How Caristo is Using AI to Reduce Heart Attack Risk (EP 133)
The Harry Glorikian Show
For March 12, 2024
Final Transcript
Harry Glorikian: Hello. Welcome to The Harry Glorikian Show, where we dive into the tech-driven future of healthcare.
What if I told you that when radiologists look at CT scans for the traditional signs of coronary artery disease, meaning plaque or narrowing of the arteries, they only catch 20 percent of the people who actually have a high risk of a heart attack?
And what if I then told you that there’s a new AI-based test that can see what radiologists can’t in the other 80 percent of patients—meaning, subtle signs of inflammation in the fat cells around the arteries? Inflammation that could easily be treated using existing drugs to lower heart attack risk?
Well, if you’re someone with a personal or a family history of heart disease, you’d probably want to get that test yourself. I know I do!
My guests on the show today are Frank Cheng and Keith Channon, both from Caristo Diagnostics in the UK. Cheng is Caristo’s CEO, and Dr. Channon is the co-founder and chief medical officer.
And under their leadership, Caristo has introduced a test called CariHeart that applies machine learning to the data in a three-dimensional CT scan of the heart.
The test looks for otherwise invisible signs of inflammation in the fat tissue around the major coronary arteries, and then it predicts the chances that the patient will suffer a heart attack in the next eight years.
Doctors can use that information to decide whether a patient needs to take a cholesterol-lowering drug like a statin or an anti-inflammatory drug like colchicine.
Caristo’s test is being used on an experimental basis in the UK, and it hasn’t yet been approved for use in the US.
But it’s a leading example of the way AI, put together with fundamental advances in our understanding of human biology, is really beginning to change the practice of medicine.
Caristo’s test isn’t intended to put cardiologists or radiologists out of work. It’s designed to help them be more effective.
And given that cardiovascular disease is the number one cause of death around the world, any technology that can help catch signs of coronary artery disease earlier could save a lot of lives.
Which is why I wanted to talk to Frank and Keith, and get the whole story behind Caristo’s test and how it works. So, on to our conversation.
Harry Glorikian: Gentlemen, welcome to the show. It’s so great to have you both here.
Keith Channon: Right. Thanks, Harry. Yeah.
Harry Glorikian: And, you know, I’m glad the time zone worked out. I mean, you guys are in the UK. I’m. I’m here in Boston, and and we all got this to work. Which is, which is which is great. I, I’m really excited to have you guys on because, uh, the whole subject matter of cardiac health is, is, is personally important to me. After, you know, my, my father died of, uh, a cardiac event, which, you know, I, in this day and age, I think he could have been managed and, you know, had lived a long life. But, uh, unfortunately, back in ’73, it wasn’t exactly managed the same way it is today. So, um, and in reality, I was telling Frank, I was saying earlier, before before you joined, uh, Doctor Channon that I should probably have this test myself. But let’s give people sort of an idea, because I think I’m jumping ahead of myself here and people need to catch up with where I am mentally. Can you tell us a little bit about the company, maybe how it started and then the maybe the scope of the problem you’re trying to address. And so I’ll, I’ll hand it off to both of you because I’m sure you guys have done this before. So whichever one of you wants to go first.
Keith Channon: Yeah. Thanks so much, Frank. Well, I’ll start off. My name is Keith Channon. I’m the chief medical officer at Caristo Diagnostics. And Caristo Diagnostics is a spin out company from the University of Oxford here in the UK. And it’s based on the commercialization of a really exciting technology that came out of scientific discoveries in our research, based in the University of Oxford. And that’s related to cardiac CT scans, CT, coronary angiograms, which I’ll come back to. And that’s now a routine test to detect heart disease. Um, and I’m sure we’re going to talk a lot about that, but I think the bigger context here, you’ve already alluded to, and that is that cardiovascular diseases remain the number one killer worldwide. Um, there have been some big improvements. So public health, reducing smoking, cardiovascular fitness, the advent of cholesterol lowering drugs and aspirin, and modern treatments for heart attack. They’ve all had a huge impact. But despite those advances, it remains the world’s biggest killer. And of course, some of the improvements of public health, on one hand, have unfortunately been replaced by a whole epidemic of other public health problems. On the other hand, including the global epidemic of obesity and diabetes, all of which ultimately manifest themselves through an increased risk of cardiovascular disease. So it’s still a really big problem despite the advances. Now, one of the advances I’ve already mentioned is that increasingly we have sophisticated technologies, including CT imaging, that is now recommended internationally in both the US and and in the European nations as a first line test to detect coronary artery disease in people who either have chest pain or who are at risk of coronary artery disease.
Keith Channon: Now, coronary artery disease is the very common problem, where the coronary arteries become furred up by the deposition of cholesterol over many, many years, and ultimately, the progressive narrowing and blockage of those arteries results in symptoms such as angina. But it also, of course, without warning, can result in the deadly consequence of a heart attack or a myocardial infarction. And for all too many people, the first sign that they have coronary artery disease is a heart attack. They don’t get a warning. And that’s because although we talk about narrowed or blocked coronary arteries, and that usually happens over many, many years, an artery can go from being just a little bit narrowed or not even narrowed at all to blocked in a very short time. And that’s been a major conundrum and challenge in the whole field of of cardiology. So to go back to cardiac CT scans, this is a wonderful technology. You can go in a CT scanner for just a few seconds or a minute or two. And with the latest technology, the CT scan can provide a beautiful three-dimensional picture in great detail of your coronary arteries. And that means that we as cardiologists, we can detect whether there is furring up, narrowing, blockage of a person’s coronary arteries. And that’s why it’s become such a standard test. And the number of cardiac CT scans is going up very quickly worldwide.
