transcript

Speaker 1:
[00:00] How should we really think about imaging today?

Speaker 2:
[00:02] These miraculous devices we've built are important to understand because they're going to change our lives. Maybe we don't need to wait for a doctor to have already found a problem.

Speaker 1:
[00:11] I kind of want to talk about this whole idea of false positives, which is something that people will push back on.

Speaker 2:
[00:16] I think everybody should have a baseline.

Speaker 1:
[00:17] You talked about this moment we're in, which is comparable to the invention of the telescope. Not just an incremental change, but more of a quantum change.

Speaker 2:
[00:24] Wait a second, if we can see this stuff, maybe we don't need to wait for a doctor to have already found a problem. I think that's this cusp that we're on where medical imaging is really changing.

Speaker 1:
[00:34] My guest today began his career at Harvard and MIT, spent years as Chief of Innovation and Radiology at NYU, and has developed imaging technologies used to guide the care of billions of people worldwide. He's now Chief Medical Scientist at Function Health. This is Dr. Daniel K. Sodickson. Dan, welcome to the podcast.

Speaker 2:
[00:53] Thank you so much, Mark. It's great to be here.

Speaker 1:
[00:55] I'm just in awe of you. As I was preparing for this podcast, I was like, wow, this dude's got quite a pedigree, and he is rethinking how we think and apply imaging. We're going to talk about that today, which is this new craze of full-body MRIs, what's the deal with it, should we be doing it, what are the benefits, what are the risks, what are we looking for? We're going to cover all of it, and we're going to talk about how to be proactive about your health. We're going to talk about some of the changes in AI and medicine, and some of the things that are happening on the horizon that are pretty sci-fi, and wild out there, like maybe MRIs everywhere, in your chair, in your bed, whatever. I don't quite get it, but you went to Yale undergraduate, studied physics, you got your humanities, bachelors in humanities as well from Yale, then you got your Harvard Medical School degree, MIT degree in physics, PhD in physics. I'm like, you kind of been around, you were the head of MRI imaging at Beth Israel Deaconess Medical Center at Harvard, and now we're working together, which is so amazing. For those of you listening, Dan and I are part of a company called Function Health. You might have heard me talk about it on the podcast. Dan is the Chief Science Officer, I'm the Chief Medical Officer, and together, we're co-directors of what's called the Medical Intelligence Lab. We're going to talk about what that is, what it means, and why you need to care about it, and how it applies to you and your body and your health and your long-term outlook for well-being and how you can live 100 healthy years with proactive health care, and that's what we're about. It's really empowering you with the data, the information, the knowledge to actually live a long, healthy life, and you feel good now. I always say, want to feel 100 percent, live 100 healthy years, so that's the goal. I want to zoom out. You just wrote a book, it just came out in October, The Future of Seeing, and it's really about the lenses we look at the world through, from the macrocosmic world of stars to the microscopic world of cells and microbes, to all the imaging that we now have access to. We're just extending our capacity and our vision. I'm curious about what inspired you to write this book? What are you hoping people understand from it? How has our ability to see changed over human evolutionary biology?

Speaker 2:
[03:05] Great question, Mark. First of all, let me just say the I is mutual. But no, the reason I wrote the book was that there's this sort of weird paradox in imaging now. We lead more imaged lives than we ever have, right? I mean, you can't walk down a street without being imaged by a whole series of cameras.

Speaker 1:
[03:25] That's right. Facial recognition.

Speaker 2:
[03:27] In utero on, and yet the mechanisms of imaging are more hidden than ever. How many people actually understand how an MRI machine works or a radio telescope works? So imaging kind of has an image problem. And what I wanted to do was give imaging back to people, to connect it to the biological vision that we evolved, to remind people that we're actually all creatures of imaging. And these miraculous devices we've built are important to understand because they're going to change our lives.

Speaker 1:
[03:57] We really only understand the world through our senses, right? And the ability to extend our senses, to look under the skin, it's pretty remarkable. I mean, vivisection, which is the human, the section of the body, was done in some ancient cultures, but often was not because the body was considered sacred, like in Chinese medicine, they never did that. So they would kind of had to intuit how things worked without actually knowing anything about what's happening on the inside. And now, we had the crude imaging with X-rays back to the turn of the last century. People used to go get shoes and they get X-rays to look at their feet, which was a bad idea. Or people had radiation X-rays for their acne on their face. Bad idea, caused a lot of cancer. So we've kind of had this sort of interesting history. And I'm older than you, but MRIs were kind of a new thing when I was in medical training. It was in the early 80s and it was just kind of coming on the horizon, our ability to kind of look deeper into things. We first had X-rays and then we had CT scanners, we have ultrasounds, we have MRI machines. There's other kinds of imaging out there as well. How should we really think about imaging today? Because we see a lot of kind of hype out there. Kim Kardashian goes to get a scan and everyone's like, oh wow, what is this about? I want a full body MRI. How should we be thinking about this?

Speaker 2:
[05:20] So I think first of all, you're absolutely correct that extending our senses is a really fundamental thesis of imaging. And I would argue that every time we extend our vision, we invariably expand our minds. We saw that from the Copernican Revolution. Basically, it was the results of imaging devices that forced us to reckon with the fact that we're living alone on this little rock in this vast universe. X-rays completely took the world by storm, like you said, as did tomography later on. And so I-

Speaker 1:
[05:56] Tomography is what?

Speaker 2:
[05:57] And forgive me, yes, tomography-

Speaker 1:
[05:59] It does not allow us to speak physics.

Speaker 2:
[06:01] Exactly. It's CAT scans, MRI machines, PET scans.

Speaker 1:
[06:06] Complicated assembly of images from different sources.

