Episode Transcript

This podcast has been paid for by Olympus Corporation of the Americas. The views and experiences shared are those of Neviana Terzieva and Slawek Kierner, Olympus employees. Olympus makes no representations regarding the accuracy or applicability of the content and disclaims all liability arising from its use. Product performance may vary and techniques, instruments and clinical decisions are unique to each facility and practitioner. Always refer to the instructions for use and applicable labeling for guidance, risks and cautions. Products may not be available for sale in all regions. Hi everyone, this is Tom Salemi of DeviceTalks. Welcome back to the OlympusTalks podcast. We’re going to be talking about AI today and its impact on medical devices. We have two experts from Olympus who will be sharing their story and helping us understand how Olympus is incorporating AI into its products. We’ll be talking with Slawek Kierner, he’s the Chief Digital Officer at Olympus, recently joined. We’ll find about find out about his path to Olympus. And we also have Nev Terzieva. She is the Director of Marketing of GI Digital Solutions at Olympus. Welcome to OlympusTalks, the podcast that brings you to the forefront of medical technologies as we explore advancements in innovations in GI. This eight episode series features talks with healthcare professionals, patients and Olympus subject matter experts. Listen as they dive into various aspects of GI health focused on improving patient outcomes through best practices. Stay tuned for conversations designed to educate, inspire and inform. Welcome both to the OlympusTalks podcast. Welcome Tom. Great to have you both here. Great to understand looking forward to understand how AI fits into medical devices and we’re going to unpack the the definition of AI as well. But first I’d love to understand how you found your way into the MedTech industry and to Olympus. Nev, could you kick us off? How did you find your way into our great industry? Thank you. Great to be here. Actually this is my first time being part of a MedTech establishment and my professional life has been entirely in software and intersection of software in the healthcare space in the US so I’ve been able to see the evolution of healthcare technology and how it has the opportunity to change both outcomes and also the experiences of all healthcare workers evolved. So obviously these days technology is focused on AI. In particular, software has migrated away from just being digital. We’re trying to exponentially increase our effectiveness and have AI be infused in our work days wherever possible. So really excited for the opportunity to fuse software and MedTech devices and and have the really the opportunity to see a fuller cycle of the clinical care. That’s great. Does MedTech feel a bit differently than different than other industries in terms of how it integrates and works with software. Is there a greater resistance here than maybe in other industries? Because I think we see ourselves as slow adopters of new tech. How are you viewing things from your perspective? Definitely more touch points. Typically, software industry partners and companies are somewhat isolated from the direct touch to the patient and then direct interaction between the MedTech device, the physician and the patient. In this particular case, the continuum of care or the episode that we have the opportunity to change is includes so many more touch points from the physician, the patient, the software, the device itself, the use of the device and of course the cloud is a consideration these days. So it’s a much broader spectrum of what I’m used to and I’m excited about it. That’s great, great insight. Slawek, how did you find yourself to Olympus? As I mentioned at the top, you’re fairly new to the company. Give us a little rundown of your path here. Thank you. Thank you, Tom. And first of all, I really want to thank you and Nev for bringing me on the show and I truly appreciate this opportunity to unpack what AI can mean for the healthcare that our company provides. Now, as you said, my name is Slawek Kierner. I’m a new Chief Digital Officer at Olympus and my path has been shaped here by working with the largest data and AI platforms in the world. I started by leading data and AI within product development at Microsoft, helping build and scale a product which hopefully many of you use, which is called Power bi. That then turned into a large portfolio of products which today we call Microsoft Fabric. From there I joined Humana. Humana is one of the larger healthcare organizations here in the US obviously also a large payer. And over there I helped move that company into the cloud and among many things, we created, I think the first healthcare large scale AI platform which we called Florence AI that hosted hundreds of models that helped in our operations and in population health. And then from there I joined Intuitive where I spent several years developing large part of Intuitive digital product stack that included our simulator products, the Intuitive hub, the AI suite that I helped launch alongside the new Da Vinci 5 platform which is now scaling around the world. And that experience helped me understand what does that mean to bring AI to bring technology to the hands of surgeons. And I hope to apply all of this right now at Olympus with an immense scale. All right, wonderful. Well, that’s a lot of information, a lot of experience we can draw from for the sake of this conversation. I mean AI is this all encompassing term that means a lot of things to a lot of different people. Slawek, maybe you can kick it off and then Nev bounce back to you. Let’s define AI for the purposes of this conversation and help me understand what are some of the principal problems in MedTech that it can help resolve. Yeah, thank you, Tom. First, I think at the bottom of it, AI is math. So I started this 35 years ago. People