Ep7: Radiology’s AI Revolution and the Future of Digital Medicine, with Sofinnova Partners and deepc

[00:00:00] Henry Peck: Hey everyone, it's Henry. Welcome to Emerging Medtech Today by LSI. Today I'm joined by Simon from Sofinnova Partners and Franz from deepc AI. Together we discuss the emerging field of digital medicine, Sofinnova's new investment strategy focused on digital medicine, why radiology is ripe for AI innovation, and how deepc AI plans to build the AI operating system to support radiologists. Enjoy.

[00:00:28] Henry Peck: Simon, Franz, thank you so much for joining us this morning. Simon, we'd love to start with you and a brief intro to yourself and Sofinnova.

[00:00:35] Simon Turner: Yeah, thanks Henry. Thanks for having me. My name is Simon Turner. I'm one of the partners at Sofinnova Partners. Very briefly on my background. I'm basically been focused in on this entire health tech space for the last decade, background both within business development and early stage investing. And with my partner, Ed, and we've put together this new digital medicine strategy here at Sofinnova.

[00:00:54] Simon Turner: Maybe Sofinnova in brief, in a couple of words, we're one of the oldest life science investors that you'll [00:01:00] find out there. We've been around for 50 years, backing entrepreneurs and teams as they're looking to change the status quo within health sciences. And what we've done throughout the years is build up our network, but also our strategies. So right now, what you'll find is actually, we have multiple strategies operating within the health science space, as well as greater life sciences. And the idea that we've always had is you need dedicated teams and dedicated strategies to invest into these extreme spaces.

[00:01:23] Simon Turner: So we do biotech, medical devices, both early stage and a bit later, as well as more bespoke funds, like our industrial biotech team, looking at sustainability side. And then of course us with the digital medicine franchise.

[00:01:35] Simon Turner: Thank you. And we also have Franz joining us. Franz, I'd love you to introduce yourself and deepc AI.

[00:01:40] Franz Pfister: Thanks so much, Henry. Great to be here with you and Simon today. So, my name is Franz. I'm the CEO and one of the co founders of a company called deepc. My background lies in medicine. I'm a medical doctor by training, but also in technology. I have a degree in machine learning data science.

[00:01:58] Franz Pfister: And a couple of years back, we [00:02:00] looked at the space of healthcare AI and figured out that there is like a myriad of different solutions already out there in the field of radiology AI, that's medical imaging and tailing CT, MR, x ray, mammography scans, and we figured that there is a plethora of different solutions out there in the market, and it's very difficult for hospitals to adopt those solutions because simply there are so many out there, but then at the same time for the respective companies that develop those great solutions, it's not super easy to access the market. So instead of developing yet another of those AI applications, we decided to focus on infrastructure and platform development. And that's what we do with deepc.

[00:02:48] Henry Peck: Awesome. Simon, let's talk about the new investment strategy that you co lead at Sofinnova. You recently launched the digital medicine strategy, and I think it's important to baseline first. Can you explain to us what digital medicine means to you? [00:03:00]

[00:03:00] Simon Turner: Absolutely. It's a good question because we throw around so many terms in this space, be it health tech, tech bio, e health, digital health, et cetera. The beginning stages was really this e health space that started about 10, 15 years ago. Now we started to actually sub segment that a bit better.

[00:03:16] Simon Turner: When we think about digital medicine, what are we looking at? We're looking at tools and technologies that can be applied into the health sciences that will ultimately impact patient and patient outcomes. And coming from Sofinnova, we look at this with a very critical eye and really focus in on the medical application perspective of it, which is why we've focused on this being called the digital medicine strategy rather than a digital health one.

[00:03:38] Simon Turner: We don't necessarily do more consumer grade types of approaches, etc. Many of the apps that you'll find are more well being designed. This is not to say they're not good investments, it's just it's not where our expertise lies. We like to dig down into that next level where we have clinical efficacy, we have proof that there is an impact on patient outcomes.

