Modality.AI | A Turning Point in ALS Trials

To drive a turnaround in trials for ALS and other neurological diseases, we at Modality advocate for new tools that increase understanding of the course of disease and response to treatment. Together with MGH, Everything ALS, and, more recently, Verge Genomics, Modality’s platform is building the most comprehensive audio/visual dataset of people living with ALS. In this 45-minute call, we discuss advancements in digital endpoints in clinical trials and the potential to accelerate clinical research. Our special guest is Karl Kieburtz, MD, MPH, Co-Founder and Chair of Clintrex and principal investigator for numerous multi-center trials in neurology.

Creators and Guests

Host

David Suendermann-Oeft

David Suendermann-Oeft is the Founder and CEO of Modality.AI, creator of the world's first self-driven neurological assessment. A scientist, engineer, founder, and former tenured university professor, he has 15 years of experience leading industrial and academic R&D departments in AI, machine learning, speech and signal processing, NLP, multimodal technologies, and dialog systems with a publication track record of over 170 books, patents, and articles. David held C-level and senior management positions in industry and academia in the US and Europe, has built enterprise-grade speech and language technology systems (totaling over one billion interactions), and serves as an evaluator and innovation expert for the European Commission and the National Science Foundation.

Guest

Karl Kieburtz, MD

Co-Founder and Chair of Clintrex and principal investigator for numerous multi-center trials in neurology.

What is Modality.AI ?

Welcome to Clinical Trial Insights and Innovation, the interview series where we talk to experts in medical research. Each episode features conversations with professionals who are actively involved in designing, conducting, and analyzing clinical studies. Our goal is to provide insights into the process, challenges, and latest developments in clinical trials and research studies. Subscribe now.

DSO: 00:00

Share my screen here. And I think I can start the webinar on that very important day, which is the 21st of June. It marks not only the beginning of the summer, this year as usual, but, but also, the ALS Awareness Day. That brings us to a very important topic that we wanna discuss today with our good friend, colleague and mastermind, Carl Kieberts, who is the president of Clinrex and a professor of neurology at University of Rochester. I will provide a more comprehensive introduction into his background and what we will talk about, today in just a few minutes.

DSO: 00:47

But it is specifically about, the conduction and conducting of ALS clinical trials and the use of digital endpoints such as modalities in these interventional studies and, how we can learn from, the the past execution of such trials for what we can do in the future. But before doing so, let me just go, through a brief agenda for today. It's a 45 minute session. We wanna introduce, you, to, specifically a few, few pieces about Modality's background and then start the conversation with Karl. We also allow, of course, for questions, by the audience.

DSO: 01:32

So please feel free to put them into the chat, and, we will close at 11:45, Pacific today. So just brief overview. We are talking about clinical trials, specifically in ALS and pain points in these clinical trials. We know that many of these studies fail. It was just at NCALIS last week and we, saw the presentation of top level, top line results of 5 studies that failed in ALS.

DSO: 02:02

And so, these failure rates are very high, phase pre studies, for example, over 95% in ALS, studies have failed in the past years. So and that is due to many factors. One of these factors are inadequate assessments in clinical studies. And, this is also because clinical studies today use endpoints such as the LSFRs R that have many, characteristics that make them not ideal endpoints for the studying of effects in, in these clinical studies. And there's also no not very reliable lab tests that are available today, but more about this in our conversation.

DSO: 02:45

Today, Modelli specifically has built a platform that uses a virtual guide to interview patients. And we can tell from what patients say and do, how they progress through disease. It's a digital, biomarker, or digital endpoint that is being used in clinical studies and clinical trials. We look at speech language, facial movements, body movements, cognition, and also at what's called the prop the patient report of problems, where we ask patients to describe their most bothersome problems in their own words. And that's actually delivered on anyone's device, usually at people's homes.

DSO: 03:19

So you use your laptop, your smartphone, your iPad, and you talk to Tina in the convenience of your home on a regular basis so we can actually measure disease progression and treatment response in these interactions. Just very quickly, we're working with some, of the big pharma companies, biotech companies, you know, sponsored by the national government here in the United States and the other foundations. And they have a pretty significant pipeline of diseases we're looking at, specifically in ALS, and Parkinson's disease. We're highly published and have worked with almost 80,000 patients over the past years in, validating, verifying this technology. And we're not going to talk about this technology only today, but, definitely about the, opportunity in digital biomarkers, with Carl today.

