Kathleen Wessel (Host): Advancements in AI are leading to incredible strides with groundbreaking technologies in medical imaging. AI brings with it a whole new set of complexities that healthcare leaders need to unravel and understand the true potential of AI in medical imaging. A

hello and welcome to a HA Associates Bringing Value, a podcast from the American Hospital Association. In this series of podcasts, we speak with a HA associate program, business partners check in on their efforts and learn how they support a HA hospital and health system members. I'm Kathleen Wessel, vice President of Business Management and Operations at the A HA, and today I'm joined by the North American CEOs from two of United Imaging companies.

Jeffrey Bundy, PhD, chief Executive Officer at United Imaging, and Terence Chen, PhD, chief Executive Officer at United Imaging Intelligence. Jeffrey and Terrance, welcome to the podcast.

Jeffrey Bundy: Yeah, thanks for having us, Kathleen. Excited to talk about this really, uh, fast moving topic.

Terrence Chen: Thank you, Katherine. Yeah, really excited to be here.

Kathleen Wessel (Host): Thank you so much for joining me. Let's just dive right in. One of the hottest topics in healthcare innovation is ai, especially when it comes to medical imaging, rapidly bringing innovative patient care advancements to our members. Most of us are familiar with ai, whether it be chat, GPT, co-pilot, some of the other various tools that are out there.

Can you tell our members what is ai? When is it part of the medical solution?

Jeffrey Bundy: Yeah, great question. It's pretty much everywhere, just like we see it in our, in our private lives. Wasn't too many years ago where nobody talked about this thing called chat, GPT, and now everybody talks about it. And similarly for us, AI touches the entire patient journey from the moment a patient walks into an imaging suite, through the scanning process, through reconstruction and uh, visualization interpretation, we're touching pretty much that entire process.

Terrance's group has made and developed a 3D real-time camera, for example, That is in all of our imaging suites. And what that does is it tracks the patient's as soon as they're in the room or on the table and does automatic positioning for the technologists to, to speed up that part of the process, prove reproducibility.

So pretty much their AI tools from that positioning part of it, speeding up acquisition times, improving image quality, and even quantifying physiological parameters. Our business model here at United Imaging allows us to roll those things out very quickly to customers, which I can talk about a little bit later.

And then the great thing for customers is Terrance's team doesn't stop.

Terrence Chen: Yeah, thanks Jeffrey. As Jeffrey mentioned, a NL plays a role across the entire patient journey, supporting physicians, technicians, and all of us as patient's in helping us to make more informed different decisions and enhancing the overall quality of care. So, it is not just about conversational chat bots. AI is fundamentally transforming workflow efficiency, diagnostic accuracy, and clinical outcomes.

In addition to what Jeffrey mentioned about using a camera system to improve scanning and interventional workflows, AI are greatly used in many healthcare solutions. For example, AI can quickly and accurately analyze medical images to improve clinicians' reading efficiency and accuracy. AI system can analyze vast amount of patient data to help doctors in diagnosing disease and even risk prediction.

And based on patient's genomic data, lifestyle, patient and family history, AI can help doctors develop personalized treatment plans. It is also increasingly used in interventions or robotic surgeries for guidance and safety. AI can also help health management and monitoring, for example, through wearable devices. So one thing we know for sure, AI will continue to grow and be deeply embedded across more and more medical solutions.

Kathleen Wessel (Host): You know, you've hinted a little bit at some of your background. Can you share with members a little bit about your professional journeys And what brought you to United?

Jeffrey Bundy: Sure. I think we kind of both came at around the same time. My background is I'm a proud, uh, Purdue Boilermaker engineer undergrad. I went to grad school at Vanderbilt and actually studied in imaging and quantification and analysis of cardiac mechanics using MRI. So really right in the center of this area, AI was not one of the tools I used at the time.

I joined one of the other industry players about 30 years ago. I led a couple of their businesses for them over the years, And then saw an opportunity to be a part of a company that's really trying to change healthcare, do things different ways from a technology standpoint and a business model standpoint.