shouldn’t be scared if you know math, you know AI in many ways. AI is not that complex as people think. But obviously now we are applying it with MAP in a very sophisticated way and there is a lot of applicability of this to healthcare there. I would say that when we think about the application, it is going across all of a quintuple aim. So it is a technology that can help improve patient outcomes, that can certainly improve healthcare costs, it can improve experiences for patients, for providers, for healthcare professionals, and it can help us increase access to healthcare, which obviously all of us are suffering from. I hope we’ll be having time today to unpack how AI can actually help in each of these domains. That sounds great. It’s a great start. What are some of the problems that you’re seeing that AI can help solve in MedTech? I see great potential in reaching something that I call a blended intelligence. There are naturally things in healthcare that a machine will never be able to replace the humans for. Starting from the patient touch to really having a more contextual ability to care for the patient. Now what a machine can do for us, regardless of whether it’s machine learning algorithm or generative AI, they can complement the human intelligence and be able to assist in the edges of decision making and serve up information for the healthcare professional to confirm. I really see great potential there. We all know that healthcare workers experience a tremendous amount of burnout right now. There’s so many more patients entering the healthcare system and the regulation that all healthcare workers have to deal with outside of just the core care of the patient is very burdensome to them. So I, I see that potential of alleviating some of the more mundane automatic tasks, of course, and then in addition alerting healthcare professionals to perhaps areas of opportunity, examination, diagnosis that they otherwise could have missed. Yeah, I think if I can, if I can add, I like you how you framed it into this blended intelligence and I think it is important for a long time we also have been using this term augmented rather than artificial or as I was saying, math. I think what we have in front of us is obviously a huge opportunity and a large adoption problem where some people might be a bit afraid, especially when you apply something that might be feeling as artificial into contact with a human body into healthcare. And we shouldn’t be scared. That’s where I think this blended augmented, like really explaining what it is and in the concepts of this, it doesn’t have to be in any way frightening, but to the contrary, it can actually help the surgeon. I think patients are certainly, well, I don’t know if they’re open to it or not. As someone who’s using AI a little bit and is also getting frustrated with interaction with my healthcare providers, I’m looking forward to some sort of tool that can help build that bridge and make it easier for them to talk to me. So I’m not taking too much of their time and, and I’m getting the information I need. How do you see the reception in the medical device industry more broadly are engineers and others who are in the business of building metal devices. I imagine they see the opportunity to incorporate AI into device development as well, Slawek. Yeah. I think there is obviously a lot of interest, a lot of openness to adopt AI in forms which can help in the product development process. When we think about writing code, writing test scripts, many of these things have been in the past requiring many more software developers and right now a lot of this can be automated with tools that come right part of your IDE and help obviously not only faster but also better code and better documented code which in the end also turns it into, into code and in the end medical products which are safer and more controlled. So I think there is a lot of openness and while obviously we need to proceed with caution because we are in a regulated industry, so we are looking at all of these technologies with appropriate care. But I can clearly tell that I think the team, the talented team that we have is very open to use AI also within the product development process. And I haven’t touched yet the whole topic of QMS, which again means documentation that medical device company needs to create to have this controlled development process. And that’s also a lot of work. That’s why it takes three, four, five times longer to develop a medical device versus similarly complex piece of technology which is not regulated. And that’s where we also see, I think, opportunities to again bring AI to improve the quality, to improve controllability, but then also remove some of the busy work which lengthens this process unnecessarily and then through this lower the cost of devices and make them more quickly available to to improve this innovation cycle. And Nev, how are you feeling? What kind of response are you seeing from your colleagues in your new industry? In MedTech, what’s the attitude? Maybe compared to other industries, are you seeing a receptivity to warm reception to the possibility of AI? I think the reception is definitely warm. There is a little more trepidation about the actual introduction and duly so. Slawek said. This is a very regulated industry compared to software. MedTech carries a lot more burden of proof and for obvious reasons in the software world most solutions are considered just reference points. When it comes to surgery and actual interventions, the burden of quality is much, much higher. I do like the fact that people are more measured about introduction because in my history of software and technology introduction into the healthcare, sometimes there’s a lot of excitement. But the use case is not that great. And physicians and other health care workers have the opportunity to explore technologies that sometimes don’t lead to much benefit at the