Episode 148: Navid Alipour - AI In Healthcare, CureMatch, And CureMetrix

we take that input from the NGS panel and we have our database that we're always updating, we're always curating, we have a clinical science team led by a two science officer that's always updating this, always shining the apple and then the algorithm will match and say based on this patient's specific cancer and there's no other cancer like this person, this individual's cancer here is the recommended three drug, two drug, one drug combinations that will have, they'll have higher efficacy. Welcome to the People of Pathology podcast, I'm Dennis Strenke, on this podcast we explore pathology, laboratory medicine and forensic science. Today we're going to step a little bit outside of pathology and talk about AI in healthcare in general. My guest is Navid Elipore and he's the CEO of AI Med Global. We're going to talk about two of the companies that he co-founded Cure Match, which utilizes AI in oncology and cure metrics, which applies AI to radiology images, specifically mammograms. And as we're talking about these, I think you'll see how these technologies can be easily adapted to pathology in the future. All right, here's Navid Elipore. We're going to be mostly talking about AI and AI as applied to medicine and healthcare. And I'm curious kind of how you got into this world because I mean your background is in political science and business and venture capital, which is a little bit closer, but still, how did can you kind of take me through like how this all led to AI? Yeah, no, absolutely. And without boring your listeners, you know, one thing kind of led to another to another. I wish I could say it was all masterfully planned. It wasn't, you know, I ended up going to UC San Diego, which is one of the birthplace is one of the epicenter's really of artificial intelligence. So they're just a natural talent pool here. I do have a long business background. I don't have a, I'm not a data scientist. I don't code. And I'm not a doctor either, although most of my family on both my side and my wife's side are. And healthcare has always been kind of a, a, a passion to do something in it on the, the finance side, the business side. And, and so we had started a, a small venture fund and we weren't investing directly putting our own money in as well into startups here in San Diego. And, uh, the long and short of it is we were approached by some scientists out of UCSD that were literally rocket scientists that don't work for NASA. You know, normally when they come to investors, they say, you know, here's my business plan. Here's an investor deck. Here's, you know, who we are and what we want to do, they don't have any of that. They just said we're these AI machine learning experts. Anywhere you can apply to make a prediction or recommendation or forecast or detect anomalies that don't belong. There's an opportunity to bring efficiencies and increased revenues. And, and we said, look, that's all great. San Diego is a healthcare town, a defense town. What can we do with those protocols? And, and the interest of time we said, you know, in, in the healthcare space, you know, can we detect breast cancer better than existing computer assisted detection technologies? And, and they're very comfortable and said, yeah, we can do that. So we literally co-founded, uh, the first company, cureometrics, under this venture studio model, uh, which we didn't invent the venture studio model in a gentleman by the name of Bill Gross back in the late 90s, uh, started idea labs up in Pasadena by Jet Propulsion Laboratory. And a lot of VC funds used the studio model to start companies. So we started to put the first money in and, uh, the rest of history, so to speak. So that's the genesis really of how we started the first company in applying AI in the healthcare space, um, specifically for breast cancer first. Okay, so getting into AI, I mean, it was always going to be into healthcare for you. It was not always, um, I, for me, it was in interest for sure. And, you know, and just to take it back a step, um, artificial intelligence has been around for some time. Um, but it's had its fits and starts, right? And I think it was really in the early 20 times that it, it truly became sustainable. And it's only been building and building and accelerating. And you have, uh, you know, people, uh, certainly better known than I am, like Andrew Ning saying it's a next electricity. It's going to have a bigger impact on us than the internet itself. And, uh, you know, some big, big proclamations like that. You know, I'm not going to say that's right or wrong. I do think it's truly going to have a huge impact on our lives bigger than the internet itself. I won't quite equate it to, it's bigger than fire or electricity, as some of them have said. But the reason it's sustainable now, and it's not going to go away, comes down to three factors in my mind. One, there's more data to process. Uh, everything is connected to the internet. Whether it's your phone or your watch or your a pacemaker or, uh, Tesla, a Tesla is an IoT device. It's a, it's a, it's a, a node is the engineer say it's connected to the internet and it updates a software every night, right? And so all these machines, they're all generating data. Well, data is useless unless you can process it and find