Okay. We'll get started with our next session. I'm Andrew Tsai, Senior Biotech Analyst at Jefferies. Thanks for joining me today, and it's my pleasure to have the Definium team. To my direct right, Brandi Roberts, CFO, to her right, Rob Barrow, CEO, and to his right, Dan Karlin, CMO. Welcome, all of you.
Thanks for having us.
Thanks, Andrew.
To help level set things, can you please give a brief introduction about Definium, what you're trying to achieve? Milestones would be helpful as well.
Yeah.
Thanks, Dan.
No problem.
Two assets at the clinic. The furthest along and the one that's garnering a ton of focus right now given upcoming readouts is our DT120 program. It's a form of LSD that we've been developing. It's an ODT formulation that we've been developing for generalized anxiety disorder, major depressive disorder. We have a breakthrough therapy designation for the GAD program and are rapidly approaching three pivotal readouts, two in GAD, which will be in the third quarter, and our first readout coming later this month, which we're anticipating for the Emerge study, our first pivotal readout in MDD. We saw in phase II data a dramatic, rapid, durable 12-week response after a single dose of drug that was represented by one of the largest absolute changes and placebo-adjusted changes, not only in our field, but in all of psychiatry that we've ever seen.
To have that kind of response profile that put about half of patients into remission after a single dose of DT120 12 weeks later is just a real shift in how we can think about treatment if that ultimately plays out in our pivotal data. Tons of excitement to have these three readouts come together so closely makes for an exciting summer.
Okay, fantastic. Then bigger picture, SPRAVATO is in the marketplace. $2 billion run rate, maybe you're doing $5 over time at least. It's interesting because, I don't know, you could technically be approved fairly soon in GAD, MDD shortly thereafter. All while Compass is potentially getting approved first in TRD depression. How are you thinking about leveraging the SPRAVATO network infrastructure and when Compass comes, how do you think about leveraging that, if at all?
Yeah, I think the thing when we look out at the world and the number of treatment centers, and these are just psychiatry offices that have registered and are able to deliver esketamine. The ability and the desire for psychiatry to adapt to new treatments cannot be overstated. For a drug that requires up to 56 doses a year in SPRAVATO, every one of those treatments being a two-hour session in the clinic is an incredibly high burden for patients and for these sites. For that to have now gotten the traction is just symbolic about the level of need that's out there and the desire for change in the treatment landscape. What we've seen is vastly different from, again, other drugs in the category, certainly from the durability of SPRAVATO.
To see a multi-month dramatic response and efficacy after a single dose of drug is just something that really stands out in terms of patient benefit if it ultimately pans out. That said, the centers that are out there in the world delivering esketamine today or delivering ketamine historically and still today, certainly are places that when we do research with these groups, have the highest expressed interest of anyone in adopting our and other treatments like it. They're the ones that we know where they are, who they are, what their incentives are, what the economics need to look like. It's a very logical starting point when we think about getting out into a commercial landscape that we would want to go there and engage there first.
By no means, I think this is part of the sort of perhaps lack of understanding of some of these dynamics sometimes, is that that is not an upper bound. It's the starting point. It's sort of the launch pad that allows us to go out and market. It's going to take time to build that market. It's going to take time, of course, for all of us to really get there and establish this and the scale we think is possible. To have a lot of the sort of administrative and logistical aspects that are going to need to be in place for a drug like DT120 now operationalized in a world at a much larger scale than it was a couple of years ago, is just really exciting to us and gives us a sort of springboard to think about in the future.
Great. Like you said earlier, MDD data first in late Q2. Because you're here, I infer it's end of the month rather than next week kind of thing.
It would be a strange thing if our data were coming right now.
Yes.
Yeah, we're happy to be here at the Jefferies conference, and after this, we'll be laser focused on getting data out the door.
Yep. During the R&D day, you shared a set of what to expect in the top line release. Thank you for sharing that. Maybe we can go through one by one, but one of them is that you'll share the entire 12-week blinded phase, even though the primary endpoint is week six. Is it your expectation . You powered the study, by the way, 80% powered it to show a five-point delta. That you'll see durability from week six to week 12 full stop on an absolute drug basis?