And I'm just excited to be here, to be able to bring great technology to every zip code across North America and to help other's bring it to other locations around the world.

Terrence Chen: For me, uh, my journey into the world of AI in healthcare truly began with my roots. I earned my PhD in computer science from UIUC, specialized in computer vision. While my passion for technology led me to ai, my family background always pulled me towards its real-world applications. 'cause my father was a surgeon and his dedication to healthcare certainly influenced all the kids.

My sister end up to be a pharmacist and my brother is a molecular biology scientist. So after graduating, I joined one of the big medical device companies where I became the senior director for their research center. We extensively applied machine learning and AI technology to solve complex medical imaging problems.

But later I started to looking for an opportunity to push my expertise and ambition even further. And this class led me to United Imaging Intelligence, the AI subsidiary of United Imaging. This was a place where I found a sheer passion for change And the drive for revolutionary breakthrough in healthcare.

Kathleen Wessel (Host): I just impressive backgrounds to you both. Uh, Terrance, I think your journey reflects several other's, uh, with family. Once you're in healthcare, you tend to stay in healthcare. It draws you in and commits you to the cause. You know, one term that keeps coming up, and can you help us understand what does Denoising AI mean?

Terrence Chen: Sure. Yeah. Why it's easy to associate denoising with simply making image clearer, which is one of AI's well-known capabilities. And in such context, AI is trying to understand the underlying distribution of signals within images. So this allows you to identify and remove noise by smartly filling in, missing, or distorting information.

And restore the original data. But here, as everyone learns, AI technologies and tools are widely used and improve different aspect of healthcare. Jeffrey's team is using a little play on words as we bring these tools to market. So by Denoising ai, I think we are trying to demystify AI hype, set realistic expectation, highlight ing capabilities, and promote transparency of AI tools for informed decision-making.

Jeffrey Bundy: I would agree with Terrence. I mean, we're using a little bit of play on words because Denoising AI is, You know, is something that it actually does. But we know that the people in our industry are inundated with ai, all kinds of, you know, hundreds of companies, hundreds of applications.

And they often tell us they don't know where to start And what are the things that they need to do to start. And the other thing is, there's a lot of tools out there that are not being rolled out and being used. Clinically very often, and word says, well, we're taking two main approaches in order to change that trend.

One is we have a, a unique business model that we call all in configurations, And what that means is when our equipment is delivered, every single software package that we make is delivered with all the software. All the AI capabilities or other capabilities are provided with the system so that no matter in whichever zip code you live in, the equipment is delivered fully equipped at a as a high end piece of equipment.

And another thing is, as I mentioned before, Terrance's team doesn't stop. If you buy a scanner in 2025 and something new comes out in 2027, it takes forever for that to roll out to an installed base. But we have something we call software upgrades for life, which is then putting in the hands of, of somebody who bought a scanner a few years ago, the latest, greatest technology on a continuing basis, such that AI tools that come and enhance a workflow or make a job easier are continuously delivered to, to our customers.

Another really important part of the thing is we have an applications team that does training for these technologies. Make sure that customers know how to use them, how to employ them, because I think that's one of the bigger challenges. People being unclear on how they work and, and a little bit scared about implementing them.

And I think that's the thing that we're trying to take away. I think Terrence's, word of demystify is a really good word to make it something that people can comfortably use every day to make their jobs better.

Kathleen Wessel (Host): I can imagine, as you've got, You know, even new team members coming in, like all of those resources continue to help and make sure that teams are up-to-date, up to speed, You know, switching a little bit. And members are continuously looking for innovative strategies to address workforce challenges. You know, how can AI and all its capabilities impact the burnout that radiologists face every day knowing when a human needs to be involved?

Jeffrey, why don't we start with you.

