Signature Series: The Future of Wearables in Cardiac Care | LSI USA '25

Amir Soltanianzadeh  0:05  
All right, so thanks everyone for coming today. I have the pleasure to introduce Dr Asif Ali, and we'll be talking about the future of wearables in the cardiac space, and what that means for the future. You know, Dr ASAF Ali is an incredible cardiologist out of Houston, the Houston cardiology consultant group, and runs many groups in terms of research on devices and their effects in cardiology. How you think about inclusivity data, AI, and so this is really a informal session of us having conversation of what were wearables like in the past, what are they looking like in the future, and what are we most excited about? So I'll just start off with Assa, if you want to talk about a couple of things that you're most passionate about, about the wearable space right now.

Asif Ali  1:10  
Yeah, I think I'll answer that through a publication we just did at the American Heart Association, their 100 Year on actual digital wearables. And I think one of the areas that we don't touch upon enough. So the study was looking at PPG. And PPG is photoplex monography. It's the common red, green infrared wavelengths that you see on your devices. And if you remember back in COVID, when we were using pulse oximeters to look at oxygen saturations, especially in people of color. And there were individuals for actually hypoxic. Their blood oxygen levels were low, but the devices that we were using were not capturing that. That's a big no no, right? And so literally, there are individuals using devices that were not studied on, and we call it the Fitzpatrick scale, which is the different skin tone. So Fitzpatrick scale one to six. And so in our studies that we presented at the American Heart Association was looking at how do we use PPG to do cuffless calibration, free blood pressure monitoring across all different types of skin tones. And my conclusion in that poster, and it was presented in published in circulation as an abstract in his paper, was that if you're going to sell to people of different skin colors and skin tones, then you need to make sure that you include them in your research, and they're already inherent biases in AI, and we're going to perpetuate them if we do not include in our machine language models all the people that we're going to sell to. So if you're going to use devices, you better include all the players into those into those arena. So inclusivity is as a mainstay that we really need to make sure all the stakeholders are included when we do our research right.

Amir Soltanianzadeh  3:12  
And kind of the elephant in the room that both of us were talking about before this talk was, let's, let's talk about the bad rap that a lot of wearables have gotten, which is a lot of new information, some of that exciting, some of that just more data. And so how do we see a lot of the future of these devices coming into where wearables are actually part of a real standard of care, treatment, measurement, better understanding of the personalization of healthcare, instead of just sick care when we're waiting all the way til the end, when patients only options are in hospital or procedural therapies. So how do you look at wearables as being the new age of the continuum of healthcare. Yeah,

Asif Ali  4:05  
I think you have to again address health equity and health disparities to make sure we're equitable with everyone. So making devices that are almost seamless, you know, the more complex, the less likely of adherence. So you want to make the biggest impact in healthcare, people have to use your device. They have to adhere to it, and you need to have a stickiness and to be able to do that. And I think there's been validated studies around this, the human computing part to make it easy, almost seamless, like we're sitting on a chair, or, say, you're sitting in a car, we can take a wait. I mean, there's, you know, just in the day to day churn of life. How do we access that data? But it's not just access and data. How do we actually make those into action items? So how do we actually talk to a patient, right with those devices? Just to give an example. Amp. I thought it was very interesting. There was a group from UK where you plug in a device that looks at pedal edema, like, basically fluid going down into your ankles. Really simple, just plug it in your bathroom. You walk by it, it says, Hey, you have a lot of ankle swelling. It literally has something that will alert you that says your ankle swelling has gone up. You need to increase your lasix. So it's not just diagnostics. It's actually, how do you create action items and therapy? So it's it, and that's where I think AI can actually help us bridge some of those signals and processes. So I just give you another example which jpm they were talking about this with one of the CGM devices. You have a CGM device. It's monitoring your glucose. You have type one diabetics who have insulin pumps. You know, how do we connect the data from diagnostics to therapeutics and make it seamless, and then be able to make that information accessible to the healthcare provider? Right? There's some examples of practical ways of looking at digital AI,

Amir Soltanianzadeh  6:00  
yeah, because we were talking a lot of actually, the similarities between what's been done pretty well in the diabetes space with what's happening in cardiac obviously, we're in the treatment and monitoring space as well, but I think we saw a bit of this in DBS recently with The announcement of those closed loop therapies. So I think wearables might be an interesting future in which, hopefully their closed loop nature in either all in one or being a companion with another device, whether it's an implant or a different type of therapy, I think maybe that's where they'll become a bit more ubiquitous in actual standard of care.