DSO: 04:08

So I was promising to you that we will, introduce Carl a little bit more, comprehensively. Carl is a professor of neurology at the University of Rochester. Has, worked in Rochester for many, many years and, specifically with with many patients, of course. But in many other capacities as a researcher, there's hundreds, hundreds and hundreds of papers that are associated with his name, specifically in Parkinson's disease and other neurodegenerative diseases as well. He's the founder and president and chair of Clintrax, which is an organization that has worked with over 150 organizations, specifically from industry, to develop products in CNS disorders and on various stages of the development, but specifically also in the clinical stage, which we will talk about today.

DSO: 04:58

He is the PI of over 50, clinical trials, that were sponsored by the NIH, the FDA, but also, of course, by the industry, has a lot of experience in that specific field, experience about pain points, but also effective solutions. And, he has as an adviser to various US governmental institutions like the FDA, specifically, but also the NIH, the Veterans Administration, and others, work with regulators directly, and also understands not only the pain points of the pharmaceutical industry, but also those topics as they are looked at by the regulators as well, which will be topics of today's conversation. And now, without further ado, I'm gonna head this off. I'm we've stopped sharing my screen here and, wanna say hello. Carl, how are you doing today?

Karl: 05:55

I do. So good. Good. How are you?

DSO: 05:57

Excellent. I was really looking forward to to this conversation. It's gonna be fun. And I wanna start asking you, about, you know, one of the pain points I I discussed. You know, it's been perceived and it's actually reality.

DSO: 06:12

That clinical trials in neurology are long and therefore costly. They need many patients. They are prone to failure. Compounds don't seem to work. You know, how much truth is there, and specifically in ALS, to, to to this.

DSO: 06:30

So so what what is is happening with clinical trials? Are clinical trials failing, or is it that the that the drugs aren't working?

Karl: 06:39

Yes.

DSO: 06:41

I think

Karl: 06:42

it's it's a little bit of both. Well, you know, CNS drug development in general, as you said, is is associated with a higher failure rate, a longer run time for trials, and, more cost. And, there's a couple of things that intersect here for ALS in particular, but also shared with other neurodegenerative disorders, is that we don't know exactly what the biology of those illnesses is. And we we we have a lot of hints. ALS is typical for that and having we know individuals with SOD one mutations, for example, can have a genetic form, but we know that that's kind of a minority of people.

Karl: 07:26

So the underlying biology that's behind the neurodegeneration and the disease remains a bit enigmatic. When we don't understand biology well, we're less likely to be able to bring forward highly potent therapies. It has happened in ALS. There you know, there's a a list of now approved drugs, but the list of failed drugs is much bigger. Some of that is it is undoubtedly due to the fact that we've tried things out of some degree of ignorance, out of hope, but out of also some degree of ignorance, and they they just don't work.

Karl: 08:05

But there's another piece, which is the way we do trials may be contributing to the failure, but also to the length and the cost of those. And, there's a the things we've chosen to do in trials in ALS are not without reason, but they they do contribute to these things. So for example, the typical kind of outcome measures we use, like the ALS functional rating scale revised, the ALSFRSR, is a very classic kind of, clinical outcome scale we would use in other areas and specifically in ALS. And it has a lot of features which might contribute to the problem. One is it has to be done by an expert who is familiar with the illness.

Karl: 08:55

So there's a demand on who the operator of the scale is. It generally requires an in person assessment of people, not entirely, but in general. So there's expertise, there's time, there's commitment to facilitate the evaluation. This is true of a lot of neurologic diseases, but it's particularly true in ALS. And that contributes to the time and cost because of the intensity of that kind of evaluation, the requirement for a clinical expert, and that these scales evolve over time, not slowly, but not exactly quickly.

Karl: 09:30

So, that's kind of a long winded answer, but I I think ALS does capture 2 things you mentioned. 1, drugs may not be working because of an incomplete understanding of the illness, but also how we measure their effectiveness is not optimal yet.

DSO: 09:48

Yeah. It seems to be, so important, to be able to to measure things in a much more optimal fashion. So so to be able to answer, actually, the the first question, which is, is the drug walking? So, so the better we measure, the the better we can answer that question. And, but without measurement, we cannot actually handle that first question, properly.

