Episode 4 of First in Human features Matt Schwartz, CEO & Co-Founder of Virgo. First In Human is a biotech podcast that interviews industry leaders and investors to learn about their journey to in-human clinical trials. Presented by Vial, a tech-enabled CRO. Episodes launch every other Tuesday.
Simon Burns: Matt, thank you for joining us today.
Matt Schwartz: Yeah, pleasure to be here. Thanks for having me.
Simon Burns: I’m Simon, co-founder and CEO of Vial. Pleasure to speak today with, Matt CEO Virgo, who’s doing a lot of exciting stuff in the IBD space as it relates to clinical trial Matt, give us a quick sense of your background and how you got to start virgo.
Matt Schwartz: Well, thanks again for having me. My name is Matt Schwartz. I’m the CEO and co-founder of Virgo. My quick background, I’m a biomedical engineer by training, and prior to starting the company was a product manager in the minimal invasive spine surgery and robotic surgery spaces.
Started out launching a number of instrument retractor systems implants for minimal invasive spinal fusion, and then spent about a year and a half at Intuitive Surgical as a product manager on the Da Vinci Robotic surgery system. And was fortunate enough to travel the world, supporting surgeries. Saw some of the highest echelon of technology in the surgical space especially when it comes to endoscopic surgeries.
And was always fascinated to see that it was really difficult. Even with a $2 million robotic surgery system, it was very challenging to get video off of these systems. And around 2015 I became very interested in machine learning, specifically computer vision. Started taking courses online, and it became immediately apparent to me that if the field of endoscopy ever wanted to move things forward with AI, the first thing that was needed was to build better data infrastructure on capturing the video from endoscopic procedures.
And so that was a jumping off point for Virgo. we spoke with endoscopists from a number of different fields across all types of surgery. Pulmonologists, gastroenterologists and ultimately found that in GI in particular, there was a latent need from gastroenterologists who wanted to be recording their own procedure videos at scale so they could use their video data for research projects, training initiatives, quality improvement initiatives, and even just day to day care every now and then referring back to a video from a particularly interesting colonoscopy or ERCP potentially even being able to send videos to outside physicians that they could refer patients to. So if you had to refer a patient for surgery, doctors would want to send the patient with a video of their initial endoscopy. And so it was really a great fit.
Ended up starting the company officially in 2017. Two great co-founders Chief Customer Officer Ian, who, we go back to our undergrad together at Vanderbilt. He spent his career, started out at Epic comes from a healthcare IT background, and then our CTO and Co-Founder David, met him when he was an engineer at Dropcam had gotten acquired by Nest and Google and had a great cloud video infrastructure background. So that was the starting of Virgo.
Simon Burns: One of the things that I’ve always found so fascinating chatting with you is just how different and clinical trials are this massive space and just how different your world of clinical trials are. GI trials their own are different, but you work in IBD and that’s incredibly different and nuanced in some way. Maybe for someone who knows oncology trials or derm trials or something else. What makes IBD trials so unique and so challenging?
Matt Schwartz: one of the main aspects of IBD trials, and there’s actually a lot of trials even beyond IBD in the GI space, is that they’re very procedurally oriented. So for an IBD trial in particular, in order for a patient to qualify for a trial, initially, you have to have them go through a colonoscopy as part of the screening process, and they have to meet these very specific endoscopic scoring criteria from that colonoscopy. And so even just to get into the trial, you have to actually go through a procedure.
And then over the course of the trial as a patient receives whatever, either the treatment or the placebo arm is, they get these additional colonoscopies because one of the key endpoints for the trial is mucosal healing as judged by visualization on endoscopy. And it’s not the type of thing where you can, give the patient a drug and kind of send them on their way. And you check back in two years and see how they turn out. The patients have to keep coming back to this very procedurally oriented trial. And so it definitely brings in some nuances when it comes to planning for these trials, actually working with clinical trial sites that have to be able to perform endoscopy and actually treat these patients as they would for standard of care. So there’s a lot of interesting tie-ins there that, get the physicians really involved.
Simon Burns: It’s also interesting how you’ve had to build a company that’s simultaneously a software company. You do a lot of AI ML software, not, not even easy software. This is hard software. You’re a hardware company, right? You’re deeply embedded in system design and in a core of the metal hardware and then ops, of course, right? You’re very much an operational company. If you think back to starting the company, what would you have told yourself as a young founder going after the problem? Key lessons Where to focus, where to deprioritize.
Matt Schwartz: I probably would’ve told myself I was a little bit crazy just because we bit off a lot to start out with. We felt like it was really important to take existing endoscopy systems and make them [00:05:00] internet-connected. And in order to do that we really needed to build some sort of a hardware solution that we could attach to existing endoscopes.
