Arterys Medical Imaging Jumpstarts the AI Revolution in Radiology
Harry Glorikian: Hello. I’m Harry Glorikian, and this is The Harry Glorikian Show, where we explore how technology is changing everything we know about healthcare. Recently I came across a news story saying that a healthcare AI company called Tempus had acquired a medical imaging company called Arterys.
Mergers happen all the time in healthcare. But this one caught my attention, because back in 2018, I had a former physicist named Fabien Beckers on the show. Beckers was the CEO of Arterys, and had co-founded the company in 2011.
What was interesting about that 2018 interview was that Arterys had recently won FDA clearance for a software platform that could help radiologists analyze MRI images of the heart. That platform was groundbreaking in two respects.
First, it used a form of AI called deep learning to automatically locate the contours of the ventricles of the heart, which is one of the things radiologists do when they’re trying to diagnose heart failure and other conditions. The technology can help radiologists spot pathologies faster and more accurately. Over time Arterys went beyond the heart and applied it to MRI and CT images of all sorts of tissue, including the breast, chest, brain, and lungs.
The second thing was that the software was entirely cloud-based, meaning doctors could access it over the web, so hospitals didn’t have to maintain expensive on-premise software or hardware. It was the first time such a platform had ever received the FDA’s clearance.
And if none of that sounds so surprising now, it’s because there’s been a quiet revolution over the past five or six years. Today it would be hard to find a health tech company that isn’t using AI and cloud computing in some way. But Arterys was one of the pioneers.
So this week we decided to dip into the show’s archives and bring that you that original interview with Fabien Beckers. For me it was fun to listen back to this conversation, because it reminded me that just a few years ago, the idea of sharing radiology images over the web was still pretty new and radical.
It’s also fascinating to me that Arterys wound up as part of Tempus, in what’s been called “perhaps the biggest acquisition in the history of medical imaging AI.” Tempus is a big company backed by more than a billion dollars in venture capital investment. It offers a range of AI applications in healthcare, and they’re not just aimed at doctors and patients. The company also works with life sciences companies to help them develop and test new therapies. So technologies from Arterys could ultimately make medical imaging into a bigger part of the way new therapies are discovered and tested. That’s a trend we’ve been talking about on the show for a while now, and it will no doubt continue.
Beckers left Arterys in 2020 and today he’s the head of digital pathology for Verily Life Sciences, which is part of Alphabet. But right now let’s listen back to our conversation from October 2018.
Harry Glorikian: Fabien, welcome to the show.
Fabien Beckers: Hi, how are you? Very happy to be here and thank you for giving us some time today.
Harry Glorikian: Fabien, tell me about Arterys.
Fabien Beckers: Yes, so we created company six years ago now and been working since then. And and the vision and the mission that we have is really to try to transform healthcare and to make it truly data driven. And we really realized today that the main issue [for] clinical diagnosis as a whole, and especially medical imaging, comes from the fact that I think today’s infrastructure is really what you call a pre-Internet mode. You have all these computers lying around trying to be interconnected. And therefore I think physicians don’t have all the power and all the right tools that they need to really be enabled as we think this should be and could be. And so what we built up is instead of building more workstation and more computers that we deploy on site, we are trying to bring medical imaging to the Web. And we think that we can bring it the power of the Internet to really help physicians. And so instead of being a local machine, it’s just use a browser. You just use your browser to get images, the data center cloud while preserving data privacy. And then therefore, you have in the background enormous and unlimited amount of computation to do the work that you can do. But then now in seconds rather than minutes or hours. And so yeah, the AI to basically prepared the case for the physician and the piece can override and change anything they want. So they still save enormous amount of time, up to 70% in some cases. And then the report goes back within a clinical workflow to whatever reporting structure or whatever data they want, and it can be shared with anyone they want also, because it’s a web-based platform. And so by doing that, we say we’re going to equalize and democratize quality of care around the world and enable it to have better accuracy and better productivity at lower cost for health care and medical imaging.
What is medical Imaging?
Medical imaging is the use of various techniques to create visual representations of the interior of a body for clinical analysis and medical intervention. These techniques include X-ray, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and nuclear medicine. They are used to diagnose and treat a wide range of medical conditions, and to monitor the progress of treatment.
Harry Glorikian: So you started in cardiac and now you’re moving to other areas within oncology. But what are the ultimate goals of the company?
