I think we'll get started. Welcome, welcome everyone to the Cantor Global Healthcare Conference for twenty twenty-four. I am Charles Duncan. I'm a senior biotech analyst, some would say the senior biotech analyst here at the firm, and I'm very excited to host this conference this year. We have a fantastic array of companies, many one-on-ones, many investors, and it's been a really interesting last couple of years, but last year in neuroscience for neuro innovator companies, and neuro is where I focus my attention, and so I'm really excited to introduce. The first presenting company is Alto Neuroscience.
Alto is a company that unfortunately, at this time, we do not cover, but it is a company that I've known since it was private, and I think that they have a very interesting approach to ushering in the future in neuroscience. So it's a pleasure to introduce Dr. Amit Etkin. Amit, thank you for joining us.
It's my pleasure. Thank you all for coming, and look forward to sharing a few slides here and then, and having a discussion afterwards.
For certain, we will be interested in your perspective on what Alto is doing, and, you know, frankly, I'm really excited about what you're doing because it's what has needed to be done for quite some time, and that is to bring the idea of precision medicine into the development of neuroscience, neurological drugs. And Amit has been at this for quite some time, and finally, he has a readout here very shortly that is going to frankly not only inform next steps for your lead candidate, but also possibly be transformative for neuroscience. So, Amit, why don't you tell us a little bit about Alto and your approach?
Perfect. Thank you. And really taking off on that idea, the potential for transformation in neuroscience is what keeps us really pushing hard, as you'll see, across multiple different fronts. The disclosure here on forward-looking statements, just to kind of make clear, that readout that's happening in October, data are blinded, still on track to report out in October, but data are blinded. We don't know the results as yet. Our view really is, as you said, Charles, understanding how to take a precision approach, who the right patient is for any given drug, has transformative potential in psychiatry, much like this approach has had in other areas of medicine. The question is how we get there, and we've been doing that diligently for a while now.
This is sort of a snapshot of what's going on. We've dosed over eight hundred patients in trials using our platform to identify the right patient, obviously screened many more using that platform. The patient populations we target are massive across very prevalent disorders. We have five Phase 2 trials going on, of which three are Phase 2bs, taking patient selection approaches with cash to go well beyond those readouts, and so that positions us pretty well to address this precision thesis. So what is that precision thesis? Right now, everything in the clinic is trial and error. I'm a psychiatrist. I can tell you firsthand, it is not fun. Patients don't like it. Outcomes are bad.
It doesn't actually tell anybody that we know what we're doing as a field, and yet we do know a lot, and one of the ways to start looking at the heterogeneity between patients, really heterogeneity in treatment response, is with the lens of biology, and so that's really where we anchor, and ways to measure brain activity that we could scale into the clinic and have been doing in our trials, to be able to get past what seem like small effect sizes because of a subpopulation responding, and use a biomarker to identify those stronger responders and design drugs for them. This is, of course, not a new concept in medicine. It's a new concept in psychiatry.
These are citations mapped for precision, either oncology or psychiatry in treatment over the years, and you can see the inflection points in oncology that have really been place markers in developing precision oncology, and now you would not be going for an all-comer oncolytic drug, you would be taking a precision approach. Yet, where we are as a field is roughly 15 years behind. We're kind of where that inflection curve was in 2009 in oncology, which, as we all know, has been very successful. Grounding in biology for us also means grounding in things that tell you about brain function for obvious reasons, but also are scalable. Tools like neurocognitive measures, which are web-based, self-administered tests, like we're using for ALTO-100 . EEG measures, we've actually been doing EEG in patients' homes already. That's being used for ALTO-300 to stratify.