Keith Channon: And that’s great for the people with coronary artery disease who have significant narrowings or blockages in their coronary arteries, because the CT scan detects that and those people can get treated really very effectively based on the result of the CT scan. Now, the problem is that as cardiologists worldwide have done more and more cardiac CT scans, it’s become clear that the minority of people have narrowings that are significant or blockages. In fact, almost 80% of people who have a cardiac CT scan for a good clinical reason, they don’t have significant narrowings or blockages. And currently most of us have, cardiologists in accordance with the international guidelines, we consider that those people don’t have a major problem. If they’ve got some flaring up, we’ll often give them good advice. Of course, we tell them to stop smoking. We tell them to lose weight and exercise. If they’ve got a high cholesterol, we put them on a statin, but we kind of don’t really focus on them as a major problem. So those 80% of people, they get managed with good advice, maybe some tablet treatment, and it’s the 20% with the narrowings and the blockages who get seen by the cardiologist. And, you know, a lot of effort is put into treating their coronary artery disease. So the big breakthrough for us came through some scientific research that made us realize that, in fact, those people who don’t have major narrowings or blockages, some of those people are at significant risk of a heart attack, even though their arteries on the CT scan are not narrowed or blocked. And that’s because the CT scan only images, the visible structural narrowings or blockages. It’s not able to really look inside the artery and see what the nature of the artery is. And we know now that the biological process that drives the transition from coronary artery disease to heart attack and death is inflammation. It’s where the artery becomes inflamed and angry and unhealthy. And that can lead to an abrupt blockage due to the formation of a blood clot. Now, Caristo’s big breakthrough is that it found a way, using artificial intelligence techniques, to detect inflammation from a CT scan that is invisible to the cardiologist’s eye or the radiologist’s eye, by a very sophisticated, computer based analysis of the CT scan image. And that means that for the first time, we have a clinically applicable tool that can detect and quantify inflammation in the coronary arteries. And we can come back to this. But the amazing thing is that many people have high levels of inflammation. And they’re at high risk, even though their arteries physically don’t look very narrowed or blocked. So that’s really why at Caristo were so excited about this discovery. And we’re very excited about CariHeart, which is the name of the product that can now deliver this technology to patients. It’s already approved in Europe and Australia, and we’re working with with US colleagues towards US approval.
Harry Glorikian: So I’m going to have to ask my doctor to get me a prescription for the UK? Is what you’re telling me. If I’m going to have to come over? Which by the way, I don’t mind. I love coming to the UK for all sorts of good reasons, probably most of them, which is going to cause more inflammation because I’m going to go out and eat a good restaurants and do all that good stuff. But. Help me understand. Is it? I know cardiac issues. So I’m going to put myself front and center. My doctor will tell me, even though my family had a history of heart disease, he’ll be like, okay, let’s do a stress test. Maybe we’ll do, um, a, you know, echo of the heart, right? And just to get an idea of what’s going on and of course, you know, regular blood tests to look at cholesterol and everything else. Right. You’re saying even with all those, I could be a ticking time bomb, theoretically, and I should want to get more information. I’m just trying to paint a picture for the average person out there that might be listening to this and understand the current dynamics versus what’s technically becoming possible.
Keith Channon: Right. So you you’ve got exactly to the point, Harry, which is that the overwhelming majority of tests that we as clinical cardiologists rely on, they’re tests that ultimately try to detect arteries that are narrowed and are causing a restriction of the blood supply to the heart muscle. And that’s really important. And we absolutely want to find people who have that problem because they are the ones who need immediate treatment. And they need not only tablet treatment to improve all the things you’ve mentioned, like cholesterol and blood pressure. They may also need those physically narrowed arteries to be fixed, either with a stent procedure or a bypass operation. Those people are already detected very effectively and efficiently using most of the routine cardiac tests. But if those tests detect narrowed arteries, which are the end result of the disease process. They don’t detect the disease process itself. The invisible driver that’s been going on for many years in people’s arteries, microscopically in the wall of the artery. And that’s inflammation. And it’s the inflammation driven by cholesterol and smoking and diabetes and all these other factors that ultimately lead to the wall of the artery to become thickened and narrowed. If you wait until the artery is thickened and narrowed and blocked, that’s a long way into the disease process.
Keith Channon: And that’s why most people, uh, begin to experience angina and have heart attacks and may die when they’re in middle age or when they’re elderly. But what about if we could turn the clock back and start to detect the disease process, and those people who might have invisible risk much earlier? Well, first of all, it would save lives because we’d stop them having heart attacks and dying. And second, if we could find who those people are, it’s a really fantastic opportunity to begin treating those people early on in the disease process before it leads to narrowings, blockages, and unanticipated heart attacks. So this is a real shift in the way we understand what we call coronary artery disease. Cardiologists have all been brought up on arteries that are narrowed or blocked as being the thing we should focus on. And it’s a fair criticism because we should have been focusing on the disease process. And the reason we didn’t do that is we didn’t have the technology to detect the disease process and Caristo’s CarHeart technology for the first time.
