Welcome, everyone. It's my pleasure to lead this fireside chat with Edgewise Therapeutics. I'm Joe Schwartz from the equity research team at Leerink Partners, and it's my pleasure to be joined by Kevin Koch, CEO, and Behrad Derakhshan, COO. This has been a very exciting 12 months or so. It's nice to see things starting to come together. Now that you're well on your way in the clinic, maybe we can have you start us off, Kevin or Behrad, with just a quick level set, your vision for everything.
Yeah, there's a lot. This is way low. Yeah, turn that down a little bit. It is a great year here. We had a great year last year. We talked about advancing sevasemten and our first report on EDG-7500, our cardiovascular asset. This year is even, I think, more exciting than sevasemten at the end of the year. We had some data on the release of the CANYON data. That is very exciting. We have a number of additional milestones coming up this year on both of those assets. Maybe chronologically, we can go through them and talk about where we are. EDG-7500, our cardiovascular asset, which is a sarcomere modulator, we reported exciting data in a single-dose obstructive HCM in September of last year.
This coming month, by the end of the month, we should have 28-day data that will describe a couple of different things. We'll have a minimum of 10 patients in obstructive HCM and a minimum of 10 patients in non-obstructive HCM. Those patients will be in two or three different dose groups. We'll report very much the same data we reported from the single-dose obstructive HCM data. We'll report data on resting gradient, Valsalva gradient, changes in NT-proBNP, data on E-prime, a diastolic measure of function, which we observed in the preclinical models that we looked at. We'll also get some really new feel and function data. Changes in 28 days of New York Heart Association and changes in KCCQ scores at day 28. I think it's an exciting data set.
We are setting the expectation that we'll see at least what we saw or greater that we saw in the obstructive HCM patients. I'm looking forward to putting the data out, and it should be fun. Now, let's go to the next milestone. We reported in December positive data in CANYON, which was sevasemten, our Type II myosin inhibitor for the treatment of muscular dystrophy. In the Becker muscular dystrophy population, which is an adult form of muscular dystrophy, we reported statistically significant inhibition of the primary endpoint, which was creatine kinase lowering, which is a measure of muscle damage. We also showed statistically significant changes in TNNI2, which is the on-target biomarker for muscle damage, which is specifically found in fast muscle fibers.
Even more important to the data set, we showed a strong trend for NSAA, which is the North Star measure of functional benefit in patients, which is the primary endpoint for clinical trials. Exciting in that we were, even though we were underpowered, the trial showed a 1.1 difference between placebo, which is clinically meaningful. I think the most exciting part about it is that we had significant natural history data from the literature, and the placebo group matched the natural history data very closely. We believe that data package, because of the high unmet medical need in Becker, no treatment, the trend in North Star, statistically significant benefit on the key mechanistic muscle damage biomarkers, and the natural history data, and of course, the strong safety profile of the molecules amenable for us to file for accelerated approval.
We'll have essentially a conversation with the agency, and we have a meeting set up early in the second quarter, provide some feedback of that information from the agency at the end of the second quarter, and we'll continue to hope for a dialogue with the agency to continue down the path toward accelerated approval for this asset. We also reported just in the last week that we have completed enrollment of GRAND CANYON. It was a placebo-controlled study in Becker, proposed to be 120 patients, a two-to-one randomization in Becker. It turns out we over-enrolled that study, 175 patients, tremendous interest in actually taking the drug. I think something more important for the agency and how you think about the profile of this drug, of the 60 or 80 patients that have been on the drug, 99% have rolled over into the open label extension.
I think the drug, based on that information, is providing benefit to patients and has a strong safety profile. Very nice package to send to the agency, and we'll see how it goes.
Great. That's a wonderful start. Thanks, Kevin. Maybe we can start with 7500, and I promise you we'll get to sevasemten. We'll leave enough time there. The efficacy so far, obviously, only seen SAD data in patients so far. What would you hope to see after 28 days in the OHCM cohort? While we're on the subject, NHCM, how would you guide us to think about early signs of efficacy there?
Yeah, so we saw with the single dose, we saw about a 60%-65% dose-responsive change in resting gradient, which I think is exciting and an important number. We also saw about a 55%-60% change in Valsalva gradient. They measure slightly different things. We would expect that these types of measures would deepen over time at some level, but that will remain to be seen when we provide the 28-day data. We will provide the data by dose as we did in the obstructive HCM disclosure and should be dose-responsive. I think also we are very keen on the evaluation and understanding of what NT-proBNP predicts. We believe that that particular endpoint is associated with wall stress and the diastolic effect we see with the drug. We think and we know from data with the cardiac myosin inhibitors that the depth of the proBNP response highly correlates with peak VO2.