Speaker 2:
[06:09] And really what it means, it comes from a weird Greek root, which stands for the writing of slices. And really that's what all of these modern imaging devices are doing. They're slicing through the body every which way without making a single cut. And I think when you ask how to think about modern imaging, that's really what modern imaging is doing. It's capable of basically dissecting the body without ever cutting into it. And it's become this integral tool in medicine that people use to diagnose disease, to guide surgery, all of that. But as you said, its use is starting to change. And people are realizing, wait a second, if we can see this stuff, maybe we can see it early. Maybe we don't need to wait for a doctor to have already found a problem. I think that's this cusp that we're on, where medical imaging is really changing.

Speaker 1:
[07:04] Yeah, because most doctors will only diagnose you when you have a symptom. Oh, I have a stomach pain. Maybe we should get an MRI of your stomach, or I've got a head pain, or I'm losing vision, or I can't walk. Maybe we should get an MRI of your brain. And what you're suggesting is that that might not be the right way to think about things. That's a proactive, preventive way to think about imaging. And we talk a lot in this space around what we call P4 medicine, which is Lee Roy Hood's vision, who's a systems biologist of how we need to think about health, which is preventive, it's predictive, meaning you can kind of predict where you're going. It's personalized, so it's pretty different. And it's participatory, meaning we all have to kind of participate in our health, not just passive activity. What's happening now is that the speed of imaging, the application of AI to imaging, innovations in imaging, the deflationary cost of imaging, are all starting to get ahead at the same time. And you talked about this sort of moment that we're in, which is comparable to the invention of the telescope, in terms of our understanding of the technological changes, not just an incremental change, but more of a quantum change. Can you kind of unpack that? Because I think most of us just think, oh, we go to the doctor, we get imaging, we got a symptom. But you're talking, and even me, I'm the doctor, and I'm still curious about it, because I don't understand what you're thinking about, how this change is so revolutionary.

Speaker 2:
[08:23] Let me attack that basically through a bit of a personal story, because I started out in kind of a traditional way of thinking about imaging. This is the tool you use to open up the body for inspection by doctors once we want to find something wrong. And what happened is over time, as I worked more and more on optimizing, for example, these MRI machines, making them faster, making them better, I realized that a lot of the time they were being used to chase after symptoms. That we were telling people remarkable important information, but we were telling them too late. Like, oh, gee, I'm sorry.

Speaker 1:
[09:00] Sorry you've been on medicine.

Speaker 2:
[09:00] Have an advanced invasive cancer. Is it any surprise that radiology departments don't get as many philanthropic donations as say surgery departments? We're the people who tell you you're sick. Right. And then we hand over to someone else to fix it.

Speaker 1:
[09:13] Let's see.

Speaker 2:
[09:14] And it started dawning on me that maybe there's a way we can use these tools, our kind of best tools for visualization, first rather than last. But that involves overcoming a few obstacles. First of all, they're big and expensive, right? So a lot of people say, oh, we can't do imaging early because it's going to rack up medical costs. Then also, we have this weird problem that we see too much. If you put somebody in an MRI machine, you're going to find a little ditzel here or there. There's always going to be something you find which raises a question. And so this raises the whole big question of false positives.

Speaker 1:
[09:55] Meaning that you see something on there that looks like something, but it's really nothing.

Speaker 2:
[09:59] Exactly.

Speaker 1:
[09:59] Could this be a tool or? You get worried about it and then you chase it down and it creates worry and costs and interventions and…

Speaker 2:
[10:06] All of those things. And those are the entirely understandable reasons people haven't used imaging proactively in the past for fear of running up those costs and creating that anxiety and giving people sort of these unnecessary tests. But as a physicist and a designer of machines, I started wondering, well, can we drive those false positive rates down? They're not God given, right? They're not somehow attached to the devices. They have to do with the way we use the devices. And so what occurred to me after some time is, well, the problem is that we're not actually putting these images in context. We're used to getting these images and then looking at them that day, seeing what we see and saying, ah, you know, we're the, we're the wise philosophers, you know, peering at the images and saying, well, this is your future.

Speaker 1:
[11:00] Yeah.

Speaker 2:
[11:00] But if you want to predict the future, you should know the past. So what if we had previous images? What if we had a whole series of images over time? Then we could say, you know what, I see this thing here, but I know it's normal for you. And in fact, radiologists do this all the time. If they see something and it hasn't really changed from last time, they might say, you know what, I'm not too worried to come back in six months, come back in a year.

Speaker 1:
[11:25] We call that an incidentaloma.

Speaker 2:
[11:26] Yeah, that's right. That's right. But we can actually understand incidentalomas if we've seen them before. And so this notion of using imaging over time and interpreting it in context became a kind of a revelation for me.

Speaker 1:
[11:43] Yeah, it's kind of a big leap, right? Which is so outside of our traditional thinking in medicine, which is to minimize diagnostic test, to rely on history, and to kind of wait until people are symptomatic or, as you said, in advanced stages of disease before we actually do something, which is kind of too late. It's kind of getting too late to the party, and then you often can't really help people or they have to go through a lot more ordeal in terms of treatment and expense and pain and suffering. What you're suggesting is that there's a way to use these technologies in a different way.

Speaker 2:
[12:20] That's right. That's absolutely right.

Speaker 1:
[12:22] To use them in a way that measures things over time in a longitudinal way and allows you to see the changes over time, and then the imaging becomes faster, smarter, better because it keeps tracking your biology over time, just like a lab test.

Speaker 2:
[12:37] Yes.

Speaker 1:
[12:37] We do this with lab tests. We check your blood sugar and maybe it's rising. We'll then intervene early, hopefully, or you'll see your PSA, which is a prostate cancer test that made you slightly creeping up. We watch it and we can see over time how the change happens. We do this, but in imaging, that's not something that really is done, and you're suggesting that's something we should do.

Speaker 2:
[13:00] Absolutely. In fact, it's interesting you say that it's not done, because traditionally it hasn't been, but quietly, there's been this paradigm of imaging surveillance, say for tumors, that has been building up in medical circles and hasn't necessarily been getting a lot of press. But if I go in and I have a moderate risk for prostate cancer, I may get an MRI every year and be followed, that MRI will be interpreted in context. And when there's a sudden change in the findings, then my doctor might say, oh, you know what, we better go to biopsy, we better check it out. So people have actually realized this paradigm, but because imaging, people tend to think of as a snapshot. Somehow that perspective hasn't pervaded and people still say, well, you don't want to do it in people of low risk. You know, only do it in people with well-established high risk. And my argument is, but most people out there in the world don't have a known risk. Shouldn't we be casting a protective net around them too? If we can figure out how to make sure we're not raising a false flag all the time.

