Dr. Jonathan Ng, CEO of Iterative Health: “Not using computer vision is a barrier to health equity. But will it replace GIs?”
Dr. Jonathan Ng, CEO of Iterative Health:
“Not using computer vision is a barrier to health equity. But will it replace GIs?”
I asked Jonathan Ng, founder and CEO of Iterative Health (formerly known as Iterative Scopes, an AI in GI startup) why investors recently gave them $150 million in funding. We didn’t exactly get to the answer but I can fathom a thesis.
Iterative Health’s investors include the founders of Flatiron: Nat Turner and Zach Weinberg. A digital health company that advances cancer research. Pharma company Roche acquired Flatiron for $1.9 billion in 2018. Other investors of Iterative Health include Johnson & Johnson and Eli Lilly amongst others. Further key team members from Flatiron have been joining Iterative.
My reading of the play here is straightforward. It’s possible that investors want to repeat their thesis of Flatiron for Iterative but in gastroenterology. Big Pharma is funding AI in healthcare to develop more targeted therapy for the drugs they develop. Therefore the central premise of Iterative’s offerings today seem to be on better recruitment for clinical trials. Physicians get an ancillary. Pharma automates patient recruitment.
I came away super impressed with Jon’s unwavering vision for AI in GI: equitable healthcare. When I asked how Iterative is different from Medtronic or others who were also working AI in GI. He repeated his vision – he said the desire for equitable healthcare differentiated them.
While on a medical mission in Cambodia, it occurred to him that not all doctors are made equal. And that means patients can’t get fair access to healthcare. Computer vision and AI could change that. He took action.
Jon moved to MIT to study MBA and build on this idea. In a very short period, he developed the foundation for Iterative that went onto pave the way for AI in GI.
Here’s Jon changing gastroenterology for good. It’s reasonable to say that the future of GI hinges on the execution of his vision.
◘ The fascinating story behind starting up Iterative Health
◘ “Not using computer vision is a barrier to health equity”
◘ Why did Iterative Health enter the field of gastroenterology?
◘ How does Iterative Health differ from the other AI companies in GI?
◘ Iterative Health raised $150 million. What attracted investors?
◘ What is their next milestone?
◘ “Right now today there’s an unacceptable level of disparity in (health) outcomes”
◘ “Being able to minimize the barrier for GIs is to be able to access information”
◘ Will AI replace GIs?
◘ How sophisticated is computer vision in the endoscopy room right now?
◘ “For healthcare, it just takes way too long to understand our impact on patients”
◘ Future of AI in GI
◘ “Even our colonoscopy screening recommendations are generic for the population”
◘ Iterative Health’s operational model
◘ Partnership with One GI
◘ Jonathan’s views on the controversial NordICC colonoscopy study
The Transcribed Interview:
Praveen Suthrum: Dr. Jonathan NG, CEO, and founder of Iterative Scopes. Thank you so much for joining me on The Scope Forward Show. I’ve been waiting to have this chat for many months now, so I’m glad that we scheduled it. So let me first introduce you to our audience. So you’re the CEO and founder of Iterative Scopes, a pioneer in the application of AI-based precision medicine to gastroenterology with the aim of establishing a new standard of care for the detection and ultimate treatment of GI diseases. The company is a start-up that’s fun out of MIT, and you’ve recently raised $150M in Series B financing to accelerate the development of your core algorithmic innovations. So that’s exciting and it’s really wonderful. So anyway, John, welcome to The Scope Forward Show.
Dr. Jonathan Ng: Yeah, thanks for having me, Praveen. I’m excited to be here and sorry it took us so long to get here.
Praveen Suthrum: No, you’ve been busy raising the money that you did. So, first of all, congratulations to your team. I think it’s very exciting, not just for you as a company and as a start-up, but for the industry itself, because it’s very telling on the interest that GI as a space is attracting and the direction of where investments are flowing. So it’s very telling. So congratulations to all of you. But John, I want to ask you first about the backstory here. So now you got the idea. I learned while you were on a trip in Cambodia and you were looking at doctors there trying to detect tumours and their inability to find it. And then the idea germinated. I think at that time you were doing your MBBS in Singapore. So please tell us all about it.