And sleep and activity measures to tell us about circadian rhythms. When we measure these things, and we do it systematically in our trials, in part to build our platform, we also think about the commercial implementation as we do the R&D. So we're actually looking for the simplest signal to scale on, rather than trying to use everything and fit very complex models all at once. And where that puts us across our pipeline is, as I mentioned, five different Phase 2 trials going on, but also different disorders: depression, schizophrenia, bipolar disorder, PTSD. And even when we're looking at depression, we have three different drugs right now in depression; it's different subsets of depression patients. So really taking a targeted approach that can tile better options and standard of care and better options, frankly, than an all-comer approach across different subpopulations.
Top-line data, as was mentioned, coming in October this year for ALTO-100, which is what we'll talk about in more depth today, but two different antidepressants, ALTO-300 and ALTO-203, coming around the corner, first half of next year. So the bottom line for where we're aiming for here with ALTO-100 is a targeted approach focusing on what we call the neuroplasticity hypothesis in depression. Neuroplasticity is the ability of the brain to adapt and change with internal or external stimuli. It's thought that that's impaired in depression.
What we're doing is that logical next step of finding the patients for whom it is actually impaired, which are, as always, a subpopulation of the whole depression population, and targeting them with a drug identified based on its ability to enhance neuroplasticity, focusing in particular on the hippocampus, which is a brain structure that ties together cognition and mood. And so you can actually identify people with a plasticity impairment by identifying them as having an impairment in learning and memory, under which plasticity is really the core operation. The drug ALTO-100 was phenotypically identified based on a screen of hippocampal neuroplasticity. It works through well-identified pathways like BDNF or brain-derived neurotrophic factor, which is a growth factor that's been at the heart of everything we've known for over two decades on cognition, plasticity, and mood. But doesn't target that receptor directly.
It targets a different target that controls BDNF release. This makes it a potential first-in-class compound. That itself is really exciting. But what we see across preclinical data is that it does a lot on plasticity, both immediate release of BDNF, immediate effects on synaptic plasticity, and then over time, days to weeks, it leads to neurogenesis, the birth and differentiation of neurons in the brain, and even, as you see in the bottom right, an increase in hippocampal volume, which we know is reduced in these patients with depression and poor cognition. Our approach at clinical development, though, really is what matters here.
So here we have taken a Phase 2a study that we ran, divided it into a discovery subset, which is where we finalize the definition of our marker, here a measure of memory, and then prospectively tested it on an independent, locked, and blinded part of that data set to make sure we could replicate our ability to enrich response. Keep in mind, these patients with poor cognition, in general, have a worse outcome with treatment as usual, with standard of care. They tend to be more chronic, more disabled, and just generally lower functioning. And so what we're looking for is the opposite of that, is a drug that actually helps these people disproportionately because it targets core pathophysiology.
And what you see here is data from that prospective replication, 93 patients in whom you get a better response to ALTO-100 in those memory-impaired patients relative to those who don't in a single-arm design here, where you can see that the effect size between those with and without the memory marker is substantial. It's roughly double what the all-comer drug placebo difference tends to be in depression.
It's substantial relative to other neuropsychiatric drugs. I mean, that's in the strong category of effect size.
Exactly.
Yeah.
Exactly. And that I think is a really important point here, is that what we're looking to advance is drugs that will make a difference clinically. And so our line for success Phase 2b, is aiming to find the right people, advance the right drug, and not just advance any drug just 'cause we have one.
Not, not incremental, but impactful.
Not incremental. Exactly. And as you zoom out from those data, you can actually see that it's not just those data, it's a lot of data that support this thesis. Starting with the left, a reanalysis that we did on Phase 2 all-comer study, we found that poor cognition patients respond substantially better. Again, a larger effect size as you were saying, Charles, and a dose-response relationship. In the top is a discovery and prospective data sets that I just showed you.
Mm-hmm.
And the kind of bottom left is yet another prospective replication. Here, we did the entire trial in a decentralized manner, so all through telehealth, literally sending patients just a link to the cognitive battery to do at home. And you see again that poor memory patients respond better. And on the bottom right is making sure that it's not just a nonspecific signal. Here, we looked at two different placebo data sets where patients had also gotten the same memory test, so we can score the same biomarker, and placebo was not better in patients with poor memory. In fact, maybe slightly worse. So you put it all together, that's a lot of clinical data going into our Phase 2b, let alone, of course, what will come out of the Phase 2b.
