I'm Gary Taylor. I cover healthcare facilities and managed care here at Cowen. My pleasure and my interest, actually, to have RadNet, who is a leading national provider of freestanding, fixed-site diagnostic imaging services and related IT and AI solutions. The company has a network of 366 outpatient centers in seven states, and now two reportable segments, imaging center, and its newly formed digital health segment. So today we have Greg Sorensen, who's the Chief Scientific Officer, and Sham Sokka, who's the Chief Technology Officer. And they're gonna do presentations and a fireside, but maybe still have time for questions at the end. So take it away.
Thank you. Thanks so much, Gary. It's a pleasure to be back here. Thanks for your attention. Thanks also to those on the webcast dialing in. Sham and I are gonna share this presentation, so we'll both be up here and kind of bop back and forth between the slides. Our goal was indeed to kind of give you a brief overview, especially of the digital health business, because of this new segment reporting. I think many people were able already to tune into the earnings call, and the transcript's available on our website, et cetera, from last Friday's earnings call, so I don't want to repeat too much of that. Instead, take this opportunity to really dive in to why we think the time is right for us to talk about our digital health business.
So that's why Sham and I are here, as opposed to Mark or somebody else from our team. As usual, forward-looking statements, so the usual disclaimer has to be present. And here's the agenda of what we'd like to talk through. I wanna give you at least just to level set some information about RadNet. And then, as I said, we'll dive right in to talking about digital health and why we think there's such tremendous opportunity here. So most of you, I hope, are familiar with RadNet. As Gary said, 366 imaging centers as of the close of the 2023 calendar year. More already since then, so that's why the plus. I think 400-ish isn't a bad number to carry around in your head.
We have about 800 radiologists that are contracted or work with our centers. We basically are one of the largest, if not the largest, outpatient imaging center chain in the world. This is our focus. It's what we do well. We think we're best in the class - best in class at it. As a result, we have some great partnerships. We have 25, 26 joint venture partners, where the hospital chains look to us to help them grow and deliver a good outpatient imaging care. We do about as many mammograms every year as the country of England, so we're at really a national kind of scale provider. We do a lot of other kinds of exams, chest CTs, MRIs, chest X-rays. Pretty much anything you can take a picture of, that's what we do, PET, ultrasound, et cetera.
We actually are a little unique that because of our scale, we also take risk in a couple of our markets, where we actually go on a PMPM basis for outpatient imaging, for that groups want to sub-capitate to us. We're very, very interested in expanding access to care, and I'll get a little bit to that more in a moment. And now, as I said today, what we really wanna shift gears to, after one or two more introductory slides, is talking to you about our digital health division, which we're calling Deep Health. We'll tell you a little bit about where that name comes from and where we're headed.
We think not only is AI very valuable for radiologists interpreting pictures, but it's also super valuable for the entire operations that we do and that our customers externally want us to do. As Gary said, as of the close of the fiscal year, calendar year, we were in seven states. We have since announced last week a definitive agreement to move into the Texas market. So that's why Texas is now added to this map, at least with white letters. We haven't closed that definitive agreement yet, but we don't imagine, or at least right now, we don't see much risk to that, and so I think that by Q2, we will be in an eighth state, and we're excited about that. And I would like definitely emphasize our partnerships.
Those joint venture partners bring a lot of ballast or evenness to our revenue, and to our patient outreach. They give a sense of comfort to referring docs because it associates us with the quality providers in many markets. We also have many other partnerships that I've listed here on this slide, ranging from other small companies and big companies we're co-investing with, to societies, to even patient groups or sports teams that we provide a lot of services for. Interestingly, if you think about what outpatient imaging is all about, our biggest single clinical area is musculoskeletal imaging. So all those sports people, you know, injuring themselves or having a sore knee after just running too much or whatever.
And so we do a lot of sports medicine, we do a lot of musculoskeletal work, a lot of back pain, a lot of the things that you're not sick enough that you need to be in an ER for, and that you'd want some imaging for. That's where we are. Just before I switch to the DeepHealth concept, I did wanna mention again the equitable care thing. One of the investments that we are actively pursuing, and we've announced already, but I'll just highlight again, is our interest in expanding access to underserved populations. Nowhere is the data clearer on this than it is for screening mammography. Screening mammography is proven to save lives. As I said, we do it already at scale, and we've seen big benefits in our screening mammography from artificial intelligence.