end. Unfortunately, of course that’s the cost of innovation that we carry in any industry. For me, having a measured approach and being more conservative in the use cases in which a technology can be applied at the end is to the benefit of everybody. Yeah. And just to maybe build on, I think, what Nev said, I think, you know, there is this kind of large amount of regulation which clearly is applicable to those pieces of development which we do, which becomes a medical device, and obviously creates all kinds of risks if it might malfunction. And that’s where we need to be much more measured and much more kind of described in our approach and make sure that everything is working well. But then when you really think about the problem that we are trying to solve from providing better care to our patients and improving outcome, there are a number of pieces of that process which are not necessarily a medical device, but which look way more similar to health software that EPIC or Cermer or other vendors provide, which enjoy faster innovation cycles. And obviously probably this is where we should start applying AI in these pieces of development process. So there’s a little bit of. I really like to think about this as a continuum and then different types of controls need to be applied also in application of AI, depending on how much risk certain part of a product provides. Let’s talk about some of the specific technologies you’ve got at Olympus, the extended depth of field tech. We can talk about OLYSENSE™, talk about some practical and direct applications that you’re working on at Olympus where AI is helping to enhance the performance of the devices. Yeah, perfect. So Let me pick a few. So obviously we are, we are building the core technology and the core new medical devices like the new scopes, for example, this EDOF scope, which improves clarity and precision and it really helps physicians stay in focus and control and kind of see the full field of view. Then you have our CADDIE™ device and I think Nev will talk more about this. But that is a very specific application of AI to help help detect polyps. So the CADDIE™ product that we have launched right now is part of the OLYSENSE™ that’s probably the most sophisticated and direct use of AI directly wearing the contact with the procedure of the patient. Nev, let’s talk a bit about the polyp detection aspect of it. How does AI, I mean, this is an area where Olympus is obviously a leader in a lot of work. A lot of physicians are using your tech to help folks identify things early. How does AI help that process, help them help the physicians do their job and possibly improve diagnoses or the success rates of diagnoses. Sure. Well, we are very excited about the opportunity to impact colorectal cancer screening in particular. And I’ll spend a couple of minutes describing the context here. But in the US recently there have been a lot more patients entering the colorectal cancer screening funnel, if you would, for physicians. The clinical societies just changed the age requirement for your first colonoscopy. So I’ll use this opportunity as a public service announcement. If you’re over the age of 45, please schedule your colonoscopy. The reason for this adjustment is truly because we’re seeing a rise in, in colorectal cancer incidents, particularly in younger patients. And that’s very concerning. So you have a higher prevalence of a disease conjoined with the fact that the healthcare system is already overwhelmed and physicians are already overwhelmed. So this is a perfect opportunity for technology to assist. And the CADDIE™ device that Slawek mentioned is a machine learning algorithm that, that in real time helps endoscopists detect colorectal polyps that could lead to cancer. With all the technology that Olympus and other vendors are providing globally for the colorectal cancer screening initiative, still the majority of the detection happens between two pair of eyes and a monitor that’s an endoscopist watching all day long, of course, with tremendous help with the visualization products that we have and others to make sure that the patient is not at risk for developing colorectal cancer too early or at all. So our product has the ability to serve up to the physician areas where they may not have seen a polyp in presence. And again it goes back to that idea of blended intelligence. We don’t want physicians to over rely on any technology whatsoever. We would like them to still be the main, if you want the main conduit for healthcare decisions and diagnosis in this setting, however, technology has the opportunity to augment them. And if most of us these days have had already interaction with AI of some kind and we’ve all realized that if you, you ask a nonsensical question or an incomplete question, you’re going to get a nonsensical answer and an incomplete answer. So what we’ve done differently for our AI algorithm is to really curate a data set. And a data set for a machine learning algorithm is just as important as the question that you’re asking. For generative AI tools that, that we are exposed to. We’ve taken care to, to train this machine learning algorithm we with incidents of polyps that are harder to find by physicians. And the theory here is that by exposing the algorithm to the visualization, visual aspects of these harder to find polyps, then the algorithm is more capable of assisting physicians to find those particular polyps as well. So obviously the verification data set needs to, to match prevalence, et cetera. There are technical, you know, specifications there we can go into. But all of it to say is our focus in introducing AI in this particular case has to do with again, not just providing a general tool, but making sure the tool is focused in an area where endoscopists and healthcare workers are particularly struggling with. So Nev, I asked you earlier about how the medical device industry feels about AI and I talked about from a