Guggenheim 3rd Annual Healthcare Innovation Conference - EDGEWISE THERAPEUTICS INC

Morning Guggenheim's third healthcare innovation conference. My name is Moritz I'm part of the therapeutics team and it's my privilege today to host Kevin Koch CEO of address therapeutics Kevin thank you so much for making time for us I want to jump right into your recent news last week You published the design of your phase 3 program for obstructive and non obstructive HCM Could you just give us a brief overview of the trial design and the highlights? Yeah, so we had a very productive conversation with the FDA in regards to the trial design. And as you know, we published and presented data, the 12-week data, on the drug class. And we also, this last quarter, talked about the mode of action of the drug. So let me give you a little bit of a highlight. The mode of action is highly differentiated from the cardiac myosin inhibitors. In fact, we target a regulatory protein, which is within the lever arm of the sarcomere, which is the contractile element, and the CMIs bind directly to the ATPase, which binds directly to the actin filament. So they affect the enzymatic activity. They can drive the contraction down to zero. Obviously, that has consequences for your heart when you eliminate all contraction. We have a mechanism that regulates that contraction and changes the rate of that contraction by binding a regulatory subunit on the lever arm. So that makes it a very different drug and the outcome very different when you're looking at endpoints like ejection fraction. So that's data we put out. We published a paper. I think it's one of the very nice paper on this mechanistic possibility. You also, in that paper, we showed that there are multiple different ways of interacting with the regulatory chain and that different ways provide different outcomes. So in that paper, we showed that the enantimer, so the mirror image of 7500, 7,499, looks almost exactly like a CMI. So we had exquisite control of the pharmacology by small changes in chemical structure. Okay? So how is that translated to the clinic? In June, we reported that under multiple different measures, we did not decrease ejection fraction. And what we showed, and we showed the FDA this, the PK versus the PD versus, and so the concentration of the drug relative the effect on the systolic endpoints. And we saw no relationship with over now, I think, a thousand different measures, no relationship with the concentration of drug relative to ejection fraction and multiple ejection fraction-like measures of systolic function. So the question is, and what is the challenge for the CMIs in hypertrophic cardiomyopathy, is that what you've seen is a dose and concentration-related decrease in ejection fraction with those drugs. That has led to a REMS and a black box warning for heart failure and has limited their ability to have uptake into the community cardiologists because of the need for repetitive echoes to keep the drug safe. Now, they're highly efficacious drugs, no doubt about it, but that is a significant liability. And what it does is cap the utilization of the drug in the market. So we think that that caps the market into perhaps, let's say, $2 or $3 billion. Whereas if you look at the number of HCM patients, there's, say, 200,000 symptomatic patients. You could have a much broader market if you could move the drug into the community cardiologist community. Right. So we think that this mechanism has that possibility. So what we put out last Thursday was with our conversations with the FDA, we requested not to have any monitoring of the patients to get to the target dose, which is 150 milligrams. And what we asked was allow the PI or allow the physician to determine what is the appropriate dose for each individual patient. That was accepted by the agency. So if you remember back in June, we had an initial conversation with the agency in kind of the March timeframe. And they said that our original design was reasonable, but they wanted to see more data. So we gave them the additional data. It, in our minds, remained reasonable, and that we could move the drug up to the target dose based on tolerability. So that was a very important win for the company. We had a lot of discussions with the agency. In fact, I would say we had no discussions about more monitoring. We had discussions about how do you define tolerability? And also, instead of stopping at the low dose when the patients have benefit, maybe how do you get them to a higher dose so that you really see effects on things like peak VO2? So that leads me to the next response of what else was in the press release. We had proposed, and I've said this to some of the investors, a Z-score, which is basically a composite of the peak VO3 and VEVOCO2. that is a I think we could drive greater power we think that might be a more representative endpoint for benefit but from what the agency spoke to us basically would be hard to interpret and say what number equals of what efficacy level so we decided to measure them separately and by measuring separately we ended up taking our conversations with the EMA and with the Japanese Regulatory Society is that they both preferred Peak VO2 as a primary endpoint. The agency also preferred Peak Vo2 as primary endpoint and allow things like KCCQ, other functional measures, and things like VEV over VECO2 to be measured as secondary endpoints and looked at separately. So that was after, you know, trying to, you have to try and coordinate three different regulatory agencies simultaneously to get them all in line. And so that was the lowest common denominator of everyone really wanted peak VO2 and believed in peak VO3. So that's how we got to peak VO4 as the primary endpoint. So I think that was important. The third important part of this protocol was we had originally thought that we could run one trial and use the statistical analysis plan to essentially take both populations into account and derive a p-value and a powering calculation off of that 600-patient trial. the agency preferred and told us directly that they would like to see statistical significance in both the non-obstructive and an isolated obstructive HCM population so what that means is well if I have to hit stat sig on both I might as well run them as as separate trials under essentially an umbrella protocol. So we have one protocol. You have