Dr. Junaid Kalia: Good morning everyone. I'm excited again for an amazing episode of Signals and Symptoms with our core experts, Edward Marx and Harvey, and an amazing guest, Dr. Anthony Chang. I mean, if I have to start talking about him, it's going to take a long time, approximately a whole hour of what he has done for physician community in general to make them interested in AI.

To further ado, I'm going to let Harvey reintroduce Anthony and then talk about AI Med and then we'll go from there.

Dr. Harvey Castro: All right. Well, first of all, I want to say I'm honored to have him here. He's also a great mentor friend. We've met, we've talked in text and we geek out a lot on this AI Med and really big pictures. What can we do in humanity? So I'm really excited about this conference. I am going to literally come from my TEDx stage, literally to get to San Diego, to get there, to be there. And unfortunately, I can only be there for a few hours, but I'm literally going to fly all the way there from Dallas just to be there because unfortunately I have to be in Miami, but I'm blessed to have met him. I know this sounds like we all are excited about him, but he is a true leader and thought leadership and mentor and friend. So let's just jump right in. Before we start, Anthony, what do you want to tell us also about this amazing conference? Just to put it out there to everybody.

Dr. Anthony Chang: Well, thank you so much, first of all, for the accolades, but, you know, I'm humbled by all of you because I'm just trying to convene and gather smart people to transform medicine, particularly clinicians. I don't want clinicians to be left out of any discussion for the future of medicine. It's ironic. I just got off the zoom call with my publisher for the Journal of Intelligence-based Medicine.

And despite the challenges trying to create a publishing platform for particularly younger investigators from developing countries are trying so hard to have a stake in AI and healthcare. And the good news is I'm able to convince my publishers to continue to give them an opportunity to publish. The world, even though I have a Harvard and Stanford background, ironically, I'm very much supportive of everyone else too, in terms of having publishing opportunity and AI and healthcare, Junaid, is not about publishing in Lancet or having a publication in New England Journal. It's much more about bringing the AI capabilities as a resource to rural hospitals in the US, to developing countries around the world. And I'm a firm believer in that. I'm actually going to go back to China today to emphasize that. That it's not about the elite institutions having a lot of publications. It's much more about transforming medicine around the world and global health with this resource that's amazing.

So speaking of which, so education and global health, both things I know both of you believe in are going to be a big part of the meeting AI Med this year. So the AI Med is going to have, it's been in existence for about 11, 12 years before AI got popular as a topic. And we were expecting over a thousand people in San Diego with lots of breakout activities. So we let the registrants decide on what topics they want to hear about up to a month before the meeting. So we already have breakfast workshops on the hot topics. You can probably guess what they are. Ambient intelligence, agentic AI, digital twins will all be represented. And then we also have for the first time, a major presence from the chief AI officers. We have about, I think, close to three dozen chief AI officers coming. So they'll have a closing day Q&A from the audience, but they'll also have several meetings throughout the meeting. So it's a very, very exciting meeting. I would say arguably the best one yet because of the diversity of speakers from all over the world. We have a company that's going to be nominated and win the company of the year, it's actually not from the US. So I'll keep you in suspense. And just amazing mix of clinician users, hospital administrators, investors, entrepreneurs. I think it's the most balanced in terms of attendees. It's not just C-suite executives, but also we're proud to have 150 abstracts that will be presented throughout the meeting. We have a Shark Tank event and then an awards night to award, to give people recognition that I think deserve recognition. So very, very exciting three days. And then we have the American Board of AI and Medicine, a short version of that course, the day before Junaid on the 9th of November. We'll have a special edition of the ABIM that's now reaching 3,000 past attendees, including both of you. And I just got back from China giving an ABIM course there. And we're going to be heading into Canada, I'm happy to say, next year, as well as probably the UK. So we're now international, which I think is what clinicians really have needed is some validation that they are at a certain level. Obviously, AI and health is way above even the certification levels. But at least it gives them a head start on their journey.

Dr. Harvey Castro: Awesome. Awesome. Hey, I just I know you mentioned digital twins and I'm going to start off with the question for digital twins. I know you're a pediatric cardiologist. I know you have two amazing young ladies and I know that ties into your passion of health care and AI.