patient perspective. Patient perspective. I would have to think that physicians do appreciate the backup. Do you have any sense of how the clinical community feels about these new tools coming in? I imagine they’re welcoming of the help. For the most part. I think physicians are duly protective of any technology that enters their direct patient care space. So there’s certain sentiment that, that the technology needs to of course prove itself over time and it needs to be non-intrusive. That’s something that they’re highly sensitive to and we absolutely understand that. Of course some physicians feel that since they’ve been practicing all these years without this technology and their outcomes are already great, what is the additional benefit really of being able to introduce yet another, another device, another software application, another potential delay in the patient care. And we recognize the sentiment as well. There are physicians that are of course absolutely excellent at the level of patient care that they’re delivering at the same time, I’ll go back to the clinical gastroenterology clinical societies again and mention that they recently raised the number one clinical quality metric for colorectal cancer prevention, which is a denominator detection rate. They raised the benchmark from 25% to 35%. If you look at that, it tells you that there’s still room for improvement. Nationally there’s many excellent physicians. However, at the same time we’re combating a disease that is growing in incidence and societies are recognizing that and one would hope that any kind of improvement is considered and given due consideration. Now let’s just kind of move to the look ahead. We’re standing at a really fascinating time for MedTech with the inclusion of AI, Slawek. I’d love to hear from you and then we can get your take as well. What are the future opportunities for Olympus? What does AI do for Olympus’s strategic direction? You’ve got such a broad footprint touching a lot of different areas. What do you see as the opportunities for the future? Yeah, thank you Tom. It’s a great question and I hope like I said earlier, I think AI can be applicable to all parts of our quintuple aim. So obviously starting with better outcomes which were really well described by navigation. But now beyond that when we look into this physician burnout, which again like when you bring AI to augment the procedure, which we can obviously do now, that’s already the first step. But when you look at the future direction we certainly can help create clinical reports and notes more effectively. So here the technologies that we all started to use, including class language models or or transformer architectures, voice related experiences can help elevate these pain points. And beyond that, I think the bigger issue in healthcare ecosystems has always been in interoperability. How do we bring all of these products into hospitals, how do we integrate them with with EMRs? And only this tight and good integration can really be meaningful in helping elevate this burdens. So here again I think we have this huge opportunity at Olympus because of this footprint that we have, but also the fact that we have actually worked with hospitals on the digital side for a very long time in our GI side, in our surgical side of business and have our anchor products and VaultStream integrated with many EMRs opens with amazing opportunity to bring these new tools and also have them right away faster integrated and then have experience flow to really help surgeons and GI physicians in elevating the unneeded unnecessary work. But there are opportunities beyond that. Again when you are integrated in the hospital we obviously have our processing machines which are helping run the GI suites. We also have tremendous experience in endoscopy. I’d say I’m just getting excited here because when I entered Olympus I had seen so much knowledge of how to do it effectively. Let’s just talk a bit about how you’re delivering AI. I know Olympus has a big presence on the cloud. How does that give you advantages in working with physicians and getting them the solutions that they need? From my experience in software, I firsthand saw how having the ability to deliver software over the cloud enabled both industry partners and also healthcare organizations to realize instant efficiencies. For those of us that have been around long enough to remember that all applications were entirely on prem server based and every healthcare institutions had to invest tens of thousands of dollars of licenses for the software and also compete to get really good internal IT resources to maintain these huge healthcare systems. Once the cloud became the go to solution for software outside of healthcare, it did take a little longer for healthcare to be open to the idea. But the advancements of healthcare security and also the fact that we have reliable cloud delivery partners in the US and globally I think has, has brought great opportunity. When I first started working in healthcare it it was actually surprising to me that healthcare was a little bit behind, let’s say the financial industry when it comes to technology introduction and efficiency. So we are excited about the opportunity to deliver a system over the cloud for the sake of these advantages to us and to our customers. Whenever our technology is upgraded, it can reach our customers instantly without us having to deploy folks to each ASU or each hospital in the room. And the hospital administration themselves can focus more on the actual healthcare and less on the IT burden and cost that they used to have to spend spent. Now of course every cloud software provider has to have the proper certification and security safeguards which we are confident we can have. Yeah, and I absolutely agree and we are taking security very, very seriously and current execution frankly doesn’t even store any PHI data. And obviously like you heard