[00:03:55] Simon Turner: So what we've done with this is broken it down into three sub segments, if you will, [00:04:00] within digital medicine. If you imagine the patient in the center point, and the patient for us, regardless of the strategy is always in the center. We'll go upstream of that, we'll focus on understanding the patient and their disease journey, and then we'll actually go downstream towards how we can actually impact the patient.

[00:04:15] Simon Turner: So, this upstream component, we call that enabling technologies. And enabling technologies, they're dedicated really to development and or adoption, dissemination of AI and also all these algorithmic approaches within the health science space. So, if you will, they're picks and shovels. They're tools, they're technologies, they're underlying infrastructure that's required before this can happen, deepc, being an absolutely excellent example of that, and I'll give a little bit of an analogy at the end as to how and why we invested in deepc.

[00:04:43] Simon Turner: The next piece, analytics, this is really the precision medicine side of things. We see today that it takes too long for patients to be diagnosed. We're at a very responsive approach in medicine rather than something that can be proactive.

[00:04:57] Simon Turner: So we're looking for these approaches which are able to [00:05:00] screen, identify patients earlier, reduce the delays that we see in terms of disease diagnosis and final actual treatment application. And also being able to eventually get to a point where we can get personalized care. So, this notion of the right treatment for the right patient at the right time, but also conforming to their lifestyle, conforming to their personal needs.

[00:05:20] Simon Turner: And then the last bucket, we're into the treatment side, and this is where we're looking at approaches that can leverage new and existing technologies to ultimately improve patient outcomes and modulate disease. So fundamentally change the process that a patient is on and hence improve outcomes. So it can be as standalone treatments like digital therapeutics, potentially, but also in combination with existing modalities.

[00:05:41] Simon Turner: So it could be hardware devices being supported by this type of approach or even more classical biotech, et cetera. But we ourselves, we won't invest in hardware, we won't invest in molecules, for example, we have existing strategies that do that extremely well. So enabling technologies, analytics, and treatments is the kind of way we've subdivided it.

[00:05:59] Simon Turner: [00:06:00] And maybe one of the interesting aspects of this is, as and when we've been developing the strategy, we sit in a very interesting intersection because of course, it's somewhere between life sciences and also technology. So we are a bit of a gateway in terms of seeing these interesting novel technologies that may come from other industries or be totally new and homegrown within health sciences, bridging into this space and really revolutionizing it.

[00:06:22] Simon Turner: And we see this in terms of expertise coming in from the automotive industry, in heavy engineering, in silico simulation systems that were developed in totally different sectors and industries now being applicable into the health sciences. And when you think about there are multiple ways of investing into these types of technologies, we break them down into three core ones, if you will. There is existing technology with a new channel, and that is, for example, Netflix. You've got the same existing content that we've known for years, but now suddenly on demand in a way that is totally new to the consumer. And that's something tech investors do absolutely fantastic.

[00:06:58] Simon Turner: Those types of approaches, they've got it down to a tee. [00:07:00] The next is, more classically, what we've been doing at Sofinnova, which is new technologies and existing channels. So we'll develop totally novel therapeutic approaches, even new modalities, if you will, but then we can plug those into the existing channels themselves. This could be pharma or medtech who again have these incredible sales and marketing organizations, they take care of that entire commercialization piece.

[00:07:21] Simon Turner: The last piece though, that's where we see the kind of potential and the disruption here. It's new technologies with new channels. And exactly to that point, the new technology we were looking at for a number of years, actually, is all of this AI and radiology that's going on. And we've seen dozens, if not a hundred vendors building these fantastic AI solutions to do analysis. But we came to the conclusion that that's not necessarily the key bottleneck. The key bottleneck is actually the new channel, which today, there isn't a good one that existed at the moment that we started looking at this. That's when we said, okay, let's leave the new technologies. Let's now go for the new channel. And that's what led us to funds, to deepc and really saying, [00:08:00] yes, we're going to focus in on this. And this is definitely the project that we believe has a lot of potential here. And I think to Franz's credit what they've been able to do just in the last seven ish months is testament to how much the company is developing because there is such a push and pull for these types of approaches.