Keith Channon: It doesn’t replace the detection of narrowed or blocked arteries. But I’ve already said they’re in a minority of the people who have a scan. It opens up this whole new area of detecting a substantial number of people. Maybe 25 to even 40% of people undergoing a coronary CT scan, they have high inflammation without knowing it. Many of them don’t have any visible narrowing or blockages in the arteries. But our recent research work, which was published and which had a late breaking presentation at the recent American Heart Association meeting last fall, that high profile research showed that people who had high inflammation score in their coronary arteries, even if they had no visible detectable coronary artery narrowings or furring up, they had a 10-fold higher risk of a heart attack or death in the next ten years, compared with their age and sex matched control people. So this is an incredibly big effect. And for the first time, we can pick it up. And one final comment before we move on. This is not a special scan that you have to go and have in some super specialized imaging center. Caristo’s CariHeart technology can be applied to a routine cardiac CT scan acquired in exactly the way that cardiac CT scans are done day in, day out, in thousands of cardiac centers around the world. That’s one of the really exciting things.
Harry Glorikian: First thing is, I’ve got to bug my my doctor about getting a CT scan, which is one we have not done yet. But now let’s say the technology that you’ve created. before we get into the details about why it’s super easy to detect the inflammation. And here’s the question. So how do you then manage that patients? Are there obvious ways to act on that information? Is it statins? Um, do we know how well they work or, you know, is there other treatments or interventions that that say are available? Uh, because they don’t fall into that 20%, let’s say.
Keith Channon: Yeah. Great question. Well, this is a great opportunity for Frank to come in, maybe because Caristo is already not only answering that question, but starting to form some strategic partnerships which take advantage of the technology. But perhaps I’ll just give a comment first. And that is that drugs like statins, we know that they already work in part by reducing inflammation. Now their main effect, of course, is to lower cholesterol, particularly the so-called bad cholesterol, LDL cholesterol. And that is a driver of inflammation. So statins do that. And if a patient is found to have high inflammation in their coronary arteries, then of course one of the obvious things to do is to, uh, escalate and maximize their statin treatment to make sure they’ve got the absolute biggest bang for from the buck that they can from that treatment. Um, so that’s the first thing to do. The second is that, again, this is this is a really exciting time for Caristo because right now inflammation is being appreciated as, if you like, the next treatment target in coronary artery disease. There have already been clinical trials done which show that inflammation is indeed an effective target to improve people’s long term outlook and reduce the risk of heart attack and death. There are trials such as the Cantos trial that used a very powerful monoclonal antibody drug, called canakinumab to inhibit inflammation and a much more well known drug, colchicine, which Frank will talk about in a minute, is now also through a couple of trials been shown to be effective. So this is really gaining momentum. And the pharmaceutical industry are very interested and excited about a new generation of therapeutics that are going to target inflammation. But the key thing, Harry, is that you need to know who to give those drugs to. Because they’re likely to be costly. They’re very powerful, which means they might have side effects. And if you give them to absolutely everyone, it’s not going to get the best possible effect for the health economic investment. So CariHeart, for example, could detect the people with the highest inflammation. And it’s those people who we think have got most to gain by receiving these types of treatments. In addition, you can monitor the effect of the treatment. Now that we can see inflammation, because if you come back two years later and have a repeat scan and the inflammation is damped down, then that’s great news and it’ll correlate with an improved outcome. So Frank, do you want to say a bit more about the inflammation and this really exciting relationship that we’re forging with, with companies who have both current and new drugs?
Frank Cheng: Absolutely, absolutely. Keith. So, Harry, um, you know, CariHeart’s detection capability of, you know, coronary inflammation has been noticed by all the pharmaceutical and biotech companies around the world, especially those who are working on anti-inflammatory drugs for the heart. Right. So about eight months ago, um, Agfa Pharma, which is an emerging pharmaceutical company based both in the US and internationally, had received FDA approval of their anti-inflammatory drug, specifically indicated for the heart disease. Their drug’s name is called Lodoco. I guess you know, it stands for low dosage colchicine. So Lodoco is the was the first drug that ever received FDA approval specifically to fight coronary inflammation, which is one of the significant tools that Doctor Channon and others can now use in the United States to add to the arsenal of tools against this very, you know, severe disease, right, which is called a coronary inflammation, you know, on top of aspirin, on top of statin that, you know, Doctor Channon already discussed. So Agfa Pharma and the last month had announced that we have joined forces and we have teamed up to, uh, conduct a clinical trials together in the US and outside as well, to specifically test the combination of our diagnostic, which is CariHeart, but give pharmas a Lodoco and use that as a combination. So, you know, both identify the right patient to give that drug, but when needed, on top of statin, as well as other things like aspirin when needed, when indicated, uh, use this very potent anti-inflammatory drug as well for those people who needed them and give a pharma and Caristo have also decided to educate the clinical field in the US and in EU as well, related to the severe, let’s just say, the severity of inflammation and its long tum impact to people’s health, especially cardiovascular health, over time as well, because we’re about the only two companies right now around the world that really are just purely focused on inflammation. I guess you know, Lodoco drug is not, um, is not alone. As, uh, Keith had indicated earlier, there, there’s a good pipeline of additional anti-inflammatory drugs for the heart that is being developed by big and small pharma or biotech companies as well. Uh, fortunately, Caristo has been involving have been involved in a lot of these, um, you know, kind of drugs in its both research stage as well as development and clinical trial stage as well. So, you know, stay tuned. More drugs with the help of, you know, CariHeart will hopefully be approved by the regulatory authorities around the world very soon.