from my past, I’m a big proponent of the cloud. The beginning of this journey at Microsoft and helping them move Humana into the cloud. And that was a time back then, roughly five, seven years ago when it was a new topic and it was very hard to convince people. And I think hospitals were at the kind of last frontier here because obviously their mission is so critical and you want to maintain the security and large hospital networks in the US still required a lot of the products to be very independent and run on the edge within their own walls. But that has changed. And I think Nev to your point, all of this benefits of not needing to have a software license, not needing to patch your servers and to maintain large IT teams have enabled rethinking and the new kind of security architecture where you are fine with running your workloads in the cloud. But now I’m kind of turning a little bit back and that’s how I think technology history is kind of moving. Which is like if you really want to be an integrated platform that enables many different workloads, you honestly need to have almost like a tribute architecture. So you need to have your cloud component that allows you to move fast to leverage the newest capabilities that are driven by the cloud service providers. And we are there and we’re actually starting from there right now in this current version of our software. But then as we do this we also bring our OLYSENSE™ hub that has computing capability within that device. So our OLYSENSE™ hub has GPUs that enable also running AI on that hub. And as we kind of look into the future, we have that opportunity and we’ll be unlocking it as we need. And finally, when someone wants to be well integrated with hospital systems, EMRs which in large majority right now are on prem data center based in the hospital, you also want to have your footprint there because that allows you to listen to fire messages, that allows you to really integrate to to that patient workplace which you then want to light up on your eventually within your experiences. So again we are lucky enough that we have the whole stream service that are deployed and integrated with EMR in, I think, more than 1,000 hospitals, large institutions here in the US and we can tap to that. And then the innovation I think comes now from how can you optimally integrate between that on prem data center, Compute we have the edge. Compute which is right next to a patient bed but not necessarily always connected to power and then the cloud, but it’s always powered and so much more powerful and scalable and secure to your point. And that’s I think where the future will take us. The last word to you. I think you joined our industry at the right time. I feel like there’s a lot of good news coming down. We’re lucky to be in a position to have such opportunity for impact. It is rare when a single company has such a wide reach and a deliberative introduction of technology can impact healthcare workers and patients so quickly. Just on a personal note, I feel like I’m hopeful that the new generation of physicians won’t be so hesitant really to enter the space and in the taking care of patient daily life. As we’ve recently seen, being a physician used to be about taking care of the patient. And then over the last couple of decades, I’ve heard a lot of providers saying just the paperwork and the administrative work and everything else that comes with drains us. And it really takes away from that the reason why I became a doctor. And I’m hoping that with this new wave of technology for healthcare, we can have some folks that are reinvigorated for the core mission of medicine, why they became physicians to begin with, and nurses and everyone else that touches the patient and allow them to connect to that again. Just like any new technology, takes a while for it to settle. I remember when we first started digitizing medical records, there were some physicians that said, I would rather retire than use an EMR. I was just going to dictate for another two years and I’m going to retire. So every technology carries, of course, its risk and its impacts in different areas of health care. But I’m very excited of what we’re the precipice of, not only in healthcare, but in general, what AI can bring to us in our professional and personal lives. I’m a techno optimist for life. That’s fantastic. No, it’s definitely great way to end this conversation with a lot of. We’re talking about all the great technologies, all the great tools that can come, but again, it comes down to the physicians and the patients. And I think that’s a great way to. And a grounding way to end this conversation. Thank you both for sharing your thoughts and experiences on where we’re headed. Thank you, Tom. Pleasure to be here. Absolutely. All right, well, that is a wrap. Thanks so much for joining us on this episode of the OlympusTalks podcast. Thanks, of course, to our corporate partner, Olympus for working with us and making this great podcast series. If you’d like to find the other episodes of OlympusTalks, please visit www.olympusamerica.com/podcasts. Of course, we’d also love you to follow our DeviceTalks Podcast Network, you know, so you don’t miss a future episode of OlympusTalks. Also, connect with DeviceTalks on LinkedIn. Connect with myself, Tom Salemi. I’m editorial director and our managing editor, Kayleen Brown. So we’d love to be part of your future MedTech conversations and help you find additional episodes of OlympusTalks. Well, that’s it, folks. Thanks again for joining us on this episode of the OlympusTalks podcast. The CADDIE™ device is not intended to replace a full patient evaluation. Nor is it intended to be relied upon to make a primary interpretation of endoscopic procedures, medical diagnosis or recommendations of treatment/course of action for patients. The CADDIE™ computer-assisted detection device is limited for use with standard white light endoscopy imaging only.