Yeah, I think the trajectory of the drug response is something that is very important and very interesting. What we saw in phase II, as I said, was durability with really no degradation of that effect 12 weeks after a dose of drug. We would love to see that again. The dynamics, if that were not to happen, the dynamics of why, of course, would be important. We feel, based on the data we have in hand today, very confident in the study design and the ability to demonstrate a durable effect. That has translated into a program where we're really seeking to prove more of a durable effect than, I think, anyone in our field. We believe a single dose of this drug has the horsepower to drive that rapid and durable effect, and that's why we've designed the studies we have.
That's why we've set the endpoints the way we have. In GAD, which is coming in Q3, in both studies, primary endpoints out at 12 weeks. We're really pushing the bounds even of what's feasible in these populations, given it's a monotherapy trial with no other background therapies.
Yep. In MDD, well, TRD Compass, when they reported the phase III data, the efficacy on an absolute basis looked way different compared to the phase II. Is there a risk from this happening, is the root of my question for your program in MDD. Why should we feel confident it'll look similar to your own phase IIb subgroup analysis?
Obviously, we'll have the data very soon. We'll be able to just look at it and see. We feel incredibly confident, not only based on the phase II data, based on the history of research in this field and with LSD in clinical settings. There are always study-to-study variations, and that can show up in different ways. We designed our phase II study to be probably the highest bar we're putting the drug up against. It's a five-arm study. It's the only real comprehensive dose response study that has been done in this entire field, ever, to our knowledge. No one's actually looked in a population with a disorder and seen the clinical dose response on validated endpoints. Some people have looked at pharmacodynamic endpoints and how much people feel the effects of the drug.
That's a very different thing than showing the actual dose response on efficacy. That study design, the way we conducted that study, really translated into a pretty outsized placebo effect, something that really exceeded our expectations. We still beat that by a margin that was more than double the separation that most drugs are able to show. That gives us a lot of confidence going into study design that is a bit simpler, that I think could reasonably be expected to deliver a placebo response that's a little bit lower. Of course, drug effect tends to be additive on top of placebo effect.
Numbers can change between studies again, but everything about the nature of the study designs and the way we've progressed through this program gives us a high degree of confidence going into the phase III readouts that we should, if we see a clinically meaningful difference, be able to detect it.
Yeah. Sometimes investors are pretty focused on placebo-adjusted MADRS. For me, I kind of appreciate remission rates, and I think from at least phase II-B in GAD, you did show pretty high remission rates. In MDD, I'm just curious, what would a five-point placebo-adjusted change, in your view, translate to in terms of a remission rate in MDD? Yeah.
Yeah. That can certainly vary. Means, of course, hide a lot of valuable information about response patterns and about the consistency of that response. You can end up with means that not many people are close to if you have a bimodal distribution in any given study. We are also very interested in the sort of absolute magnitude of change and achieving certain thresholds. Now, remission in MDD is, for recent studies, defined as a MADRS of 12 or better in most studies. The interesting thing is that as we think about the treatment modalities that we're studying in DT120 out in the world, the way we've designed our phase III studies is to treat patients upon the return of moderate or worse symptoms. That corresponds to a MADRS of 20 or greater.
The reason for that is that functional impairment and the highest burden of depression come in when MADRS scores are at or above that level.
Yeah, there's a little bit of a gap between 12 and 20. We're going to be, of course, wanting to see remission rates at sort of both of those thresholds, or achievement rates at both of those thresholds. If we can get as many people into 20 or better as possible, that's a great outcome. It's improving functional activity and the ability to function in these patients. Remission, getting people better is always the goal. The more we see there, also an endpoint we're really interested in.
Great. Another piece of data you'll consider sharing is the open label extension portion to kind of help us elucidate time to retreatment, the durability aspect, basically. The question is, how much follow-up, in terms of months or weeks, can we expect for the average patient?
Sure. Yeah, for every patient, we can kind of expect the same thing, so long as they stay in the study. What we've done here is design what we call the extension phase of the study. You said an open label extension, and there is an open label component, but these aren't conventional open label extensions because we don't treat someone automatically when they enter that extension phase. We've got this 12-week primary period where there's no open label treatment opportunities. That's the primary observation period, as you noted. At week six , we'll check the primary, and then at week 12, we'll look again. We've got some earlier looks as well
Once people enter that open label period, or the extension period, if they have moderate, 20 or worse symptoms, then they become eligible for an open label treatment. The advantage of a paradigm like that, unlike a conventional open label extension, is that we get to watch people in their initial allocation. Right, people are randomized, they're double blind, they're controlled through that first 12 weeks. If someone doesn't get an additional open label treatment, they remain blinded and controlled for that full year of observation. That gives us a tremendous amount of durability information about that first treatment. There are two big other things we get out of that extension phase. One is it keeps people in the primary phase. Right?