Hundreds of cancers that we'll tell you about, we found through AI. However, the data are also very clear that certain populations are underserved by screening mammography. We know, the CDC reports that somewhere in the neighborhood of 65%-70% of women who could be getting mammograms at least every other year are getting them. Many people, including the USPSTF and others, now recommend that those screening guidelines get expanded, and so the CDC numbers don't include the 40-50 group. There's even lower numbers than 65% that are getting mammograms in that age group. And of course, many of us think annual mammography would work, and so if you were looking at annual mammograms, it's a lower number. But more importantly, there are populations of certain racial groups that are not getting enough screening mammography.
Part of the reason that the draft guidance from the USPSTF around screening mammography changed last year to start screening at age 40 is because they recognized, finally, publicly, that Black women get breast cancer at an earlier age, and they get breast cancer that tends to be more dangerous. For reasons we don't completely understand, Black women die from breast cancer at a much higher rate than white women do. And we can see that those benefits need to get to those women, and one way we are attempting to solve that problem is bringing the screening services to them, and hence our pilot sites at Walmart, where a different group of patients can get access to mammography easily, and in a rapid way.
Happy to talk more about that, but I think it's an important part of what's driving where we're going with RadNet and why, candidly, AI is so important. We would not really have been as comfortable as we have been doing the Walmart initiative if it weren't for AI. So, let me now shift gears and turn some time at the podium to Sham, who's gonna talk through some of the capabilities that we already have and that we're building. Take it away, Sham.
Thanks. Thanks, Greg. So, and good to have me, good to meet you all. I joined RadNet in June of last year from about 20 years in with Philips running various software businesses. You know, about this time last year, RadNet really said, "Look, we have all of these digital assets. We're seeing real efficiencies by using these assets, whether it's AI or operational assets, but we really think there's a huge opportunity in both scaling that up, leveraging cloud and AI technologies, but also creating a real growth product business around it." So I then joined RadNet essentially to start to put this digital division, digital health division together. We started that journey in middle of last year.
Mark had launched it at the Radiological Society of North America meeting this past November, December, and then now we formally are reporting this as a separate segment. So what is DeepHealth, right? So DeepHealth really brings a couple of key capabilities together that we think are really unique in the healthcare IT and informatics market. One is that because we have the vertical integration capability with RadNet itself, we can look at the whole process of end-to-end care and rapidly build applications and tools that could immediately impact end-to-end efficiency.
So we have a portfolio of solutions, what we're gonna walk you through, called the Operating System for Radiology, that essentially now allows us to bring these solutions and these learnings that we have from RadNet into the broader market at large, right? In addition, we also have this really interesting clinical test bed, right? So where we can bring our clinical AI algorithms in. For example, we have our breast cancer clinical AI tools, our lung cancer clinical AI tools, our prostate cancer clinical AI tools. We develop them, but we can also generate, you know, the outcome proof point. So one example I want to just call out is, since we've introduced our cancer detection program in breast cancer, we're essentially finding more cancers, and we're limiting recalls, so patient callbacks for false positives at the same time.
It's really a very dramatic change in that relative to standard practice, and we see that happening, and that's just in one area, just in breast cancer. We see now the opportunity to really bring that for other, more and more applications. The third element, what we can now do at DeepHealth is just the scale, right? So the scale of RadNet, you know, 10 million studies a year, we have customers that that do another 6 million additional studies. So we have this customer base where we can actually now start to bring the next generation of solutions after testing it and validating it in the RadNet enterprise.
So this sort of combination of capabilities, I think, makes us a really unique software business, a SaaS business in the healthcare sector. A little bit about us as of today, just to give you a sense of where we are today and what we're gonna leverage going forward. So we have about 300 customers globally. So we're in the U.S., we're in Europe, we're in smaller markets like South Africa, Israel, where we actually have pretty interesting market share positions. For example, in Israel, we're the market leader in radiology information system, RIS, which essentially, think about it as the EMR for radiology, right?
So we have RIS PACS solutions in the U.S. with over 200 customers, some of them the largest outpatient imaging centers after us. For example, in Boston, Shields uses our RIS platform, right? We've really historically catered to that outpatient segment, and we think there's a tremendous opportunity to drive informatics solutions to that segment, given where also the care is going, right? As you think about here in the United States, but also in Europe and other markets, outpatient is a much lower cost setting, so the systems are driving care more and more to the outpatient setting, number one. Number two, care is more and more, as you know from covering other things, about healthy patients. And where does that happen? That happens in the outpatient setting.