Asif Ali  6:43  
I think your companies kind of touches upon that, because you're looking at neuromodulation and atrial fibrillation. So again, you know patients are coming with proxy malefib. So that's the diagnosis. But what are we so that's great. AI is really good at the diagnostic part. But really what I'm interested in some of the things that even you're working on is even predictive analysis. So for example, we know we use a scale called the chats too fast score that predicts your risk for stroke. You're working on neuromodulation. But I also think AI can go from Population Health to more personalized. And so if we look at a chat Stu vascular that's for population. But how does that apply to say, a female so, like, Women's Health is a big issue that we talk about, but to make it very practical in this sense of atrial fibrillation, for example, most of the machine learning models were based on 65 year old white males. So if you're gonna again, I'm going back to this whole inclusivity, back at the American Heart, which is if you're gonna train your models, which are were based on, so the previous data was based on mainly a cohort that may not represent the cohort that you're trying to create these models. And so I mean, we can talk about digital twins, and how do we go around that? But the crux is, you know, where are you acquiring your data, and what are the data sets that you're training your models on? And in particular, for a fib and EKG historically. And so that's where I think making sure the models that we're collecting the data are applicable to the individuals that we're applying them to.

Amir Soltanianzadeh  8:28  
And I love to pick your brain. I don't know how much we've talked about this in the past, I think maybe in one of our prior calls. But how do you see the provider? Let's call it broadly, but cardiologists of the future interfacing with wearables where, you know, there are interventionalists, but also those, I think, uniquely, that are seeing a lot of these patients in the clinic are trying to manage them remotely. How are you able to manage maybe not just one wearable, but the whole ecosystem of remote devices, whether on the therapy side, monitoring and how should innovators out there think about what the provider is thinking in terms of how to manage it the burden on the provider, not over alerting and providing value to all the stakeholders, not just to the patient as

Asif Ali  9:24  
well. And that's the million dollar question, interoperability. I mean, you know, having agnostic dashboards that can help integrate within the EMR. And I mean, there's death by data, right? And there's a lot of noise and signal. I mean, you know this because you're in the A you're in the AFib space. So it's about getting accurate data. I always say that MD means medical doctor, we know that, but it also means make decisions, and if we don't have good data in we can't process that. And you know, there's not a day that goes. I've still practiced as a clinician, so there's not a day that goes by that a person comes to my office and says, I have some information I want to share with you. But I think there's a big gray zone, and it's very blurred, and it's a challenge that I actually think industry needs to take over. In fact, we just had an AI ethics committee meeting with American Heart, and we will be in DC next month talking about this very fact, which is, you know, how do we create standards, standardizations in the AI data, there's a really great article. I hope you guys can read it. It was September 18 in 2024 and I'll have to remember the actual title. But essentially it was in nature, nature medicine. And basically what Nature Medicine said is, is a is the not AI all AI health tools with regulatory authority or clinically validated. So basically, it showed in 500 AI projects that they were they valid, that were validated by the FDA, did not have publicly available clinical validation data. So just because you have clearance does not equate real outcomes. And so that's where I think the partnership between healthcare providers and industry really need a partner. And furthermore, how do you like what we just said? How do you take independent datas and bring them together? And I think AI can help with the signal signaling processes to help bring that together, make it interoperable, so it's actionable. At the end of the day, we can, we can have intellectual discourse about how great AI is, and it's such a powerful tool, but without real outcome, clinical outcomes, and in that nature, article, over half of those only, I think 20% actually had randomized control trials, which is sort of the gold standard, right? So I have no issues looking retrospectively at the data and creating your models, but you have to go prospectively to make sure that the clinical validated data actually exists and publish and and then, you know, push it out to the market. Yeah. 100% that's a mouthful,