DSO: 10:06

So so you're pointing out the LSFRS R is, still, the measure to go the gold standard, when it comes to as an output measure outcome measure first and primary outcome measure in clinical trials alongside, survival, sometimes as a combination, which is potentially even worse measure because it's binary and requires people to die, which is, terrible. So given given these down these these negative qualities, of of these measures, why why are they being used? Why why is it not that, there are better measures which are today in use?

Karl: 10:44

Yeah. It's, I'll just say, it it seems a little crazy, but this this whole field of needing to provide evidence around the approval of drugs only harkens back 60 years, in the United States to this Keith Alber Harris Amendment, which created the need for the FDA to be able to find substantial evidence of efficacy For the approval of drugs before that, there was a different standard. Now it's the substantial evidence of efficacy. And the substantial evidence of efficacy is provided by experiments, clinical trials, which provide evidence of impacts on how a person feels, functions, or survives. And so we've developed scales that measure those things.

Karl: 11:38

Survival's fairly easy, but

DSO: 11:43

a both

Karl: 11:45

irrevocable and undesirable outcome. We don't want our patients to die, but that is an outcome that does happen, of course, for all of us, but, and we can measure the time to that, not only that it happened, but what we're largely met, left with are measures of how people feel or function. And the ALSFS, FRS R is a measure of function. And because it provides functional data, it's able to be the basis of approval for new drugs and has served as such on more than one occasion, often combined with, as you said, with mortality data and the so called joint rank analysis. So there's a reason why we use these scales, because they fulfill this regulatory need.

Karl: 12:33

But because of that nature and the way they were developed, they have intrinsic variability and sometimes a lack of intrinsic applicability to the individual. They are scales that exist for ALS, not for a particular person's ALS. And so people progress in different ways on those kind of scales, and then there's all this additional contribution of variability around how the scales are actually completed by the people who use them. So there's an operator characteristic, and then there's an intrinsic characteristic of the scale itself. But the reason we have scales such as this is because they give us information about function.

Karl: 13:17

There are more precise things we can measure, But the problem with is you there's a there's a tension between precision and meaning, meaningfulness in the FDA sense, meaning it measures how a patient feels, functions, or survives, whereas we could measure, you know, very precisely a lots of things, heart rate, for example, skin resistance. There's there's biophysical, biochemical, biomarker measures, which could be quite precise, but their meaning their meaningfulness regarding this other set of things, fuel, function, or survive, is uncertain. So connecting those two things, precision measurements with meaningfulness, has been

DSO: 14:03

a big challenge, particularly in neurodegenerative diseases. Which makes, one question. When you look at the endpoints that are being used in typical interventional trials now in ALS, We mentioned primary endpoint, usually ALS, FSR, or or and or survival. And secondary endpoints, often things like, maybe, muscle strength or force micro capacity. Or what we see a lot now, neurofilament light and other fluid biomarkers that are being investigated, sometimes as tertiary endpoints also.

DSO: 14:39

And one would also question how those, for example, are related to function or even feeling of survival. They are being deployed now. How do you see in the context of what is being done today with these end points and the emergence of digital end points, that can potentially answer these questions while providing answer questions about, how patients feel and how they function, but combined with much stronger statistical properties. So so how do you how do you see the field evolve and and and the opportunity, of using those?

Karl: 15:23

Yeah. So a 2 part answer. The the first part is, yes, indeed, there are secondary outcomes now. Slow vital capacity, manual muscle testing of strength, you know, slow vital capacity does relate to time to ventilation and potentially to survival, so it fits in with that kinda FDA framework. Manual muscle testing or strength is an intrinsic feature of the disease, so we think that if there is an improved function and improved survival, we should see improvements in that.

Karl: 15:58

But it is, again, theoretically, a more precise kind of measure of the phenomena of the illness without necessarily having intrinsic meaning because we it has to roll up to some kind of function for survival. Digital biomarkers, present the opportunity beyond manual muscle testing and so forth to actually look specifically at motor performance in very precise ways that we couldn't perform. We did in the past, we've spent a lot of time standardizing rating and training of raters to try to get quantitative capabilities in place. But as we're doing right now, we you know, and it has really accelerated over the last decade, is the ability to use other modalities to assess precise motor performance, that does not require a human interface. So if I tap my fingers like this, I can tell you, you know, is that a little slow, very slow, not slow at all?