Why is that important? Because peak VO2 is the primary endpoint for obstructive HCM trials. I think the depth of that response and with this 28-day data can actually be indicative of what we might see in a phase III study and could we expect it to hit statistical significance in obstructive HCM. Now, for non-obstructive HCM, they do not have a gradient to go by. All you really have are biomarkers like BNP and looking at ejection fraction. Again, with both of these therapeutic indications, we would expect our drug to have no association of dose or concentration with changes in ejection fraction.
That's a key point of differentiation from the CMIs is that we have not observed ejection fraction lowering, and ejection fraction lowering has driven the REMS in place for mavacamten and some form of REMS, whatever that might be, with aficamten that we'll find out about in September. We think that continuing to have a safety benefit in where we do not see changes in ejection fraction is a key differentiator for our mechanism, which is novel. The depth of the response on efficacy should be very similar to what we might expect to see with the CMIs.
What do you think you need to show in order to not have a REMS ultimately? Is it no ejection fraction drops below 50% ever? Is there any leeway on that front? It probably depends how it's measured. Love to get your thoughts on that.
Yeah, you know, it's an interesting question because the placebo rate is 1%-2%, right? And with the CMIs, they're seeing 5%-6%, something like that. You can tell a difference in the phase III. Obviously, that running a phase III will be much more definitive if you didn't disassociate from placebo. At this early stage, we've treated on the order of 80 patients. We've not seen any patients go below 50. All the patients have been within, I guess I would call the noise of the placebo. An ejection fraction measure is kind of ± 10%, at least based on our data and the placebo responses we saw. I think it's the more data you accumulate, the strength of your story that you're not really related, the concentration or dose of the drug doesn't really affect the ultimate ejection fraction change.
It will take time, and it won't be definitive, I think, till we get out through the phase III.
Right. Okay. Very helpful. Will you be providing the same kind of analysis and quantum of data when you report the next set? Will it include OHCM and NHCM together?
I think we are at this point, we don't have all of non-HCM data at this point in time. We will probably by the end of the week or next week. At this point in time, though, I think we're going to report both sets of data. I think they sort of support each other in some ways. It's important that I think this drug, I think the non-HCM market is growing. I think the potential in that indication is also indicative of the potential of the mechanism in heart failure. That, I think, would garner tremendous interest if we could demonstrate some benefit in the non-HCM population. Of course, because we have a second-generation drug that we're planning to file an IND on in the second quarter, that becomes really a cardiovascular platform that's addressing HCM and heart failure with two different agents.
I think it's important to show that as early as we can.
Interesting. You mentioned several biomarkers that we should focus on, I know NHCM in those data sets. What about functional outcomes like New York Heart Association class and KCCQ? 28 days is not a long period of time.
It's a short time to show that data. I think everyone wants to compare our mechanism with the CMIs. Twenty-eight days is short. However, I could cite the six-month data from the CMIs. Two things of note, they do have deep responses in gradient. However, the changes in New York Heart Association, like a one-point change, about 55% or 60% of the patients had a one-point change, and that included placebo. There is about 30%-40% of the patients that do not have a one-point change. That is important to note because even though the gradient went down, they did not feel better. For KCCQ scores, really for both drugs, they had about a 12-14 point change, but five of that was placebo. Placebo adjusted, the KCCQ scores are between six and eight points. That is clinically meaningful, but is that enough?
The question for our drug is, can we do better even in 28 days? Our mechanism of affecting not only the early aspect of contraction systole, but the relaxation of the heart. Imagine that the heart fills with blood, relaxes the left ventricle. It's more oxygenated blood. It releases that blood, that oxygenated blood goes through the body. What would be the effect on the patient? The patient might feel better. I don't know if it's going to affect the gradient, but the patient might feel better because they simply have more blood. The question is, will that translate to a better feel and function measure? Is that part of what's missing in the CMIs?
We will see how that data plays out, but we would hope that we could get more patients into what I would call the normal range, meaning can I take New York Heart Association patients from three or two, which is symptomatic, to take them to one, which is asymptomatic? Can I take their KCCQ scores instead of 65-75? Can I take them 65-90? Can I take them up to a level where they get normal? And then thinking about the biomarkers, can I really get a patient to normal with the biomarkers, not just an improvement, but what does normal look like? I think normal is what every physician would want their patients to have: normal symptoms, normal biomarkers. Normal is normal. That is where the goal is.