Anthony Chang: He means daughters, by the way, not girlfriends.

Harvey Castro MD MBA: Yes. Yes. Yes. Two amazing daughters.

Junaid Kalia MD: Harvey gets us all in trouble.

Dr. Harvey Castro: No, but Anthony knows me well. I love his daughters. They're great children. So with the digital twin aspect of it, I'd love to hear your thoughts on digital twins.

Dr. Anthony Chang: Yeah. As a pediatric cardiologist for 40 years now, incredibly, we're leading the AI track at the World Congress of Pediatric Cardiology and Cardiac Surgery in Hong Kong coming up in December. Year after year, time after time at the World Congress, we're always grappling with the same challenges, right? What to do with this lesion and what to do with this situation. And I think I bring in what Einstein used to say, which is, you know, insanity is doing the same things over and over, expecting different results. So I think we've reached insanity officially with all the registries and all the randomized control trials. We haven't been able to come up with answers. So I'm proposing, and I think it'll be probably a little bit shocking at the Congress that we don't abandon the registries, but we try to amplify the benefit of registries and having a lot of data by coupling it with digital twins. So there'll be, I think relative news as far as I know at the World Congress. I want the pediatric cardiologists around the world to start considering the future of clinical research, which is digital twins. I call it N of 1X research. So you create copies of the same patient and apply it to what's available, combining statistical with a mechanistic model to really get the answers to the questions that we've been asking for decades, but haven't come up with any answers.

So as you know, Harvey, my own daughter was debated amongst the surgeons I know in terms of what to do with her heart lesion. And I wish that we had a digital twin for her back when I had to make that difficult personal decision, as well as a clinical decision. Because I think that would have really, really benefited the decision-making progress. To basically have virtual copies of the physical person, and I anticipate this will be by the end, by another decade or so, will be very routine that we all have our health digital twin and then have projections in terms of what happens if we have certain situations and conditions, and what happens if we go on different medications. That will be routine within a decade. So until we get to that point, I think we all have to really, there's definitely a hype about digital twins. And I hear people talk about digital twins. And when I say, what's digital twin about this digital twin? It's not digital, nor is it a twin. But I think they're just using it as a catchphrase. And Alfonso and I are actually working on digital twin. We know how hard it is. So it's not to be taken lightly as a convenient buzzword, but it's more of a philosophical change in how we look at clinical medicine, which I think I'm envious of the younger generation. They're going to have all of this at their fingertips. It's going to be truly exciting.

Dr. Junaid Kalia: Very well said. Again, just to remind people, audiences, do register. The link is in the description. And make sure that you select the topics while you are attending them. I promise you there are people that I've met there before that has helped me for free. That's insane. This is not just given that Anthony puts this large hearted, generous aura over the whole conference, people come in.

Dr. Harvey Castro: I get this question a lot and I'd love to see what you think. What is your opinion? What are the biggest barriers to AI adoption right now in healthcare, Doc?

Dr. Anthony Chang: I think clinician awareness and education and expectation and understanding. I think it's the... I always half-jokingly say the human factor is what gets in the way, but also what will get things done in AI and healthcare. Ironically, something that's technological is going to require humans. One of the comments I made in China was there are 26 Chinese astronauts, and most of them are engineers. There are not 26 clinicians who can speak data science that I can find in China. And I was not in a small city, I was in Beijing. And I asked the big audience, how many of you consider yourself clinicians with an AI education or a data science background? And not one hand went up. So I think it's just the lack of education and awareness, as well as a sense of what's possible and what we should look out for. And I think there's perhaps excessive fear about it too, because, you know, agentic AI should be sort of autonomous. And that word autonomous scares a lot of people, understandably, but also we need to get over the fear of something.

And when Jesse Ehrenfeld, as you know, the past president of the AMA asked me at an AMA meeting, it was kind of a fireside chat situation. And he said, what keeps you up at night? And I said, it's actually not what you think. It's not AI going crazy. It's us not using AI, the non-use of AI is what keeps me up at night. Because if you think about, I recently had to deal with a relative with a complex medical situation, fortunately, I'm an insider, I can navigate the system, but I can't imagine someone who's not an insider in healthcare trying to navigate our healthcare system in the US. And it's just as bad in China, you know, it's not better in other countries. So I think the non-use of AI is what keeps me up at night and what worries me.