Karl: 17:07

I mean, that's a human interface. Whereas a digital assessment of that can tell me exactly, you know, how many hertz I'm tapping it with with great precision. So the human interface is qualitative and approximate. The digital interface is highly accurate and quantitative, and reliable and repeatable. So the digital and and that's just an example of a digital interface to create a digital biomarker, but we have a lot of and and as you alluded to in the modality assessment, some of those are directly relevant to ALS.

Karl: 17:44

So for example, the ability to speak clearly, the speed at which you generate speech, the speed at which you breathe, and your ability to breathe and speak simultaneously, measurements of facial motility. These are all things that can be impacted, in in ALS and in other diseases, but we're talking about ALS. So we've created this capability over time of being able to capture these phenomena with great precision and with great, ease of use. So, typically, to get a expert who's been trained in a rating scale to see you do certain things, you would have to go to see them wherever they are. They might not be near you, but through to that question about,

DSO: 18:42

you know, how clinically meaningful these things are.

Karl: 18:45

Right? Yeah. I'll get to that in a sec. So there there so you can do this, you can do this ubiquitously, anyone who has an Internet access, and and in case the modality assessments, all you need is an Internet browser and a Internet connection, and things like eye blink and clarity of speech, and speech production can be measured with great, precision. But then you're stuck with this issue of, so what?

Karl: 19:12

Like, what does that mean? You've answered it very carefully. My old mentor used to tell a joke about neurologists. And the gastroenterologist were always thought to be, you know, diagnosticians who give you very precise answers that don't help you much, until a joke about a guy who was in a hot air balloon. He came down out of the fog and asked a farmer on the ground, where am I?

Karl: 19:40

He said, well, you're about 20 feet off the ground heading north northwest. The balloon guy says to the other guy, that farmer must be a neurologist. And he said, how do you know? He's a 100% accurate and a 100% useless. So there's a you know, you can be very accurate without being very helpful.

Karl: 19:59

So precise measurements of movement have to be contextual. They have to be given, they have to be able to move into the realm of clinical meaningfulness. So they alone, while precise, reliable, repeatable, need this context. So, and I think, you know, you kinda glossed over it at the beginning, but part of the modality assessments is this prop, this, seemingly very simple thing to say, but it's actually a complex thing, which is patient reported problem, is to ask people in their own words, what bothers them about having their ALS? Like how does their ALS bother them?

Karl: 20:48

What what what does it What do they experience that's problematic? This is kind of turned around from how we usually do scales. We usually do scales when we tell we get together a focus group, we ask people with ALS what bothers them, filter it through, and and then come up with things that we we understand them to have said, and then ask people with ALS, how much does item a bother you, item b bother you, item c, because these are the things we understand to have been perturbed by having a in having ALS, by having had these interviews with groups of people. But everybody's different. It was it was felt like if you asked an individual what bothers them about their illness, that would be so idiosyncratic, so solipsistic, you could not get generalizable information about that.

Karl: 21:41

But that's not the case. So the prop asks people in their own words, what bothers them the most, and then uses natural language processing and other techniques of machine learning to put those verbatim into analyzable domains, and further queries about that problem. How does it influence your ability to do what you wanna do? How does it impair your function? And what can you do to make that better or worse?

Karl: 22:14

So that kind of here, not my endorsement of a list of problems coming from some external source, but here's my report of my individual experience, lumped into a domain that could be lumped with other domains, and the functional context of that problem. As as the person is saying that, giving that verbatim, these other kind of quantitative measures are being captured simultaneously. So you have this simultaneous acquisition of what a patient is saying in their own words about what bothers them, and its functional consequence with these very precise performance measurements. Now, that doesn't necessarily give you the answer as to what does the precise measurement mean in terms of function, but it allows you to start simultaneously accessing both of those things. The precise measurements and the things the patient says is bothering them in a idiosyncratic way, but also in a generalizable way, and the functional consequences of that.

Karl: 23:31

So generating those data simultaneously over time starts to create the database that allows you to understand the functional meaningfulness of the very precise data. So collecting that conjoint dataset over time in a observational way, and also over time in an interventional, starts to give you the information that allows those very precise measurements to become clinically meaningful in the FDA sense.

