JEFFERIES GLOBAL HEALTHCARE CONFERENCE IN LONDON - STOKE THERAPEUTICS INC
Andrew Tsai, Senior Biotech Analyst at Jefferies, and it's my pleasure to have the Stoke team with me. To my direct left is Ian Smith, CEO, and to his left, Jason Hoyt, Chief Patient Officer. Welcome both of you. Thank you, Andrew. And so maybe to begin, can you please discuss your pipeline, the Stokes story, your lead program, and milestones over the next six to 12 months? That would be helpful. Yeah, thank you, Andrew, and thank you for those that are joining us both on the webcast and here in person today. Stoke Therapeutics is a company that focuses on genetic medicines and genetic diseases. Our primary focus at this point is a medicine for Dravet syndrome. Dravet syndrome is a genetic epilepsy, and we have a medicine that goes to the root cause of the disease and addresses a protein that is not expressed in the sufficient levels, and we upregulate that protein to go to the route cause of disease and potentially provide clinical benefit with our medicines or even nursing. That medicine is in Phase III study that I'm sure we'll talk to today, Andrew. Thank you. Beyond zirivinersan, we also have a next medicine, it's an autosomal dominant atrophy, which is a genetic progressive loss of eyesight, where patients with ADOA also move unfortunately to being legally blind. And our hope with that medicine is actually to restore vision. So not just to stabilize vision, but potentially restore vision for those children and older adults, younger adults, have actually lost their vision. It's a very similar approach to the DRAVE approach, where we upregulate a protein that is insufficient in these patients, and given the success we've had in DRAV, we anticipate that the same approach in ADOA may result clinical benefits and may restore vision for these patients. And then we also have a number of other earlier stage efforts based on mainly haploinsufficient diseases. What that means is they do express protein, but you don't express the protein to the level that they need for it to be a healthy child, mainly. And so we hope to address those as well. They're primarily in the neuroscience area or neurology area, and I look forward to talking to you about those in the future. We just came off our third quarter conference call about two weeks ago, where we updated the investment community on our progress with our Phase 3 study. I look forward to doing that today and taking all your questions. And once again, I thank you for joining us today. Great. And so maybe to start to level set the market opportunity for Zerva-Newson data, Phase 3 data, pivotal data, third quarter of next year. Existing drugs for epilepsy reduce seizures, obviously, but none of them benefit neurodevelopmental cognition. Your drug seems to do that. So maybe first, can you speak to what that means in terms of pricing power? How much could that price be if you did show cognitive benefits? That's a complex question, Andrew, because I would say it's not when you asked the question as in, can you help us understand the pricing power if we show cognitive benefits, I would deposit that we've already shown cognitive benefits. And so this answer deserves kind of a little unpacking of what I mean by that. So we've studied Zoriva-Nurson in a phase 1-2 study and over into an open label study. The patients that were in the phase 1, 2, 90% of those patients chose to roll over into an OLE study. And we've been able to study those patients for a further four years after the phase 1-2 study. And studying those patients für four years, what we see in the data is a continued reduction in seizures. At the doses that are more consistent with the phase 3 dosing, we see seizure reductions of 80%, 85%. So that itself is very important in terms of what this medicine benefit may provide to the patient. But what it also allowed us to do was to study the cognition and behavioral benefits over a four-year period. And that data has, each year, over many domains of what's called Vineland-3, has been able to show that we've improved receptive communication, expressive communication, personal behaviors or personal interactions, and also interpersonal skills. And so all of that data has been collected from a body of patients ranging between 60 and 70, 75 patients, I believe it is, over a period of four years. And so, all of dat data has actually been taken together, both the seizures and the OLE data, the cognition and behavioral data from the OLE, and actually shared with a payer. We've done a lot of research as well. I'm going to ask Jason to comment on that in a moment. But the payer looks at this in terms of the totality of the data, not just as in do you need to hit your end point in the 52-week Phase III study. And so we already believe that we've shown that with this total body of data that this medicine, because it goes to the root cause of the disease, There's a plausible mechanistic pathway in terms of addressing NAV1.1, the protein that needs to be expressed. This medicine has both seizure reduction benefits as well as cognition and behavioral benefits, and that's the position on the value of our medicine. I don't know whether you want to add, Jason. Yeah, no, I would just say that, you know, when you think about promoting a drug like zarivinersin, obviously you're tied to what's specifically written in your label. But I think when payers are evaluating coverage policy for a medicine like Zarevenersen, they're going to look at the totality of the evidence, right? And I think the label is one piece of that, but they're also going to work at the published literature, what's in the public domain in terms of peer-reviewed presentations at scientific meetings, and look at holistically what this means. And so when we put different scenarios in front of payers