Speaker 1:
[14:12] It's interesting. I kind of want to talk about this whole idea of false positives, which is something that people will push back on. And I want you to kind of explain this because you've written a lot about it and you've talked a lot about it. And I think we're touching on it now. I think it's important because my personal belief is that with the radically deflationary costs, with the potentially ubiquitous nature of these imaging technologies, which we'll talk about soon, with our ability to collect large personal health data sets from your lab testing to your medical history to gathering your EMR, to wearables, to all the omics, your genome, your proteome, your microbiome, your metabolome, to gathering imaging data, be able to aggregate that in a platform, a technology platform that allows you to track your biology over time and moving your biology online is a revolution that we've never seen in medicine before. You and I, as doctors, we see patients, we get, and I had a patient like this yesterday, who's got a chart from here and a chart from there and a lab from here and a lab from there and an imaging test from there and an imaging test from there and a scope from here and a scope from there. And I'm literally having to aggregate all this. I'm having to gather all this data. It takes hours of my time or my team's time to get it ready for me. It's not very user-friendly for the doctor or for the consumer or patient. And what we're seeing now is with Function Health, which is I think why you kind of left your big job at NYU. You had a big fancy job there. And joined Function Health as a Chief Science Officer because you see the future. And you go to book all the future of seeing and you see the future in a different way, which is where medicine is going, which is a proactive, longitudinal, large personal health data set tracked over time. They can understand that biology just doesn't change overnight. It's a continuum of dysfunction. It's slow and progressive over many, many decades sometimes that we now can see. For example, we can tell on imaging, and maybe you can talk about this, changes that can predict Alzheimer's decades, decades before you forget your keys or you have a symptom. Should we be doing that? And people are high risk. There's ways of actually seeing changes that are really important on all these datasets, whether it's your blood sugar, your blood pressure, or your cholesterol, which we're familiar with, or whether it's other things like if you have a low vitamin D, maybe you're not symptomatic. And by tracking stuff over time, we can start to really understand the human body in a way we've never done before. I'm just setting this conversation up because I want to dive into this false positive conversation. I had a conversation with a friend of mine the other night. She's like, I don't want to know. I've got Alzheimer's in my family. I don't want to know if I have the gene for Alzheimer's. I'm like, explain to her, look, you might have a risk gene. So ApoE4, which is a risk gene for Alzheimer's, is common. And if you have this gene or two copies of this gene from both your parents, you're in a much higher risk of getting Alzheimer's. It doesn't mean you're going to get it. It means you're at higher risk. And then you go, okay, I know, I can be product about every other single thing that we know may influence the risk of getting Alzheimer's from my diet to my exercise routine to my sleep management practices to my stress regulation to the right nutrient levels that I need to make sure I maintain the right hormone levels I need to maintain if I'm a woman or a man. Like so much you can do. But she was like terrified to know. I'm like, no, no, this is not a predestiny. This is a predisposition. In that way, I think we can remove some of the fear by realizing this with our scientific knowledge now, there's such a moment for empowerment around knowing your own data. Given that background, and then I want to dive into the medical intelligence framework because I think the longitudinal scanning is the answer to the positive and maybe there's more. But it's also the answer to understanding your health in a better way and it's understanding how to apply the advances in AI and medicine and science to you personally through what we call medical intelligence and our Medical Intelligence Lab at Function Health. So to take us through a skeptic's view, I'm like Dr. Harvard here and I'm like, oh, this is expensive, it's too much to do, you're going to get all these red herrings, you're going to chase down all these things, you're going to cause unnecessary suffering and worry and anxiety. Why should everybody get an MRI every year, like a full body? We do it, you do it, we do it for ourselves. We want it for our families, I just ordered it on one of my staff members tonight because I think he needs it. But why is this so important and how do we get out of this just sort of fear mode or these worry mode about too much information?

Speaker 2:
[19:00] Even that framing is interesting, isn't it? Too much information, right? I mean, there's sort of this sense that, oh my goodness, we'll see too much, we won't know what to do with it, so let's just close our eyes. And there was a time when that was appropriate, right? I mean, people often say in medicine, if a test isn't going to influence your treatment, your plan, your decision making, then don't do the test. And that's actually entirely legitimate. But as you were gesturing towards, we live in a very different time than even just a few years ago. Now we live in a time of big data and AI, when we can collate a large collection of data and we can use AI to connect it over time, to look for subtle changes, for subtle patterns, at a scope that's hard for a single human mind to do. And so I think, you know, my recommendation isn't just go out and get a traditional MRI and have people read it in the same way they always did, looking only at today. My recommendation is establish a baseline for yourself. And I think it's up to us in medical intelligence and up to the broader community to figure out how we deal with this multifaceted data. And I'll give you just a couple of examples coming from work in my NYU lab before I made the jump to function. So we took an AI model and trained it to predict your risk of clinically significant prostate cancer in five years' time based on today's images. Did an okay job about as well as humans, huge false positive rate, like 64% false positives. So not very good at predicting.

Speaker 1:
[20:47] On the MRIs for prostate cancer.

Speaker 2:
[20:48] On the MRIs for prostate cancer. So not a very good prediction five years out. But then we did something interesting. We took that same model and we fed it last time's images. And a year before and a year before, and we also fed it some blood tests and some clinical data. And lo and behold, the more prior information and the more diverse the information we gave the model, the more the false positive rate dropped until it was below 10%. So an order of magnitude reduction in false positive rate just by incorporating context. The second thing we did.

Speaker 1:
[21:23] So in that sense, more information helps you make better decisions.