So we talked about effect sizes, and part of the kind of next logical question there is-
Yeah
... what does that mean for your powering-
Mm
... and design for the Phase 3?
Mm.
So in the light blue is our estimated placebo-adjusted effect size from the enrichment just due to drug in Phase 2a, puts us over point five. But if you look at the standard of care effect sizes in all comers, gray is monotherapies, dark blue is adjunctive therapies, we're studying both with this drug, they're all around point three. Phase 2b powered at a Cohen's d of point four, which would put us above standard of care. And we'd really consider anything above point three, something that is worth advancing to Phase 3. and obviously, when you power at point four, you can detect significance at a level below that, and that's really guided by the clinical need here.
This population really has nothing that works well for them, and so, aiming to be both better than standard of care and targeted in this case with a drug that's really well-tolerated. The design of the now completed Phase 2b, reading out in October... is patient selection up front. We're including both patients with a poor memory phenotype and those without-
Good.
but powering within the poor memory group.
Yeah.
We're using an enrichment approach, which means that targeted population, that's where our statistics are, but the comparison to the good memory group is then a qualitative, is there some degree of greater effect size and enrichment, and that follows the FDA's enrichment guidelines.
And the primary endpoint is on the poor-
The MADRS.
On the-
In the poor memory.
On the MADRS in the poor memory.
Exactly.
Yeah.
So, you know, to that point, right? FDA understands what that outcome means.
Yeah.
We all understand what a MADRS outcome means, has obvious face validity as a standard approval endpoint. What we're doing is effectively just better defining the inclusion, exclusion criteria, and that is the enrichment concept and guidelines. And then the way we do that here is with a double-blind period that is much like every other randomized trial.
Yeah.
Six weeks, one-to-one randomization. In addition to having studied placebo as a treatment, we're also mitigating placebo response as best we can, using what we would consider to be best practices. One-to-one randomization to not change expectations. We don't assess patients too frequently. We have an external rater here at MGH do a separate baseline interview to ensure that there's no incentive for score inflation from sites. Importantly, and this is really, really fundamental, we kept sites, patients, and our own clinical operations staff blinded to the biomarker status for any patient, and even blinded to the ratio of those with poor memory to good memory. We only disclosed that for the first time at our Investor Day last week. That's really to really run the most rigorous study we can and not inflate expectations, which can have placebo effects.
This is the study population. We're focused here on our mITT population, which is those people who meet criteria both based on the MGH interview and then an analytic cutoff at the baseline for the trial, which is done at the site. Then we split into the poor target population here, poor memory, and the qualitatively compared good memory population. And then we expected around two-thirds of those patients would of all patients, would have monotherapy treatment with ALTO-100. That's exactly where it came in. And so we have a key secondary analysis called out there, and that was through FDA feedback on the importance of looking at that group.
But one of the things that we would like to be able to argue is if you have enrichment, if you're seeing a difference in poor and good memory, you need to show they actually start in the same place. And we've argued, you know, historically for years, that cognition doesn't map onto severity cross-sectionally, it maps onto chronicity longitudinally.
Huh.
And so let's see if that holds true in our baseline data, and we disclosed this last week at Investor Day. And you can see really, really nicely, the poor memory group and the good memory group do not differ. They don't differ in age. We took age out of the equation when we normed the data, which means that it's not just poor memory or older patients, right? They don't differ in severity, nearly to the decimal point, identical severity, and they don't differ on other demographic factors. The other thing you can see here is part of our design and our execution. We run our own trials. We have our own clinical operations team. We don't use a CRO. That ensures higher quality data, allow us to operate with much greater capital efficiency, roughly half the cost of industry averages.