and say, what will be the most important piece of evidence to you as you think about writing a medical policy on the other side of phase three data and approval? they say the long-term open-label extension data is more compelling from their perspective than a 52-week endpoint because they're thinking about a chronic lifelong treatment for that patient and what it's going to mean from a safety and efficacy perspective over the long haul. And so what does that mean for price? So from a pure pricing perspective, I think what we've been using from an analog perspective are other genetically targeted disease-modifying rare disease assets, drugs like Spinraza, another intrathecally administered antisense oligonucleotide for a neuromuscular condition in SMA, or the exon skippers from Sarepta, you know, another genetically targeted antisense product for DMD, and, you know, even the likes of Debut for Rett syndrome, I think, are the three analogs we've been using most recently as the right range to be thinking of for a drug like serifiners. Makes a lot of sense. And in terms of the patient prevalence in the U.S., how many patients are there officially diagnosed, and how did you come up with that number? Yeah. So from a diagnosed perspective today, what we've been saying is there are about 6,000 immediately addressable patients at the time of a potential approval. And you can come at that from two different ways. You can either come at it from an epi perspective, where we think there are about 16,000 total patients in the US. Of those 16, 000, 6, 000 would be 25 and younger. 25 and young matters because more often than not, the pediatric provider is the one diagnosing a patient and often follows a patient all the way through to the point where they transition to adult care. But because of that bond that's been established between the caregiver, the family, and the provider, there's a reluctance to transition to adult care, so often they stay in the care of a pediatric provider into the mid-20s. And that's the group that's most familiar with Dravet, most familiar with the cutting-edge technologies that are coming in Dravet and, you know, staying up to date with the literature. And so those would be the most likely to be, you know, in the care of the right provider that's going to be most familiar at the time of a potential approval. If you also look on the claims database side of things, there are about 6,000, just less than 6, 000 total patients that have ever had an ICD-10 code ascribed to them for Dravet syndrome. So there are About 6,00 addressable patients that go across all the different ages with Dravet, right? There are about 40 to 50 new patients being diagnosed every month. So 500 to 600 new patients coming into the claims universe every year, and that's across all ages. We see about 30% of those are adult patients being identified. So not only is the age of diagnosis moving earlier, you're also seeing adults finally coming to the right diagnosis, which is encouraging. Very good. And even within that 6,000 patient pool, you mentioned ExxonSkippers, Exxondas is like a million dollars per year. So that alone is a multi-billion dollar blockbuster. Okay. So point well taken, you cannot take away your existing data as it relates to payers, although you still need to succeed in phase three to get this drug approved. Fair. The drug needs to hit its primary endpoint, high-confidence primary and secondary endpoints. But when you say succeed, Andrew, I do want to define that as in you need a P-value in your primary endpoint. Got it. And so you mentioned earlier back in phase one, two, at the phase three equivalent dose, 80% to 85% seizure reduction. So maybe talk about what exactly you saw on those individual violence scores, what kind of point benefit, and what do untreated patients show over time? Yeah, so thanks for the question, Andrew. It's a fairly long and maybe detailed answer, but the best data I could point to would be data that was presented at a medical conference in August last year at EPNS, where we were able to identify patients that were dosed at levels that were similar and consistent to what we're dosing in our phase three. That's approximately 225 milligrams, 230 milligrams of drug. And when you look at those patients, there was about 13 of those patients. And over a period of 68 weeks, we saw vinyl and cognition and behavioral benefits of points of anywhere between 8 and 11 points. And so that was very compelling. And just to give a reference to the 8 to 11 Vineland points that we saw, we also ran a natural history study over a 24-month period and measured Vineland. And when you went to these same domains within Vineland, measuring expressive, receptive communication, interpersonal skills, personal skills, what you actually saw was no improvement in the patient. There was sometimes a deterioration, in fact. But over that body of patients in the natural history study, I think it was around 30 to 35 patients, they basically were kind of level throughout that period. And just to help people understand Dravet syndrome itself, this is a disease where, yes, there are seizures that these children are affected by. But they neurotypically develop through the ages of around 18 to 24 months. once they get to that age and fortunately they stop developing neurotypically they flatline in terms of where they are in life even though they may progress on to be a 10 year old but they still neurotypyically they are a two-year-old and so what our natural history study showed was we on average the the age of the patient in that study was around 10 years of age and it showed those patients that when measured over a period of 24 months that they actually would continue to unfortunately be stable in terms of they were still like a two-year-old or an 18-month old. And so the