those nuggets of value. And so we have higher compute capacity. We have, you know, stronger computers and we have, uh, the, with a cloud, we have the ability not to scale and ramp up very quickly. So in cure metrics, for example, we don't have to have our own data center and our servers and that would have cost, you know, tens of millions of dollars to purchase back in the day. We're 100% on Amazon's cloud. And we could ramp up and process 10 million milligrams tomorrow and then ramp back down if we need it, right? So that, it's a, these factors allows the algorithms to be trained and to get better and better and better. That's machine learning. And we can get into different types of AI tools. There's obviously natural language processing. There's something called novelty representation reasoning, which we use at cure match are other digital health company. But with machine learning, it's, you know, the more data you feed it, the better it gets, right? And so to process all that, you need the data, you need higher compute capacity and to keep the costs down, you have the cloud. And so I think those three factors are why this growth and artificial intelligence and this acceleration exponentially is happening. Yeah, that's definitely the case. It seems like, you know, not only in healthcare and medicine, but in pretty much all areas, there's been this explosion of the use of AI. Absolutely. And, you know, I get excited talking about it, but healthcare was just a natural progression of my interest aligning with, you know, these are opportunities that we saw of these scientists and then also being in San Diego and, you know, having the medical school here and the talent here in the biotech industry here and just the ecosystem was ripe for that, right? I, you know, we joke in San Diego, we do hard things here, whether it's healthcare or B2B, cybersecurity, defense and so on, said, I'm not going to climb credit for it, that, you know, if you want to do a dating app or deliver pizza to you faster app, you know, go to LA or San Francisco. And so here in San Diego, healthcare and the hard sciences, there's definitely huge talent pool here, so it was just natural for us to see what we can do. And that, again, that convergence is happening. And, you know, we have big companies like Illumina that are headquartered here, that, you know, that converges to software and life sciences and genomics and, and our bodies are genes are software, right? And at the end of the day, you need higher compete capacity to process all that and to detect earlier and to treat better and get the right rights to patients earlier. And I mean, the 21st century is a century that I think that's going to be defined by, specifically this, the bioinformatics, healthcare, artificial intelligence, convergence of the life sciences, to the end of the day, for longer lives and, and allows us to live longer and within higher quality lives. Yeah, that, that makes a lot of sense. And you mentioned a little bit ago the increase in the amount of data that's out there. And I think in medicine, I think people are realizing that there is data in areas where maybe they they didn't think there was as far as like, it thinks like, you know, patient demographics, you know, you know, that are still anonymized, but even like lab data or like disease progression and things like that, that's all data that can be analyzed and used by AI. Yeah, 100 percent, 100 percent, Dennis. And I mean, the, the simple fact of the matter is that at the end of the day, if we're going to live longer, the two top causes of mortality are cancer and heart disease at a global level, right? Cancer, heart disease, they don't know borders or boundaries. Right. We're blessed in the US with, with incredible, you know, technologies and teams and money to, you know, come up with new treatments. And of course, there's lifestyle changes. And you know, that's another complaint of our system, those that we spend more money, but we're not necessarily healthier and that's, that's a whole nother conversation. But the bulk of new technologies or significant pieces are still made here in the new drugs and the new medical devices, the new digital health companies, we're at a cornerstone in history and, you know, historians will look back in the 2020s and I think it'll be truly one for the books. Yeah, I happen to agree with that, actually. Now, there's kind of a flip side of AI because, you know, everybody, when you talk about artificial intelligence, everybody thinks of like terminator and stuff like that. And it's right. There's always this sort of, you know, fear of artificial intelligence will replace people, and it'll take away jobs, especially like in healthcare. That jobs that a lot of us do, so there's always this fear that the machines will take over whatever the matrix kind of thing, stuff like that. But I tend to think that the use of this stuff is not to really replace people, but to make people more efficient. Like what are your thoughts on that? - I agree with you. I agree with you in a couple of different comms to make there. One, we can't stop technology from advancing, right? 