So things like screening applications, things like surveillance applications are really happening in well patients. So we're really driving that, trying to pull on that globally. The last thing I would say is, you know, all informatics or software solutions are local, because they take local. You know, how we get paid here is very different from how it gets paid in France, and workflows are different and so forth. So we're also partnering in these regions to take our solutions that we might have developed in one market, but really customize and configure it locally. So you see a list of our partners, both with large OEMs like GE, who have global footprints, but also local players like Incepto, whose 70% of the AI applications go through Incepto in France.
And so we can start to bring their local scale as well. That's the way that we're thinking of really going to market and bringing our propositions together. So, what are the opportunities for DeepHealth? And maybe this is where I'll hand off to Greg, who will maybe give a little bit of background of the kind of propositions that we're going after.
Yeah, you know, we certainly see that RadNet itself is well poised to, because of the uniform IT we have across all these 370+ imaging centers, it's all uniform. We have an opportunity to get to take advantage of Gen AI, clinical AI and other tools, cloud, et cetera, and make our imaging services much more efficient. And that is certainly the core of what consists of the, you know, the extensive revenue that RadNet generates today. We also see that digital health could enable us to expand into new and new areas, and to do better the things that we're already doing today other than imaging. So, as we've mentioned, as I've talked about before to this audience, cancer screening is such a value creator for the healthcare system.
To find a cancer at stage one instead of stage four is super value for the patient. They live longer. They get less toxic chemo. It's valuable for their employer because they, they're back at work sooner. It's valuable for the healthcare payer, whether it's the government or their employer, because it costs less, right? So all of this is a win. Those are the areas we've been investing in, and precision screening, I think, is gonna make that even more possible. You look at all of the blood tests that are being developed, these MCEDs, the multi-cancer early detection tests. If those are positive, what do they need? They need a PET scan, a CT to localize the cancer. So as we think about the idea of precision diagnosis, imaging plays a role.
It's a natural next step for us, and that gets you to integrated diagnostics, not just screening, but bringing genetic or other kinds of information together with the imaging. To do that efficiently and at scale, we need a digital backbone, where that information can have a common grammar, a common way of interchanging information, and of gaining insights and providing insights to the doctors and the caregivers through that. And that, of course, when you talk about insights in healthcare, that naturally leads to clinical trials. RadNet does act as a service provider for more than 100 clinical trials every year because they need our imaging services.
But right today, we don't have the infrastructure or really the bandwidth or the teams to actually push clinical trials as a bigger opportunity for us, and we think that could be a potential opportunity. There's actually quite a bit of activity, as you all know from this conference, in using imaging and related diagnostics to gauge whether or not a drug or a novel intervention is successful or not, and we think we could play a bigger role in that. And so that's kind of the big vision of where we might take things with digital health. There's plenty of near-term challenges we have. RadNet takes 50,000 calls a day to schedule its millions of exams, and that happens at a call center. That call center could be more efficient.
We don't have enough of those patients scheduling online. We have doctor burnout issues. We have staffing issues. All kinds of things that you can see on here that I won't go through. All of those are opportunities where cloud, Gen AI, modern technology could help us gain real serious efficiencies. And when we, as RadNet management team, saw that opportunity, we're like: Look, there's so much to do here besides the pixel AI. Yes, the pixel AI is enabling us to find cancers earlier, and it's enabling better care, and it's enabling new revenue streams. That's all great, and we want to do more of that, but there's even more we could do across the whole level of service that we provide. And so that's where this idea that Sham mentioned of kind of an operating system comes out. And so let's talk about that, Sham.
Yeah. So, Greg, thank you. I think the summary of that is you see this great demand for outpatient imaging, both with what we talk about as well imaging, but also because it's a lower care setting. So we're getting this influx in demand. At the same time, what Greg just talked about is this significant burden on the supply, right? Physicians, technologists, the staff is, you know, under tremendous pressure to deliver on that growing demand. So if you come into a diagnostic enterprise today, right, you have these sort of six steps of the diagnostics, right? So you have everything from, you know, scheduling the patient to doing the acquisition. Then you need a radiologist to report on him. Then the data is then transmitted to a referring physician. Decisions are made, right?
Then you have collection, billing, operations, all those different pieces. Today, every one of these blocks is its own technology stack, right? So at RadNet, we have to basically cobble together a Frankenstein of an IT system to drive operations, right? Now, in some areas, we've consolidated, right? So we have one RIS, one PACS, so we've consolidated it, but still, we're driving hundreds of integrations in our existing environment. So if you now want to really optimize the system, right? The data doesn't flow really between all these subsystems. So we're in a situation where it's incredibly inefficient or very difficult to really drive and use digital data to extract value.