Dr. Jonathan Ng: Happy to share that one. It’s quite an interesting story. So I think since I got my start in healthcare pretty early on, just by chance, I happened to visit Cambodia for context. Pretty much first generation college, no one else and family had doctors. But I was fortunate to receive mentorship from a friend’s dad who was a doctor. And he said, John, can you come to Cambodia and carry my bag? Go around. Medical mission happened to be on that trip. And that was my first experience of healthcare outside of Singapore, which is very much a first of all healthcare system. And frankly, it was pretty shocking how the state of health care was back then. It was back then emerging from the Khmer Rouge, from the Pol Pot regime and it was feeling at a very fundamental stage of like. After the Pol Pot regime there were basically two doctors left in the country serving patients and for a population of 12 million people because Khmer Rouge decided that the best way to prevent an uprising against them was to kill off all the intellectual people. And so I arrived in the country that observed how kids were not making it through for really basic reasons. And back then, the under five mortality rate was about 20%, 25%. And I was like, this is totally unacceptable. You can’t know this and not do anything about it. Long story short, I end up in a deep rabbit hole, end up spending about 14years in Cambodia, building all kinds of facilities, open its first paediatric cardiothoracic surgery units, unit awards, for instance, reconstructive units, getting involved in a couple of establishments, establishing a couple paediatric hospitals in rural Cambodia. But as part of this process, we were trying to train Cambodia’s first generation of surgeons. And I would frequently bring in my best tutors from Singapore, from the US. From places like the Brigham Boston group of hospital, and bring them to try and impart knowledge. And frequently we would struggle with this thing where and this thing plays on my mind over and over again, where I would be sending in the operating theatre. My mentor would be standing next to me, and you’d be like, look at this huge tumour. You would describe all the details to me and all the ways in which we could treat it. I’d be like, yeah, that’s a massive tumour. Obviously, I’m lacking all these other details because I’m not as experienced, I’m not as good. And then there was a session with trying to train, and he would be like, where’s the tumour? And we were like, oh, then if you’re not even seeing it, if you’re not even identifying it, you can’t treat it, right? And that sets off then series cascade of decisions that essentially ends up with the patient receiving a very different outcome. There are many ways to go wrong, only one way, pretty much one way to go right in, and the patient would essentially receive a very different outcome. So when I reach kind of like my 12 13 hours by then doing a lot of regional work, and it got pretty frustrating, to be frank, in terms of my inability to feel like I was moving the needle. I felt like I was spinning in circles, right, constantly trying to move this needle. And so I said, yeah, I’m going to take some time off. I’m going to go back to grad school. By then, I had my immediate I was go to a place which essentially was very forward thinking and very a hub for innovation. And that updates MIT.
Praveen Suthrum: Sorry. Just to time this a little bit, you were doing your MBBS, and you were in Cambodia at the same time, so it was happening.
Dr. Jonathan Ng: Yeah. So I started working in Cambodia when I was pretty much 15 years old. So this is the way that yes. And then I continued through military service, continued through med school, continued through medical training. So I was working in Cambodia throughout this period.
Praveen Suthrum: Fascinating. Okay. Yeah, please continue. So now you’re in MIT, and then?