And yeah, as an early stage company trying to build cloud video infrastructure plus hardware to do it in a highly regulated, space with very long sales cycles with hospitals hospitals like all of the hardest things you could ask a startup to do, looking back, I don’t know that there’s a ton that we could have done differently other than, me and my co-founders, we were also first time founders, and so we had that added level of difficulty. Maybe having chosen some other slightly lower degree of difficulty startup to build some chops so that when it came time to do early fundraising for Virgo, maybe would’ve had a little bit of a leg up getting things off the ground but we just felt like it was the right strategy to try to build hardware and software together if we’re gonna build a complete solution. I think it’s paid dividends it was definitely tough going in the early days, but we learned a ton about what it means to go through hospital procurement at top tier academic medical centers in the country.
And now it’s a core competency of the company is we know how to do that in a lot of cases, actually better than some large established medical device companies. You’ve gotta pay your dues at the start. I don’t know that there would’ve been a better solution for us.
Simon Burns: It’s always a balance on whether you think a point solution’s gonna be the right, path, or whether you have to go full stack. And I think in this case you probably had to go full stack. It sounds like that was…
Matt Schwartz: Yeah, I think so. You know, it’s interesting, even in the early days, I talked to investors who would ask me, ” why don’t we just build our own endoscopy systems?” And maybe that was a, pathway as well. But, that would’ve been, I think, the ultimate degree of difficulty to try to actually build really complex endoscopy systems. The leading endoscopy equipment manufacturers in the world, companies like: Olympus, Pentax, Fuji Film, Stryker, and Rigid Endoscopy, they’ve got a ton of experience under their belt when it comes to optics and, I don’t think we would’ve ever been able to fully integrate to that level but I think we picked the right level of full stack integration.
Simon Burns: One of the things I’m always fascinated to, chat to healthcare founders about is what other problems they run into. Healthcare is so deep. You go down one little nuance and before you know you’ve stumbled across three or four different problems around it. I know for us, we run into all kinds of things and in doing trials we go, that’s a cool problem. I wish we could have a little more resources. We would go after that one. Well, what is that for you? Are there problems you’ve run into and you go, “I really wish in a different world we’d be running after that problem, too?”
Matt Schwartz: Yeah. I mean, I’ve got a couple of pet ideas I would love to be tackling. I don’t know that any of them are super related to Virgo and what we do specifically, but just kind. Concepts that have come to mind. I think everyone is always bumping up against EHR systems and the pains that come with them. And I’ve long thought that EHR; it’s actually a great product in general for what it’s intended to do, which is to be primarily a billing record. And, in that function it works incredibly well, but when you try to take that and turn it into a clinical tool, it can be really challenging. And, you know, everyone that tries to apply NLP or other whatever type of machine learning to EHR, I often think it’s a little bit of a fool’s errand. And so I’m always interested in opportunities to capture new source data at the point of care in lieu of trying to work with the medical records.
Like opportunities to do in-procedure audio capture could be really interesting in clinic audio capture. I’ve always thought the company, Augmedix that they used Google Glass early on to record clinic encounters and automate note taking was really interesting. Not necessarily just for. The fact that they were potentially reducing the time doctors have to spend on charting. But I thought the actual raw audio capture of the physician encounter could be really interesting from a machine learning perspective. So I’m always a fan of, like, placing more sensors of any type into the healthcare environment, capturing that raw data and figuring out what to do with it down the road. I think of what we do at Virgo as that, specifically applied to video data that historically wasn’t getting captured.
That’s something that’s always really interesting to me. The other kinda like, out there idea that I had and kicked around would be some sort of like very low intensity magnet. MRIs at home. I don’t exactly know how to make this work. There’s a really cool company, Hyperfine, that’s doing mobile MRIs that could be used in the emergency room. It’s lower quality imaging, right now but the fact that you can do it in more locations is really interesting. And I think if there a way to safely get an MRI every day could you apply machine learning to that and detect things earlier? Just a, an out there idea that I like to noodle on from time to time
Simon Burns: Wow. Fascinating audio first EMR and miniaturized MRI. That’s pretty, you know…
Matt Schwartz: Like an MRI in your shower. So, every day you step in the shower, take a shower, and you get an MRI when you step out.
Simon Burns: Okay. These are not easy problems.
Matt Schwartz: No.
Simon Burns: This is not consumer social . Alright, let’s move on to a segment. “Overrated, Underrated.” I’m gonna send you some topics you’re going to give me your first Overrated, Underrated. The use of ultrasound in IBD clinical trials. Increasingly a hot topic. Where do you stand? Overrated, underrated on ultimate impact.
Matt Schwartz: Yeah, I think, probably currently underrated, but definitely gaining a ton of popularity. This is something that’s been coming across our radar a lot recently. I know there are groups that are looking to define how intestinal [00:10:00] ultrasound can be used as an actual clinical endpoint in IBD studies.
Also, how intestinal ultrasound can just be used for clinical care to potentially reduce the number of colonoscopies IBD patients need to go through. We also think there’s really interesting opportunities in using intestinal ultrasound as a patient recruitment mechanism again by reducing the number of colonoscopies needed for screening patients into trial. So it’s something we’re super excited about. We’re starting to work on some Virgo-a-fied capture mechanisms for intestinal ultrasound and think there’s some exciting stuff to come there.