Can we get to that dose and exposure in a safe way without seeing ejection fraction changes?
Joe, as people contextualize what we show at 28 days, one really important piece of information, because the natural thing for investors, strategics, everyone to do is to say, I'm going to compare SEQUOIA and EXPLORER data to fixed doses of 7500 at 28 days. Now, that's a great comparison if you want to get kind of caught up in efficacy. Remember, we're not dose-optimizing patients yet. We're looking at the minimally efficacious dose for a given patient because we haven't dose-optimized. For patients from SEQUOIA and EXPLORER, those patients have gone through a titration protocol based on an optimized dose for that individual patient. They are getting the max benefit they can get, and probably some of them not even enough because of an ejection fraction liability. We're just starting. We're saying, look, we've got certain doses that have a certain effect in the Part B, the open label.
That's where we're really going to start kind of thinking about, okay, how do we get individuals to a dose that's optimal to maximize for efficacy without having to worry about ejection fraction potentially?
Right. Yeah, that makes a lot of sense because I think we know that most mavacamten patients nowadays are on 2.5 mg and 5 mg. And we still only have 10,000 patients of 250,000 diagnosed and maybe closer to a million out there.
Yeah, I mean, Ted Abraham gave a really good example yesterday. He was talking about, he has 1,500 patients, of which 200-250 are eligible for mava, and he's only got about 60 of them on mava. That tells you he's got, and that's, I think he's had to accommodate by adding echo staff to really get those patients on. That's a really important figure because that shows that there's a bottleneck even at the centers of excellence that are high-volume centers to get patients on the drug.
Yeah, that's a great point. We heard very similar feedback in our panel yesterday too. It's a great segue.
I think it's a key aspect of the market is many of the patients reside in the community practices. For a community practice, they have to rely on their personal echocardiographer to make the call as to whether or not a patient is going below a 50 threshold. That puts a lot of risk on that individual's cardiologist practice that perhaps they don't want to take. Add to that having to manage the patients, interact with the pharmacies, keep on a schedule, and the number of echoes to get to the actual dose that's optimal for the patient. What we're hearing from the community docs is they largely send them off to the centers of excellence because it's too much of a hassle for a relatively small number of HCM patients they may be treating. The centers of excellence have capacity limitations as well.
It is not that the drug is not effective and it is not that there are not people trained to deal with the REMS. It is that it is not commercially beneficial for many cardiologists to actually personally treat the HCM patients. That is the goal of our drug: if you do not have the safety signal, you have a simpler titration to get them to an effective dose, you would move that asset into the community cardiologists and expand the market dramatically.
As far as the profile, the target product profile that would be required in order for that to happen, what is the minimum amount of titrating that would be required in order to have something that would be readily adoptable?
I think the echo itself is not an issue. It's the number of echoes and the need to have them in a regimented structure and having to do them for safety. We believe that one echo somewhere in a dose ranging to get to the optimal dose, plus utilization of either feel or function measures or something like an NT-proBNP, would allow us to titrate to the target exposure with a minimally kind of difficult strategy and would really be effective that way. I think the target product profile would get you to something that would be easy for the patient, easy for the physician, and with a strong safety profile that would allow not to worry about how much drug you were giving the patient.
Yeah, I mean, I think that's right. I mean, think about it this way, Joe. Someone comes in, you give them a dose, and then really leaving it at the discretion of the physician if they ever do an echo to optimize for efficacy, or if they just wait till the patient comes in at their usual schedule of echoes, which is six months to a year, because you naturally have to have those echoes just to track disease progression. Really leaving it at the hands of the physician saying, "Hey, you're free to dose this drug without any concerns of ejection fraction, and then you decide when you want to put an echo," rather than being regimented on when to perform the echo because of a safety concern. Because remember, mava's four-week and eight-week echoes are required safety echoes.
It's going to be really hard to move away from those. I think that's the bottleneck. It's those initial echoes to get to an efficacious dose that's limiting expansion outside of the centers of excellence.
Okay. This has been great. You anticipated several of my questions. Let's turn to non-obstructive a little bit more. We hear this is a really heterogeneous patient population with great unmet need. How do you see this landscape? We talked a little bit about what the 28-day data might show us. How have you thought about the enrollment for patients that are in this serious cohort and how representative they will be of this big market opportunity?