Speaker 2:
[21:27] Exactly. So instead of, gee, we don't know what to do with it, let's close our eyes. The idea is, let's incorporate everything we know. Now we need to build the models to do it. But the example I just showed you shows that it is in fact possible. Even in a pretty simple prediction model to incorporate context. And you know, I mean, as a master of functional medicine, right? Context is everything.

Speaker 1:
[21:49] It's everything, yeah.

Speaker 2:
[21:50] You can't just look at one organ system in isolation. You also can't just look at one time point in isolation.

Speaker 1:
[21:56] You're looking at the patterns in the data over time.

Speaker 2:
[21:58] Exactly.

Speaker 1:
[21:59] Patterns in the data over time illuminate the real issues and whether there's something to do or not to do.

Speaker 2:
[22:04] 100%. As a, you know, academic or former academic, I need to say, you know, this paradigm is still evolving. So it's not like every MRI you get is going to be put in context in this way. But in the future, that's exactly what we're aiming at. We want your MRI to be, you know, hand in hand with your blood tests and your genetics and your proteomics and all of this, because that rich context is going to eliminate many of those false positives and give you the guide you need.

Speaker 1:
[22:37] And that's what we're really building a function, is a place where you can get access to your own biology. Before you had to go through this firewall of doctors and insurance companies and, you know, maybe they would order it, maybe they wouldn't order it. You wouldn't be able to really know what's going on with your own biology. You have a dashboard for your car, why wouldn't you have a dashboard for your body? And we're talking about establishing, you know, like a thousand point sensors, you can go to your, take your fancy electronic car in and hook it up to these machines and they just run through all these tests. And I'm like, this is amazing. We don't have that for our body, you know? And we don't have the dashboard that tells us how to navigate what's going on in our life. And so we're often at the effect of things rather than being at the cause of our life. You know, in proactive way, empowering ourselves with the knowledge information to prevent disease and to find things early and to actually reverse things before they become problematic. One of the things that's also happened is the ability, I think, to really improve the speed and the access and the cost. So can you talk about that? Because, and I remember going to get my knee, I had a knee issue because I jumped off a golf cart and I kind of tore my meniscus and I was like, oh, my knee is sore, I'm going to go get an MRI. And so I went to get an MRI and it was like 2,500 bucks for my knee. And now we're talking about 499 or 999 for a whole body MRI. So how is that taking us down the road to making this more accessible and affordable? And also, you know, how do we, how do we think about using that?

Speaker 2:
[24:08] So here's the really interesting thing. In the future, and I think it's actually pretty near future. The more we image you, the faster we can scan you next time.

Speaker 1:
[24:20] Is this actually the same machine?

Speaker 2:
[24:22] No.

Speaker 1:
[24:22] It doesn't.

Speaker 2:
[24:23] The faster and the cheaper, we can scan you next time. And I'll give you one other example that came out of work from, from my lab. Basically, what we found is, if we've only, if this is the first time we're seeing you, we need a requisite amount of data. We need the scanner to gather a certain number of views of the body to create those slices we need. But if we've seen you before, this time we trained another neural network whose job is to take those different views and assemble them into a set of images. And we tried taking a drastically reduced set of views, 20 times less data, 30 times less data than you would need for a traditional image. In other words, 20 or 30 times faster. And we found that the neural network, if it had your prior scans, could generate a perfect high quality image, 20 to 30 times faster with 20 to 30 times less data. Why? Because we already knew the rudiments about you and your anatomy. All we needed to look for was change. So once we have that baseline, not only can we predict your health better, but we can also scan you faster. And it turns out another thing we tried was, what if we use worse data? What if we use data from a low power MRI machine? Or maybe from an MRI machine we might build into a seat that would otherwise give pretty lousy looking images. We did that simulation and we found that actually we can get away with much worse data.

Speaker 1:
[25:56] Interesting.

Speaker 2:
[25:57] If we have that prior information about you. So once again, context is everything. If we have the context, maybe we don't need these big multi-million dollar tubes.

Speaker 1:
[26:07] Yeah.

Speaker 2:
[26:08] Once we've seen you at least once, maybe we can put something in a chair, in a bed, in a CVS, in your home at drastically reduced cost. So more imaging paradoxically allows cheaper imaging.

Speaker 1:
[26:22] But just you have to use the same machines. Like you had Siemens one machine or GE another machine. How do they gather that data from the past?

Speaker 2:
[26:30] So there's a logistical challenge of how do you bring your past images from another machine into today's machine so that it can do this. But that's just logistics. I mean, nowadays we have digital image transport systems and so on. But what we found is it doesn't need to be, the image doesn't need to be exactly the same last time as this time. In fact, we used different contrasts last time and it still informs your imaging this time. So this is actually part of something I think you may have referred to it before that I call the Everywhere Scanner Vision.

Speaker 1:
[27:06] Yeah.

Speaker 2:
[27:06] If we have enough information about you, if we've done the advanced imaging, the advanced blood testing upfront, then for the interval scanning, maybe we can use cheap scanners, maybe we can even use constellations of wearable devices on your clothes, because all they need to do is measure change.

Speaker 1:
[27:24] That's amazing.

Speaker 2:
[27:25] Which means we can move health care, not only more proactive, but also make it more continuous.

Speaker 1:
[27:30] Crazy. It's so like, what was the guy's name? It was Bones from Star Trek.

Speaker 2:
[27:34] Yes. The guy with the tricorder, the scanner. I would just tell you the tricorder has been like a holy grail for imagers forever, right? Because it's this tiny little hand-held device, you wave it in front of somebody and you get everything you need. I actually think, and I talk about this in the book, I think the tricorder is a bit of a trick. I think it's actually not its own device doing imaging. I think it has access to all of the records that Starfleet Academy had on you. And all it's doing is looking for change.