Amir Soltanianzadeh  12:26  
yeah. And I think, you know, something interesting you said there there's kind of a tension with wearables, and I think them often being easier to go to market and get cleared than invasive procedures or implants. And so I think sometimes there is a rush to look at the endpoint of the FDA clearance like you're saying, as opposed to, well, let's give it the same rigor as if we had a very long road ahead, because you do no matter what. And so let's think very carefully about how this fits in. Why are we measuring this data? How is that going to actually, then impact patient compliance, which is one of the biggest issues as well with wearables. In terms of, okay, you can have the best insight in the world, but if no one's going to use it, you kind of don't have a useful device. And so, you know, we've had to deal with that. A lot of my colleagues have had to think through those challenges as well. How do you look at patient compliance? And that also ties into what how we thoughtfully designed for different populations. How do you look at, especially at Houston, you see all types of patients, all ages, all races, all genders. How do you look at patient compliance and really devices that have succeeded and not succeeded as much?

Asif Ali  13:53  
Yeah, so it's a great question. I'm actually partner as a chief medical officer a company called Tabia. And Tabia means behavior behavior change, right? And if you want to make the biggest impact in healthcare, it's behavior change. I mean, people have to adhere. And I think there's two parts here. One is we have not as a whole in the medical industry done well on the gamification of devices. How do we get people to, you know, by I don't mean to plug in Starbucks, but they do a really good job on gamifying me. Yeah, you have stars. You know, the more stars you get, the more I mean it the psychology. And, in fact, it's so interesting. I just had a amazing call with Dr Sonnenburg. He's at UT. UT undergrad, Pat Dunn he's Program Director of American Heart and Frank Larkey. And these are all PhDs in human in health literacy and behavioral change. And I really think it'd be so interesting that if in all medical. Devices that we're working on. If there was a layer that understood the psyche of the actual patient, what is, how do you how do you talk to them to best get them to adhere? We don't, we don't do that. That can be gamification, but could also be the the psyche of that individual. What actually makes your device more sticky with them, right? It may be different for different genders, different ethnicities, and that's where health literacy comes into play. And I don't think we spend enough time addressing we spend a lot of time on our products and go to market, and yes, I mean, like on the pathology, but I'm talking about the psyche of the actual individual who's taking that in I don't think we address that human side of it as well as we do on the AI computing part of it.

Amir Soltanianzadeh  15:47  
I actually completely agree with you, per usual. I think there's a lot wearables and devices in general, especially remote devices, have to learn from the consumer world. I think consumer health is having a little bit of a moment right now. I can tell you how many of my friends have some form of, like, an aura ring, a whoop, and some of that you can attest to just being like, okay, that's really great consumer marketing, and it's a cool thing. And then there's an aspect of, well, the gamification is actually very closely linked to their health, because they're thinking a lot more consciously about what they're eating, about how much they're exercising, how much stress level they're having. I'm now getting into debates with my friends of like, what's your sleep score, right? And I think this whole aspect of wearables can be a huge turnaround moment for us to take a page out of consumer health. Make health cool. You don't get that opportunity with the invasive and in the hospital type of devices. I think we have an opportunity here to make health really cool, to make patients excited to learn about their heart or whatever other disease we're working on here. And I think that's that's kind of a call to the innovators to Yes, think about FDA and think about payers, but also think about how your device is exciting, how it how it drives the patient to want to look after their own care. Because if a patient's engaged, not only are they going to use their device, their outcomes naturally are just going to be better than a patient who's like, okay, yeah, this is something my doctor told me to use. I guess I have to use it. 100%