We're on the journey, and that's what DeepHealth is really fundamentally driving, is to create an environment where we can bring all of this data together in one environment, and we're doing that, and we're doing that on cloud. That's actually one of the key solutions we're building. We're putting all that data together in one environment. Essentially, from that we create one solution, and that one solution is what we're calling our DeepHealth OS. So think of it as a diagnostic management platform, right? Maestro is our brand name for all the operational sort of assets.
So think of it like a conductor for the factory of imaging, if you will. And then Sage, providing wisdom to physicians to guide diagnoses, so more sort of augmentation type of tools, right? So these three product lines will essentially drive the future of DeepHealth. Now, on top of the OS, it's one data platform, but what we're now working toward is creating workspaces for the different stakeholders in the enterprise, right? So if I showed you before, the different tools were based on stages of the workflow. Now what we've done is we've put all the data together into one and created workspaces or work spots for the individual stakeholders. So data across multiple—what was in multiple systems now are put together in front of the one user.
So for example, if you take a radiologist, everything they need to do is in their environment, and now I can inject AI to make them go faster, right? For a technologist who's running a scan, you know, one-third of the technologist's time is not spent scanning. It's been chasing down all the information to make sure the right scan is done. Well, in the technologist's workspace, I'll consolidate that information with some of these Gen AI technologies, summarize the data, show them where the gaps are, and now bring predictive techniques to tell them, this is the protocol that should be done, right? So if I now squeeze that 30% time to 10%, techs are one of the most expensive assets and resources we have, I can now run more scans with those technologists, right? So like that same thing, re-registration, patient engagement.
Now, less than a quarter of our patients are scheduled online. That means 78% of our patients are calling someone. Every time they call someone, it costs us $12 to schedule that patient. If they use the digital app, it costs us anywhere between $1 and $1.50, right? So, and we're—I would say we're at the state-of-the-art on RadNet in terms of digital scout scheduling, right? So putting these digital tools into the workflow will dramatically drive efficiencies of our core enterprise, the core RadNet business. And as we build this, and it's one of the reasons I'm here, we're building it not as an IT implementation, but as products, as these workspaces on this platform. So now we can bring that to the rest of the industry to scale, right? So that's fundamentally what DeepHealth OS is.
We're on the journey, so I'm just showing you some snapshots. This is not a vision. We're actually built these assets. We've actually launched some of them, some of these assets, at the Radiological Society of North America. I was in Europe last week for the European Congress of Radiology, where we again talked about the market entry into Europe. So we're very much along the way in building this. This platform is cloud-based, but can work with on-prem capabilities, and we're leveraging much of the AI that you're hearing about, especially with the generative AI technologies, to drive efficiencies in the workflow. Now, in addition to these operational sort of resources, our clinical, you know, Greg talked a lot about our volumes in, well, imaging, which is a lot of screening.
We've also now consolidated our investments in clinical AI in breast, lung, prostate. These are the big three cancers as many of you know, affecting women and men on the prostate side. They're in various stages. Mammography is a very mature screening protocol. Our goal there is to take women, currently, we're at national compliance of 60%, 63%. We're trying to drive that up to 75%-80%. It's already covered. There's a lot of capacity for growth in breast cancer screening, and tools like this are critical because we don't have enough radiologists. If I don't have AI to help do that with high quality, with more accuracy, I'm gonna be able to do more breast cancer screening, but not with the accuracy, right?
And so that's what the AI tools are letting us do. Lung cancer screening, which is the one in the middle, again, nationally approved, but less than 5% of the eligible patients in lung cancer screening are getting screened. So again, huge opportunity for growth there. And prostate cancer, where all the clinical folks sort of accept it's the best way to assess prostate cancer risk, but we're in the very early stages, non-reimbursed yet. So we're now building a program, and we've actually started pilot programs in New York to drive prostate cancer screening. So one of the reasons that we call our business DeepHealth is we really think these solutions will drive health versus really being much more imaging-centric. They're imaging-centric, but they're really about cancer risk. We see a portfolio coming in the neuro space-...
And some of the other areas where we're really looking and surveilling the population for escalation and risk management of these diseases, right? Just to show you this, we're not just doing this in the United States. You know, it's fascinating. Europeans are actually ahead of us in deploying lung cancer programs. So we are the leading technology in the UK lung cancer program. They now are expanding that to the full country based on pilots we've done with them early. The idea is basically they have lung check centers, so diagnostic centers, and people that are at risk show up to those centers, get their CT, get their diagnosis, and they come back every two years.