Speaker 1:
[28:10] Yeah, that's amazing. That's amazing. So people understand that they can get blood work and know a lot about their bodies and a lot of people have joined function as members and are learning so much. And we're seeing so much in the population that people are discovering that saves their lives from cancer, or figures out they have autoimmune disease, or figures out they have other problems that are really fixable. How does imaging differ from blood work? And what are we looking for? People are looking for my cholesterol, my blood sugar, my hormones, or my vitamin D level, or whatever, my blood count and my immune system. But what are we actually looking for? And how is it different from blood work? And then last question, intro. How do you think the two together are better than either alone?

Speaker 2:
[28:55] So I think blood tests give you biological and chemical context, right? It's the various biomarkers that your body is producing that tell us about the biological functioning in systems. Imaging is spatial context, right? I mean, if we were just undifferentiated bags of chemistry, then blood tests would be enough. We wouldn't need to know anything more. But we all know that bodies are sort of these complex bio-weavings, and it matters what's where, when. And so imaging, as I see it, is what puts all of this chemistry in context, in spatial context, which leads naturally to the question of synergy. Like, you want both, right? You want to know what's where, and you also want to know what's the biological functioning in each position. And when you've got both, you sort of have this magic mixture. So what do we...

Speaker 1:
[29:50] Structure function.

Speaker 2:
[29:51] Structure function, exactly. And so what do you look for in imaging? Well, you look for tissue that's out of place, right? A tumor that might be growing where it shouldn't. You look for derangements of the brain that tell you hints of Alzheimer's disease, things like that. Things that... There may not be a circulating counterpart. There may not be something that was spit off and sent into the bloodstream, and so you can measure it in a blood test. But you can see it in situ. You can see it where it is.

Speaker 1:
[30:22] You have an aneurysm. There's no blood test for that.

Speaker 2:
[30:25] Exactly right. I think this combination of biological, biochemical, and spatial context is really cooking with gas.

Speaker 1:
[30:35] We are both part of Function Health and we do imaging as part of the offerings we have. I think it should be part of the, ultimately just part of the thing that everybody does, which is not just the blood or put all the imaging. What are we finding? Tell us some stories about what we're finding. Because you've been working with Ezra, which was a company that became part of Function for a long time, and you've seen a lot of stories. And we're seeing crazy stories of people, what are people discovering and what are they finding?

Speaker 2:
[31:02] Absolutely. And I'll preface it by saying, I know there are going to be some physicians out there who say, any story I come out with, it's just an anecdote. It's not randomized controlled trials and so on. I'll get back to that later because I think there's an answer for that too.

Speaker 1:
[31:16] Anecdata.

Speaker 2:
[31:17] Anecdata. But no, I mean, the obvious things clearly, we have found tumors that people didn't know they had.

Speaker 1:
[31:23] And early before it kills them.

Speaker 2:
[31:25] Exactly. And that's the key. I mean, you know, our friend and colleague, Emy Gahl likes to say, we already have a cure for cancer. It's early detection. Yeah. Because most cancers, if you catch them early enough before they've become invasive, they're that much easier to get rid of with radiation, with chemotherapy, with surgery. So we have found certainly prostate cancers, brain cancers, kidney cancers at such an early stage that they weren't giving anybody symptoms. That was the whole point. But what that meant is these people could then go in for therapy right away, long before these things would have been discovered.

Speaker 1:
[32:04] Yeah.

Speaker 2:
[32:05] And it is saving their lives. There are any number of other kind of body areas where you can pick these things up. Ezra had a particular focus on cancer, which is sort of obvious because early means life.

Speaker 1:
[32:17] But we can see the change in the brain function structure.

Speaker 2:
[32:20] Absolutely.

Speaker 1:
[32:20] We can look at brain size changes, we can look at the structural pieces of the brain that change over time, it could be linked to different diseases like dementia.

Speaker 2:
[32:27] We can also see interesting things in terms of body composition, fatty liver, cardiovascular health, right coronary artery calcium scans have been shown actually with very good data to be predictive of cardiovascular risk. That we can see in a very straightforward way, combine that with some of the cardiovascular biomarkers, and again, you're cooking with gas.

Speaker 1:
[32:51] Yeah, I think that's it. I think the combination is important. I remember, if you remember, that textbook we had in second year medical school called Robinson Coatran. Oh, yes. The patho is called the pathophysiologic base of disease. And I went back, I still have my copy from 1984. Me too. We have a clonospina, we have a mammogram, we have the thing PSA we screen for, but most cancers we don't even screen for. And when you combine that and those two, and I think emerging proteomic data, which is coming, when proteomics are basically proteins that the body makes, and these cancers spit off these proteins, that we use sometimes already to detect cancer or follow a progression like alveolar protein or CEA, CA125 for ovarian cancer. These are things that we've been using in medicine a long time, but they're used kind of late, they're used to manage the disease, they're used to track progress. But if you combine these and using AI, this is the amazing thing about the data. This is coming soon. These large databases of cancer survivors, and they have biobanks where they collected their blood, they've been able to go back and say, okay, well, let's look at all the patients with lung cancer, all the patients with pancreatic cancer, all the patients with colon cancer, all the patients with prostate cancer, all the patients with breast cancer. What do they have in common within each cancer? And then they can go back and check this blood and screen five years ago and see these proteins that get expressed, and they're able to through AI make sense of all it, because if you've got millions of data points, the average doctor can't, well, no doctor, and a brilliant doctor can't, like you, can't sort through all that. And so using, like you said, big data and AI with our understanding of biology, we're entering a new era of medicine. This is the era of medical intelligence. That's what I'm talking about when I say that.