'Cause guess what, if we do, the Russians aren't, the Chinese aren't, right? The Israelis aren't, the Iranians aren't, the Eastern Europeans aren't, so we're gonna fall behind, right? And so the fact is that if we stop in the US, we're gonna become a third world country and very quickly. So we have to innovate and technology is not good or evil. It's how is it used, right? So when the automobile was invented, people said, what's gonna happen to the people taking care of the horses and the horsebuggies? They're gonna lose their jobs. And you know what they did? But look at all the jobs that were created because of the automobile, from the manufacturing to the servicing, to the oil companies and the gas stations. And yes, tens of millions of people have lost their lives have been maimed because of an automobile. But look at all the benefits that have come from it to us as a civilization. And so the automobile is not good or evil. It's how is it used? And yes, there are gonna be negative consequences, whether it's global warming or pollution or other factors. But net, I'd like to have someone argue that it's been negative for us, it hasn't. And so I think artificial intelligence is the same. Like the horse is out of the stable. You're not gonna put it back in. And if you do, we're gonna become a third world country 'cause others aren't gonna stop. So it's about how do we maintain our competitive advantage as a country? And how do we also manage the risks and be able to detect the bad actors or it's be able to put guardrails on the freeway, so to speak? - There was this quote that I found. It's, I think it's a couple years old, but it's from the American Medical Association and I'm gonna paraphrase. But they're saying that when talking about AI, they prefer to use it to call it augmented intelligence rather than artificial intelligence 'cause it focuses on like the assistive role. - So I'm glad you brought that up for a couple of reasons. One, healthcare, there will be certain jobs that will, if not disappear, be less in demand. But we have such a shortage of healthcare workers. Like that is not a concern, right? We have a shortage of doctors. We have a shortage of mammographers in the U.S., and it's even a more acute problem in the rest of the world. And of course, mammographers are radiologists specialized for breast cancer, right? We have a shortage, so we need to have tools to empower those doctors we have to be able to take care of more patients and better. And so that's where our CM triage product, that was a first and second FDA cleared, a cure metrics, we can detect that the 99% accuracy and triage of the mammograms where this batch, no anomalies detected, look at it when you have time. And by law, you look at all the mammograms, this other batch anomalies detected, you wanna get to these women ASAP, get it to your top mammographer, you're senior radiologist, and that's a huge workflow benefit, right? We have a lot of radiologists as investors or imaging center owners, and they'll say, you know, when I sit out at 6 p.m., and I'm gonna read mammograms for the next four hours, I wanna get those suspicious cases first thing, not at 9.52 p.m., and in the end of the day, the AI doesn't get tired, it doesn't need a coffee break, it doesn't get distracted, it doesn't have food coma, right? And so it's a tool to empower the radiologists to do their job better. It's not gonna replace them, in my opinion, anytime in the future. Move it on to cardiologists and oncologists, then in the day, you know, and what we do a cure match, it's a cure metrics and cure match under this AI met global umbrella, which is our kind of doing business as name. But you know, two severed Delaware C-Corp companies, we started with cure match. We started with a world renowned oncologist, Dr. Rizelle Kersrock, it's K-U-R-Z-R-O-C-K for listeners if you wanna look her up, you know, 800 plus publications had been an MD Anderson before Schumann and Morse Cancer Center, where we met her, and we ended up starting this company with her, 'cause at the end of the day, if an oncologist wants to recommend a three drug combination, there's over four and a half million combinations, it's beyond human cognition to do that. And so that's what cure match does, and we are not there to replace oncologists by any means, we're a tool, an augmentative intelligence tool, and the fact is the AMA, the American Medical Association, found out about us and encouraged us to apply for our CPT code last year, and we did that, and I'm very happy to say that we, on January 1st, the AMA announced the companies that got CPT goes across the board, and we were the first company of its kind in this sector, this digital health, augmentative intelligence, we're the first company to get that code, and it was no easy feat. We had to show the pathology caucus in the American Medical Association, the clinical validity of what we are doing, and so that speaks to the caliber of what we've developed, and then have oncologists that say we've changed their therapy decisions that but for cure match, they would have not thought of that therapy combination. And so this is truly where we get up every day, saying how can we do this one day faster, 'cause cancer has touched all our lives, including my families, and so if we can help get the best combination of treatments, first off, detected as early as possible, right? And whether it's tools like cure metrics for breast cancer, we develop other companies, of course, exact sciences, big public company