[00:08:15] Henry Peck: Yeah, you talk about that new technology, new channel. Franz, let's build on that for a minute. What is the new channel that you're operating and that's enabling you to bring something like AI to radiology at this new scale that Simon is talking about?

[00:08:29] Franz Pfister: So it's basically on two levels. On the one hand, it's on the commercial level where we creating like a harmonized contractual framework, a procurement framework, a process framework, and an administrative framework where we have one contract with a hospital or with outpatient imaging center or with the tele radiology company that allows us to be very flexible in adding solutions as we go, giving them access to more solutions as we go. So this is the [00:09:00] contractual framework that we set up once and can use end times. On the other hand, we have a technology framework, which is a cloud native platform framework, which allows us to bring on board different applications in a very streamlined fashion in a very scalable way.

[00:09:17] Franz Pfister: We have our software development kit, our developer portal, where AI developers can basically study the documentation and then submit new AI models in no time. And once those containers are registered, we can basically scale everything up to infinity due to the cloud technologies that we are using. Of course, we do all of this in a privacy preserving way with the highest cybersecurity standards to address all the challenges and concerns that we can find in healthcare, but that's basically the framework. So it's like contractually commercial, and then on the other hand, it's technologically.

[00:09:56] Simon Turner: And maybe just rebounding on that. This was something that, as part of the due diligence [00:10:00] we were doing up until the Investment, I mean, Franz, we had so many interesting discussions with clinicians and also the kind of CTOs of hospitals, et cetera, saying, look, we're pretty convinced that AI adoption and radiology is going to happen because we right now, we don't have the capacity, quite frankly, to be able to deal with this.

[00:10:14] Simon Turner: We need these types of technologies to be able to really meet the demand that we're seeing. And growth in radiology is extraordinary because it's one of the first pieces of the diagnostic workup as and when a patient comes in from a new pathology or a suspected pathology. And the feedback was continuous.

[00:10:29] Simon Turner: We love the approaches that we're seeing. We just, we need a simple way to actually be able to implement them. And then for the radiologists themselves to be able to work with them, because we're now leap years away from where we were previously, which was, here's a login, use it for application A, and then here's another login with a different portal for application B, et cetera.

[00:10:48] Simon Turner: So you need this workflow integration, this modernization, this almost AI working, as you will, in the background, and then you just benefiting from the results. So it's the actionability that's a key component of it. And that's something that [00:11:00] we found just wow, when we were looking at the deepc platform and how it could be implemented and the kind of impact it was having on these radiology services.

[00:11:08] Henry Peck: Super interesting. Simon, thinking more about your business for a moment, you mentioned that this digital medicine strategy is building on the strong foundation that Sofinnova has with its legacy strategies. Why would this type of investment with deepc AI not be covered in an existing strategy? What needs are addressed and new value is added with a new strategy specifically for digital medicine?

[00:11:29] Simon Turner: There's multiple different kind of realities as to how digital medicine is playing out. But fundamentally, if we boil it down, what are we talking about here? These are very often deep technology companies, but that have software development needs. But also processes in place. So they move very quickly.

[00:11:46] Simon Turner: They're able to actually generate revenues relatively early on in fact. In certain companies, in fact, we've seen them be bootstrapped because they're very early generating initial commercial contracts and hence growing in that kind of sense. And they also have the potential to [00:12:00] generate revenues very quickly and rapidly as and when they scale.

[00:12:02] Simon Turner: Now take the parallel, I guess, in biotech, for example. Most biotech companies, it's more of a de-risking process, if you will. So you're going from a very early preclinical stage and eventually to clinic, and then you'll get to a point where often the exit will be, for example, MNA for pharma, and you can almost see the total trajectory.

[00:12:22] Simon Turner: It's relatively linear there. And ultimately what you're doing is building value of the asset and that core asset, it could be a clinical candidate or two or three or pipeline of them. That is what will be sold in the actual asset itself. Whereas in, in the case of many of our digital medicine companies, deepc included, for example, revenue is a core component of how these companies will be valued and assessed, but also a metric that we're using when looking at trajectories going forward to these companies.