Speaker 2:
[34:33] I agree entirely. And I've obviously voted with my feet to that effect. Can I get back for just a second to the kind of clinical trials question? Because that is one of the things that gets thrown out a lot as a concern. Like okay, all of this is wonderful in principle. It makes sense, but where's the data? And should we be proceeding until we have the data? And I actually want to go back to another time in history, the 1970s, when all of these tomographic imaging techniques, MRI, CT, PET, ultrasound, were being developed. Back then, the value of knowing what was where in the body was obvious. So there were a thousand CT scans in hospitals between 1971 when CT developed and 1979 when the inventors got the Nobel Prize for it. There were no large scale clinical trials showing the efficacy of seeing versus not seeing. Now, I'm not suggesting we should throw caution to the winds. We should absolutely be gathering data as we go. In fact, big data allows us in some ways to do almost real time trials as we go. But to say, listen, I'm not going to do anything until the data is there. I think that's one extreme of a spectrum that we should be thinking about. I think there is this kind of protective instinct, which as somebody in medicine, I believe in. But I don't want to be protecting patients, protecting people from this new era that's coming. I want to figure out how we make it happen as quickly as possible and measure as we go.

Speaker 1:
[36:22] Yeah, I mean, how do we not get ahead of ourselves? But in a perfect world, we would bring the cost way down. We had a lot of people access large data sets themselves, we'd be able to track it over time. We'll be able to see where they're headed and what to do about it. And that's really what Function Health was designed to do. That's why we created the company was to empower people to be the sea of their own health, to be empowered to own their own data, to be able to have a data-driven health care and medical system, and to use big data and AI analytics to understand all this massive amounts of information. How many gigabytes or terabytes is like a full dense MRI bot? It's like a lot, right? You can't even put it on your computer.

Speaker 2:
[37:01] That's right. It's a bunch of gigabytes per person per session.

Speaker 1:
[37:05] What I want to have people understand is like, who should and when should somebody think about starting to get their first baseline MRI? Is it when you're 20 or 50 or 100?

Speaker 2:
[37:17] Right. Well, again, it's hard to point to a data-driven age because it varies for the particular thing you're looking for and all of that. I guess I would reframe it and say, I think everybody should have a baseline, a baseline scan because, and okay, maybe not well, the body is still developing when you're five or 12 or something, although there's some argument there too. But the whole point is we want to be able to measure change in your body. We want to be able to know what's normal. And so I think at the very least that reference scan, there's no reason for that not to be done early.

Speaker 1:
[38:03] Like in your 20s?

Speaker 2:
[38:04] In your 20s. As long as the people who are interpreting it aren't jumping the gun and freaking out at everything they see. So the problem with that first scan is we don't yet have the context. And so there's a tendency then to follow every lead. If, and this is another sort of paradoxical thing, if we know that imaging is going to be regular, then we don't have to freak out at every finding. So in other words, we get a baseline and we say, okay, we're going to see you again in a year or two years to make sure we've established not just one point, but a trajectory. Even just that second scan is already going to rule out most problems. So I think that we're heading to an era when people should have a baseline and a sense of trajectory relatively early so that we can establish this basis for change.

Speaker 1:
[39:02] So then how often should someone do a scan? Yearly? For example, I'm 66. I want to do one every year. Does that make sense? But what if I'm 35, do I want to do one every year?

Speaker 2:
[39:15] Again, the scientist in me is pausing because I don't have studies to point to. But from basic logic, my feeling is yes, more frequently, as you get older and changes are more likely, a little less frequently when you're very young and changes aren't that likely. If we can, let's put it this way, if we can get the cost of something like an MRI scan down enough, and if we can make sure that we're not over-calling things, then there's no reason not to have an absolutely regular scan, let's say every two years when you're younger, every one year when you're a little older. I want, and you know with the everywhere scanner vision, I want imaging to be an ongoing intimate part of our lives. Not this thing that we do just when we're worried, are we sick? I want it to be the thing that tells you you're still okay, not the thing that only tells you that something's wrong. I want it to be a safety net, not an end stage tool.

Speaker 1:
[40:21] There'll be like tools and devices and things that we can have to put our biology online in real time and see what's going on. And this kind of combined with this idea of an everywhere scanner is really very futuristic. So talk about this idea of the everywhere scanner and how it's going to change how we think about our health and medicine.

Speaker 2:
[40:40] I really think of it almost like building ourselves a new augmented artificial sensory system, right? We have multiple senses. In fact, you know what? It's not sci-fi at all. We already have continuous sensing. Yeah. We've got our entire nervous system.

Speaker 1:
[40:59] Yeah.

Speaker 2:
[40:59] We can sense temperature and pressure and pain and all of these things. These sensors are woven throughout our body. The only problem is they're really not great at giving us early warning of internal things that are going wrong. They're really good at telling us, don't touch that hot stove now.

Speaker 1:
[41:21] Yeah.

Speaker 2:
[41:22] But they're not giving us advanced warning of cancer. That's not what they evolved to do. I think what we're talking about with Everywhere Scanner and with abundant sensors is basically building that artificial nervous system that's giving us early warning of all kinds of other biological things that we just didn't happen to develop nerves for. I think the body, quite frankly, it's a remarkable piece of engineering. I think we should pay attention to what it's built. Certainly, in imaging, almost every innovation in vision that has evolved has been copied and improved upon with an artificial imaging device somewhere. Every single thing that the eye does, we can learn from and that the brain does in processing vision. Likewise, I think when we think about this network of continuous sensing, we should look at what the body's built and build on that.

Speaker 1:
[42:22] It's kind of cool. I rented a car recently and the thing just senses everything. It's like I drive under a bridge and the Google Maps turns a different shade, or I'm driving down the road and there's no car in front of me, it turns the brights on. When a car is coming, it turns the brights off. Or when every little thing, I literally took my hands off the steering wheel and said, hey, put your hand on something. It looked in my eyes when I looked away for something. I was like, oh, make sure your eyes are on the road. I'm like, wow, this car is spying on me. But it's sensing everything all the time, all around it. Kind of like a Waymo or a Tesla, which the self-driving is the same thing. And so we're talking about is augmenting the sensing of our own biology through various kinds of tools, whether they're intermittent or continuous tools, that allow us to put our biology in a different context and to understand it over time and to not have this episodic, often too late to the game diagnostics, which unfortunately with medicine, when you find things too late, it's often hard to fix. And so I think particularly around cancer, my father died of cancer, my sister died of cancer, she had cancer twice. I don't want to die of cancer. I want to live a long healthy life. And I feel like it's one of those things that we now actually, potentially with the gallery test and liquid biopsies, the regular imaging and even the proteomics that are coming, we literally could make cancer and dying of cancer a historical thing.