has colon guard for stool, or you test a detect colon cancer, and there's others, but detection is paramount. You detect earlier the odds of survival are significantly higher for any disease, of course. And then the treatment, if you get the best combination, as early as possible, you have a much higher chance of survival or living longer, at least, as the oncologist say, increasing the progression free survival or the overall survival. So cancer is not a death sentence, but it's something like someone said, equated it to diabetes or HIV, these are, you don't wanna have these, but it's not the death sentence that it used to be, people live with these diseases for decades. - Right, right, okay. All right, so you mentioned both cure match and cure metrics already, and I wanna kind of get a little bit more in depth with both of these, because they're pretty interesting. So let's start with cure match, now you mentioned, so this is in the oncology realm. So this applies AI to oncology to determine the best treatment. I'm curious, can you kind of explain that a little bit more how that works? - Sure, no, no, happy too. So patient has cancer, arguably the worst news they've ever had in their life. Patient goes to the oncologist, or gets a recommended oncologist, so they take a biopsy of the cancer, whether a solid cancer from the tumor or liquid biopsy if it's a blood cancer, that gets sent to a lab. Now, it could be a lab core, but it could be, you know, these are public companies like Garden, or Foundation Medicine, which is owned by Roche, or a lot of community hospitals and centers like Cleveland Clinic have their own NGS labs increasingly, and there's private companies as well, and all over the world, increasingly, they're doing what's called next gen sequencing. So using an aluminum machine, a thermo-fisher machine, a company out of the UK, a Oxford nanopore machine, but aluminum is the 800 pound gorilla here. They will use these machines to sequence the cancer biopsy and that next gen sequencing, the NGS panel, which is what it's called, that is our input, a cure match. So we don't need to connect to CERNOR, EPIC, we don't need healthcare records, literally all we need is a 3031-page PDF of the, from the lab, or the oncologist sends it to us, right? That is our input. And from that, basically to the labors, and I'd say it's like the 23 in me of that patient-specific cancer, and cancer like two snowflakes never being the same or two fingerprints never being the same, no two cancers are molecularly ever-ever the same. So no two lung cancers, no two breast, prostate, et cetera. So it's an N of one, as they say. And so that's where, I made an earlier comment about different tools in the AI toolbox. Machine learning is not what's used on the cure match side. It is on the cure metrics side, because you train it on mammograms that are static and the more data you feed it, the better it gets. You say, hey, Navi, here's 10 million more, 20 million more, do you want that, of course? There is a diminishing rate of return, right? Once you're 94, 95, 96, 99% accurate, how much better are you gonna get in that specific task for that algorithm? Now, on cure match, if no two cancers are ever the same, how can you train it to detect cancer optimally? 'Cause you need it to be the same. And so the technology, the tool that's used is called K-R-R-A-I, knowledge, representation, and reasoning AI. And it's the same premise behind GPS technology, let's say. So Dennis, if you and I were in the same city, and let's say you and I and all your listeners were in the same exact spot, and we were gonna map one second after each other how to get to the airport. No one's gonna get the exact same results, 'cause traffic patterns are always changing. It's not static. And so that's the premise behind K-R-R-A-I, and that's where we use a cure match where we take that input from the NGS panel, and we have our database that we're always updating, we're always curating, we have a clinical science team, led by a two science officer, that's always updating this, always shining the apple, and then the algorithm will match and say, based on this patient's specific cancer, and there's no other cancer, like this person, this individual's cancer, here is the recommended three drug, two drug, one drug, combinations that will have, they'll have higher efficacy. And so that's where we recommended treatments that sometimes an oncologist would have never thought of, and they've set this test. A great example is with Merck's key true to drug, when it first came out, it was for lung cancer. And we kept seeing it pop up on some reports where like a stomach cancer, and we have oncologists say I would have never thought of that. So we saw this years ago. Now, of course, it's more common knowledge for that drug specifically, but there's also situations where a certain immunotherapy treatment might stop a certain pathway of a cancer, but then it causes hyper progression down another pathway. So in a sense, they may think we're stopping the cancer and how it responds to this pathway, but then it flows faster. It's called hyperaggression, and it's in fact long-term worse for the patient. And so that's where our system can detect that. And we've had an oncologist at a very highly esteemed institution that knew this on a one-off basis