychiatrists and mental health care practitioners really giving it their all, a lot of what patient care looks like today is being shuffled in and out of the door and rushed through treatments and cycled through medications. That's not something that anyone particularly enjoys. The concept of being able to facilitate the process of care whereby providers have an easy way of getting set up, patients have a high level of support, and we can facilitate them actually getting access to the drug. It benefits everybody. Everyone comes away with a net positive if we do that well, and we intend to do that better than anybody.
Okay. All right. Well, looks like we are almost out of time. I should thank you both for taking the time for having this conversation with me. Thanks to all our listeners as well.
Thanks so much.
Pleasure.