Speaker 2:
[43:51] I believe so and I hope so.

Speaker 1:
[43:52] That's really the part of the mission of Function Health is to do that and to relieve so much suffering because there's so much suffering with cancer. And I just, you see it all the time and I know it in my family and I've seen my own relatives just wither away and die. And it's just, it's such a heartbreak. And it's in some ways, if we had this proactive, preventive approach to medicine, we wouldn't be in this situation.

Speaker 2:
[44:15] Absolutely. And my family has had that type of cancer history as well. And I just wish that we had had these tools earlier.

Speaker 1:
[44:24] It's giving us insight into human biology in a way that we've never had before. And we're able to then on top of that, apply the 39 million scientific papers that have been published on PubMed to filter and understand all that information. They're taking all the case studies we can apply, all the training that we've done based on root cause medicine into the system. And so when you put your data into function, you're actually putting your biology online. And combined with these large language models and the advances that those are seeing every day, we're entering an era where we're really truly being able to understand the body in a way we never had before. And look at the patterns in the data and create an early assessment and continuous monitoring over time, rather than this episodic random checking, to really know what's going on in your body. And then to be able to understand the subtle changes, the differences, to look at the patterns in data, to learn and to advance science, to help individuals with their own issues. It's really quite amazing. So I would love for you to unpack your vision of what we're doing with medical intelligence lab, where we're headed, and what we want to build in the world. Because I think this is really foundation revolutionary to medicine and science itself. I think it's going to change everything we know about human health and biology.

Speaker 2:
[45:48] I think of it a little bit like a GPS for health. And if you unpack what a GPS does, it actually has a lot of the features that you talked about. First of all, you need a map, right? That's all of the accumulated medical knowledge that you're talking about. You need to know what the landscape is like that you're navigating through, otherwise you're going blind. But more than that, you also need to know your personal history. That's your biology that you've put online, right? Because if you don't know where you've come from, you know, you don't know what road you're on, you kind of need to know that individualized information, not just the collective information. And then the key thing, which I think we're really working hard on in the Medical Intelligence Lab, is how to create that guideline that gets you where you want to go, that travels with you and makes sure you get to 100 healthy years. And that involves then taking all of these patterns that we've learned, the population-wide patterns and the individual patterns based on your data, and projecting them forward and making predictions. Hey, listen, if you just keep on steering this way, you're headed for trouble. No, maybe you need to do a little course correction, change your diet, change your exercise, go in for another test at this interval, that type of thing. And I think it is definitely a remarkable time when we can think about creating that sort of comprehensive GPS. You mentioned that big tech is already gunning for this space. It's happening regardless of whether we in medicine are comfortable with it or not. I sort of see it as both our responsibility and our privilege to try to bring the science of medicine to that endeavor, rather than just feeding lots and lots of data to chatbots, really trying to bring the collection of medical knowledge and the knowledge about integration of body systems.

Speaker 1:
[48:00] And the context as you're talking.

Speaker 2:
[48:01] And the context and your individual context.

Speaker 1:
[48:05] Yeah, that's what I mean. Yeah.

Speaker 2:
[48:06] Exactly. To this problem so that we're not just generating a nice sounding set of answers to questions, we're actually providing you with a guide. We're giving you that map to your health.

Speaker 1:
[48:18] And it's those little course corrections that make the difference, right? If you see, you know, if you track your blood sugar and you go, well, you know, it was 70s, fasting, well, then it's 80s next year. Or maybe the next year is to go like to 85, and then maybe the next year is 89. And then, oh, I'm getting worse metabolic disease, and I'm heading towards pre-diabetes and type 2 diabetes, even if I don't have the official diagnosis yet. And I can course correct.

Speaker 2:
[48:42] And in fact, going back to this model of biological senses, you had talked about the concern that some people legitimately have. Well, I don't want to be anxious all the time. I don't want to be thinking about the diseases I might develop. If you think about our senses, they evolved to protect us from harm in a similar way. We don't think about them all the time. We just get this burst of alarm if we step into a street and we see there's a car coming. But most of the time, the senses are just operating in the background, keeping us alive. That's how I see this online biology and this network of sensors in the future. It's not constant alarm. It's just waking up and giving you a ping if you're about to step into the street with a car coming, medically. That type of safety net is something that even the most squeamish people might be comfortable with. It's just for the moment, don't do this because it's going to harm you, but otherwise live your life and live your life well.