one interaction with a site, and this is all kind of the sausage-making in running a clinical trial, but you have one reaction with the site. People come in. They can be funneled off into obstructive or funneled into non-obstructive. There's two separate trials that are powered for each individual trial. That actually was, I think, much more popular within the investment community. It made a lot more sense. I think it is cleaner. But there is, you know, the need of perhaps increasing the N with that study. So maybe the other fourth important point in that press release was that we have engineered into the protocol an adaptive design. So we wanted to be able to expand the study based on if we see efficacy. It's almost like a futility analysis. If we see any efficacy, we will probably increase the ends of, in particular, probably the non-obstructives. But if you look at the obstructives, it's possible that you only need about 200 patients to actually get STAT-SIG in that population. And if everything goes in the right direction in non-obstructive, we could save perhaps six months. So it gives us this opportunity to get real data on the variability of the trial results and also the actual magnitude of the clinical differential. and so you can now readjust your sizing based on real data instead of guessing. And as many of you know, we did not measure peak VO2 in our Phase II study, so we're really going off of our NT-proBNP and some other parameters and what we see in KCCQ. We're also going off the preclinical data where we seem to have a greater cardiac reserve for this mechanism relative to CMIs. That's a long-winded answer to your question. Well, it made my job a lot easier. You answered a lot of the questions I was about to ask. But maybe one point I wanted to focus on, you touched on this already briefly, is the titration based on tolerability. So how is that going to work in practice? In practice, the patient will come in. They'll be asked different questions. We do a New York Heart Association, or we do a question, for instance, it's different for each patient, but you ask them the question, like, if you can go up the stairs, how do you feel? Are you short of breath, or how doyou feel during the day? Are there any side effects? Do you have any, you know, they'll go through perhaps whatever they normally go through with their typical patient in this space. And if they feel better, we've, in the protocol, said then you should go up to the next dose, even if you feel good at that dose, because the efficacy is dose responsive. So we also showed that. So we believe that by getting people to the, I think, a target dose or a maximal dose of 150 will benefit the patient, even if you feel good at, say, 50 or 100. So that's how we, in practice, it will be used. That's how, in practise, cardiovascular drugs are actually treated for most patients. They start at a low dose, get them to the maximal dose that they tolerate well, and then leave them there. Yeah, and you don't have the same issue of sort of trying to find a goalie lock zone between EF drop-off on the other side of it. Right. So we had told the investment community, and we had spoke to PIs about this and the KOLs. We originally proposed baseline echo, which, you know, for all intents and purposes, you need that baseline echo to enroll the patients. You have an inclusion-exclusion criteria, okay? And in the real world, that would be the diagnosis of the HCM. You do that by echo, by wall thickness, and ejection fraction. We have a 12-week echo. that's a so after you've been on maximal drug for at least four weeks you get an echo that's not read by the investigator that's read by a core lab so that is sits in the database to support the safety database for the filing it is not used to titrate and should not be on the label and then we have a end of study 24 week echo which is largely being used for efficacy The other thing we want to do with that 12-week echo is to see you have a 12-weak echo, and then you go have a 24-week eco, and you're seeing is there a change in efficacy over time with that endpoint. So that's an important measure to get ultimately, but it's more for the actual label and looking at the final results than it is. It doesn't really have any effect on how you get to the target dose with the drug. Yeah, no echoes during dose titration. The other thing to say is that in the trial, it needs to be taken at 12 weeks. It could be entirely optional for the patient. The doctor, any time, could ask, well, this person's been on a stable dose of the drug for X months. I'll get a standard echo and see what they're doing, which is how they're monitored for disease progression. Yeah. One other thing that struck me about the trial design is the duration to the primary endpoint, 24 weeks. So given you have two four-week dosing increments in there, that will mean that patients could be at the maximum 150 dose for 16 weeks before the primary end point. Do you think that's long enough for 7500 to show its full effect? we we know that so the other key aspect of the mode of action is that it's four to five times more potent on the diastolic relaxation of the of the ventricle and so that characteristic we believe drive a more rapid response. And so I think we're thinking that 24 weeks is sufficient to show an important change in KCCQ and peak VO2. So I think that's a calculation. We'll see. Mm-hm, and the other sort of key question when it comes to 7500 is around REMS. If Stratus works out and delivers everything you, you hope it will deliver, do you think that's gonna put you in a position where you could get a label without, without a REMS? I, I certainly think so. If we're, if we're not, if We're not actually running the trial with multiple echoes to get to the target dose. I don't understand how it now would end up back in the label as a requirement because it doesn't seem to make any sense. You didn't run that trial. You didn'T have a safety issue, so therefore why would you have a REMS? And for other drug mechanisms, and let's presume this is a novel drug mechanism, that's not a requirement. So the SGLT2s did not have that as part of their phase three trial, and you wouldn't have expected them to be like that. So I think it's quite reasonable to believe that you would not have a REMS. That's obviously up to the FDA, but I think that there's no reason if you didn't have