Edward Marx: What would you tell a board, like generically, like they're coming at this, they know a little bit about AI, they're the average consumer, and they're wondering how do we propel our organization, our academic medical center, how do we propel forward from a board point of view? So what would the message that you would give to a board of directors that's not, again, they have average understanding of AI and they know we need to pursue, but they're very cautious. So what would be maybe one or two things you'd tell them?

Anthony Chang: Yeah, as much as I'm an anti-war person, because they always asked me like, tell us something relatable. And I said, it's kind of like fighting a war now without drones. You know, you can obviously count on people being courageous and loyal, but that kind of mentality and dedication is going to really fatigue in a war without technology supporting them. You know, I used to say it's kind of like the stinger and Patriot missiles, but it's actually like drones. Drones is even a better analogy. The drones will totally change the perspective of you having a disadvantage of numbers or capability. So it's the great equalizer, right? So I think AI and healthcare Ed is kind of like the equalizer. If you're already a pretty big institution with a lot of reach, it's going to amplify that. If you're a smaller healthcare organization, I agreed to just this morning to speak with a healthcare organization in a rural part of the US, and they're anxious to get their AI going. So I think the desire to do this is a good start, but they have to do this in a very strategic and financially feasible way.

I think nurses are, as I said, the fastest growing segment now amongst the clinicians. And I'm hoping others will follow. So the reason to come to AI Med is, first of all, it's an amazing community of people, those of you who've been there. You feel a real sense of belonging in a sector that is just growing. And we have actually also breakout sessions with nurses or for nurse practitioners or for pharmacists. And the ABIM course is a nice bonus the day before. And also you'll hear directly, directly from a large group of chief AI officers now that are in the middle of all this. And most of them are pretty honest and candid about everything. So, and then for the first time, we actually have what we call AI med match, which is companies and organizations looking for AI and healthcare talent and people will be able to talk to these organizations and enterprises.

Dr. Junaid Kalia: If AI can diagnose patients, what are doctors for? And this is the New Yorker. And then this actually, I'm just going to be honest with you. I mean, we have built something very amazing in the background that is releasing for what we call acute stroke care, exactly multimodal time series and all of this. And we're releasing it soon, actually next year or so. One of the fears when I was presenting it like, Junaid, what are you doing? Why are you doing this? I mean, do you not want your job anymore? And now go ahead and address this hard question. And by the way, you should come. That's why I keep saying you should come to these conferences and learn because you are going to. It's not about replacing you, which again, Anthony's going to explore more. But what I'm saying is you need to learn because the way exactly you need to learn how to use a stethoscope.

Dr. Anthony Chang: Yeah, I think if you look at astronauts, they've transitioned from being pilots to being engineers. And I think I'd love to see the same transition in healthcare, that clinicians take on the additional domain of data science and really find the best use. Because the clinician with the data science background are going to come up with the best project ideas and the best solutions. Data scientists, and I have a high regard for them, are not necessarily the people that come up with the best clinical ideas and problems to solve. Clinicians without a data science background though, really don't understand how to go about answering a question with data science. So the bilingual cohort will really drive the future of healthcare and AI and make the job more meaningful.

I think in answering, I have only perused that article, but my answer to that question would be there are lots of things that clinicians are going to be able to do still. You're going to be making very difficult ethical decisions, although AI can do a pretty good job already. They can help promote critical thinking and clinical judgment. I'll give an example of how AI is not working. I just came back from China and my colleague that decided to go later than I did with the same hotel, I paid like $300 a night and she was quoted $4,500 a night. Now that's because the machine learning is saying there's a high demand, there's low supply, but that just makes no common sense, right? And you're going to find that even with the best AI tools, it's going to come up with recommendations or solutions that are just not making sense or not feasible or not logistically possible or too expensive, whatever it is. So the weight of the recommendations still needs humans who have actually lived, and who live in healthcare will have to help out the recommendations. The recommendations are just coming in because it's so stochastically makes sense, but it's not gonna have that common sense and definitely not enough clinical judgment for at least a decade or two.