because he saw it, but 80% of cancer patients are at the community hospital level, and even in this case, as an individual doctor, he was aware of this. It doesn't mean that all oncologists are. And so that's where, again, we're here as a tool, just like Excel is there for CPAs. It doesn't replace the CPA. It helps them do the job more efficiently and accurately and faster and process more data faster. Cure matches like Excel for oncologists, right? We're there as a tool for them to use, bringing that augmentative intelligence to your earlier comment. - Okay, that's really interesting. Now, this database that you're talking about, the database of treatment modalities, I mean, this is constantly being updated as new treatments come out. - Constantly, constantly. So there's roughly 350 FDA-clear drugs right now, specifically for oncology, right? As a new one is cleared, this isn't a difficult thing to track, right? It's not every day that you get FDA-drug cleared, and there's papers published and esteemed journals and whatnot. And so we have a clinical team that's led by our two science officer, a lady by the name of Ali Perlina, and she was with Craig Vettner, who, again, Craig is famous for being first human genome mapped. He started a company called Human Longevity Inc. That Ali was at. And in fact, he started the oncology division that a lab called Neogenomics Purchased, and then went to Vyome, which is a microbiome company. Naveen Jane had started that, and now there's a billion dollar plus valuation, and it's all about your microbiome, a probiotics and whatnot. But she was instrumental in developing what they had there. And so we're fortunate to have her join us a couple of years ago leading our clinical team that is, again, always curating and with new research, whether it's by others or by our own team. And so with that, and always, with AI, it's never, you're never done and say, okay, here it is. You're always shining the athlete. There's always improvements in iterations. But we have that competitive advantage where with K-R-R-A-I, it's not about throwing more money at it. It's that head start, right? And an example to give. And it's in the news now with chat G-B-T and Microsoft, right? But the premise is at the end of the day, Microsoft has all the money in the world. They could throw it in their Bing product. Is it ever going to catch up to Google search? I'm in the highly unlikely camp that I'll have. - Yeah, no. - And so that's the same thing. We have this head start. It's not about someone throwing more money at it. That's K-R-R-A-I, right? It's on machine learning. That someone says, okay, I have talented scientists. I have all the money in the world. I can purchase all the images I need. I'm going to train it and I'm going to make more accurate algorithms. I'm simplifying it. It's really not that simple even with machine learning. But it's even more challenging with applying K-R-R-A-I. - Okay, that's actually at the K-R-R-A-I is something I have not heard of before. So I'm going to have to do a little bit of reading - Okay. - You know, look it up. You'll find content on that. It's not in our lexicon as much as, of course, natural language processing and machine learning 'cause these technologies are in the news 'cause there's so many consumer facing applications. Be it our Alexa's or our series. So it's more common knowledge and common use, right? But yeah, it's definitely in the healthcare world. And again, our body's being software. If you're going to bring true precision medicine, you got to base it on that patient specific cancer, specific disease. Yeah, no two cancers are molecularly ever the same. So that's where we're bringing true precision medicine in that regard. - Okay, okay, that makes sense. That's actually pretty exciting. This is the People of Pathology podcast with our guest, Navid Alipore. We'll be right back. Labvine is an interactive online learning platform where laboratory professionals learn, develop and discover by sharing knowledge and building on each other's experience. The platform provides global access to internationally accredited laboratory specific courses and other resources developed by lab specialists like us for the laboratory industry. Labvine is free to sign up and you can use the link in the show notes to get started. Okay, whether you're working hard at the grossing bench, the autopsy table, behind a microscope or any other area of the medical laboratory, there is one thing that we all need, comfortable scrubs. The scrubs that I wear come from dress and med. This is a company in California and they've been making high quality scrubs since 1980. They have a variety of styles and colors to choose from. As a matter of fact, I just bought a set of the new soft stretch scrubs and I gotta tell ya, they are so comfortable. I might even be wearing them right now. You can check out dress and med by following the link in the show notes. Oh yeah, and while you're there, make sure you sign up for their loyalty program where every order will earn you points towards special offers and discounts. Now for the rest of my conversation with Navid Aliport. I'm a people with pathology podcast. Now, so cure match, it seems like it works on pretty much any kind of cancer, right? - Well, good question. So we are in what's called pan cancer. So yes, any