DSO: 24:05

Well, thank you very much for for providing this, extra context. And, by the way, I wanna, just remind everybody that, you are, free to provide some of your own questions, about what you hear from from Carl in in the q and a window. And while we are looking at this, I wanted to follow-up on on what you just said, Carl, about, you know, the fact that, in in this specific case, we we have, measures like, say, speech measures that are, that have intrinsic meaning, clinical meaning potentially, because, you you care about as a patient, as the people that you're that you're surrounded by, you care about communication. Others are maybe not as intrinsically meaningful, and, but they can be backed up by, for example, the prop that you mentioned that provides intrinsically meaningful, information around how patients feel and how they function and bring this into context. So the question is, it is still, something that when, when you look at the history of the LSFRS R that has been around for decades, when you look at the history of say fluid biomarkers, like the neuro filament light, that has been studied also for over 20 years now.

DSO: 25:23

And so there's a significant history of, of data that was produced on these sorts of measures that the community has accepted, that regulators are starting to accept, for example, for the NFL. So what is the motivation, and, in your case, it's a suggestion of, say a sponsor, not, not only by the federal government, but specifically so also by the industry to deploy measures such as these, modalities measures, for example, in interventional trials, not just observational studies, where there's many, which this is being studied, to provide these body of data that you mentioned, but also, in interventional studies. So what are the benefits, for example, in early stage studies, but also in late stage studies, to use these digital endpoints today?

Karl: 26:13

Yeah. So, you you just you kind of made a a general question about motivation, which I will make a passing reference to and then answer your question more, specifically. You know, really 125 years ago, it was perfectly acceptable to get around by horse. And that, you know, there was a whole set of industry and ways, and and when, you know, people, you know, at Camry Ford, started making cars, people were like, what do we need that for? Like, yes.

Karl: 26:49

You can go farther. Yes. You can go faster. But why would you really need a car? What would you use that for?

Karl: 26:55

These these were questions that were asked. Like, we've got a perfectly good transportation system right now. Railroads are the horses. Why would you need anything else? There might have been something good about that for climate change, but here we are now.

Karl: 27:08

So I think the tyranny of the now is something that we always have to face, which is we've got it. It works. The FDA accepts it. Why are we gonna do something different? Because be because we have to.

Karl: 27:23

Because we have to be able to do this more cleverly, more efficiently, more in a patient directed manner, so we get treatments to people that address their needs. We are doing a good job, we're not doing a great job. And we can only get to be doing a great job by innovating and finding things that are meaningful and make sense. And so I think we always have to strive to do things better, no matter how well things are working right now. There's a reluctance and a worry about doing those things in a regulated environment where you have to get drugs approved.

Karl: 27:59

And so using these kinds of new measures as primary outcome measures in pivotal studies, nobody's gonna do that today. It's just too risky. On the other hand, they could be extremely useful today in earlier drug development. Earlier drug development is not so much around showing that substantial evidence of it before, but trying to make decisions on company basis such as, is one dose more effective than another? Is there any sign of benefit here?

Karl: 28:37

So quantitative digital biomarkers can be very important about that, at finding how doses compare to one another or differentiate from placebo. And I think that is actionable and useful right now, and people are doing that with other kinds of digital measures. I I I think the the interesting thing about the prop too, particularly, so there's the digital measures and there's the prop. I think the prop is very useful, because we often don't understand all the facets of the intervention's impact. So, yes, we're looking at it to improve, let's just say, motor performance, but maybe it has an impact on mood.

Karl: 29:22

Maybe it has an impact on sleep. You know, minoxidil was for the treatment of blood pressure, and it made hair grow on people's heads, and that, you know, that was important. But if you're not asking what people are experiencing, you're gonna miss some of the things that are important to them. So the PROP, as an assessment tool in early development also helps you understand the context of experience, and the impact of that on function. So you start to understand the relationship between these things.

Karl: 29:59

I think that's very important in drug development. Early phase development is making choices, and it may also be on the safety side. People might start saying, you know, what's bothering you the most about having ALS could be nausea, headaches, things that they weren't having before, but are coming because of the medication. So it's a way of of capturing in a patient's own words, potentially, the adversity of what's what's happening to them. So that's a little trickier in parsing that from the the disease features themselves, but there's no substitute for asking for people in their own words what's troubling them.

Karl: 30:38

So I think these kinds of things, there's there's information about what people are experiencing in their own words in early drug development, as well as quantitative assessments of outcome measures in early development, can be very useful on dose selection, or in deciding, you know, the net risk benefit portfolio just doesn't justify going further and making those important kind of no go decisions faster and more efficiently. So, half of drug development is is stopping soon enough when something's not working and not you wasting the resources going further when it's not gonna pan out. And of those things that are working, choosing the best ones among them.