it in the trial. And so what do you think Stratus needs to show to sort of confirm that? Is it a continued lack of kind of the PKPD association that you already mentioned? Is it you mustn't have any LVF below 50? What's needed? Yeah, you know, I think when you start getting to, you Know, six, eight hundred patients, there will be patients that are going to go below 50. So the question is, how does it change relative to placebo? So that would be the main thing. But we have not seen it, so I think it's important to monitor that aspect of the drug. But I think the below 50 is a little bit of a synthetic construct of the nature of the CMIs. In reality, it'll be relative delta from the placebo. And what will your inclusion criteria look like when it comes to a VF? Oh, it will be 60. So we were conservative on that aspect because I think you want to prove and collect the data and make sure that you minimize any kind of aberrant change. So, no, EF measures are plus or minus 10% in general, just the variability of the EF measure. So we went to 60 because that would minimize any risk for anybody below 50. And since we've not seen any kind of heart failure signal with this drug, I think what you'd end up with is no deviation from placebo and minimal or no heart failure risk in the trial. And we already talked about the trial size for a little bit, but as it sort of relates to the size of the safety database and the REMS discussion, Do you think the size of the trial that you're running it would be large enough if you don't Upsize it at all from the like 200 and 400 current estimate. You think they'll be large Yeah, well, you know the one thing we have gone through is like okay so is 200 or say it's 250 or whatever the number is and obstructive is that big enough and you would use the blinded data to augment in the non obstructive piece to augment the obstructive now, we don't know what the rate of recruitment will be for the different populations. You have competitive agents in obstructive. You have one agent that may or may not have an approval in non-obstructive. So we don't know which one is going to recruit faster. I think from our internal estimates, the non-ostructive is goingto recruit faster so they may line up together, but But one may outpace the other, we don't know that until we get there. But I think, I think the, I think we can, we have certainly enough safety. Great. I want to touch a little bit more on one of the main differentiators of 7500. We briefly talked about this, but it appears to be it's, it's lucid tropic effect especially if you look at things like E over E prime and some of the data that you've published today, especially on the non-obstructive it really seems to be substantially different from anything that the CMIs have shown. So could you speak a little bit to how important is the Dastoli in non-obstructive but perhaps also obstructive HCM, and how is your molecule perhaps particularly suited to addressing that? Well, in nonobstructive, you're really talking about—well, let's go to obstructive first. When you remove the gradient, you're releasing kind of a plug on the system of flow. And that itself has a dramatic effect on how the patient feels. In non-obstructive, that plug doesn't exist. So what's happening is you have a thickened heart that relaxes very weakly. And what you want to do is that relaxation to improve. improve. So that is, if you don't fill the cavity with more blood and pump more blood, that will limit your exercise capacity. So what we think is going on in non-obstructive is that allowing the more blood to be filled, greater relaxation and allows greater blood to be pumped throughout the body. So we think that's a key aspect in nonobstructing and realistically in HFPAF. I'll get to HFBAF in a second. I do want to quickly talk about 15400 as well. But before we get there, just one perhaps final point on 7500. A common misconception that is still out there in some places is that you have traded EF drops for AFib. Can you maybe just address that point one more time? Are you convinced AFib risk can be adequately controlled in the phase 3 study. Yeah, you know, this is a lot of competitive marketing against, you know, against us, and everyone is talking about each other's AFib. The bottom line is that AFib is a significant part of the phenotype of HCM, and so unless you have a placebo-controlled study, it's very difficult to determine what is the right amount of AFib you've seen up to 40% levels in recurrent AFib. So I think with our mechanism, realistically, without a placebo control, it's impossible to tell what the level of AFib is and what should be the level. And cross-trial comparison becomes very difficult. Makes sense. So let's talk a little bit about the rest of your pipeline in the final two minutes here. You recently published some healthy volunteer data for 5400 and you announced that you started dosing in the phase two Equinox trial with data expected in the second half of next year. Can you just tell us a little bit more about this molecule? How is it similar or different to 7500 and what are your expectations for the readout? Yeah, so it was designed basically to take advantage or actually consider the patient populations within HEF-PEF. So they tend to be an older population. They tend to Be on multiple background meds of multiple classes. And so the certain metabolic properties of the drug are differentiated relative to 7500. We haven't disclosed what those different properties are. We did put out this past month, we put out some data on the PK. What you want is a very, very tight and controllable dose from a PK standpoint. And the CVs, so the variability, coefficient of variability, very low, less than like in the 10% range that we published. It's a 60-hour half-life, has about three and a half full accumulation, very reproducible drug. which led to very reproducible PKPD. So we think that particular profile, it does not have any drug-drug interactions that are going to be difficult to deal with, like genetically driven drug-rug interactions. So in this case, I think it's a really very nice drug for the hep-pep population. Great. I think we're pretty much at time already. So thank you so much, Kevin, for taking the time to speak with us. Looking forward to seeing that data in the second half of next year, and 7,500 data soon as well. Thank you. Thank you so much.