fact that our medicine can be treated in a child that might already be 10 years old but have a neurotypical development age of a two year old but you give them back gain of function And gain of function could be things like before the child's on the medicine, they may be able to process two words. That may be mama or no. And we have some patients that have been on drug now for four years, and they have the use of many more words and including sentences. They may be allowed to tell their parent that they can go to the bathroom. They have a pain in their tooth, and therefore, you know, you've got to go tothe dentist, unfortunately, for them. but at least they can say they've got a pain in their tooth, and then they're getting motor skills back as well, and so the compare and contrast, actually, of the natural history for this disease compared to what we actually see when dosed with our medicine is they gain tasks and function, and because the medicine is a chronically dosed medicine, so therefore continues to express NAV1.1, the protein that these youngsters are short of, they gain function over this chronic period of dosing. And that's what our data has shown. Now, we're involved in a phase three study, 162 patients. We're doing well in that study, and we've got to hit our primary endpoint of seizure reductions. And we believe we will reach our secondary endpoints that includes cognition and behavior measures. Perfect. And then speaking to the primary endpoint again uh week 26 or week 28 caesar reduction how did you power that on a placebo adjusted basis so maybe i could give you the overall pairing because we had a recent interaction with the the the u.s fda um a couple of months ago uh but the outcome of that uh that meeting with the fda where we went and presented our statistical analysis plan because we're in the midst of a phase three, was that the study is powered for its primary endpoint for seizure reductions or treatment difference of approximately 46% to get a p-value of less than 0.05. I'll remind people that our data so far has shown seizure reductions compared to baseline of around 80%, plus or minus. And so we feel as though we're well-powered on that primary endpoint. And on the secondary endpoint of Vineland, we are powered to show an improvement of 1.8, 90% confidence level if we have an improvement greater than 1.80. The p-value would be less than 0.05. As I mentioned, the EPNS data we showed last year showed improvements in Vineland of anywhere between 8 and 11 Vineland points. That was a dosing that was similar to what we're doing in Phase III. and it was over a 68-week period. I bring all the total powering and the statistics up of the study. It is a question that's often asked by the investors, both on the primary and the secondary. But the reason it's being asked so often is we have moved that secondary endpoint in agreement with the FDA from one that was once a hierarchical endpoint where you had to step your way through each Vineland domain in a hierarchal order. What that actually means is that if you missed your first domain, you know, on a single domain within Vineland, say measuring expressive communication, that you don't get to measure the other three or four. And, you Know, that feels a little random, to be honest, in a disease that manifests itself with multiple comorbidities of cognition and behavior. Those manifestations also evolve at different rates within each patient, and the medicine also affects them at different weights. So what we did have agreement with the FDA was a basket approach of four domains that equally get measured. And you basically measure each four separately. You add them together. You divide by four. Yes, it is as simple as that. And it gives you what's called an MC4 score. That MC4 school is what you do your statistical p-value on. and a nice output from that was also that it enhanced the powering of the study because it's more valuable data because you're measuring four different domains each time equally and in a statistical world, more data, more valuable data narrows the standard deviation because it takes out variability of the data and therefore has enhanced the Powering of Study for that important secondary end point of cognition and behavior. Makes sense. Thanks. And so one point, your power to show on Vineland, a 1.8 placebo-adjusted, give or take, not placebo, sham-adrasted separation, what do you think sham would do? You just said untreated patients, maybe no gain over time, but what do... Yeah, I mean, that's what the butterfly data, sorry, the natural history data shows. It shows that treated, looking at patients measured over a 24-month period, there is no gain. And that is what's kind of known for the disease. And so, but, you know, as you mentioned before, Andrew, with the EP&S data showing anywhere over a 68-week period, anywhere between 8 and 11 violin points, and that was only just 13 patients, you're needing to achieve greater than 1.8, you know, we're well-powered based on what we've seen in the data previously. And so because now it's a composite score of four subdomains, even if one out of four subdomaines did replicate that 10-point-plus benefit over baseline placebo as zero, if the other three show no separation, if I divide 10 by 4, that's 2.5. That's above 1.80. Is that the right? Yes. I mean, obviously, it depends on the variability and the standard deviation within the data, but that is a way to think about it. In fact, just on our Q3 call a couple of weeks ago, we put up a slide and some data to help the investment community understand exactly what you just said, because we've done a propensity match scoring and p-value assessment of our data compared to natural history. And when measured at week 68, the p-value that you attain on MC4, that's the multi-component four