Harry Glorikian: So and at some point, I would love to also get into diet as a driver of inflammatory activity. Right. Because we’re sort of doing it to ourselves if, you know, unless you’re genetically predisposed. Right. I think I’ve just noticed that when I make fundamental changes to my diet, that it has a my wife will say, you look less puffy, right? There’s a less of an inflammatory effect. And I think the technology you guys have created might have broader application beyond cardiac. But I’m just I’m I’m spitballing now. So I want to go a little bit into the technology. It’s, um. It’s called fat attenuation index, if I’m not mistaken, FAI, and I’m sort of curious sort of, if you can explain to people sort of how is it measured, you know, what is it measuring? Um, when did this become, when did it emerge as a biomarker for coronary artery disease? Um, if I’m not mistaken, I think it was the CSO of the company that showed that having less, um, perivascular fat around coronary arteries, it’s actually a marker of higher heart attack risk. So I’m not I you guys know way more about this than I do. So I’m going to I’m going to I’m going to stop there before I say something that may be technically incorrect and let you guys go from there on on those questions.
Keith Channon: Thanks, Harry. It’s it’s a fantastic story, actually. And this goes back to work being done in our department based in cardiovascular medicine in the University of Oxford, led by one of our colleagues and co-founders, Professor Antoniades. Uh, and we became interested together probably ten years or more now in the relationship between the blood vessel wall and the fat tissue that surrounds many of our arteries in the body. Now historical, uh, people interested cardiologists and scientists interested in the coronary arteries have kind of considered the fat tissue as a bit of an inconvenience and something to be ignored or even removed. And when you look at cardiac CT scans, um, usually when they’ve been prepared for presentation on the screen, the scan has been prepared digitally such that the fat tissue has been removed from the image. And you can see the heart and you can see the coronary arteries, and the surrounding fat has been removed so it doesn’t get in the way. And that’s the way the technology has been used. We became interested in the relationship between the wall of the coronary artery at a biological level and the surrounding fat tissue. There was an emerging scientific insight and publications at the time that in people with cardiovascular disease and cardiovascular risk factors, all the fat in the body, including the fat around the coronary arteries, might be having a detrimental effect on the coronary arteries through the secretion of pro-inflammatory molecules called cytokines and ones that come from the fat tissue itself, adipokines, and that these signaling molecules might be important in what we understand about the causes of coronary artery disease and its relationship, for example, to obesity and diabetes.
Keith Channon: Our breakthrough was to discover that, in fact, that communication between tissues is not only happening from the adipose tissue to the wall of the coronary artery, but it’s also happening the other direction. The tissues don’t just have a monologue, they have a dialogue. There’s a conversation going in both directions. And what we discovered is that the fat tissue around the coronary artery, it changes in response to the presence of a high level of inflammation in the neighboring coronary artery wall, and it changes in a way that leads to a reduction in size of the individual fat cells. They become smaller. And the amount of fat within each of those cells reduces. Number one. Number two, inflammatory cells enter the fat tissue. So the fat tissue contains not just fat cells, which are the cells that are specialized for storing large amounts of fat. But the fat tissue contains inflammatory cells, and it also contains an increased number of microscopic blood vessels and other changes. And these changes in the fat tissue, they’re most marked close up against the wall of the coronary artery, if there’s inflammation. And as you move away from the wall of the coronary artery, a bit like the skins skin of an onion or the layers of an onion, those changes become less marked. And we did all of this in the basic science lab, and we did it using some microscopic samples of fat tissue that were taken from patients who had consented to donate small samples of tissue when they were undergoing cardiac bypass surgery. And so we were able in Oxford to collect very large numbers of these tissue samples, and we were able in the basic science lab to undertake very detailed studies of how different molecular pathways were altered in these fat samples, including pathways that are archetypal pathways for inflammation like interleukin one, interleukin six, TNF alpha.
Keith Channon: These will be the well-known pathways for scientists in the field. And what we discovered, amazingly, is that whether these pathways were very prominently switched on or not was directly related to the changes in fat cell size and the distribution of fat cell size in the in the direct relationship to the coronary artery wall. So that was the scientific discovery. The real clinical advance was to then discover that these changes in the fat tissue, the texture, the composition, the amount of water versus the amount of fat, all of these microscopic cellular changes that we’ve seen under the microscope, and we characterized by expression of all of these different genes involved in inflammation, it turned out that if you applied a machine learning algorithm to the CT scan image, including the fat tissue, you could derive what’s called a radiomic signature. Now, we all know words like genomic and proteomic and metabolomic, where very, very large data sets of items can be analyzed by a computer. And machine learning and artificial intelligence can discover patterns in those omic data sets. Well, radiomic is where you deconstruct an x ray picture, a CT scan into its constituent data pieces. They’re like the pixels of a picture, but they’re three dimensional, so we call them voxels. And if you deconstruct the three dimensional CT scan image purely into data, you don’t try and make a picture of it, you just create a data set and you use a machine learning algorithm to find how those radiomic features best correlate with the features of the scan that correlate with inflammation, it turns out you can derive an incredibly powerful radiomic signature that is directly related to how inflammatory pathways are switched on in the fat tissue. And that was just a wowie moment, because it turned out that it was much more powerful than we ever anticipated. In other words, it was able to detect and quantify inflammation from a CT scan image in a way that, frankly, we never thought was going to be possible. And that’s really the big breakthrough that led to the technology. Now, I’m going to hand back to Frank because ss we all know, a scientific discovery in a laboratory based on a few thousand CT scans is really exciting, but it’s not a clinical product that can be applied to patients in order to on which to base clinical diagnosis, because there’s an enormous amount of regulatory validatory product development work. And that’s where teaming up scientists and the academic strengths of an institution like the University of Oxford with a spin out company and and engineers and of course a great deal of investment. That’s what’s required to deliver a clinical product. So it’s been one of those journeys that’s combined multiple phases of discovery all the way through to a clinical product. I’m going to hand over to Frank, and he can tell you a bit more about that chapter of the story.