Speaker 1:
[49:42] That's right. I think we know so much now about how to prevent disease and I think people are worried about finding out something that they can't do anything about. I understand that. But most of the time, your biology is changeable and there are early detection signs that are, as I mentioned, these biochemical changes, these then turn into pathological, early change we can see on scanning, that really give us a roadmap to what's happening with our health. Putting our head in the sand and not paying attention and not looking at our own personal data, it doesn't make any sense. You go, well, I'm not a doctor. How do I make sense of it all? Yeah, you're right. If you don't know how to make sense of it all, then it's a lot. But if you have the facilitation of a company like Function Health that provides you with the guidance, provides you with the intelligence behind it to make sense of it, to create a ranked order priority list of what you have to address to help you understand what the steps are. You can take yourself to do self-care and when you need to seek medical care, and provide that whole continuum of care for you, rather than just waiting around until something's happening. That's the thing most people don't realize is that disease doesn't just happen. It's occurring because of low-grade changes over many decades. The thing I want to end with here is our bodies are this highly intelligent system that want to be healthy. Your body is not designed to be sick. It's not a design flaw. We are providing the conditions in our current modern society for the body to be sick with the crap food that we're having available. Seventy-three percent of the food on grocery store shelves is not even technically food. It's ultra-processed Franken foods. We have enormous amount of environmental exposures and toxins that sometimes we can do things about and actually help our bodies detoxify. We have dysregulated circadian rhythms and sleep. We have excess chronic stress. We have all these things, sedentary lifestyles. These are things that we are empowered to do something about. We have nutrient deficiencies, which you can do something about. And when you actually can know what's happening early, then you can make changes that really change that course and allow your body to provide the conditions that are going to create health rather than simply wait until you have to really treat some serious disease. And this is a fundamental paradigm shift is the idea that disease isn't just some random phenomena. It's something you can predict from early indicators and then do something about. I just saw a patient yesterday with Parkinson's disease. He had warning signs from way early. He had tremendous amounts of environmental exposures from hobbies and being in the Navy as a chemical engineer and in childhood. And I'm like, this guy would be a sitting duck for some kind of toxin-related illness. Parkinson's is a well-known toxin-related condition. And yet he had to wait until he got Parkinson's for someone like me to look at his history and go, well, Jeek, we gotta get all this crap out of your system and we've got to detoxify you. And that was something that he didn't have to necessarily do if he'd been proactive and actually was able to measure the toxic load of his body early on. Same thing happened to me. I had heavy metal poisoning from China, but I wasn't sick right away. It was like this kind of slowly building up burden of toxins that then knocked me off my feet. But if I had known earlier, I could have done something about it and not ended up in this catastrophic illness. So I think we can actually see these changes over time. We can do something about them if we have the right information. And he just didn't have the right information. So that's really why I think medical intelligence is such an important concept. And in our medical intelligence lab at Function Health the science we're putting behind it and the effort we're putting behind really providing the best quality understanding information of your biology is going to change medicine and healthcare.

Speaker 2:
[53:37] Hear, hear. No, and listen, Mark, I mean, we started the conversation with the future of seeing, right? In some ways, I think in a nutshell, the future of seeing involves actually looking. Now that we have the capability, now that we have the capability to see lots of your biology, now that we have the capability to use AI and other similar tools to integrate that, to connect it to knowledge that has been accumulated over all of these centuries, now is the time when we need to start living with our eyes open and living with that kind of guidance.

Speaker 1:
[54:13] So the future of seeing, your book, which everybody should get a copy, where can they find it?

Speaker 2:
[54:17] They can find it at Columbia University Press or on Amazon, of course. I love it.

Speaker 1:
[54:21] Columbia University Press.

Speaker 2:
[54:24] I love it.

Speaker 1:
[54:25] Of course. It's a great title, The Future of Seeing, because it's not literally just about imaging. Your book is about imaging, but it also implies that the future of seeing is about the future of seeing deep into human biology in a way we've never been able to do historically, and it will transform medicine, healthcare from the outside in. Because traditional healthcare is not changing anytime fast. The edifice is too solid and the resistance is too much, and the old ideas die very hard. I think there's a book I read in college called The Structure of Scientific Revolutions by Thomas Kuhn.

Speaker 2:
[54:58] Yes.

Speaker 1:
[54:59] In this book, he talked about this idea of a paradigm shift, and that's where the word paradigm shift came from. In the book, he talks about this idea of normal science. What we believe is just so embedded that we can't unsee it. In other words, if you were living in 1400, the earth was flat. If you were living in the pre-Gaulain era, the earth was the center of the universe. This is something now that we have to understand because we are living in a totally different era where we can actually see things that we never could see before. We can look where we never looked before. I mean, look, I remember, I mean, let's see, it was 13 years after I graduated from medical school that we decoded the human genome. So, I mean, this is in a very short time, and that was a billion dollars. Now, it's $200 to decode your own personal genome. That's where we're going. We're going to this massive personal data-driven healthcare system, and I think in a way, we're disrupting healthcare because we're going to empower people to be, in a way, their own healthcare agent. And then, yes, use medicine and use hospitals and use surgery and use doctors when you need them, but most of the things that we pick up early, they're fundamentally things that are under our control. It's what we eat, it's how we move, it's how we sleep, it's how we manage stress, it's our relationships, it's our toxin exposures which can mitigate to some degree. Those are all the things that are driving the disturbances in our health and those are things that we can pick up in these early warning signs like your car. Okay, your tire pressure is a little low or your engine light is a little thing or whatever. I don't know, these sensors are amazing on these cars. And wouldn't it be great to have that dashboard for your body and that's really what we're doing with Function Health and it's just gonna get better, it's smarter. So I encourage everybody to, not just because I co-founded the company, but I encourage everybody to think about how do you put your biology online so you can be proactive about your health and not get that horrible sinking feeling in your stomach when you're in the doctor's office and they say, you've got metastatic cancer. Chris Vanderbeek, I think his name was his actor who recently died of cancer. And he didn't need to, he really didn't need to. My sister didn't need to, my father didn't need to. And I wish this technology was around then. And I think this really what we're talking about here, Dan. So any final thoughts or words for people listening?

Speaker 2:
[57:33] I think you said it beautifully, Mark. I think really my final words are, in this remarkable era, keep your eyes open, get that biology online, figure out how you can essentially have this new safety net that nobody in the history of humanity has had before. Yeah.

Speaker 1:
[57:53] Amazing. Well, thank you, Dan. Thank you for your work. I'm excited to work with you in building the Medical Intelligence Lab and keep function evolving and helping it to actually help millions and millions of people. I think we're just getting started. People should stay tuned. You can learn more about Dan's work through his book, The Future of Seeing. Go to functionhealth.com to learn more. It's only a dollar a day to join as a member. That will give you a deep dive on your biology and you can get a full body MRI scan as a baseline through that website and even more deeper scans if you want for other things. I'm really excited about what we're doing together. I think combining the ability to gather your history data, your EMR, your wearables, imaging, lab data, all putting it together and helping people understand their biology is really revolutionary. I'm super excited about it.

Speaker 2:
[58:41] Thank you so much Mark. It's a pleasure and a privilege to talk with you and to work with you.

Speaker 1:
[58:45] Amazing. Well, thanks Dan. If you love that last video, you're going to love the next one. Check it out here.