Yeah, there's a couple of things that are different about non-obstructive that are important. You don't have the gradient to go by, which makes it a little more challenging. It's harder to separate them from HFpEF patients. You have to be really cognizant of the criteria of non-obstructive HCM, and what's the fibrotic level of the patient. Obviously, you will eliminate the ATTRs, but you need to be really thoughtful about things like the thickness of the ventricle wall. They have a number of other comorbidities that you have to consider. For all of those things, you probably have a more heterogeneous population, and you don't have a marker for how high you should go. For us, we don't have an ejection fraction signal from the data we've seen in obstructive HCM, so what's the right dose?
I think at the moment, what the expectation for that population in our treatment would be, we'd want to continue to see no changes in ejection fraction relative to the concentration of the drug as we go up, and we'd want to see deepening responses and dose-related deepening of responses on proBNP. I think how we might end up thinking about this is, can I get them close to normal? So normal on NT-proBNP below 200, 150, we have to talk about exactly what good looks like in that trial. If good looks like below a threshold of BNP, and that over time, and it may take more than 28 days, but perhaps it can be seen in 28 days, things like the KCCQ score moving, that would be really positive.
It's a lot less information on that particular indication, and we're going to have to do a bit of exploring to know what does good look like. The opportunity is there because even though they are treated with beta blockers, they're even less effective than in obstructive HCM. There really isn't a good treatment at all in non-obstructive HCM.
It's also from our research, it's the largest growing segment of HCM patients. Traditionally, it's been two-thirds of patients are obstructive, one-third is non-obstructive. The pool of patients feeding into the non-obstructive is growing, and we're getting to kind of closer to 50-50. I think that's driven by family screening, cardiac MRI, AI-based tools. I think the proportion of NHCM is going to make up a higher proportion of the symptomatic HCM patients that go to seek treatment.
Our tact at the moment is to take phenotypically derived non-HCM patients. We are looking closely at, do you have a pathology-driven mutation to get a more homogeneous look at the disease? Because of what Behrad just said, that families are now doing the genetic screening, you are finding more gene-positive non-HCM patients. That might be a growing market, and it might actually create an opportunity to demonstrate a really oversized clinical benefit in the smaller population of the genetically driven patients and then expand into the more phenotypically driven patients. Have not decided yet. We will have to look at the data.
Yep. Very interesting. The next trial for last question on 7500 before we get to sevasemten , the next trial in both OHCM and NHCM, is that like further dose ranging? Do you think either of them might not need further dose ranging, and you might be able to go into like a phase II-B?
I guess you could call our Part D, which we're amending right now, is how do we get to the optimal dose for each individual patient? Which metrics do we use? How do we minimize echoes? Can we look at these endpoints out at kind of the 12-week timeframe? We're still looking at, by the end of the year, having a substantial number of patients out to 12 weeks that we could evaluate to give ourselves a really good understanding of how we would run a phase III. The goal would be by the end of the year, we'd have that data, we'd be able to go to the agency with a phase III proposal based on the dosing schedule as well as the safety monitoring, and then initiate the phase III trial sometime in the second quarter when we have drug supply.
We will have drug supply, phase III drug supply in the second quarter of 2026 to initiate the phase III studies. Phase III studies, we're marking in about 18 months. Could be a bit shorter, but 18 months, I think, is a good number. That puts us out into 2028 with the phase III data. The base case today is to start both studies in parallel. There may be some staggering based on logistics, but I think there's reasons to believe we should be aggressive in both indications.
Great. My question on sevasemten is, based on the over-enrollment of GRAND CANYON, what are the latest powering assumptions? How does that improve powering assumptions? Is that something that might make the FDA feel good that there's more likely than not to be a success in?
It certainly helps and increases the powering. We were very well powered above 95% with 120. Let's presume that some of those patients either will drop out or perhaps we'd make a certain cut of some of those. The analysis set might be slightly different. It gives us a lot of leeway. We would be highly powered to see statistical significance over this 18-month period with this data set. I think what's more important about being able to recruit the 175 patients is just the enthusiasm for getting on our trial. Becker patients have told us, "No one's ever paid attention to us." Now you're putting all your efforts, and they're all really pleased to get on to this drug and get onto the trial.
I think coupled with how many patients have rolled over into the open label extension, it gives me a really good feeling that not only will we be able to find patients in this orphan market. There are about 13,000 patients in the G7 in the major markets. We can find many of the patients. They will come, and we'll find a way to get them on drug, and that they'll stay on drug. They believe this drug is helping them. That is an exciting aspect for a disease where there is no therapy and no one's paid attention to them in 25 years.
Excellent. Unfortunately, we're out of time. Great discussion. Thank you both so much.
Thank you.
Thanks, Joe.
It's great. Thank you all for your attention.