Asif Ali  17:35  
agree with that. And I think consumers, you know, there's a lot of gray zone right now in consumers who are taking in these products, some of the products that may be FDA approved or cleared for certain aspects of that product, but not the whole product. So there's a lot of blurring to a consumer. I think it's very confusing, and that consumer doesn't know. Should I bring this data to my doctor? Should I not and so actionable data, you know, good data, clarity of that data is so valuable. Again, going back to making decisions, if we're making decisions on on data, that's maybe not. FDA cleared. So I think there needs to be clarity for consumers. What is a pedometer? Pedometer fits off a couple of 100 steps. It's not, it's not a big deal. Maybe to some people, not a big deal. But if your heart rates off and you're doing new modulation for a fib, that's a big deal, right, right? Because you're actually taking action items on it, on a patient, right? And if that data is not good, so those algorithms have to be very clear about delivering, especially if you're delivering treatment, right, not just diagnostics, right? Yeah, cool.

Amir Soltanianzadeh  18:48  
I think we have a little bit of an opportunity. Maybe, let's talk on one more topic. Maybe we even take a couple minutes from the crowd see if there's any dying questions as well. Do you have another topic I think you wanted to talk about, or you want to take Q? You want to take Q and A? 

Asif Ali  19:03  
No, we can do Q and A's. Yeah, let's make it fun. 

Amir Soltanianzadeh  19:06  
Anyone have any questions on cardiac wearables?

Audience Question  19:10  
Hey guys. Hey Ken. So when you think about the emergence of therapeutics in the wearable space, what are some of the more exciting ones that you guys are seeing that you're you're most excited about?

Asif Ali  19:26  
You want to talk about your product?

Amir Soltanianzadeh  19:28  
I don't, I want to talk about cardiac care, but I appreciate the question Ken. I think neuromod past few years is definitely having a huge moment. I mean, there was, like, a neuromod for everything. I think that's an exciting movement in healthcare, frankly, because it's the coolest thing that got me excited from the first moment, which is, let's use our own body's natural nervous system. I think that's a really interesting area of. Um, I don't know, is there a core area that excites you the most on therapy side?

Asif Ali  20:04  
Yeah, I'll talk more about an area rather than a particular product. But I think predictive analysis is it's so mind blowing because we're really leveraging signals and pathways that we may not be able to see. I'll just give you an example of maybe a couple of examples. You know, in the EKG space, that's such a great area of innovation. And I think, you know, there's a lot of new products coming out, but we also have ubiquitous access to things like EKGs, for example. And when we look at, you know, predictive analysis, where AI can really help us with that. I think EKGs have a no brainer in that. One topic I was just talking yesterday, is just ultrasounds of the heart, right? You've heard of ultrasound. You look at valves. And so interesting that I had a I had a patient of mine who had a a it was called aortic stenosis. People know about tavers, which is a non invasive way to fix the aortic stenosis. So I remember talking to this individual and saying, You need to go get the valve fixed like now. Do not pass go. Do not pass Jon. You know, collect $200 I'm going to send you to the top surgeon in Houston, but with shared decision making, he said, I don't want to have a surgery. And I said, it's not a surgery. We're going to put a TAVR, a artificial, you know? And he started feeling better because I was giving some medication. So he came back to my office, and I just remember going, I told you have to go to the hospital, right and and I think this is a really good message to talk about health literacy and how to message to that individual, even though I told him he will die if he doesn't get that aortic valve replaced. So believe it or not, two weeks later, he's finally in the hospital, but he refuses a pulmonary function test, which was necessary to get the TAVR, and he died that night, and I had literally told him two weeks ago, you have less than a month to live, but imagine that if we had an ultrasound of his heart that had a predictive analysis to tell him, literally, you know, maybe it would have been two months, maybe it's six months. But here's the prediction of your aorta, that you need to go get it fixed, right? And then put in all the parameters like, you know, reduce the blood pressure, make sure you're on gdmt for that particular thing, right? But then we can also predict, and this is why LS company astute imaging, what TAVR we should use? What size? What's the prediction model for leakage of that TAVR device? That's amazing, because it really, it personalizes it. So we need to get away from public like, you know, we have, I call it P to P so, you know, looking at just pub, you know, public health versus personalized health. And I think that's where AI is helping us really understand the individual with the different signals. And to me, the most exciting part is that predictive analysis for your device, for example, for neuromodulation. We were talking over lunch, you know, predicting for the patient with their atrial fibrillation, we use a chats to vasc score, which tells you what your risk for your stroke is. But we can actually collect a lot more data for predictive analysis for something like a stroke with atrial fibrillation. But I think the key part is, how do you message that information to the patient, where they take action items on that and they understand the importance of adherence, which then creates behavioral change, which then helps us get better outcomes, right? I think that's a very fundamental theme that I'm seeing, and I think that's something that wearables need to address. I agree. 