[00:12:48] Simon Turner: The other thing is, I'd say one of the big differences between tech and of course, what we're doing, and Franz already articulated this pretty well, it's, we're dealing with patients lives. We're dealing with people, fundamentally. [00:13:00] So, the old adage of run fast, break things, and continuously iterate until it works, you can't really do that in health sciences, you must constantly be aware of if something goes wrong, that could be a big problem to the point where you might actually cost someone their life. And it's unforgivable to even consider that as a possibility. So the need to kind of integrate how we have safety, how we apply regulation, quality management systems, et cetera, is a critical piece of that kind of component to it.

[00:13:28] Simon Turner: So it's all of these factors that one, make it different, but two, also where we see the similarities and why launching such a strategy from a platform as knowledgeable, as experienced, as network as Sofinnova in the health science space, it makes total sense because we are able to be the doorway, the gatekeepers, if you will, for these new technologies coming into the health science space.

[00:13:47] Simon Turner: Plus with our 50 years of experience in it, we have the connections also to be able to help our company scale, to grow, to partner, to find the right resources to compliment in their development and hence hopefully turbocharge and [00:14:00] accelerate their growth rather than see them be relatively, let's say, flat lined in terms of what they've actually been able to achieve.

[00:14:06] Henry Peck: Franz, as you hear Simon talk about a key differentiator for these types of investments being revenue, how do you think about that? Because unlike the legacy strategies where growth or exit horizon has revenue further down the journey, you know, molecule gets FDA approval and acquired by pharma. Med device gets FDA approval and eventually reimbursement, and then scales are acquired by a strategic. You're thinking about revenue now, as a key driver of value and a metric of success for the company. Yet, as Simon said, we don't have the same ability here to move as quickly in digital medicine as we do in other industries. We're not building email automations. We're not building consumer games. And not that those things aren't important and complex, but they don't carry the same risk profile that healthcare and other heavily regulated industries carry. So, how do you think about trying to generate revenue quickly, show that traction, while also maintaining safety, particularly with a new technology like AI?

[00:14:59] Franz Pfister: Yeah, [00:15:00] that's actually a very good question. And I think you already gave yourself part of the answer. Absolutely right, it's a good balance that you need to find between execution speed and growth, but then also really maintaining quality because the stakes are high, like what we do, everything of what we do has the patient, or I even want to say the human, at its center, because it's not only patients, but also the clinicians and the physicians and the clinical staff that are using our solutions. So we want to make sure that everything works properly, is scalable, but we do everything sustainably. So I think balancing those two out is a key component of what you need to do strategically.

[00:15:42] Franz Pfister: And you can't expect to scale like a SAS business and a consumer area in six months or 12 months to 10 million subscribers. That's not really possible in healthcare. But then on the other hand, what we are building is kind of a network and an ecosystem platform. [00:16:00] One thing that's of course investors really like as a keyword, is 'flywheel effect' that you need to create as a platform. So everything that you do should increase the ecosystem value and that you're delivering and provide like additional value to all the ecosystem players. And I'm not only talking about hospitals or clinicians or eventually patients, but also the AI developers or channel partners or the IT of a hospital. So you need to take all of those into account and make them part of the equation to sustainably grow commercially.

[00:16:35] Simon Turner: And that's one of the things that we spend a lot of time discussing, as you can imagine, it's, how do you find additional pieces that you can integrate into a business plan, which acts synergistically? It's not just here's an additional component that we need to contemplate adding, but there's some strong synergy, hence, either in terms of adding users or providing more features for the actual users of the platform. So either the healthcare practitioners on the one hand, or the actual vendors [00:17:00] themselves who are using this as their go to market strategy.

[00:17:02] Simon Turner: The other thing is, as Franz pointed out, the safety component cannot be overlooked, because when you think about it here, we have standards already in place. And in digital medicine, frankly, it's early days. A while back, it was still the Wild West in terms of how do we do things and why should we do things. Now we're starting to see certain standards coming out. But with people like Franz and what deepC are doing, they're also setting new standards, and it's building those gold standards going forward that's also creating an incredible potential.