Ordinarily, in a study like this where we're treating someone once and watching them for 12 weeks, if they think they got placebo in that first part, or they think they got drug and they don't feel better, there's no real reason for them to stay in other than the good of science and doing something they agreed to, but it might not be what's best for them. By having treatment available to them as they get through that initial part, it gives people real motivation to stay in, and we've disclosed the results of sample size re-estimations we did. No, not in the MDD study, but in the very similarly designed GAD studies and showed that we took dropout rates from phase II, where we saw a 25% dropout in the 12-week period, down to 10% and 6% in the two GAD studies where we did that analysis.
That function of the part B seems to really be working. Same basic patient population, markedly better retention rate. The last thing we get out of the part B or the extension phase is that we have people proceeding in their controlled blinded state. Once someone gets an open label treatment, now they enter this new condition. Right? Now they're in really an open label study, and we design that open label study to mimic as much as we could real world treatment. This threshold that Rob mentioned and I talked about already, moderate or worse people with functional impairment from their illness are eligible for up to four treatments during that extension phase. We think that's pretty similar to the maximum that people would get in the real world, and probably exceeds it by a little bit.
We actually set that number for 40 weeks to be a little higher than we thought most anybody would get, in part so we would have a real curve describing how many treatments people get on average.
We'll be able to see the emergent treatment patterns when people can be treated ad lib when they have moderate symptoms or worse. We're really excited about the data that will come out of the part B across the studies. Obviously, it's a little less easily numerically summarized.
Yeah.
When we talk about a primary outcome measure, placebo-adjusted change, we're talking about two numbers that we can compare. We can generate a P value. It's a pretty accessible and assessable outcome. That successful then gives us the opportunity to start to look in much greater depth at the richness of data we'll get out of the part Bs that will absolutely be used to inform real world prescribing.
The net of that said, is it fair to assume directionally, though, if you look at the MADRS reduction overall, it just continues to go low, you see more benefits over time because placebo patients now get drug, patients who didn't respond now get drug? The net of it, I'm assuming you get a stronger benefit over time.
You retain the ability to have placebo-controlled comparisons, so long as there are people who remain in the study in the placebo group. Those people are obviously going to be the ones who are doing better because they have to be better than moderate. The comparative value as far as descriptive statistics goes down realistically. Our ability to make durability claims.
at least to scientific satisfaction, remains pretty powerful because what we do is not do that comparison independent of who remains in the study, but we start to use the number in the group as a component in the statistical analysis. It's a pretty classic survivorship analysis.
Right
that we would use.
Okay. With the phase II-B GAD dataset, again, the curves are very durable. You can kind of dream the dream about the redosing profile. What is your guys' base case kind of conservative view on the average doses per patient per year? I'm just curious.
Yeah. At this point, we don't like to speak beyond the evidence that we've generated to date. At this point, we've seen in a study that we were able to show statistical separation 12 weeks later. There was no degradation of that effect. There was no narrowing of that separation, it would be extrapolating to try to make anything beyond that. I think one can sort of assume that even if on week 12 +1 , everyone lost the effect magically somehow, I don't know how that would happen.
We would then kind of assume, on average, patients would need four doses a year. Very feasible, very practical. Particularly when we look at the burden compared to what now is, as you said, a multibillion-dollar drug that requires, in comparison to that, almost 4x as much time spent in clinic getting treatments. We're certainly going to be very interested, and we'll look forward to updating with greater clarity once we have the data in hand and once we're able to say reliably what we're seeing over a period of time.
Yeah.
I think reasonable to expect something could be a little bit lower than that.
Yeah
in terms of an average number of uses a year. Again, until we see the data, we won't know the answer.
Sure. As part of the top line, can we expect you to analyze the MDD patients with comorbid GAD, and you have the HAM-A data in that subgroup of patients in that top line release?