Frank Cheng: So, Harry. Um, so the company got spun out of Oxford University about six years ago, and obviously there is a few rounds of venture funding. So in total, over the last six years, the company has raised about $30 million USD. And all that money had went into mainly R&D. So, you know, finishing up the research, you know, kind of the algorithm research to turn that into a product. And we then develop that into a medical device, which then got CE mark, which is the licensing here in Europe and UK that you need to have in order for a technology to be used clinically. We also, along the way, have gained the right to approval in markets such as Australia. And right now, as of now in UK and the rest of Europe alone, there um, hundreds of on an annual basis, there are hundreds of patients. Um, so CT scans being analyzed by CariHeart and then the report then got shared either with their cardiology team or radiology or other kind of medical specialty team, in order to then enable the health care professionals to decide what to do to manage the patient specific situation. Then the same care product is also being worked on, uh, in clinical trials that supports um entry into other markets as well, such as the United States market. So we have multiple clinical trials going on right now in order to gain, uh, you know, FDA approval for different aspects of the CariHeart technology on inflammation, on plaque analysis and a combination of the two. Um, so we anticipate all of this good R&D work. And also, you know, non US regulatory approval kind of work will eventually result in us entering to the US market later this year or the early part of next year as well.
[musical interlude]
Harry Glorikian: Let’s pause the conversation for a minute to talk about one small but important thing you can do, to help keep the podcast going. And that’s leave a rating and a review for the show on your favorite podcast player.
On Apple Podcasts, all you have to do is open the app on your smartphone, search for The Harry Glorikian Show, and scroll down to the Ratings & Reviews section.
Tap the stars to rate the show, and then tap the link that says Write a Review to leave your comments.
On Spotify, the process is similar. Search for The Harry Glorikian Show, click on the three dots, then click “Rate Show.”
It’ll only take a minute, but you’ll be doing us a big favor, since positive ratings and reviews help other listeners discover the show.
Thanks. And now, back to the show.
[musical interlude]
Harry Glorikian: I want to jump back here for a second because I was going back, and. Looking at the story, sort of trying to go back in time to say 2020. And I don’t remember seeing anything distinct around AI, and I believe back then the focus was reverse engineering standard CT scans to see to show water content of the perivascular fat cells. So am I making the assumption that sort of incorporating the machine learning that you guys sort of have brought now this together, that was the spark that really made this easier and made the interpretation of that data more readily available for someone to then make an actionable decision.
Keith Channon: I think. I think there are three things in that recipe, Harry. The first is that CariHeart is calibrated on a biological insight. It’s it was originally discovered and the fat attenuation index was calibrated and created based on the most fundamental readout of gene expression. This is not, that’s where we started. So I think that’s the number one thing. It’s related to the very, it’s calibrated on human biopsies that were characterized at in the most detailed fashion to show inflammation. And if you start with that, you’ve got a very good biological tool. Secondly. Even if you’ve got that, and even if it’s a wonderful tool to tell you whether there’s inflammation, it doesn’t really tell you whether it’s clinically powerful or not. So the next thing that we did is we validated that against clinical cohorts. And I’ve already mentioned, but we probably skipped over it. This was fortuitous, um, for us and for Caristo, because the technology can be applied to very, very routine CT scans. So you don’t need a special scan. And that’s great for everyday life. But it’s really important because you can analyze CT scans that that have been acquired over the years as, as long as the CT scan has been done on a 64 slice scanner or later, which essentially includes pretty much all CT scans from the last 10 or 15 years. That scan can be pulled out of the archive. The digital scan can be sent to Caristo, and a CariHeart analysis can be done on that old scan. So rather than recruiting people who are having CT cardiac scans today, analyzing their scans and then following them up for 15 years to see what happens to their heart attack risk, you can do that experiment as quick as you can. Analyze the scans. So we were able to clinically validate the technology that can quantify the molecular aspects of inflammation against clinical endpoints, including heart attack and death. And we did that in a super quick time because we were able to use existing data sets. And Professor Antoniades’ study, called the ORFAN Study, based at the University of Oxford, is in the process of analyzing more than 100,000 CT scan images from the UK and around the world, and the first cohort of that vast number of CT scans was presented at the American Heart Association in Chicago last November. And that’s what caused this really big impact, because in the first 40,000 scans, we start to see an extremely powerful validation of the CariHeart technology. It’s able to predict heart attack, risk and death in totally new cohorts.