And so we wanted to make sure execution was as we had wanted it. And what you can see, for example, is that the site baseline, Visit 3, is very similar to the MGH MADRS score, and that suggests that, as designed, there's no incentive here for score inflation from sites, which is a common driver of placebo response.
Yeah.
So everything looks good in terms of, the analysis population here, the execution. You know, really coming down to how does this then come together analytically and statistically, and decision-making-wise with respect to Phase 3? our primary outcome, as was mentioned, is in the poor memory group with a built-in step-down test in the key secondary analysis group here, which is monotherapy. That's to be able to potentially support just a monotherapy perspective for Phase 3. there's no alpha spend if the primary is significant. It's a built-in step-down test. We're also going to look at one additional layer of greater cognitive severity, a Z-score of minus one, so one standard deviation below healthy or below. Our inclusion is at half a standard deviation below healthy and below.
What we're not gonna be doing is just fishing around for data, playing with severity and so forth, like has traditionally been done. We're trying to call our shots here so that people can really interpret our results very clearly and very transparently for how we're gonna make decisions in moving the drug forward. As I mentioned, we're gonna be looking at good memory as a qualitative comparison, and then there'll be a variety of other things that come out of this trial. Not everything will be top-line data. We're gonna try to make top-line data as clear as possible for how our decision-making lays out. So where are we? You know, at this point, I would argue that we've tried to turn over every stone we could possibly find before doing this real kind of powered direct test of this drug.
We have a target population that's linked mechanistically to the drug, that we understand how to identify, and we can do even with a commercial-ready web-based tool that we've given investors access to coming into Investor Day.
Yeah, and it was actually easy to do. It was a humility sandwich, the outcome, but that was easy to actually conduct.
Exactly.
I can see it being done by patients.
Exactly. So that was the goal, right?
Yeah.
Is a 15- to 20-minute web-based test. You do it at home, you get your score right away, and so you can actually imagine from a commercial perspective, right? Put yourself in the shoes of a patient. You get this information, that changes how you interact with a doctor. That will drive penetration, that will drive change in practice above and beyond the usual way we do it, which is, of course, in medical affairs, educating doctors, and so forth. One of the things we did actually, and we shared at Investor Day, was also research with doctors themselves.
Mm-hmm.
The, you know, it was very, very clear that people are ready for precision. People want, speaking as a psychiatrist, likewise, right? They want to know what they're doing and why. And-
It surprises them. Sometimes it surprises them that a patient would be in or out, right?
That's right. So that's a really interesting point and perspective, which is we, as clinicians, really only have access to what we get told-
Yeah
... by patients, and then what we observe and cognitive deficits, when they're extreme, can be observed, but that's why you really need this test to look at the sort of more mild to moderate range of severity and quantify it, because you're actually going to miss a lot of the people otherwise, if you just use clinical judgment.
Yeah.
The fact that it's easy to do, it's something that...
Yeah
... a patient can access for free outside of the doctor visit, also means that you don't have to worry about how it integrates within clinical practice.
Now, cognitive tests are notorious for a practice effect, so how do you remove the potential of that? You give patients only one shot, or?
Yeah, so they'll have one shot-
Mm-hmm.
... but we'll also know if they're reentering.
I see.
We have alternative forms, which decreases practice effects. You want to essentially get them to the point where there's not a practice effect to doing it, so they get some instructions and some practice initially.
Sure
... so they're familiar, and then what you're measuring becomes more stable. We also know about the reliability of this measure, which is really important. So across two months, both our measure and the traditional way to measure memory has a reliability correlation coefficient of 0.8, so it's really quite stable over time. And so even if there's a little bit of learning, it doesn't necessarily change your scores particularly.
You constructed the test in a way, you and your team, to not only ask questions, get at an answer, but also retest it within the test-
That's right
... to validate that the person wasn't just guessing.
That's right. So, we don't actually have the slide here, but you can look on our Investor Day deck.
Okay.