They're deploying it at scale, and, you know, we were just sharing the results of that in Europe last week. Again, the physicians cannot deliver this program without the AI, because you don't have the scale and the volume can't be delivered without tools like this to overcheck and oversee the quality, right? So we're quite excited about the growth opportunity there as well. Finally, a quick view on, you know, really how we see the rollout. 2024 is a build year, and you'll see that also in our financials. A lot of these solutions that I showed you that we've built and prototyped, we're spending the beginning of 2024 piloting them in RadNet. So again, going back to the vertical integration, for our software business, it's I can't tell you how valuable that is.
When I was in Philips, we had to really recruit sites. It would take 18 months just to recruit sites to validate these things. So now I can literally roll over to my partners on the RadNet side, put things in, validate it, assess results. So we're in piloting form there. In the second half of the year, we're injecting the AI enablement, so the gen AI tools, the automations in patient engagement, right? The conversational AI and patient engagement, the ability to kind of drive some feature extraction from reports for auto coding, right? All these kinds of use cases, we're bringing that in in the second half of this year, and the idea is in 2025 to scale this across RadNet and to our customers.
So our ambition is that by end of 2025, we're fully on cloud on this AI-enabled operating system. We'll be the first provider of any size like us to be really on a modern cloud environment, on a 100% modern cloud environment. And then the idea is that to bring these solutions more broadly to the external market. And I think I just want to hit one point out there in terms of the opportunity. Less than 2% of the healthcare solutions are on cloud. And so we think by being the first in this space, in the radiology and diagnostic space, and showing that at RadNet, we can become the leader in this area, disrupting the conventional players, right?
So, that's actually one of the opportunities for the business growth, and we're gonna do that over the next 3-5 years. So with that, let me hand it off to Greg.
Yeah. So just to finish up, I thought I'd share with you the guidance that we released as part of our earnings release last Thursday afternoon, just on the digital health specifically. On the left, you can see how, with the new segment reporting, the performance on both the revenue and adjusted EBITDA basis was for this now digital health segment. So that includes our RIS PACS business, the old eRAD business, and the clinical AI business, the three DeepHealth, Aidence, Quantib all consolidated together. And we've also given some guidance for 2024 based on what we see in our own pipeline and customer interest. And I guess the two obvious things here are, it's more profitable than many businesses, and it's growing relatively rapidly.
So, definitely, I feel the pressure to now deliver on that. But this is the guidance that our management team feels comfortable with, and so, we wanted to share that with that. With that, I think just we were gonna just leave this slide up just to remind you of kind of our overall strategy. But we've got a minute, or I guess not even a minute. I don't know, Gary, if you wanted to ask a question or two, but we'll leave it, or we can end a minute early. Your call.
Yeah, I have a quick question, and I think probably we'll take questions aside-
Great
in case another
Sure
... coming in. I just wanted to ask on the clinical AI part of it. Are you yet using any of that tool without a radiologist putting their physical eyes on something, or is it still that's the backup, and they can move faster with more confidence because they know they have this on the back end, whether we're talking mammography or what you're doing in the UK with lung?
Yeah, it's a great question, and the answer is no, not yet, for regulatory reasons. The data we have suggest that the AI is ready to do that and to take over at least some of the image interpretation for docs, and that would actually be better for patients, better, better accuracy. However, especially with mammography, there's a law that says a radiologist has to look at every mammogram. And so we've talked to the FDA about it, and they're like: Yeah, well, we'd love to give you approval in theory, but in practice, there's this law that says it's not possible, so go talk to Congress and get the law changed.
Is Europe ahead on that part?
Europe actually is. There is one group that has an autonomous chest X-ray clearance, but it's not really being widely used, but I think that's coming. And we can do. We will be able to do this autonomous in non-mammography places. So I'd say watch this space. It's an area that the regulators are, as you might imagine, super nervous about. So it's gonna take some negotiation no matter what the domain. But at least in, for some tasks, like, "Is there breast cancer here, yes or no?" The AI does seem as capable, or in our data, show more capable than humans at some very narrowly defined tasks.
Maybe, maybe one quick thing about Europe. Most screening applications in Europe are double reader, so you have to have two radiologists read, right? So our lung product is the first screening product in Europe that's actually single reader, because AI, you know, essentially-
They're hoping to move their mammo to just AI plus one reader rather than two readers. They're not ready at all about going to zero readers.
Yeah. They're still at 2 readers for most of screening. So that's, I think, the good factor.
The starting points are different. We're both moving in that direction. We're just starting different places. Yeah, super interesting times.
Well, awesome. Great. That's RadNet. Let's move the questions-
Questions
... to the sidebar just so we can-
Thanks, everybody
... Yeah, thanks.