Fabien Beckers: The ultimate goal of the company is really to think about…so we started because we felt, first of all, it’s a really important clinical need. So our mission is really about patients and helping physicians, and we take that extremely seriously. And so we realized early on that cardiac is probably one, first of all, there’s extremely strong need on helping newborn and pediatric patients that have heart defects where today diagnoses were made using echo, and sometimes you don’t get the right quantification for the physician to make the right diagnoses. And we felt that was a very important, clinical need. A problem we need to work on. And so after that, we went and enabled an augmented product for cardiac because we said, okay, if we can tackle the hardest problem in medical imaging, we can probably build a platform that we can address for everything else after that. And that’s what we built. And now we really truly believe that with our conversation with physician and the couple hundred physician we have on Arterys today, that really there’s a potential for medical imaging and health care to move to the Web and to the cloud like any other industries and Arterys hopes to play a large part in really helping that physician to happen while respecting all the need that the health system has and all the constraint they have from a medical imaging device or regulatory standpoint.
Harry Glorikian: As a cloud based system, how have you designed the, or are imagining the business model from a use case perspective? And and how how do you think about charging for a software product like this?
Fabien Beckers: Yes, that’s a great question. So at the moment of where we are of any again, we’re taking learnings from other industries here and it’s very clear that SaaS and cloud-based have been adopted by other industries successfully and most easily, it’s as a SaaS subscription. And I think in health care you have a slight variation because in some ways they have sometimes large multi-year budgets and we can do that and be more of what you call a CapEx expenditures. But from other hospitals that tend to have more on OpEx we also flexible to enable that as well. And one of the value of a web-based system is not just doing the same as you used to do before better. You also open new possibilities that they didn’t have in the past. For example, like today, if you have a low cost system, you tend to upgrade every six months, every year. Now we upgrade every two weeks. You work very hard to connect and share and track with other physicians outside medical imaging or outside the institution. Now, as a web based platform, it’s easy to do. So you have compute, you have other data that can be leveraged as well. So it opens new avenues that could not be possible, that are really critical for a health system or for helping patients and doctors. And so that’s why I think the SaaS subscription fits potentially better. It’s a lower burden on a hospital that can tie that to their P&L while reducing their costs, increase their productivity and get better accuracy across.
Harry Glorikian: So it’s interesting, right, Because, you know, many conversations that I’ve had with hospital IT and so forth, cloud based systems, you know, always, to a certain degree, make everybody very uncomfortable. So how did you how did you sort of get over that issue or bridge that that that gap when when you were introducing this?
Fabien Beckers: And you’re very right. So if you really look at the why, if you understand structurally what has happened in health care, you will understand. And you also you may ask a question like why is how the last industry not to have really fully adopted the cloud and the web. And I think you really like put your finger exactly where it is, it’s all about patient privacy. And so when we started five, six years ago, when we arrived with really great solution that could only be done cloud based and very cost prohibitive fashion on a local basis, hospitals were very excited about a solution, but said we can’t move to a cloud because we don’t want to send patient health information or PHI to the cloud. And so from the get go, we realized what you really enable and bring the value to the cloud, to a health system, you have to solve the data privacy point. And so we spent four and a half, almost five years now working on a unique data privacy solution that allows hospitals to to send us images but doesn’t have any PHI on it. We strip out all the PHI, but in the browser for the physician, the options will reconcile for that physician in that session, the pixel data and the PHI at a time of the read. And so what it means is that by doing that, they will have access to the PHI, but we have no risk of taking the PHI outside of hospitals. And by doing that, we really bring the value of the cloud into computation you have on a true cloud native, real web based platform, to health care while making them reassured and secure on their private patient data.
Harry Glorikian: And so is this a real time interaction that’s happening or is it an upload, analysis, download?
Fabien Beckers: Nope. Nope. It’s real time. So it’s like think about analogies in other industry. Like think about Netflix. We’re basically creating kind of Netflix of medical imaging. Of course, the patient data comes inside the browser for the physician whenever he or she looks at the images, but the PHI comes from the hospital. So the same as if you were inside the wall of the hospital and pixel data come from the cloud so that we are bridging the two problems into one working solution that integrates and is interoperable to the workflow and health IT system.
Harry Glorikian: So one of the big, you know, milestones for you guys, because I remember reading it, was was the FDA approval for the product and now it looks like there’s been multiple FDA approvals. I’m sure there’s a whole series that you have in the pipeline with them now that you had your first one cleared. But can you talk a little bit about that FDA approval journey [and] the implications for the for the company and your experience with with the agency?
Fabien Beckers: For sure. And I think what is really important as I think some people really tackle AI and medical imaging recently, I would say last two or three years, since deep learning came nascent, as a tech problem. And I’d offer is we really strongly disagree with that vision. We really think that it’s profoundly a medical imaging device solution that happens to have software and it happens to us some form of advanced analytics that can be put to bear. But it’s at the core, a medical imaging device that is supposed to help physicians taking care of patients for very critical conditions. And so therefore, if you apply the process of a true medical imaging device 510(k) process, then you basically can get clearance and therefore you think that adopted by hospital. But I think the real landscape is the following one.