You can see the overall cognitive impairment of this group. So it's not just memory. We're selecting on memory, but memory brings with it cognition overall, and these folks did about 13 different tests, 12 in addition to memory, and you can see that across the board, their impairment on average is almost a standard deviation and a half below healthy, which is the level of cognitive impairment we associate with schizophrenia, which we think of as a cognitive disorder.
Yeah.
And in depression, this group is that impaired. By extension, the group that has good memory is stone-cold normal on memory. So these are really divergent groups across measures, stably so in terms of cognition, yet have the same baseline MADRS, so we can compare the trajectories.
Let me ask you, how did you come up with this biological hypothesis that you know, hypothalamic activity or function is really a driver, if you will, or a basic pathological feature of depression?
So, it's not just us. The field as a whole has-
Right
... had this neuroplasticity model of depression, focusing on hippocampal plasticity. We, in my lab at Stanford, have been publishing on this subpopulation and their poor response to standard of care and greater impairment for a long time. But also even when I was a grad student working on neuroplasticity in the hippocampus, this concept was already there and developing. So it's a matter of putting these things together, and honestly, it's a matter of saying: "Look, we have hypotheses mechanistically. We have drugs that were identified, like this one, specifically for this function," and yet never in the past have we actually connected the two of identifying the patients who have the demonstrable deficit-
Mm-hmm
... and the drug that's meant to address it. We kind of wave this wand and hope that an all-comer-
Mm-hmm
... approach will capture-
Mm-hmm
... enough of those people.
Mm-hmm.
But the bottom line is, most of the time, it hasn't-
Yeah
... and that's why we are where we are.
Yeah
... as a field.
Because of heterogeneity, and that's the real innovation here, Amit, is that the hypotheses have been around. I'm sure that you could come up with a similar one for GAD, as an example.
Yeah
... Generalized Anxiety Disorder. But having a test that specifically helps you to stratify patients, that's the key here, and that's what makes this more precision medicine.
Exactly.
Yeah.
So if you look at the pipeline, right, to kind of bring us into that bigger picture, you know, we're taking the same approach around cognition and identifying patients who respond better to ALTO-100 and bipolar depression. So the same mechanistic rationale, the hippocampus neuroplasticity, and so forth, is there in bipolar depression. The only treatments we have for bipolar depression right now are all antipsychotics. Tremendous side effect burden. So taking that, you can already jump to a different disorder, where you have the biological rationale and advanced care there. ALTO-300, a drug called agomelatine, a known antidepressant, really well-tolerated, has plenty of data in all comers, approved in Europe and NCE here in the United States, and there we've developed an EEG biomarker driven through machine learning that we think relates mechanistically to the drug's action as well.
And so all of these tools come into play across disorders, across mechanisms of action. I mean, for ALTO-100, it's a potential first in class, right? So it's not just known mechanisms, it's totally novel mechanisms.
Yeah. Yeah, so in that latter one, ALTO-300, that's a readout for next year, that EEG-based mechanism really helps you to identify a safe dose, a dose that's active.
That's right. So, we have, we know all-comer efficacy is there at the-
Mm-hmm
... twenty-five milligram dose, actually just the same as any other higher dose. We know it's safe in terms of tolerability and not having liver tox issues, which higher doses have some elevation in LFTs and been reversible, not been an issue, but you can eliminate that with twenty-five milligrams. You have a really nice setup for the drug, enhanced with a biomarker to get greater efficacy, and even more so with a biomarker because it's physiology-based, you can take that biomarker into animals and now really understand mechanistically why is it that those people respond better? How does that relate to mechanism of the drug, which has not been done before as well. So there's a lot of benefits to taking a precision approach.
So three readouts in the next year. It's going to be an interesting year. I wish you luck.
Many thanks.
I appreciate you taking time. Unfortunately, we're about out of time. I was going to ask you the last question, which was, what are you most looking forward to? But we'll come back to that-
Absolutely
... and talk in October.
Absolutely.
Thank you, Amit.
My pleasure. Thank you.
Thank you to the audience for your interest.