As we analyze all the data, absolutely, we'll have those, and at some point, we'll share them. I think we don't like to conflate outcomes here either when we're talking about a population like MDD. These are patients who are coming, of course, with the primary concern of depression, who have been diagnosed with a depressive episode currently, and that have elevated MADRS scores. Our focus for the top line readout is absolutely going to be on the depression scores more than anything, and functioning, the things that tie into that. We certainly will very soon thereafter have additional data in GAD. That's going to be the most representative.
That's true.
of what we see in GAD.
It's just a week or two later or something. I know. Yeah. Anyway.
Quite a bit.
Yeah. In any case, back to MDD then. When I think depression versus GAD, my gut, for whatever reason, suggests maybe that people are more prone to suicidality. My question is, do you expect no imbalances in suicidal ideation?
Yeah, this is worth dissect out. I'll turn it over to Dan in a second. I think it is interesting that we think of suicidality being associated with depression in part because it's baked into the disease definition, right. Using the severity scale, item 10 on the MADRS is suicidality characteristics. It's part of defining that disorder. An interesting piece, and exactly what you're saying, I think for a long time, the risk of suicide and suicidality in GAD has been underappreciated. Underappreciated is not the right word. It's been overlooked. Our team presented a paper, I think it was last year at this point, showing that actually the rates are quite similar across GAD and MDD.
There could be some selection bias in terms of when people show up, and when someone shows up for clinical care, and how that's associated with MDD and all these sorts of things. I think in psychiatry generally, suicidality is something that is always looked at closely, appropriately, and is going to be present. I don't know if, yeah, you can talk a little bit about that.
Maybe beyond the disease states, because for 30 + years now, we've just had this sense that MDD is the serious diagnosis, and that's in part because of SRI marketing and what SRIs were able to help with. As little as they were able to help, were able to help with more than GAD, so that suicidality as a serious outcome became more associated with the serious diagnosis. As Rob said, everybody in severe neurotic distress, whether it's the product of a major depressive episode or the product of chronic anxiety, develops at some point, if they're sick enough, some degree of suicidality. The important point for our drug, and I'd say for the whole category of drugs, is that there is not evidence that these drugs induce suicidality or worsen suicidality.
When you deal with psychiatrically ill populations, of course, we try to rule out people who are in danger because people who are in danger from suicide shouldn't be in research. They should be in care, and not in the kind of care where you're randomized into placebo sometimes. That's really important just from a patient protection standpoint. Inevitably in these populations, some suicidality, usually some sort of passive ideation, "I wish I hadn't woken up this morning" kind of level, might crop up. We have not seen in our data or anybody's data evidence that that's a product of drug taking.
Right. No drug-induced suicidal ideation is what we should expect. Yeah. Okay. You also share whether patients are discharged at the five-hour mark or closer to the eight-hour mark. Can you explain how that fundamentally works as protocol to dictate the timing?
From the end of phase II on, we worked to develop an end of session checklist, and that checklist has been incorporated into all of our phase III protocols, and in fact, has now back been incorporated into ongoing safety work phase I protocols. We've got a ton of experience using this very simple end of session checklist that doesn't actually require clinical training to be able to apply. In the phase III studies, we, at hour five, begin using the checklist, and that has allowed us to, despite the eight-hour minimum monitoring period that's protocol specified, which you have to do in research, right? You need a protocol-specified minimum that doesn't automatically differentiate what allocation somebody got at the outset. If we said, "Okay, you can leave at four hours if you're ready," then about half of people would probably be ready at four hours.
That's sort of a source of functional blinding that we wouldn't want and that we can prevent. What we will have for the top line is hour by hour counts of percentages of how many people were ready to leave at that hour based on that checklist that we've agreed on with FDA. In one group, we would hope that it's pretty close to everybody ready to leave at five hours. After we're unblinded, it's okay to look at that.
Yeah. Okay. Very good. Patient experience, a company like Compass will say there's preparation and then integration afterwards. Are you doing anything like that? Can you kind of describe the patient journey for LSD?
From the outset of our efficacy research, we have done everything in our power to isolate drug-only effect, to make sure that wherever we could, we removed the impact of any sort of therapeutic or dyadic intervention. What we do in our studies is pretty much what you would do in any drug study, except of course, we're just giving drug for one dose at the outset, and that drug has a pretty remarkable, and at times dramatic set of adverse events that people need to be told to expect and be ready for, and what it might feel like. The informed consent process for this study involves education about that. People need to be ready for what they might feel, and we can reasonably tell them that they may or may not feel a number of things, and what might happen while they're feeling those.