Keith Channon: And then the third thing. So the first thing is it was created on a molecular understanding. Secondly, it was validated on patient cohorts with long term clinical follow up and outcomes. That’s that sets Caristo apart from any other of of the AI driven, um, CT scan imaging companies. But thirdly, Caristo had to invest through its engineers and its R&D team an enormous amount of laborious work to make sure that the technology is reproducible and validated on multiple vendors, platforms and technology. When the scan is repeated on two different scanners, when the patient has a repeat scan on two different scanners one week and then the next week, what’s the effect of age and sex and all those things that could potentially alter the signal? If you want a product that can generate a number that tells you what your inflammation score is, and we call that the FAI score, the fat attenuation index score, then you have to be very, very clear that that is a trustworthy clinical readout. There’s no point saying, well, your arteries look like they might be a bit inflamed. That’s no use. We need something that’s absolutely rigorous and that we patients and regulators can have absolute confidence in. And that’s why CariHeart, the clinical product is such a huge advance from even from the original research which showed the proof of concept.
Frank Cheng: Yeah, I want to also just jump in to answer Harry’s question from the other angle. Right. So Harry, earlier you touched on the role of AI, right, since 2020. Right. So, you know, I’ve been in the AI field in healthcare for the last at least 15 years, right? So I observed the migration of AI’s role from being a supplemental tool to health care professionals in stage one to replacing either part or all of healthcare professionals work not because health care professionals cannot do the work, but just for efficiency and cost reasons. Right? AI may be a little bit more efficient with acceptable quality and reliability. Then the third phase, which which really is, is a, you know, with CariHeeart being a perfect example, is AI be able to do something that human cannot do. Right. So this invisible inflammation or fat attenuation index that Keith was talking about earlier, is not something that a doctor can do with human AI or with another tool. Right? So this is really we’re really CariHeart is really, uh, helping us to, to bring AI to a whole new level where you can quantify and visualize things that human AI or in other tools beats AI or non AI tools just cannot do simply as simple.
Harry Glorikian: So is the business model software as a service?
Frank Cheng: Today. Yes, yes. Today. How we’re deploying this in our current clinical market is that, um, you know, we essentially we sell the service in the form of a report, right? So on a on an individual basis, that is the business model that we find it currently easy. And let’s just say and also easy to understand and easy to adopt for healthcare professionals in our current clinical market. That obviously is very different than a pharmaceutical clinical trial model, where it sometimes is not per report but per project.
Harry Glorikian: So maybe you guys can describe, what does a Caristo diagnostic report look like? Can you maybe you can walk through the main elements like the the FAI score and the, I think there’s two, there’s the FAI score and the CariHeart risk score if I’m not mistaken.
Frank Cheng: Yeah. Keith, why don’t you go ahead?
Keith Channon: Yeah, that’s a great question. So all of this amazing information is returned to the physician in a 2 to 3 page report, and it has a few different parts. First of all, it it gives some some color map representations of how the inflammation is distributed around the three main coronary arteries, which I think we found is very helpful for both physicians and patients in sort of understanding the concept. And, you know, apart from all the technical aspects, after all, this is this is a tool to improve the health of people’s cardiovascular system and everything we can do to try and communicate the concept and help both physicians and patients understand that. So the first part of the report is a is a color map representation of the coronary arteries and their surrounding fat tissue, and where the inflammation is highest. The second part is a quantitative report of the FAI score for each of the three main coronary arteries. Most people have three coronary arteries, the left anterior descending, the left circumflex, and the right coronary artery. Those are the three that most cardiologists will talk about and report. And for each of those three arteries, CariHeart analyzes and produces a FAI score that is expressed for that individual patient per artery and is shown on a graph of centiles. In other words, where does that person’s right coronary artery fall on the distribution of people of the same age and sex? Are you on the 20th percentile. In other words, you’re low down in the distribution of inflammation? Are you in the 50th percentile meaning in the middle, or are you on the 75th, the 95th, or even the 99th percentile, meaning that artery, compared with other people of your age and sex, has got very high inflammation? What we know is that if you’re on the 75th percentile or highest, typically your risk of a heart attack or death is increased by about two and a half fold compared with other people of your age and sex. If one of your FAI score is on the 95th percentile, you’ve got an even higher risk of heart attack or death. So the FAI score per artery is the next part of the report. The third part of the report, as you mentioned, is that the FAI score of the arteries is then combined in the CariHeart algorithm with other demographic factors, with your with your clinical risk factors. And that generates a calibrated estimate of your actual cardiovascular risk of death in the forthcoming eight years, based on your FAI score and your clinical risk factor profile. And the final part of the latest version of CariHeart is as, as Frank said, we’re not ignoring the presence of furring up or narrowing, which is called plaque. CariHeart also delivers a fully comprehensive characterization of all of the plaques, meaning the volume of the plaques, the composition of the plaques, the degree of narrowing in every artery. So plaque analysis and plaque characterization is now a fully integral part of of the CariHeart report as well.
Harry Glorikian: So I should definitely have my physician order a CT, and then I’ll email it to you guys because this is….
Keith Channon: And and that means and that means that the, you know, the Cariheart report is therefore unique because it’s the only clinical report from a cardiac CT related to coronary artery disease that not only characterizes the plaque and distributes it, it’s unique in reporting these parameters related to coronary inflammation. No other clinical product is able to do that.