Dr. Jonathan Ng: In MIT, I come there with an open mind, trying to explore all the labs. And one of the first few technologies that I see is the use of computer vision and autonomous vehicles. And I look at it, I’m like, this is fascinating. This technology can identify cats, dogs, humans, cars. And even at the point of time, they were like, we can integrate. If a human is walking in a certain way, we can kind of predict whether the person’s going to turn left and right, dodge it, start running all these complex intuition. And I was like, that’s super cool. But also, why don’t we have this in medicine? Why can’t we have this in healthcare? Because to me, that represented a major barrier to patients receiving equitable health care, right? As equitable in outcome as possible. And that’s, in a nutshell, what I’ve been working through towards for the past, I’d say now, 20 years now. And I was like, this feels like a big one. This feels like a big knowledge gap that back then, we were in today. Still, we’re trying to move medical knowledge. Sure. Medical textbooks. We know that doesn’t really work. It doesn’t represent real world, because we did then we wouldn’t need residency. We do residency, which is essentially apprenticeship, which is nearly a one to one teaching model, which is not scalable and doesn’t really fulfil what the world needs in terms of better health care, more scalable health care. And then we also do conferences, which, unfortunately, I’ve been to a number of these, and for a good number of them, I think, with the ability to educate once a year, the feedback loop isn’t short enough. The examples we use aren’t real world enough, and they’re not really geared towards, like, very fundamental, like, questions like, is this a tumour? Is this not a tumour? Right. People expect for it to be covered in medical school, but it’s also that gap exists very much in a real world today. And I kept thinking to myself, damn, this is awesome technology that we can use. Now, I’m not a gastroenterologist by training. In fact, I was training to be a surgeon, but I kind of landed, worked my way backwards into fascinating technology. Where can we use this? And so they started looking at places which I’m more familiar with and plastic surgery and orthopaedics and cardiology. But I did landed in GI really, because I think for a few reasons. Number one, I think GI doctors are just awesome to work with. I think when you deal with poop and stool all day, I think he just can’t take himself too seriously and it’s just straight out. We all know start-up as a grind. We all know that you try to innovate the grind and you really want to go through that grind with people you enjoy working with. There are set of specialties which I will not name on this record, that I will not try to innovate with. Right. Like they will bite your head off before you try to move and change anything. I think beyond that understanding that GI isa very visually based disease, naturally it’s like surgery but with a natural camera already. Whereas most surgeons don’t operate with a camera or don’t have a primary view of a camera. All these factors played into it and I said why don’t we try it in GI? I think from the very get go and still today I have a broad mandate, I have a broad vision, but in terms of using this technology to close in on disparities and health care outcomes and I think that accounts for what you alluded to a lot of our successes. I think many people see this potential as well.
Praveen Suthrum: If I may ask, how does it differ from the AI solutions that are already there in GI? For example, there’s GI genius from Medtronic, there’s Satisfai Health and there are a lot of start-ups waiting at the gate of FDA from what I learned. And there are newer and newer publications all the time. Even Google has got its interests, they’ve written a paper on computer vision and detecting polyps. So how does your solution actually differ and what part of it is similar to what’s out there?
Dr. Jonathan Ng: Yeah, great question. Sorry. I won’t claim to know every single solution out there. I think the biggest differentiator we have is really the vision behind and why we’re doing this. I think that will eventually we keep doubling on why and we’re doing this in terms of reducing health disparities and health care outcomes. We have a very practical role take on the application of the AI and GI and I think eventually that just leads us into more differentiated space, slightly different angle than say someone who is trying to just sell more snares or trying to do it for maybe sell most products. I’m not sure. I think for us the fundamental reason is very critical and it’s what drives every single decision we make here at Iterative.
Praveen Suthrum: Let’s go back to the name, Iterative Scopes itself. How did you come up with it and get a story here?
Dr. Jonathan Ng: Yeah, essentially I was having too many drinks of my professors at MIT, you know, and trust us to come up with overly complicated names, but I think why ‘Iterative’ firstly, I saw it as a very much a learning journey in the practical sense of, I don’t have it all done today. I expect to have an Iterative process. Iterative learning around this. I think the other part is also the whole machine learning aspect of this. It’s like it’s not one algorithm at this point in time, tens, hundreds of algorithms that are all learning on repeat. And the more times you iterate, the better you become. And so that word really stuck in my head now, the ‘scopes’ part. Here’s some news on this. I think by the time this podcast is out, you’ll get the news that’s what the change? I think to me, it was defining a narrow initial path of GI. First, and to be very clear, the flip side did not realize that would do is I think 80% of the market thinks we manufacture hardware and like, okay, I probably made the mistake. Everyone’s asking me, so where’s the scope you manufacture? I’m like, I don’t know, man. We don’t manufacture anything. But we’re about to fix that. And stay tuned for some news. It’ll be pretty fun.
Praveen Suthrum: I’m looking forward to that news, but I can guess what it might be based on what you just said. But somewhere, John, the message seems to have resonated with investors, with digital health investors. So somewhere the community seems to have taken to your vision. Can you talk a little bit about that? You’ve raised a huge sum of money. I’ve not seen $150M plus being raised in digital GI.I think your start-up is leading that game. What attracted investors to your company?