measure, was significantly smaller than those on the individual domains. And that's where you get the extra powering from. And then, in fact, we also have data where we measured against month 12. Month 12, the patients were suboptimally dosed. they have roughly just more than half the drug we're delivering in the phase three. But when you look at those four domains, you actually miss on a couple of those domains in terms of a p-value of less than 0.05, and you hit on one of those domain. Still, when you measure the MC4 statistical output, that gave you a p value of still less than .01. So it's because of the combination of all the valuable data that also leads to a greater amount of data which reduces the variability and narrows the standard deviation and therefore enhances the power. But the most important thing with this measure is actually moving away from what could have been random in terms of the hierarchy of assessment of these individual domains and being able to look at it four different ways that are all important to caregivers, caregivers being the parents, but all important to the caregivers and the prescribers, the physicians. And those are the four domains we're measuring today. And may I ask, you recently met with the FDA to change to MCI-4. Why did it take until now to get that type of alignment? it's um see i i i look at these things broadly as the the fda is your friend and what i mean by that is the friend is that you need to work with the fta to explain the disease first of all so they have an understanding and you also need to Work with the FDA to explain them what to them what your medicine is doing you know starting with plausible mechanistic pathway root cause of a disease, and then sharing data to truly educate them about your medicine. And when we had the phase three agreement with the FDA, which was over two years ago now, or roughly two years ago, the FDA pushed the company into a hierarchical assessment of these Vineland endpoints. Company accepted that at that point, rather than challenge it, rather then talk about the disease, talk aboutthe medicine. They accepted that. Anyway, we progressed forward. And over the last two years, importantly, this time last year, and you'll probably remember this, Andrew, but this time this year, we went to the FDA to see whether they would support an NDA filing on the basis of our Phase 1, 2, and 3 years worth of OLE data. They pushed us to continuing to work with our Phase 3 program. Okay. Disappointing outcome for us and maybe for patients. But, yeah, we're focused on our Phase 2. But one of the really important outcomes of that meeting was increasing the understanding for the FDA and the education of what our medicine is doing in this disease and also what the disease is. So when we went to the FDA just two months ago, we were concerned about this hierarchy. We were powered that we were still confident to hit it, but rather than measure it on a hierarchical assessment, it was more appropriate to measure the benefits of our medicine by looking at multiple endpoints of cognition behavior. When explaining it that way to the FDA, they accepted the MC4 and supported it. So that's where we ended up. It was a great outcome for, I'm going to say it's a great outcome for the patients because it allows us to run the study and measure the outcome from that study in a more appropriate way to measure the benefits of our medicine. Mm-hmm. Thanks. And, you know, it's ultimately it's an benefit-risk equation. Benefit's clearly potentially there or is there. Risk, how should we think about safety AEs, SAEs, any color on that front would be helpful. Well, obviously the phase three is ongoing. As of two weeks ago when we had our Q3 call, when we updated the investment community, the progress of the trial, we had recruited 162 patients and we had zero dropouts. And of those 162 patience, 130 of the 162 were already through week 24. The importance of week 24 means that you've already had three of the four doses in total in the 52-week period. And so this was two weeks ago, and the primary endpoint in this study is at week 28. So two weeks before, we were four weeks away from 130 patients being through the primary end point. So the progress of the study is very good. And the second part of the question, I'm sorry? Oh, the safety. I'm Sorry, yes. So one, I can tell you the interpretation of zero dropouts is the drug is being well tolerated to date in the phase three study. But the other thing in terms of data supporting the profile of the drug is the four years of OLE data. These patients have now been on drug effectively five years. And there was approximately 90% of the 75 or 80 patients that rolled over into the OLE. And we've consistently measured these. And that profile coming out of four years' worth of data, which is now probably over 1,000 doses of this medicine, certainly was over 900 recently, maybe getting over 1 ,000 now, has been shown to be well tolerated. I think it's very rare that in drug development that you have a genetically designed medicine that has four years of safety and efficacy data, and you're running a phase three study. Now, we do have to run the phase three studies. It has to be successful for this medicine to be approved, but it is a unique situation to be able to have this valuable data to help confidence in understanding how to design a study, but then what the output of that study may be. It's great. And then, to me, zero percent discontinuations also means you've increased your study power as well. That is correct. Okay. I think I like the chances here of the positive outcome, Q3. So thank you for the update on Stoke, and wishing you best of luck, and thanks everyone for listening. Yeah. Thank you. Thank you, Andrew.
- Company
- STOKE THERAPEUTICS INC
- Keywords
- STOKE THERAPEUTICS INC