Harry Glorikian: I keep mentioning my physician because he actually listens to the show regularly, so I know that he’ll get the message. Um, let me ask you a different question, because I’ve seen so many great technologies sort of start making their way into the world. And you see the data, but there’s this thing called the established medical rubric if you want to call it that. And do you find any pushback from them? Do you find any difficulties? Um. You know. Arguing pro or con, and they’re just going to manage their patients as they’re doing them. Or are they being receptive to this and saying, this is adding to my arsenal and being help me manage my patient, uh, more effectively.
Keith Channon: It has a great question. Our experience is that we we don’t experience pushback, but we experience, probably the most frequent question is, how do I use this amazing new information that you’re providing for me? Because of course, this is so new. It’s an area that has not been available before. And so the clinical guidelines and the way that we as cardiologists have all been trained, and the lifestyle interventions and the drug treatments that are out there, they’re not yet attuned to be deployed in a way that is cognizant of coronary inflammation. So the biggest challenge for us is one of education. It’s it’s education even for expert physicians. And it’s definitely education for patients and for people who are interested in their cardiovascular health. And that’s much more exciting. And it’s much more fun because it’s not a pushback. It’s an opportunity. And there’s a real need for us to educate and to inform and to show to people that there’s a new technology here which could be transformative. And we want you to work with us to help test it and validate it. And so to answer the question, how best can CariHeart be deployed in clinical practice to make the most difference to people at risk of heart disease? We’re doing that in the UK, in the National Health Service. We have one project which is completed, one which is currently running their pilot projects. They typically involve 800 to 1000, 1500 patients, and they are, for the first time in real world evaluations, testing how the availability of CariHeart makes a difference to clinical decision making. How does it make a difference to the recommendations that physicians are giving their patients? And what does that really mean for the way patients are being treated? And so the insights we’ve gained from those two studies are really very exciting. So if you if patients undergo a cardiac CT scan as part of their routine clinical care, and if those CTs are then sent to Caristo’s CariHeart analysis, and we ask the physicians to record their recommended treatment for that patient, and then we make available the CariHeart report to them and say, well, does that make any difference to you? What do you think now that you’ve seen their CariHeart score? What’s striking is in fully 45% of people who go through those types of CT scans, when the CariHeart report is revealed to their physician, there’s a change in recommended management. 45%. So this isn’t 5 or 10%. One in every ten. This is almost half of the patients. And as you were saying earlier on, Harry, the commonest thing they do is change is escalate the statin treatment. The next commonest thing they’re now doing is adding colchicine, which is the agent that Frank mentioned earlier on, which is available in Europe as well as in the US, or other ways of escalating treatment or indeed changing lifestyle. So the early indications that this that this is not a pushback, this is like all new technologies trying to find, if you like, the place in the diagnostic pathway where patients and physicians are going to most benefit from rolling out Carey Hart.
Harry Glorikian: It’s interesting because I you know, I’m not used to, at least here in the US, right? I’m not totally on top of everything over there in the UK, but CT being routine, let’s say for a cardiac scan. So um, at least I if it is, I’m, I’ve missed it somehow in my discussions with everybody but. Do you see something like CariHeart making the CT become more routine for, let’s say, somebody who’s generally healthy like me? Because I can imagine pushback from the insurance company saying, why are you ordering this if there’s no underlying indication?
Keith Channon: Yeah. Well, I’ll I’ll start off. But, you know, Frank’s got a lot of insights into this as well, because the way that, you know, diagnostic technologies are deployed in health care systems is, is a very complex, um, complex set of decisions. But you’re right. Cardiac CT scans are already recommended as a first line test for people. But you are right that what we call the installed base, in other words, the number of CT scanners that are out there and not just the number of scanners, but the experts who are able to, um, report the images is is probably limiting at the moment. This is an area of medical imaging, especially cardiac imaging, that is not well developed, not as well developed as some of the other imaging platforms, but the number of CT scanners, the number of centers who do cardiac CT scans, and the number of patients undergoing scans is is increasing at an enormously rapid rate. So I think we’ll see a real change. We are seeing a real change right now and that’s going to continue. Um. So right now, Caristo, our role is to is, is to analyze and get more information from existing CT scans that are done now. And, you know, there’s an enormous number of CT scans that are done on out there. And our mission is to get the most information from those scans to detect the invisible signal that can make a real difference to patients, because right now, 80% of patients who are having a scan, hhealth care resources are being used up to have those scans. The patient is undergoing the CT scan and they’re not gaining much benefit from the scan. 20% are because they’re detected as having narrowed or blocked arteries and they get treated. But the 80% of people our mission is to get more value and more health care improvement and better outcomes for the 80% of people who have a CT scan and really don’t currently benefit much. But I’m going to hand over to Frank, because the way that this technology might drive the greater deployment of CT technologies, I think is a really interesting question. And it’s one that, you know, medical technologies get rolled out with different waves and trajectories depending on the technology that comes along. And we think, Caristo’s, you know, a really exciting part of that in the field of cardiac CT.