Dr. Jonathan Ng: Honestly, I wish I knew. I think the breadth of the vision in terms of and I think we have such a strong team in place right now. And I keep saying this since early days. It’s easy to mistake. Everyone has to enter the market at some point. But I think the quality of talent and the vision behind the company will eventually, you know, define where each company goes. And I think for clarity first, I think the market, the GI market is big enough to host multiple AI companies. So I always tell my folks that the competition is not really, say, GI genius or Satisfai or any of that. It’s really ourselves. It’s really making sure that we do what’s best for our doctors and making sure that we understand our customers as well as we should and doing the best for our patients. We’re only fighting ourselves with this one. And so I think what’s the quality of people that we have internally, depth of which we have bottomed out our vision in terms of putting together a comprehensive strategy and our initial successes and being able to execute against them, I think that has all attracted quite a decent amount of money.
Praveen Suthrum: Did you plan to raise this amount when you were at the end of Series A or did it just happen? What is the story here? How did that come about?
Dr. Jonathan Ng: I think it’s a bit of both serendipity and I also acknowledge that a lot of it’s been a huge privilege in terms of this whole entire journey myself. I arrived in the US only four and a half years ago and two years to spend it in school. So it’s been incredible the past four years, three years running this company. I think everyone has plans. It was certainly, I think, a mix of serendipity in terms of meeting the right folks who understood the vision, also hitting milestones, and also being able to show that we were bringing in the right folks. And with bit more money, we could basically strap on some rockets and take it to where we wanted to.
Praveen Suthrum: And where would that be? So what is the Holy Grail for the company? What is the next milestone or the next big milestone from here?
Dr. Jonathan Ng: I think the Holy Grail is if we basically are able to use AI to analytics to basically provide every single doctor with the information they need, and they deserve to be able to treat their patients somewhat equally. Right. I think right now, today, there’s an unacceptable level of disparity in outcomes, whether you choose to accept it or not. Whether you choose to acknowledge it or not. Right. I’ve worked in countries where we struggle to even get basic calls into place. We don’t even talk about identification of serial polyps or any of the store or the level of severity of the IBD right. And so being able to push this information into every single endoscopy suite, being able to minimize the barrier and the activation energy for GIs to be able to access this information and be consistent about it across the board, I hope that this will be a great equalizer in terms of outcomes.
Praveen Suthrum: That’s fantastic for patient outcomes, but there may be gastroenterologists out there who might be thinking, hey, if a technology like this is going to do the job that I do currently, what would I do? So how would you respond to that or any thoughts there?
Dr. Jonathan Ng: Personally, I don’t think we’re trying to replace anybody here. I think we’re providing information so that folks can do what they do best, which is treat all of us, sign up to medicine, to healthcare, to treat our patients, to help someone. Right. And why would you turn down being able to knowledge and data that is able to help you to drive better outcomes? We’re not saying we’re taking over, for example, the mechanics of driving a scope. We’re not intercepting the polyp for you. We’re simply making recommendations, making it easier to have a common identification of polyps, common identification of certain procedures, common assessment of certain diseases, so that we have good understanding of how these patients should be treated. That’s the way we see it. We’re not going to take away what GIs do today. I think that’s impossible.
Praveen Suthrum: But let’s go back to that analogy of self driving cars that actually prompted you in this direction. You may be familiar with the five levels of self driving or autonomous vehicles. Somebody’s given this classification. I find it quite interesting. I even wrote about it in Scope Forward. So the thinking on my question to you is that what if that were to be applied to computer vision in the endoscopy room? So if you were to apply from level 1 2 3 4 5, which has increasing levels of sophistication where would you land when you apply computer vision to endoscopy?
Dr. Jonathan Ng: We’re right at the start right now. It’s very rudimentary what we’re doing. And I think for healthcare, it just takes way too long to understand our impact on patients. Like, imagine this early to put cars, these cars on the roads. We’re just barely standardized. And roads are pretty straight. Roads have, like, standardized stop signs. Roads have crosswalk. So people actually go across. Now think about anatomy. There’s no crosswalk. Everyone’s anatomy is different. If you had bowel surgery before this, it’s completely different. I think it’s going to take us tens of years to get there. The totally autonomous situation. Everything told them is an identification of patents. And just being able to support.