DSO: 31:21

Yeah. It's interesting you say this, because, for example, we conducted a study with, Johnson and Johnson and the Veterans Administration on mild cognitive impairment. It's not ALS, but, it it illustrates the point that you were just making, where we compared health controls with people with mild cognitive impairment and asked them, not only going through facial and speech measurements as we discussed, but also the prop and found actually, when looking at the prop results that the risk of falling had multiplied by a significant factor, something in the order of magnitude 10 or so, between people that have mild cognitive impairment and the ones that are, that were age matched and gender matched controls. And so that is something that no one would have expected beforehand. You know, when when things, about cognition and, you know, measure your cognition, measure your, these sorts of related condition related, things, but but no one would have bothered, asked people in a scale, say, about, you know, how often do we fall.

DSO: 32:27

And and that's I

Karl: 32:30

I was just gonna say the prop was also deployed in Parkinson's disease, and, it turned out motor and non motor problems were, you know, the top ten things, 5 are motor and 5 are non motor. And and the single most common non motor feature is pain. And any textbook of Parkinson's will tell you that pain is not a big problem in Parkinson's disease. It's, yes, it comes later, or yes, it's but it's the most common non motor problem. That who would have known?

Karl: 33:03

If if you're not asking people directly, because you already have adopted this mind state, that that isn't necessarily a big problem here, so you don't inquire about it when so sort of a natural closed problem.

DSO: 33:16

Yeah. It's it's very interesting how you described, how actually the scales were developed, asking patients in the early days, asking patients, you know, what bothers them. Then we pick the most important things, and then we define the scale, and then we forget about what really bothers them. We we stop asking. And there's there may be also the development of of treatments and so on.

DSO: 33:38

There may be many other things that's, that now bother them more as these early symptoms, for example, are being addressed already. That's really interesting. You had a question by the audience here, about the prop and specifically around whether Tina actually initiates, the prop also. It's a good question because I was talking about, Tina, the virtual guide that actually does the assessment. So there is no clinician involved, so which is a significant advantage on multiple levels.

DSO: 34:08

Not only is she available 247, so you can do your assessment, you know, anytime, it is comfortable, according to the schedule of activities, of course. But also she's always providing the same experience. So there's a very standardized way of doing these assessments. So it's not like every investigator or every, you know, site has a different way of conducting these assessments. But but no.

DSO: 34:37

Tina is conducted everywhere the same way in a standardized form. And she asked these questions. Not only does she go through standard exams like the diagnosis task in ALS or things like sentence to the durability probe, but she also asked patients to describe their problems in their own words. And she takes, as much time as as you want, listening to you, which is also different from, most of the experience people have when they see their clinicians, who notoriously have less time than they should, listening and and caring about you. I I wanted to follow-up on on what you said about, specifically about the benefits of using assessments like these digital assessments in in early stage, studies.

DSO: 35:26

For example, for dose identification, identification of signs of treatment effect that can can inform you of whether and how to move forward. Maybe also which specific groups of individuals are affected more than others by the treatment effect. But how and that's really curious also. We we are involved in phase 3 studies. We are being approached by sponsors of phase 3 studies in ALS.

DSO: 35:55

There is a significant interest also in late stage deployment of these digital endpoints. How do you see, the benefits of using, for example, modality, as digital endpoints in later stage trials?

Karl: 36:10

Yeah. It's it's interesting. At least in since we're talking about ALS and we're talking about neurologic diseases, at least in the neurology division of the FDA, since we're talking about drug approvals, there is a precedent around a notion of the most bothersome symptom. So in migraine, migraine approvals used to be based on you you had to have measures of nausea and vomiting, photophobia, and pain, headaches. So these, these core features of migraine, and you had to the outcomes were measuring the impact on all those things.

Karl: 36:51

And the division shifted to having individuals with migraine identify their most bothersome symptom of migraine. It had to be you know, it couldn't be rash, for example, because rash isn't part of the clinical constellation of migraine. It had to be from a a set of things. But one could imagine the most bothersome problem that one reports with ALS on the prop becoming the problem that is the target of intervention. And because there is this regulatory history of elevating the most bothersome problem, to the most bothersome symptom in the language of the FDA, the prop may be useful there.