Simon Burns: Awesome. GI as a therapeutic area gets a lot of talk, nash, boom, sorts of different things. It’s been particularly difficult for biotech founders to crack. What’s your sense? Overrated, underrated level of attention and, therapeutic air investment happening in GI?
Matt Schwartz: In GI? Definitely underrated. I’ve spent my whole career in healthcare and prior to Virgo never stopped to think about GI. It’s maybe one of the least sexy areas of healthcare possible, especially thinking about colonoscopies.
But the really interesting thing is that for a ton of patients, GI is actually like their first encounter with chronic healthcare use. And whether it’s motility issues, IBD, acid reflux, eosinophilic esophagitis. There’s all these conditions that become chronic and have a huge impact on patients’ lives. And, I think you’re seeing more and more of that from a pharmaceutical investment standpoint. I think it’s just starting to get the attention it deserves from a technology perspective, and getting investors involved. So yeah, I think it’s still underrated, but is becoming increasingly important in the broader healthcare landscape.
Matt Schwartz: It’s been interesting in IBD trials in particular, you know, as the DCT craze became very hype-oriented. I think it’s maybe a little bit overrated or at least just not properly applied to GI-specific trials. And I think a lot of this comes down to the fact that as I mentioned before, these trials require patients to go in for procedures and no time in the near future patients can be getting colonoscopies at home.
So the key I think is defining what you actually mean by decentralized clinical trials. Like, is this truly something where you want patients to just be able to go through a trial on the comfort of their home? Or are there other flavors of decentralization that you can take advantage of? And that’s what we try to focus on.
So I think historically, DCT is not really applied well in the realm of IBD trials. But I do think there are decentralized principles that can be applied. And that’s, largely what we’re trying to do is take your typical GI clinical trial site. And, actually increase its footprint with technology and data capture.
And so take like an academic health system and how do you get their affiliated community hospitals and outpatient endoscopy centers to become referral sites internally within the same health system. And, if you wanna call that, “decentralization,” fantastic, but think it’s more about looking at a really specific problem and how do you solve it the best way possible.
Simon Burns: It’s funny, I guess often with these buzzwords, a lot of infighting over definition. I’ve heard several companies, even within the company, they can’t agree on decent terms.
Matt Schwartz: Yeah.
Simon Burns: But it does seem like hybrid is the future. Some version of hybrids increasingly the marginal visit being pushed to a phone call. Alright. Let’s talk about use of ML and a AI in clinical trials. Seemingly synthetic control arms is the first place this is coming out. What do you think of external control arms, synthetic control arms to remove the need for control.
Matt Schwartz: I think it’s obviously really exciting. In principle, I’m a bit of an evidence-based medicine stickler and love a good randomized control trial. But, I think there are definitely ways you can leverage real world data potentially to just inform better trial design going forward. And that, something that we are pretty acutely focused on is how do we leverage the data that’s flowing through our system to glean these insights that can help a trial sponsor, maybe design a trial for a more specific indication reduce the number of patients they might need to bring through a trial. I’m maybe a little bit skeptical that we’re ever gonna be able to fully create synthetic control arms, but I love that people are pushing on it and look forward to seeing data proving that they’re viable.
Simon Burns: Last question for you. take us through a, a dream state here. We’re five years out, the clinical trial of the future has been deployed. It’s operational. What does it look like? What should we be looking forward to for the future of the industry?
Matt Schwartz: One of the things that we’re really keen on this came from looking at the IBD trial space in particular, but I think it applies to basically any clinical trial, broadly. If you just look at the epidemiology of IBD. There are so many patients with this disease, and yet it’s incredibly difficult to enroll IBD clinical trials. Your average IBD clinical trial site recruits one patient per year for IBD studies. And so if you’ve got a phase three trial that needs 1500 patients, you’re looking at onboarding 300 trial sites around the world and still waiting five years. And, so it’s not uncommon to see these phase three trial programs that take 4, 5, 6, 7 years sometimes get canceled altogether. And, I think the dream state of the future is piecing together a lot of technology components and service [00:15:00] components. I don’t think there’s any one silver bullet that’s going to solve this stuff.
We’ve been doing a lot of partnership work to try to piece different solutions together. I’d just love to see a world where, across the spectrum of disease, patients who are interested in participating in clinical trials can fully be optimized into the right trial. And you just see these trials fill up at the rate that you’d expect based on the prevalence of the disease, and that’s clearly not the case right now.
I think there’s a number of reasons why IBD trials are particularly difficult. But I think this applies really to any trial that has an endoscopic component to it, and probably even more broadly outside of endoscopy.
Simon Burns: Awesome. Well thank you for the tour de force and all things clinical trials and IBD. Really appreciate the time. Matt, thanks for jumping on.
Matt Schwartz: Yeah, my pleasure. This was fun. Thanks, Simon.