cancer. We are the most useful is frankly where the cancer is the most complex, the nastiest. It has more variance. There's more pathways. So it's harder to hit all of them. That's where we're the most useful. Some cancers, and by the way, there are certain cancers that are just naturally or the mortality rate is much higher. Whereas cancers like prostate or breast cancer or Hodgkin's to the odds of survival are very good if you can catch them early enough. But we all sadly know people that were very young and died of breast cancer or prostate or Hodgkin's. And in those cases, that was a nasty cancer. It was more complex. It had more variance. And so that's where the NCCN guideline and standard of care didn't work. And so that's where we come in, where we're more useful on those complex cancers, where if a cancer doesn't have as many variants, we'll validate and say yes, the standard treatment that is given for Hodgkin's lymphoma is the best one for this patient. So we'll validate it in those cases. - Okay, I see. All right, so then let's talk about cure metrics, which you mentioned earlier as well, 'cause this is specifically for breast cancer as far as, and it analyzes the radiology images. So how did this come about? - So cure metrics was, in fact, the first one we started, getting cure match came after the fact 'cause my business partner got cancer and that's how we met Dr. Kerr's Rocky the Morse Cancer Center. So if he didn't have cancer, this company would never exist. But cure metrics was the one where these literally rocket scientists that had done work at Los Alamos and at NASA and were in UCSD at the time came to us and we started that company first. And it is specifically for women's health, not to say that as we mature, we can't apply our AI to other modalities, to detect other diseases that would help men and two, whether it's from X-ray or MRI or ultrasound, but mammograms are for women, right? And so we're women's health focused at cure metrics. And so right now the two products we have are for detecting breast cancer earlier and better and for detecting heart disease from the mammogram. So that's something that no one else does. So not to take anything away from breast cancer, but our golden goose, so to speak, where we're gonna save the most lives is on our heart disease detection. And so how do we do that? Well, in all our bodies, male and female, as we get older, we develop calcification in our arteries. Now some people develop it faster genetically, some even demographically, some people, and so there's calcification that then builds up. Now we detect it from the mammogram and the arteries and capillaries and the breast tissue, it's called microcalculation. specifications. So we can detect it and we can score it. And if it's at a certain level, it's called this back, it gets a back score, breast or serial classification score, then that's where this goes, it can be used to get that patient to a cardiologist to do that EKG, to do that stress test, to get on a statin medication, which a statin alone reduces a risk of a cardiac event by 50 to 60%. So where this is valuable is that, you know, with women, heart disease, heart attacks are called the silent killer. 65% of women die on that first heart attack. The first one is the last one, whereas us men tend to have chest pain, we have shortness of breath. You go to the doctor earlier and they tell you how bad you've eaten and, you know, you don't exercise and change your lifestyle and get you on medication. And hopefully, you know, it's early enough where you don't need any surgery or you can delay that. But with women, because they're asymptomatic, they are, you know, Lisa Marie Presley was in the news recently, right? 54 died of a heart attack. Now, you know, we don't know her other health conditions. But sadly, I, I know, you know, 58 year old woman, you know, a triathlete dropped of a heart attack, didn't know she had heart disease. So imagine a woman goes in for a first mammogram and she's in her early 40s, let's say. And so she's 42. She's not a triathlete, but she's in decent shape. She eats relatively well, not obese again, doesn't know of heart disease. And she goes in for that first mammogram for breast cancer, but now two for one, no extra radiation, no extra discomfort. She also gets a breast arterial calcification score. And they say, you have calcification. You need to go see a cardiologist. Now, at the age of 42, she's seen a cardiologist to do the EKG, the stress test, get on medications. And that's going to allow her years of time to delay or eliminate a cardiac event. So instead of just, you know, being a walking heart attack and, you know, at 48 or 52 or 54, if she can address it at 42, you just added decades of life. And so that's where what we're doing truly, it's with urgency and, you know, eye-family members that for me again, it's very personal. Because if you can detect, of course, breast cancer, but heart disease earlier, there's a lot that can be done, whether it's lifestyle changes or getting on heart disease medications or getting that medical procedure that's needed. You know, for me, that's what drives me every day is that, you know, we're going to be able to impact, you know, countless lives at a global level. Because again, we're 100 percent cloud-based. We could go into entire