[00:17:29] Simon Turner: Because, one, you're providing your best practices as a framework for others to then quickly get to, but secondly, you're always pushing yourself to create the next standard and hence providing a way of also, one, showing that you take this extremely seriously. Secondly, also making sure that you are using your best practices. And thirdly, also creating certain defensible moats, of course. Because you as the leader in the pack, you'll constantly be going further than anyone else and hence maintaining that advantage.

[00:17:57] Henry Peck: It sounds like this idea of AI safety plays a [00:18:00] major role in how you're looking at the broader AI landscape right now, as you try and parse through all the hype and excitement of AI, you mentioned that part of the genesis of this strategy came from looking at novel technologies like AI years before this past year where AI has become the hot thing of the season. Talk to me a little bit more about how you're making sense of what you're seeing in AI today from the investor perspective and then Franz on your side, how you're thinking about what's practical and implementable that's going to drive revenue improve outcomes today versus what may be future thinking around how AI is going to impact your business.

[00:18:38] Simon Turner: I'll give the higher level and Franz, who is way more technical than I, he can probably drill down into that piece of it. But from our perspective, let's just think over the last, okay, probably now we're getting to 12 ish months, but how transformers and LLMs and now even generative AI has suddenly, totally caused this kind of wow effect of, there's an incredible power here at play, [00:19:00] we're beginning to see first use cases being developed just generally. It was amazing, I was sitting in my former university in one of the student cafes, just sitting there writing up some emails. And I looked over and I saw probably about five or ten student laptops open. And you cannot imagine how many of them had ChatGPT open. With a page or paper on one side and chat GPT on the other, and they were writing questions, taking the answers and then beginning to paste it in and then rewriting, et cetera.

[00:19:27] Simon Turner: And it's incredible to think 24 months ago, this was not on anyone's radar, really. It was very few people in the sector who even knew about this. And I'm talking to general tech, let alone health sciences. But now we've seen such a revolution happen. It's important, though, for us to consider the implications of it. One, it can facilitate an incredible amount of tasks. It could be, for example, the way that we're seeing commoditization of coding happening, or at least the speed of coding, so these types of different factors.

[00:19:52] Simon Turner: And that's a huge benefit. But then we start imagining, how can we start seeing this in routine implementation, if you will, in the healthcare system? And again, [00:20:00] incredible potentials here. But what happens if a junior doctor, for example, asks a health focused transformer or LM, how much, I don't know, norepinephrine should I be prescribing for this patient?

[00:20:14] Simon Turner: And just there and then, hallucination takes place, and suddenly, oh shit, I've killed the patient. These are the types of safety concerns that we really have and because ultimately, again, we're dealing with people with patients and impacts that can be quite truly catastrophic here. So we need to take that seriously. And that's one of the pieces that we're looking at quite a lot. It's thinking, how can we develop these approaches with best practices to avoid the major risks that we could otherwise accept in certain other industries? Because again, we're dealing with patient lives.

[00:20:43] Simon Turner: The other thing is taking a higher step back when we look at how kind of commoditization of coding is happening, the way the kind of speed up of coding tasks are being developed now, we also have a couple of just fundamental pieces of the puzzle when you're managing and developing an AI solution in healthcare which still remain. [00:21:00]

[00:21:00] Simon Turner: One is data collection and data quality, because very often you still need relatively bespoke training sets of data, or at least very high quality data sets to be able to train your models on, and even validate them then later on. And the second is again, this AI safety, this regulatory, this quality management system component, which is still one of the major hurdles that companies need to overcome. So we're spending a lot of time looking at one, how are new technologies dealing with these pieces of the puzzle?

[00:21:25] Simon Turner: And secondly, if there are technologies or approaches that are actually solving those pieces and could be levered as services, hence to the rest of it. So again, many of these enabling technologies, but then also that will translate into analytics and ultimately the treatment side of our digital medicine strategy.