Amir Soltanianzadeh  24:04  
Do we have any other burning questions at the moment besides Ken?

Audience Question 2  24:14  
So another question. So when you start to think about AI on top of ECGs, and you're starting to see a lot of publications, for example, Vivek ready and Josh Lampert out of Mount Sinai using 12 leads to predict lots of different things. And there's a company, I think last week, out of Korea. Can't remember the name of the company, but they published something about predicting a fib a year ahead of time. There's other ones that talk about, you know, identifying diabetes and other overlapping disease states. So what you're talking about is becoming a reality. How is that going to impact your practice? And where do you see this kind of leading if you look a year, two years, five years ahead, would let. Love to to get both your perspectives on that, but in particular, yours. Dr, Ali,

Asif Ali  25:05  
yeah. So you know, when we look at EKGs, I mean in general, people are like, okay, it's an electrocardiogram. Ai, wow. There's so many signals. And you really have to have subject matter experts and key opinion leaders to see within the EKG what are we trying to define? So we can now look at EKGs and do predictive analysis if you have coronary artery disease, there are EKGs from the age of 12 to 18 that we collect for sudden cardiac death in athletes, which is a big area of my research, there's EKG detection in acute coronary syndrome that can very much early on, predict before you have a STEMI or a non STEMI. So one thing Ken is, you know, always starts with asking the right question. You have to ask the right question. And I believe that's the partnership between key opinion leaders and and individuals who under who are subject matter experts, even with an EKG, which you would just think, Oh, it's just an electrocardiogram. There's so much involved in just that, data analytics and in different cohorts of different types of patients, which, again, goes back to making sure that your machine learning models are inclusive in a variety of different types of patients within the question that you're trying to answer, I think the future is that EKG machines, depending on what you're trying to answer, will be able to provide provide that data, and I think it's going to be pretty ubiquitous in all EKG monitoring systems. And you know, Ken, I know you're big on wearable devices for like holters and MC Ts. I think that's just going to translate even further. Again, predictive analysis for strokes and other areas like you said, you know, impaired fast and glucose to diabetes, there's so the the real thing that amazes me in AI is looking at the signals that we don't even realize are there.

Amir Soltanianzadeh  27:15  
If we don't have any other questions, if we do, that's great. I wanted to take a quick little poll of the crowd here. So just curious, and this doesn't have to be medical device based, but even just consumer health devices, how many people in the crowd have some form of a wearable? I'm really just curious. It could be an Apple watch. Okay, so that's for those on video. I mean, that's 90% of crowd. And so also, you know, we already have ubiquity there on the consumer side, I'd say, Who here would have interest to use a wearable to better personalize your therapy, whether it's an interventional procedure, medical management, or learning about something that is more on patient behavioral change most people. So I think that's a, you know, just in closing, I think pretty remarkable call to all the innovators out there in the especially in the remote therapy, remote monitoring, space wearables, AI, there's so many devices out there. Really think that the consumer, the provider, even now, with payers, they're really on your side, if you think very thoughtfully through all the stakeholders, how you're really delivering value to the healthcare ecosystem. So I think it's the most exciting time to be in healthcare in recent history. So it's awesome to have you know, providers like Dr Asif Ali that are thinking not just like an MD about their specific patient, but how the field moves forward. So really look for partners like yourself and and find that partnership with the physicians who are trying to think for you know how we're going to move the field forward. So really appreciate your time. That was awesome. 

Asif Ali  29:18  
Yeah, that's great. Thanks. Great job.