Frank Cheng: Today, CT angiography, you know, is a key mentioned is the first line test for people who present with stable chest pain. Right. We call this kind of usage of CT angiography is in the secondary prevention market. Right. So people who have symptoms come to their healthcare professional according to the society’s guideline, they receive a CCTA. Right. Um, you know, so, um, how big is this market? Just by focusing on secondary prevention market alone around the world, we’re looking at 30 million annually, 30 million scans a year. Okay. The advancement of technology, such as CariHeart adding new clinical insight and the possibility of getting new clinical insight out of these CT scans, will naturally increase the saturation of this market all the way toward 30 million scans a year, because today, we’re nearly probably not quite even one third of the volume toward that 30 million. Right. So everybody, physicians and clinical professionals interested in analyzing CTs using very hard kind of technology will naturally push the utilization of CT scans toward that 30 million eventually. That is not even including the primary prevention market down the road, right? Far down the road. What does primary prevention mean is people like yourself, Harry, who are interested in understanding your risk, long term. Regardless of whether you have symptoms today or not, right? Proactive. You want to get something done that includes a CT scan as well as maybe on top of that, a CariHeart analysis as well. That primary prevention market is also going to really be very, very significant for years down the road. That may require the CT equipment cost to come down a notch or two from today’s price. Um, you know, together with other things as well, infrastructure wise. But both secondary prevention market will grow as well as the primary prevention market will grow. So that means a lot of CT scanners may be deployed around the world, enabled by the availability of technology, such as CariHeart.
Harry Glorikian: The first thing that crosses my mind are places like South Korea. Right. We’ve got a friend of ours that’s that’s, uh, have been working there for a number of years. The amount its costsr you to get proactive care to prevent things is I mean, it’s it’s actually cheaper to fly there, get some of those things and take a tour of South Korea and fly back. And it would be less expensive than doing it all here. But that said, um, yeah. Is there an established, say, reimbursement path at in the NHS or in the UK or, or even in the US based health care system for software as a service, such as the CariHeart test? Or is that something you’re working with them on to help develop?
Frank Cheng: Yeah. So let me jump in. You know, just to give you an update both from UK perspective and the US. So in the UK, uh, the pilot project that Keith referred to that is being used within NHS uh right now they are on special reimbursement approval basis. So those uh, those costs for CariHeart are being reimbursed, being paid for by UK government under special approval kind of mechanism. In the meantime, we’re also working with an organization called NICE, which is the National Institute for Clinical Excellence. Who is the more likely, you know, the body or the authority that can recommend the things that go into both guidelines and even national funding mandate. So we’re working with the process with NICE to make CariHeart reimburse on a more regular basis rather than on special project basis. Okay. In the US, uh, you know, I personally was involved in a technology prior to me joining Caristo and, uh, software as a service already reimbursed well by Medicare as well as private insurance company. So the prior technology that I helped manage before, prior to me joining Caristo was autonomous AI that interpreted fundus images to detect diabetic retinopathy. So in 2020, uh, AMA had created that, uh, you know, procedure as a category one CPT code. So by assigning a category one CPT code, which is 92229 already to that autonomous AI technology. And a year later, Medicare officially covered that technology nationwide in all of its jurisdiction around the US as well, which is also quickly followed by private insurance coverage as well. So you asked about whether AI is, you know, use being used as, you know, software service being reimbursed. My prior product, which the product was called EyeArt, was a good example that is now widely paid for by both government kind of program as well as by private insurance companies.
Harry Glorikian: So let me ask a question to both of you. Like, is there a larger lesson, a lesson here for, say, the future of medicine? Should we expect to see more and more AI based ways of analyzing medical imaging data to get new sort of what I’ll call advanced intelligence like from them where you’ve got something. But like you said, human beings aren’t designed to see those sorts of things. But using AI methodologies, we can extract way more information from the existing image, let’s say, or, um, to make that next call.
Keith Channon: Yeah. I think yeah, that’s absolutely right, Harry. You know, AI is going to make it’s going to change life in all of the routine ways to make human tasks quicker, easier, more reliable, better. And it’s going to be labor saving and it’s going to refine pathways. But those are just those are helping humans do what humans do. What I think Caristo exemplifies is that if you couple AI tools and the power of AI with genuinely new biological molecular discoveries, and you use that capability to bridge the gap between what the molecular understanding of disease is and the ability to then image it and detect it and quantify it at scale in people, that’s the real power of AI in this setting. And I think that’s what sets aside Caristo from other AI imaging applications, which tend to be a helpful tool, but they’re not always a game changer. This a complete game changer to understanding what is coronary artery disease. Why is it a risk to people’s lives, and what are the opportunities for rolling out completely new treatments that we weren’t able to do before?
Harry Glorikian: Well, gentlemen, it was great to have you on the show. I hope we covered everything that you guys may want to have to have talked about. I hope I didn’t miss any underlying nuggets, um, about the company or the technology in this. Uh, but you guys tell me.
Frank Cheng: No, your questions have been very thoughtful. Really, really appreciate. Enjoy your questions.
Harry Glorikian: So, uh, thank you very much. And, uh, I look forward to keeping track of this. And I wish I could upload my CT to you guys, but maybe in the future I’ll figure out how to do that.
Keith Channon: Been great talking to you, Harry. Thanks for the opportunity.
Harry Glorikian: That’s it for this week’s episode.
You can find a full transcript of this episode as well as the full archive of episodes of The Harry Glorikian Show and MoneyBall Medicine at our website. Just go to glorikian.com and click on the tab Podcasts.
I’d like to thank our listeners for boosting The Harry Glorikian Show into the top two and a half percent of podcasts globally.
To make sure you’ll never miss an episode, just find the show in your favorite podcast player and hit follow or subscribe.
Don’t forget to leave us a rating and review.
And we always love to hear from listeners on X, where you can find me at hglorikian.
Thanks for listening, stay healthy, and be sure to tune in two weeks from now for our next interview.