[00:21:41] Franz Pfister: So I can give my perspective on things. So how do we see things? I believe it's very important how you approach things at this point in time, because it's still like a very early stage of this whole industry. For us, it has been proven to be very important to be science led [00:22:00] in AI. So we are guided by rigorous scientific principles propelling us to the forefront of AI innovation and assets to build trust with all stakeholders, but also to contribute to the academic innovation and evolution of the space.

[00:22:16] Franz Pfister: Another thing is that we have early on realized that it's very important to build collaboration networks and an enablement layer as a platform. So we lead the charge in collaborative innovation, reshaping that radiology landscape, not only through cutting edge technology, but at the core of it, it's people. And we are enabling those people as customers, as partners, but it's important to take everybody on the journey.

[00:22:44] Franz Pfister: Then the third pillar of how we do things is to set new standards, as Simon said before. We are leading that innovation and we have to redefine industry standards and how things work. Like three years ago, it was [00:23:00] impossible to deploy cloud systems, quite frankly. Now everybody is taking that route that wants to really scale things up. That's a good example of how you can make that adoption journey seamless, transformative, but then also safe for our customers and partners.

[00:23:17] Franz Pfister: And being in that position of setting standards, how things will be done in the future, it's very important for us to be at the forefront of that AI adoption journey. And then last but not least, and that comes back to the commercialization component, although it's very early days, we need to earn money of course, as we are a platform company that is driven by commercial metrics.

[00:23:45] Franz Pfister: And how do we do that? We simply need to maximize the return of investment for our customers through operational empowerment and excellence already today. And we do that on the one hand, of course, through our platform because it saves them a [00:24:00] lot of resources and eventually money instead of doing single vendor deployments, which involves like a procurement process each and every time, but then with the AI solutions themselves, we are enabling the customers to save costs, maybe enhanced operational efficiency, elevate diagnostic accuracy, et cetera.

[00:24:20] Franz Pfister: So those four components, I think, are very important being science led, being collaboration centric, setting new standards to really scale up innovation, and then last but not least, focus on ROI for customers. And we see already like a lot of positive examples today, like in the field of cancer screening, for example, and mammography AI domain, where AI is not only speeding up the whole process of reading mammograms, which is resulting in a direct efficiency gain in a direct ROI, but also at the same time, bringing the quality of care and diagnostics to a completely new level and setting a new [00:25:00] standard.

[00:25:00] Franz Pfister: And you see similar pathways in CT stroke or other critical care pathways, proxy detection, et cetera. And those are only a few examples. So already today you can create this ROI, but in the future, we believe that we are awaiting like a whole new generation of AI models. Like now we are riding the wave of foundational models and LLMs and all the new technologies that Simon already mentioned. And we believe in three, four, five years, everything that we see today in the market will be replaced by even better technology.

[00:25:36] Franz Pfister: So therefore, it's important for us to be that channel to the market, which can then accelerate also those new technologies getting into the market. Having said that, on this way from now until in five years, we still have a lot of homework to do. I think maintaining safety of AI systems is extremely important. And again, [00:26:00] we as a platform company are enabling our partners, maintaining safety of the models, making sure that they have systems in play for post market surveillance, for bias assessment, for fairness assessment of those AI systems, not only during the development of the systems, but then also on the long run in life deployment, making sure that accuracy and performance is high and the value delivery as always.

[00:26:26] Henry Peck: Awesome. Thank you. Simon, we went deep there into how AI needs to be built responsibly and how companies like deepc AI are de-risking their business models now to be durable and insulated for the future of AI. Something I want to make sure we cover in this episode is the potential for AI and radiology. You talked about radiology being a major area and opportunity for AI to make an impact today. In your view, why is radiology such a ripe opportunity for AI? And do you see what's currently happening in radiology with AI scaling to other market segments?