Praveen Suthrum: That’s interesting to know. I usually ask this question towards the end, but I think I want to ask it now. So if you were to throw a stone and that way to land, let’s say five years from now, and then three years from now and then next year, one year from now, where would it land in terms of AI, in gastroenterology? Let’s start with five. Let’s start with the distant future. I’m saying five is not even too far out, but how far would we go in terms of AI and GI? Five years from now? And then I just want to follow that up with three years and then one year.
Dr. Jonathan Ng: I think five years is pretty short term for us. There’s a pace of change in healthcare is pretty short. Okay, we might have certain great polyp detection algorithms, certain features being detected, certain historic being predicted, for example, and use in certain clinical decision making. But I suspect that’s where we’re going to be here in five years time. And that’s it. I’m pretty conservative with how fast technology moves. I know there’s some visionaries out there that are like, oh, we’re going to be self driving these scopes through a colon in five years time. I think no.
Praveen Suthrum: Then let’s extend the horizon. Like, five years is short term, and that’s what you’re seeing, what is increased sophistication from that point of time, let’s push forward to ten years.
Dr. Jonathan Ng: I think being able to, for example, ingest multiple data sources that are now becoming more and more available. So, for example, we’ve got all these like Guardant or Freenome and all these liquid biopsies, for example, right? You know, being able to ingest from various tests and being able to risk adjust, being able to take a comprehensive view of a patient, being able to ingest various multi model data sets to then come up with a more balanced view of a patient more comprehensive view of patient will be an incredible outcome if we can get to in ten years time. Right? And I think that would be a great application to be able to have personalized recommendations right now. Even our colonoscopy screening recommendations are generic for the population. Folks who ask me every time, like, what age should I start scoping? And I’m like, as soon as you can afford it, you should scope. Because I have friends who have had stage four cancer at 35 years old, right? And I’m like, I don’t know if it’s not going to be you, but I certainly hope not. And the only way that it’s not you is if you get a colonoscopy. So if you can afford it, go scope. Peace. I’m going to get my first colonoscopy next year at age 35. But instead of generic recommendations to come up with true risk predictors, to come up with a comprehensive view of the patient so that we are able to confidently and individually make these recommendations, because right now, also the way we do healthcare, because the guidance is so broad, seething, it just becomes so expensive. And so there’s a sweet spot where we can understand each patient, we can customize our recommendations for each patient, and then we can in between that apply preventative measures to just keep monitoring in between.
Praveen Suthrum: So currently, what is the operational model? You’re not yet selling in the US, are you?
Dr. Jonathan Ng: So we do have a product in the US. It’s a bit of a patient identification algorithm which basically identifies the right patients for the right clinical trials. It’s currently in working in inflammatory bowel disease. It’s a pretty cool piece of technology. We’ve got quite a few sites deployed right now, and it’s been a tremendous year for us in terms of that site deployment, and we’re really excited to see what it brings for the next 20 years.
Praveen Suthrum: But you’re not using the polyp detection tool in the US.
Dr. Jonathan Ng: That’s not available in the US. So we just got cleared with the agency and it should be commercially available pretty soon.
Praveen Suthrum: Okay. And I’m assuming that would be next year.
Dr. Jonathan Ng: I believe it’s very early today.
Praveen Suthrum: Now, you’ve raised all this money. Other than building up the leadership team, how are you deploying it? What happens over the next, let’s say100 days and twelve month period?
Dr. Jonathan Ng: I mean, 100 days pretty much it’s too short. Again, we don’t think in that short term. Right. Again, the capital we have raised upwards and allows us to do that. So we’re very, very privileged here, I think. We think very much in terms of years where we want to go, I’d say we’re doubling down on GI. We’re not going anywhere. Right. We remain extremely focused and we’re excited to engage more with a broader swath of the community. For example, we’ve got great partners. We just announced a partnership with One GI, the One GI platform, which is, you know, they’re really, really awesome people. And Robbie and Zia, we love working with them, and we want to find more partners like that. Right? We want to find ways in which clinicians, who are even CEOs and administrators who are interested in moving the needle forward, not just for themselves, but for many others in the world, to engage with us and to be able to help us understand what would make a difference to the practice and allow us to create tools against those. I think that’s very much on our roadmap.