Jeffrey Bundy: Sure. Yeah. There are so many factors that lead to, to burnout and stress in today's workplace. And every customer that I talk to is challenged with getting enough radiologists, getting enough technologists, and all of those people, even administrators and You know, our customers will, are being asked to do more every day and do it with less.

And you know, we're focusing our energy on tools that we've talked about before that really. Make simple processes, easier tools that automate things that are part of the daily workflow, whether it's during the scanning part, as I've mentioned before, or during reading or interpretation and communication with each other.

And most importantly, as we bring these tools out And we have new, new things that we're working on, we're trying to get them into the hands of customers as soon as possible. Because if, if there is a tool out there. That can make someone's job a little bit easier.

We wanna get it into their hands as quickly as possible and not try to convince them to upgrade their system so that they can get that capability.

Terrence Chen: Yeah, I think This is a great question, and AI is already demonstrating real value in supporting radiologists. For example, in digital breast tumor synthesis, AI can easily offer higher sensitivity in detecting tumor masses while radiologists has superior specificity. But in our recent clinical study involving over 1000 patient cohort.

Radiologists working together with AI could improve both sensitivity and specificity by more than 10% while cutting their reading time by over 50%. This type of collaboration highlights how AI can serve as a powerful assistant, enhancing accuracy, as threatening workflow, and ultimately reducing the burden on radiologists by not replacing them.

Another example is our AI driven vision guided angio system, which automatically positions the robotic C relative to the patient And the selective protocol while avoiding collisions along the way. This capability boost the efficiency in interventional suite and, uh, helps protect both patient's And the clinical steps. So while AI still face challenges such as the back box, nature of some models, difficulties with out of distribution data, and request human expert to take the driver's seat, this example is how AI can meaningfully reduce radiologist's workload, helping to combat burnout while preserving the essential human expert as the heart of the diagnostic care.

Kathleen Wessel (Host): You know, I think there's been so much advancement in this field over the past, you know, five, 10 years. When you think of AI within medical imaging today, what does the future hold for this space and how will it impact the users?

Jeffrey Bundy: Yeah, I keep saying that the past went by a lot slower than the future is going to come at us, and I think that's one of the things that we, we see the application is already everywhere. People are using AI everywhere. We have it in many places as that some of the examples that Terrence just gave, which are already automating workflow, even voice control the system.

You know it's ai, but you don't feel like it's AI 'cause you're interacting with the system in the way that you would like to. And I think it's actually kind of nice when AI is behind the scenes running a system And the user doesn't even know it. I think that's a good implementation where it's just assimilated into the.

Workday into a workflow and a user doesn't necessarily need to know it's ai, but it's there and it's happening, and it will happen more and more. And Terrence alluded to that topic, there's been bad press over the years. I think some bad marketing, frankly, saying that radiologists jobs are gonna go away because of ai, et cetera.

We don't see it that way. We see not eliminating radiologists, but eliminating the repetitive, repeatable tasks that they do. That they don't like in their daily job, and they're just things that they have to do over and over again. So whether, whether that can be even image interpretation or just simple workflows during the day for them as a technologist.

We've also talked about some of those already. So ideally, and maybe it's I, maybe it's idealistic, but they can actually do more without feeling pressured that they're doing more. And if we do our jobs right and we're able to roll these technologies out, I think that can be great for healthcare.

It can be great for the organization. Itself And also for the employee. I don't think Those are too lofty tasks to achieve.

Terrence Chen: Yeah, I totally agree with Jeffrey. So when we started our AI company Uniting Meeting Intelligence about seven, eight years ago, it was already very clear our goal wasn't to replace radiologists. It was to empower them. So we are firm believers that professionals who use AI will replace those who won't.

A reality that we are seeing in many fields, including our own. So we use AI tools and these have doubled our software engineers productivity. Here's the essential part, right? These tools achieve the best results when used by experts. So for example, in our breast cancer detection example, the AI's findings are a starting point.