That's the extent of what we do beforehand. Then we don't do any sort of therapy during the session. The service provided by the person in the room is essentially assistance and comfort. If somebody needs a glass of water, a walk to the bathroom, snack, they help with that. After the session, there's no sort of attempt to further change or modify or influence the course of their illness or their symptoms.
Got it. Staffing wise, let's just say this was approved. Can you remind us what SPRAVATO's staffing requirement is like? How would yours be different, if at all?
The SPRAVATO REMS requires that the prescriber, or a prescriber, be present at the site where it's being administered, and supervises the administration. That supervision can be in person, though the vast majority of prescribers are not doing direct monitoring of the sessions. In general, what that means is that staff members are designated to be the monitors for the session, and they do so under the auspices of the prescriber. That is a model that absolutely will be able to be applied to 120. Though we envision a world where other models are also applicable. SPRAVATO has a physiological monitoring requirement. ketamine has long been recognized to cause hemodynamic changes, and that causes there to be a requirement for blood pressure monitoring, pulse rate monitoring. We don't see evidence of that for 120.
Now we've used this drug in pretty high doses in quite a few people in safety studies and efficacy studies, and we don't believe we're going to have a physiological monitoring requirement. That opens up the possibility of the drug being administered with observation from non-medical providers. That's something that we absolutely will be looking for the opportunity to see how can we improve access for a wide swath of people who are in need by making sure that whatever observation and monitoring requirements are in place are appropriate to the actual risks of the drug.
Got it. In terms of reimbursement for the monitoring time, are you very confident you'll get reimbursed for five to eight hours? What is the function, or what is the protocol for that?
Yeah. There are always opportunities for new mechanisms to get reimbursement. Psychiatry and care provision in psychiatry is paid by unit of time, by the hour, by the 15 minute, however it gets sliced up. When you think comparatively, this is one of those sort of surface level assessments that maybe you got to take one step further for it to all come together. If you think of a payer who's paying for someone to receive an esketamine today, they're paying for two hours a session. You say, "Oh, eight hours or five hours or six hours or however many hours is longer than two," you would be correct in those numerical comparisons.
The reality is that over the course of a year, if a patient is receiving 56 doses of drug, which is the maximum frequency on the SPRAVATO label, it's 112 hours that are getting paid for. It's getting paid for today. If we're talking about reducing that by 70% +, four doses a year, even at the outside eight hours a session, suddenly we've reduced by over 70% the amount of time people have to spend in the clinic, that payers have to pay for. Why would that be a problem? Payers love that. Saving money tends to be a good thing for them and for everyone. When we think of the comparison and sort of the mechanism to do this, yes, it's compressed into a day, but that actually doesn't have a whole lot to do with much.
We see the mechanisms that exist, new mechanisms that are emerging, as certainly being various angles at getting to this. We've done a lot of work with payers, and when we sit down with them and talk about the dynamics, there is certainly that analog out there with SPRAVATO, and they look at it and give us the feedback that, "Yeah, we see it, we understand the dynamics, and we expect to think about it the same way.
Mm-hmm. MDD reads out two GAD readouts in Q3. How much time do you need before you file for GAD?
Yeah. Typical timelines for filing in psychiatric indications have, for the last several approvals, have looked like six to nine months from data in hand. The fastest I think we've ever seen was a little over four. When bars are set, we like to try to exceed them. We're going to go as fast as possible, and we've been doing a lot of work with an incredible regulatory team that we have and our whole organization to get ready for that, so that with data in hand, we're not wasting a single day.
Right.
We're not yet ready to say exactly what that date is, but we're not going to waste any time.
Yeah. Then MDD, you have a second study that just started. The data is 2027. You would file pretty shortly thereafter kind of thing.
Yeah. I think obviously filing strategy is going to be contingent on data quality and how compelling the evidence is and some dialogue with FDA. We think given the high degree of overlap, the adjacency of these indications, the fact that we're going to be looking at endpoints across several studies that are all aligned, similar study design. There are certainly opportunities to make sure that if patients in either of those states, in a highly anxious or highly depressed state, need treatment, and we have great data that could provide them an option, we shouldn't be delaying access there either. That'll, again, be contingent on the data, and with really strong data, it's going to, I think, give us the most excitement on all of these fronts about the future of this product.
Okay. Very good. Thank you so.