Praveen Suthrum: Got it. And these partnerships with GI practices, it’s primarily for recruiting patients for clinical trials. Is that the scope of the partnership or is it beyond that?
Dr. Jonathan Ng: We’ve got a number of other programs as well, R&D programs. We’re exploring a good number of novel endpoint and novel ways of measuring disease severity and such that.
Praveen Suthrum: So, John, the last few days there’s been a flurry of news in GI as a result of the Nordic study that was published by NEJM. And as you know, the study questioned the effectiveness of colonoscopy and linked it to deaths associated with colon cancer. What did you think of it and what did you think of the reaction on both sides of the continent? There were people on the European side who commented. There were GI societies on the US side that commented. So I’m curious to know what your take was when you read all that first-year. is your product currently commercial or when is it going to launch?
Dr. Jonathan Ng: I applaud the office for putting in all this effort, and it’s a ten year study. It’s incredible how much effort has gone into it. I think it’s a much-needed study. Unfortunately, I think the title was pretty much dictate. I think there are many ways in which we could have framed the title more responsibly so that the general population who reads it actually goes beyond the title to understand what’s in the study and what the actual recommendation or findings were. I think certain things spread out, and we’ve known this for a while around, say, for example, certain GI populations having lower detection rates. And that is the key thing we’re trying to address here in terms of, yeah, maybe everyone maybe if you have lower area rates, either we should do some population studies in your population or perhaps they just have a lower propensity to get cholesterol cancer, which I don’t think is true in Norwegian population. Or we should help them with polyp detection algorithms, perhaps, and maybe start deploying some of these tools into their systems. But this is an biological difference that we have known for some time I mean, we’ve seen other studies, for example, done in Italy, where the devices approved from Italy for Italian studies, for example, that have a completely different patient population. You know, we tend to be aware that these differences do present themselves. I think historically we try to have more balanced view of studies. I know studies are how you design them, essentially, right? But essentially for ourselves, we tried to personally stay away from the design of the study and had more clinical advisors, more balanced advisors come in and help us to design it. So, for example, in our most recent publication in Gastroenterology and the foundation in September, it was a five century RCT that was done across, you know, it was MGH, it was midsized academic hospital. Below that, there was a large PE practice based, large PE rollout type practice as well as a community GI type practice. And we saw a fascinating range of outcomes. I’m sure there are reasons for that, but we want to be truthful of ourselves. We want to make sure that representative against population we’re deploying against.
Praveen Suthrum: One of the comments that came out on the US. Side was about the ADRs, and that the ADR of the endoscopy was lower. When I did that, I was thinking, shouldn’t we actually mandate using AI tools for endoscopy so that the ADR across the board reaches a certain baseline? Do you think that should happen or is it likely to happen? Any comments?
Dr. Jonathan Ng: I wish it happened. It’s why we are here. It’s why we’re present. I think it’s a two way street. I think as much as regulators and insurers and payers should regulate it into existence and to broader assistance, I think the other side of this is engaging us and for us to minimize the barrier to adoption. I think a lot of AI right now is current for would be infeasible or overly expensive for most of GI practices to adopt. And I think we as a company have a responsibility to try and in the spirit of our mission, trying to reduce or minimize this barrier to adoption. And that’s what we’ve been working towards having again.
Praveen Suthrum: John, this is great to have this conversation. I think it gave me a great perspective on how you’re thinking and how the company is thinking. Is there anything else that you wanted to share?
Dr. Jonathan Ng: No, we were just really excited to keep innovating. We’re always looking for good partners. So let us know. Hit us up if you’re keen to work with us and we’re keen to understand our products and yeah, pretty much, we’ll keep going.
Praveen Suthrum: Congratulations once again to you and your team, and I look forward to catching up again.
Dr. Jonathan Ng: Yeah. Thank you. Bye.