It takes the knowledge and experience of a breast radiologist to interpret those findings, apply their own expert judgment, and make the final. So the expert knows when and how much to rely on AI and making them indispensable. And I believe the future of AI in healthcare is about moving beyond single tasks to revolutionizing the entire system.

One example is deep integration of AI into the core healthcare IT infrastructure. This will allow AI to drive system-wide efficiency in areas like hospital workflow, scheduling, and even billing. So by integrating more types of data, AI will also help create more accurate diagnosis. And more personalized treatment plan for every patient.

Kathleen Wessel (Host): I agree with the direction that United Imaging is taking And the position that you're taking here. It'll only enhance the quality, the work-life, and really allow radiologists to operate kind of at that top level using their expertise, their skills. I think that's a wonderful contribution. Finally, what guidance and practical advice do you have for HA members as they look to implement AI solutions from an evaluation standpoint?

Jeffrey Bundy: Yeah, I think first and foremost, embrace it. I think we need a, a mind shift, both I think me personally, but all of us in our private lives as well as in our business lives that we need to change our mindset about what is ai. And a lot of us are using it as a new search engine, a new place to find information.

We're going from googling something to asking Chat GPT, uh, what the answer is, but it's much more than that. It's a way to create things. It's not just a way to find information. My private life, for example, I use AI to create pictures and to make songs, and Those are things that I could not do by myself.

There's no way. And I've created some funny pictures and some nice looking pictures and some funny songs and songs that people make fun of. But I've enjoyed that and I think in, in our business life, and Terence mentioned as well, using these tools to make ourselves more efficient every day and just looking for ways, even in, in your private life to adopt it.

So, it doesn't seem so foreign at work. I think that's my practical tip. I think there are things that we are already creating on our scanners. We're creating images, using AI on our scanners. We're creating answers and, and quantifications using AI on our scanners. So, It really is already an everyday tool and I think customers should just start to see that as it's just a normal part of our life, and look for ways to, to broaden their use of it and their application of it.

Terrence Chen: Yeah, I, I think what Jeffrey highlight is crucial, right? Embrace AI and be open to it. I think maybe I can give you a couple more example how we are already doing this and leveraging, uh, the creativity from ai. Our Chair city foundation model, uh, is a good example.

This AI model is trained on millions of pair city images and radiologists reports, so he knows how to generate reports, including all the incidental findings directly from chest CT images. More importantly, establish correspondence between the report And the images. For instance, if a radiologist harbors their mouth over a ref fracture, mention in the report, the viewer automatically navigate to the fracture And the corresponding images and vice versa.

So this feature significantly streamlines the verification process for thoracic radiologists saving them variable time during cardio intervention. For example, our angio system use AI to create virtual contrast agents. So after a small amount of actual contrast is injected and absorbed it, the AI remembers it and generates virtual contrast agent dynamically as a roadmap to guide interventional cardiologists during the PCI procedure and significantly reduce the total amount of contrast needed, lessening the burden of patient's kidney.

In another example, AI also plays a vital role in enhancing safety with scanning rooms and during interventions in our CT scanners. Nowadays, AI checks if technicians remember to cover patient's with lab protectors in the areas outside of the target scanning region. It also verifies if any part of the patient's body extend beyond the scanning table, which could lead to collisions with the scanner gantry during a scan.

So as Jeffrey noted, AI is already handling many of these smaller yet critical tasks, thereby improving their safety of both patient's and medical staff. We anticipate AI's contribution to grow even further in the future.

Kathleen Wessel (Host): Yeah, Those are great examples. Thank you for sharing those. And Jeffrey, I love some of your more creative recommendations. I'll definitely give that a shot myself. So I'm really sorry though. This is the end of our podcast. I really wanna thank you both, Terrance and Jeffrey for joining me today and sharing your takeaways with a HA members.

For our listeners, if you'd like to learn more about United Imaging And the a, a Associate Program, please visit us@sponsor.aha.org. This has been an Associates Bringing Value Podcast, brought to you by the American Hospital Association. Thanks for listening.