countries and process mammograms and detect a woman that are walking heart attacks that need to go see a cardiologist up right now. They have no idea. That's pretty amazing that it has that kind of double use like that. I like that a lot. So is there been any work with, with cure metrics to kind of apply it to other, other types of cancer, I guess, other sort of types of radiology images? Not yet. We focused on, you know, that, you know, just again, not just for health care, but in a startup world, right? Do laser focus and being the best that, you know, one or two things and then at the right time move on to other applications. So our focus has been specifically around women's health and specifically around breast cancer and heart disease. Again, at the right time, we could apply our expertise to data collected from MRIs, from ultrasound, from X-rays. That would then benefit not just women, but men who get an MRI for, you know, detecting other diseases. So it's definitely on our roadmap. So we're keeping some of that close to the vest. But right now, we're focused on breast cancer and heart disease for a woman. Okay. Makes sense. What about like other types of images? Because I mean, of course, this is a pathology podcast and pathology is full of images as well. Has there been any look at that? So we have not ourselves, but of course, there are a lot of companies that are digitizing those images, right? Instead of sending the pap smear, for example, sample, you know, by FedEx or by mail, right? There's increasingly there's digitization happening. So it's not a focus of ours, but absolutely, there's I think incredible progress being made because the first step is in this curation of data. And so when it's digitized, it's much easier to do this. And you know, this goes to the premise that, you know, we all know that now there's a promise, you know, data is the 21st century's oil. Well, that data is useless, unless you can get it out, get that oil out of the ground, refine it and turn it into gasoline. I like my analogies in case you can't tell. If you have data, you know, the AI is the refinery to then, you know, refine it and, you know, detect that oil, detect that gold, detect what is valuable and that information. But you have to, there's also saying, you know, garbage and garbage out. You have to collect that data and curate it and then be able to train the algorithms on like kind data. So that's digitization of the images is the first step because then it makes it capable to, you know, feed it to the algorithms to train it to do what you want it to do, detect the anomaly you want it to detect. Okay, that makes sense. So then, you mentioned earlier that you think that the 2020s are going to kind of be the time that AI really takes off and it seems like that's going to be true. So looking at then at the future of like AI and medicine, what do you think is going to be next as far as like the big thing? Well, I think if you put it in three buckets, DTM, detect treat monitor. So anywhere you can use it to process information to detect a disease or a symptom or a cancer or diabetes, you name it, apply AI to neonatal images from ultrasounds, you know, a woman who's pregnant gets multiple ultrasounds right during those nine months, anywhere you can apply it to data to detect something that does not belong, that is not, that is going to automate a task. So the human doctor can focus on more important things or can take care of more patients and reduce human error, you're going to have an impact, right? And so detection is the first thing that has to be done. And it could also be, again, technologies like our Apple watches or, you know, Fitbits or other wearables that now track our sleep and our heartbeat and our breathing and, you know, I know again, someone very close to me a couple of years ago, the watch kept detecting high resting heartbeat. Well, that led to going to the doctor to see what that was, because it never happened during exercise, it happened, you know, having dinner, reading a book, having a glass of wine, you know, it's a high resting heartbeat. Well, this led to going to cardiologists to doing EKG's stress test and ultimately leading to getting on medication, but for that watch, would have never happened, right? And so, and by the way, I'm not getting sponsored by Apple, I wish I was, but, yeah. Right. And then of course, on the treatment side, again, it's, you know, whether what we're doing a cure match to recommend the best combination of drugs earlier, you know, with, you know, other technologies to, you know, for applying to other diseases to develop new drugs faster from, you know, the billions of different, you know, combinations and molecular compounds to streamlining the clinical trial process. So actually, in fact, I didn't talk about it, but a cure match, we have a clinical trial intelligence platform where it works out, we don't have to develop it, we could help find the right patients for trials and the preclinical trial setup that will have a better response to that drug so that if you can get the right patients for a trial, the efficacy is higher, and it clears faster and you get the drug to market faster. So instead of taking 10 years and a billion dollars on average, you get a cancer drug to market, even if you shave 10% off that, and it's nine years