[00:26:56] Simon Turner: So, Franz, I'm counting on you to grade my answer here. You'll let me know afterwards how I do. [00:27:00] It's a good question, Henry. To your point, we're seeing innovation happening and AI adoption beginning to take place across the healthcare, the health science space, in fact. We've seen recently how pharma is saying it's going all in for AI. We've seen it in the way that hospitals are now beginning to use workflow systems to improve the way that patients are handled and treated. It's everywhere. It's pretty ubiquitous.

[00:27:22] Simon Turner: But that's just as a initial phases of this, some areas more mature. Radiology, one of the core kind of premises there, it's a bit, if you will, the segment itself is relatively digital already, in that you've got certain components which were critical in the making of it being an interesting first area to see this development of AI in fact taking off.

[00:27:46] Simon Turner: And if you will, it's one, the systems and the approaches, they've been digitized for quite a number of years already. So when you go and get a radiology scan now, it's extremely likely it's going to be in a digital format and a standard digital format called a DICOM. [00:28:00] So again, you're not dealing with apples and oranges and pears and pomegranates and things like that in terms of trying to assimilate data.

[00:28:07] Simon Turner: It's not to say there is no complexity there, because again, a GE machine might not produce the same as one that's coming from Canon, for example. So you still have variances. But at least a lot of it has been standardized, and hence the kind of heterogeneity has been brought down substantially. The second is, we've seen radiology as well just growing in general as a field, massively, because demand for radiology is increasing year on year.

[00:28:30] Simon Turner: However, it takes a long time to actually get a radiologist to be at the forefront, at the stages where they're able to do extremely effective diagnostics. So, there's a supply demand mismatch ultimately also happening. And the fact that when you're looking at the kind of expected return on investment of using these approaches, it's pretty quick to actually see the positive calculations that are being applied into the radiology field.

[00:28:54] Simon Turner: So there's these several components, which have come together to make it one of the very interesting spaces. [00:29:00] Now, when we take a step back, though, if you will, why the entire healthcare segment is ready for this, is ready for digital medicine and also a little bit of a thesis as to why we thought now is the time for this digital medicine strategy. It's, on the one hand, we have healthcare costs. They continue to grow year on year. We're expending an incredible amount of money on healthcare itself. And unless we change something, it's eventually going to be causing a massive issue. And we see so already happening in certain countries where the healthcare expenditure is just too high in fact.

[00:29:30] Simon Turner: But on the other hand, there's also an issue with the supply demand mismatch taking place. We have aging populations, rising rates of chronic diseases. The supply of actual healthcare practitioners and expertise is, to a certain degree, stagnating, or in fact, in certain cases where we're seeing extreme burnout, dropping.

[00:29:47] Simon Turner: So, unless we do something to basically, one, unburden the system from a healthcare cost perspective, and two, unburden the system from just a sheer pressure perspective, we are set to have a situation where [00:30:00] we need to almost choose what type of health care can we provide and sacrifices will need to be made.

[00:30:05] Simon Turner: Whereas up until today, we've been continuously improving our health care provisions and our abilities to actually treat the population. So it feels like we're at that very critical point now, an inflection point, of either we do this and we do this well, and we continue to see great improvements in healthcare and healthcare outcomes. Or in fact, we might actually start breaking things which fundamentally we can't afford to break. I know Franz, what's your take? This is for me, the kind of fundamentals of it all.

[00:30:29] Franz Pfister: A plus.

[00:30:31] Simon Turner: Yes.

[00:30:34] Henry Peck: It's a glowing review.

[00:30:37] Franz Pfister: Maybe just to add why this whole journey started in radiology. I think there is also like a technical foundation that we have in radiology. As many of the healthcare data pieces are still handwritten today, and many of the information pieces are not stored in a structured digital form, we have found like a [00:31:00] different foundation in radiology where digitization has already been practiced for 20, 30 years. And due to the fact that also the reporting takes place in a digital form, we also have labels for the data.

[00:31:16] Franz Pfister: So that was actually a great starting point for this whole AI innovation being pioneered in the radiology domain because the data already being available in digital form, but also the labels. Now with foundation models, of course, we can leverage also data that is not particularly labeled by physicians, but that's a different story.