Karl: 37:50

I will say that in those approvals around most bothersome symptom and the relief from it, the endurance of the relief from it, you also look at other things. You look at the other symptoms and other analgesic use and so forth. So the analogy in ALS would may might, for example, be someone says their most bothersome problem is they're short of breath. Might be that they have difficulty swallowing. So that for them is their MBS or Bothersome Problem.

Karl: 38:23

And you monitor that over time, but you're gonna wanna monitor other things too, other problems they're having, quantitative measures of motor performance and so forth. But, so I think, again, it is possible that in late stage development, prop features, per se, could be the response of or efficacy, or you can kind of adjust that a little, is use the quantitative measure as the primary outcome. And there there are analogies for this in Alzheimer's disease where you use a measure of impairment, cognition, so you might use, for example, face and voice video as a some measure there, as a primary outcome measure. And a a key or co primary would have to be the impact. So measuring quantitative movements, and then seeing whether the most bothersome problem got better or the series of problems that are reported in a property better.

Karl: 39:32

So that gives the clinical meaningfulness in conjunction with the measure of impairment. So I think, actually, the structure of the modality assessments lend themselves to being capable of being primary outcome measures. Of course, I have to generate data on that, but there's nothing intrinsically about them that stands in the way of being recognized as a primary outcome measure.

DSO: 39:57

Well, thank you. That that, gives gives good color to to that question. Just a reminder for the audience, feel free to, provide any additional questions of you. Addressed already, during our conversation. But anything, we have just another 4 minutes or so left of the regular time.

DSO: 40:16

So please take that opportunity into consideration, because, Karl can speak for, definitely for the next 4 minutes and and and that would be a lost opportunity. So, just very quickly on the last question. Because you you mentioned the 8 stage trials, primary endpoint. Yes. A lot has to happen, before any digital endpoint in ALS is going to be used as a primary endpoint.

DSO: 40:42

We, we have not seen any, actually, any non scale being used as a primary endpoint in ALS trials, as in phase 3 studies, yet. So, so that's, that's the current situation. And, and of course, we're not even talking about a primary endpoint because that's something which is beyond current sponsor interest. But how can the use of exploratory endpoint or secondary endpoints inform or help sponsors even in phase 3 studies today. And I'm specifically interested to see because the FDA, and you have a lot of experience with regulators, the FDA seems to be more flexible, specifically in ALS with respect to how they look at data that's being presented to them in the request for a potential approval or conditional approval.

DSO: 41:38

We have seen, for example, in Emiliex's case, that approval was based on the readout of a of a phase 2 study, which is an interesting precedent. Of course, the the phase 3 study results that were presented, for example, at NCALC, did not live up to that the same results and Amelix decided to withdraw the drug. But it showed that the FDA was willing to take exceptions or is flexible in the interpretation of existing data. So I was wondering, for example, in a hypothetical case, that a phase 3 study provides good results, say in the phase 1, in the primary endpoints, but not statistically significant ones. How can additional evidence, for example, swing a decision?

DSO: 42:30

Well, of course, you're not representing the FDA, but but just, you know, how how have such arguments played out in ALS and and other, diseases in the past?

Karl: 42:40

Yeah. I I would say there is this, notion and experience with so called confirmatory evidence. So, other than a primary outcome measure, things that are thought to be very important. So in the case of Amelix, they had this open label long term follow-up with survival. In the case of Hereditary ALS, there was the NFL on top of it.

Karl: 43:05

So there there can be digital or biochemical biomarker data, in addition to the clinical outcome data, which buttresses it, gives it additional credence, especially if it's kind of flirting with statistical significance, that other measures related to the disease, which were quantitated more precisely, did indeed show evidence of improvement. So I think that kind of adjunctive or confirmatory evidence, and that's kind of a high bar to be confirmatory evidence. But to provide context and color with, could be supportive of a clinical outcome.

DSO: 43:44

That's a great, final statement, actually, because we we just hit the end of today's webinar. We will answer any outstanding questions offline. We we have the names here, and, wanna really thank you, Carl, for this insightful, information, about, your take on the current landscape, of ALS trials and the use of digital endpoints, such as modalities in in, for for the benefit of early stage, and late stage, clinical research. Thank you very much, and we are, looking forward to, having other such conversations with you in the future.

Karl: 44:25

Thanks, DSO. Pleasure to do so.

DSO: 44:27

Bye now. Bye now.