and $900 million, not only are you going to have people that benefit faster, but of course, the farmer company is going to make more money faster, and at the end of the day, you've got to follow the money, right? The, you know, who's going to make money or save money, and healthcare is the same, right? And so you always have to look at how are we going to, who's going to save money, who's going to make money, and of course getting the drugs to market faster, you know, benefits, you know, those that'll, you know, get it as opposed to not getting it, if it didn't get out to market earlier. So, and the last thing in that bucket, detect treat monitor, monitor is, you know, better, you know, blood tests, better, you know, genomic tests, you know, of course, being people being proactive, better stool sample tests for, you know, colon cancer and other things, you know, and people being proactive and, you know, advice is always easier given than taking, but, you know, going in and getting monitored and seeing your primary care doctor and being proactive instead of reactive, because these technologies are getting better and better, you know, again, processing mountains and mountains of data to, to tech treat and monitor better. - Okay, I love it. That's a really positive view of the future and kind of the future of healthcare. And I think, I think, you know, we'll all kind of benefit from that. So that's a great message, I think, to end on. So, Novi, this has really been an interesting conversation. I, you know, talking about AI is kind of a, like a personal interest of mine. So it's fun to be able to have that kind of conversation. So I appreciate your time. I appreciate talking about cure match and cure metrics. So, Novi, to Aleport, thank you very much. - Dennis, thank you so much. It's been a pleasure. - If you're looking for another episode of the People of Pathology podcast to check out after this one, here's a preview from my interview with Dr. Heather Couture as we talk about AI and machine learning in pathology. - So I work with R&D teams who are looking to bring in the latest research to advance their, their algorithms applied to pathology images. You know, this research area is advancing so rapidly it can be hard for them to figure out which new algorithms to apply, you know, which new techniques to implement to solve the problems they're having. So I try to stay at the forefront of this research to help guide them and also sometimes help with implementation to advance their algorithms and to get them, you know, into the products and services that these companies are developing. - Is it hard to convince people that they need this kind of technology or, you know, with digital pathology becoming so popular now? Is it more something that they're familiar with and they're just trying to keep up? - I think probably both are true. The client base that I work with are already on board with machine learning. They might be experimenting with it a little bit. They might know they want to use it but haven't done anything yet. Or they might already be quite advanced down that line but are still encountering challenges. So there's definitely the other subs that you talked about that aren't on board yet. They're just, you know, not my current client base. - You can hear the rest of my conversation with Dr. Heather Couture in episode 63. All right, great big thanks to Navid Alipore. This was a really interesting conversation. And like I said at the beginning, I mean, this was kind of outside of pathology a little bit but of course, both oncology and radiology have many interconnections with pathology. So I think it is applicable to what we do. And you can definitely see how the technologies from both of these companies could be used in pathology. I mean, with cure match choosing the oncology treatments, I mean, that is similar probably to maybe choosing IHC stains or molecular tests to be ordered. And with cure metrics, the technology that they use in analyzing radiology images could easily be applied to pathology images as well. And I think both of these things really are happening to a certain extent already in other places. And when it comes to K-R-R-A-I or knowledge representation and reasoning AI, this is using AI to solve complex problems such as diagnosing a medical condition. And it borrows from areas of psychology and logic and other things to mimic the way that humans actually solve problems. So I like this stuff. I think it's exciting. I hope you do too. And it'll be interesting to see what these technologies bring in the next five to 10 years. As always, I'll have links in the show notes to everything that we talked about today. Don't forget you can follow this show on Twitter and Instagram. I'm at People with Path, or just find me on LinkedIn. Thank you for continuing to share the show with others. Together, let's inspire the next generation of pathologists and laboratory professionals. This show is a member of Health Podcast Network, which connects listeners with conversations and stories about health, care, and well-being. You can find a link in the show notes to Health Podcast Network. When in while you're there, check out some of their other interesting podcasts. Thanks for listening. I'm Dennis Strenke, and I'll talk to you next time. I'm the People with Pathology podcast.