[00:31:39] Henry Peck: And that's what we'll talk about next time, when you've hit all those revenue inflection points that Simon mentions, and the next generation of models are here. And we'll use that to segue us to our last topic. Simon, you're joining us at LSI Europe 23 in Barcelona, Spain, moderating a couple panels on digital medicine. And one that I want to touch on is the one you're hosting with innovators building AI enabled devices in the med tech space. What do the [00:32:00] innovators joining that are looking to integrate AI into devices with a hardware component need to know? And what are the investors that are attending need to know about that space more broadly?

[00:32:09] Simon Turner: This is a segment that we've seen brewing for a number of years. In fact, probably, let's say five or 10 in terms of new sources of data being integrated into these new medical devices that are almost therapeutic in many aspects. So what you're able to do is suddenly insert, provide a device, it could be a smart set, for example, a neurovascular guidewire.

[00:32:28] Simon Turner: It could basically be anything, if you will. And suddenly begin creating a source of novel information that we've never imagined previously. In certain cases, it could be continuous. What that's enabling us to do is to see how efficiently, effectively is this device actually functioning on the one hand. But secondly, also begin creating an incredible amount of information to now start model these patients. So ultimately, over time, what can we do? We can start building individualistic patient profiles to understand, what is the best next steps in a patient's treatment requirements? But if we take a step back even [00:33:00] further, eventually we can start assimilating enough information to start building in silico simulations of patients, or even these kind of synthetic cohorts to then start trying different approaches on to eventually start iterating in silico and modeling and seeing what will happen if we assume a different type of treatment course, how will the patient most likely respond?

[00:33:19] Simon Turner: And ultimately we see this incredible new next step of not only having a very therapeutic response at T0, but then also beginning to plan the next stages of therapy and hence improve outcomes of these patients as they go on, as they continue, basically, with these devices and these technologies.

[00:33:35] Simon Turner: So, ultimately, on this panel, we'll be delving into a couple of other pieces in depth, but I'll leave that until we get to the panel itself. But think of this as critical in being able to generate and ultimately then model how patients will react, respond, and hence how we can really deliver the best value, the best bang for our buck, if we will, for the healthcare systems as well.

[00:33:57] Simon Turner: And ultimately, what am I looking forward to? Just like the last time around when [00:34:00] I was over in Dana Point, excellent panels and discussions, but ultimately also being challenged quite frankly, that's the part I really like. Because let's face it, we all have our preconceived ideas and things. Here you've got some of the best and brightest in the industries coming together, and suddenly you've got totally different points of view coming together. You need to be ready to defend yourself, but also then to say, Oh, that's a really interesting perspective. How do we integrate that with everything we know and hence also continue preparing for this future?

[00:34:25] Simon Turner: Because something we love is being a bit thesis led. Again, it's what led us to deepc, it's what's led us to a couple of other investments in our portfolio. So ultimately, it's how we'll continue growing that portfolio, thinking, what are the next critical steps we need to think about so that we can invest there, prepare it for the rest of the healthcare system, and hence also make it something that's an interesting return on investment for ourselves, of course.

[00:34:46] Henry Peck: Well, Simon, we're looking forward to seeing you and thank you so much for joining us on this episode. Franz, great to have you on and looking forward to seeing the next steps for deepc AI and for Sofinnova's digital medicine strategy. Thank you so much.

[00:34:57] Simon Turner: Thanks very much, Henry.

[00:34:59] Franz Pfister: Thank you so much.[00:35:00]

[00:35:00] Simon Turner: Always a pleasure, Franz.

[00:35:01] Franz Pfister: Likewise. Thanks.

[00:35:04] Henry Peck: Thanks for tuning in to Emerging Medtech Today by LSI. Be sure to subscribe on your favorite podcast player so you're automatically notified when there's a new episode. For more about LSI and the Emerging Medtech Today podcast, and to continue exploring our suite of videos, interviews, and other resources, visit emergingmedtechtoday.com and find the link in the show notes.