Harry Glorikian: So how did you how did you what was your experience with the agency? I’m because this is very new to them as it is to all of us in the in the industry. Did you you know, I want to say that you guys are first. So so you’re you’re teaching them while they’re sort of teaching you in a sense, and it’s a back and forth.
Fabien Beckers: Yeah. No, I wouldn’t say that we’re teaching them. I do think that we’re trying to apply a framework that builds trust on how we are approaching this problem and try to bring solutions that respect their needs and their constraints while showing that there’s a path to bring safety and efficacy. So, for example, we built a framework on how we test and validate. We follow older guidelines regarding testing and, and, and every single validation process that we have internally. And so I do think now there’s more and more interest around the FDA on how we’re going to tackle the fact that we framework on how we’re going to be applied AI throughout the body rather than in the past there wouldb be just one organ and one function. And so, for example, our oncology care is applied to the whole body as long as of course we detail in our processes after that how to to test and validate each algorithm. And so I think there’s really ways to really work with the FDA. So far, it’s been a very good experience on how we can take a solution that is fully finalized and ready to hit clinical grade and to be put to market, once this comes to the right process.
[musical interlude]
Harry Glorikian: Let’s pause the conversation for a minute to talk about one small but important thing you can do, to help keep the podcast going. And that’s leave a rating and a review for the show on Apple Podcasts.
All you have to do is open the Apple Podcasts app on your smartphone, search for The Harry Glorikian Show, and scroll down to the Ratings & Reviews section. Tap the stars to rate the show, and then tap the link that says Write a Review to leave your comments.
It’ll only take a minute, but you’ll be doing a lot to help other listeners discover the show.
And one more thing. If you like the interviews we do here on the show I know you’ll like my new book, The Future You: How Artificial Intelligence Can Help You Get Healthier, Stress Less, and Live Longer.
It’s a friendly and accessible tour of all the ways today’s information technologies are helping us diagnose diseases faster, treat them more precisely, and create personalized diet and exercise programs to prevent them in the first place.
The book is now available in print and ebook formats. Just go to Amazon or Barnes & Noble and search for The Future You by Harry Glorikian.
And now, back to the show.
[musical interlude]
Harry Glorikian: I had seen a talk you gave or and it was referred that you were part of the World Economic Forum Steering Committee on Value Based Care. Are you still part of that?
Fabien Beckers: I am and I have to admit that with my my schedule and the travel, I’m not as involved as I was at the beginning. But yeah, I was there at the time, I think it was two years ago. Yeah, yeah.
Harry Glorikian: Yeah. So. I’m curious. You know, you you run a data driven medical imaging company and you’re you’ve been exposed now to that group. And so how do you see, what are the implications from data driven medicine on value based medicine? And I’m asking this for self serving purposes. I wrote a book on this. So I’m curious to what your your positioning is or what your thoughts are.
Fabien Beckers: Yeah. So I do think that the missing piece between the two together is having the right tool and right information and the right reporting to help drive the best decision possible. And so in that regard at the moment, I think it’s a very siloed system that we’re dealing with. It’s very hard to have the data to surface in a way to enable the physician and the institution to give the right or the best possible treatment to the patient. And so in that regard, I think web based things like Arterys, what we can provide is a layer of accuracy and productivity across the system. And so the beauty of medical imaging is that there are few vendors, the data are standardized, it’s structured, and they’re usually slightly different, but still much more homogeneous than you have in the text data for EMR. And so you can really, truly make value on that first level. And then on top of it is not aggregating those across a large pool of patients or hospitals. And then on top of that, they also start adding other data on top of it about EMR integrating to to give more substance to the analysis. And by keep adding those layers, I believe we get closer and closer to having a very exhaustive, of course, it’s never going to be fully complete, but as close as we can be to a full picture, no pun intended, to the physician about what’s happening to that patient, what would be the best treatment to enable value-based care and to enable the best optimal path for that patient.
Harry Glorikian: So if I’m not mistaken, the company idea came out of your work from Stanford. And how did that how did you get that off the ground? How did you start the companies, the company itself.
Fabien Beckers: For sure. And I think it came out of Stanford very personally. Actually. I had myself a health situation I had to deal with, and I was surprised. I came to the U.S., to Stanford, to do some mission type work. When I had personal health issues myself I discovered how variable was the clinical diagnoses process and how people had very little information data to drive the decision, I really thought that was very interesting situation and therefore me worth investigating. And the more I would learn, the more I got into it, the more I realized how structural was not just a one-off situation. And then the further I went, then I start thinking about potential solution to bring more data driven element to one area. And then by adding other physicians, we realize together that the problem is not about another piece of software. The problem is at the structural level. Health systems today are based on the local on-premise workstation approach and therefore are kind of like in shackles in some way or by default it’s very hard for data to flow. So the compute capability we put to bear to the physician and the whole system is kind of a holed up situation. And so we felt that it would be worth trying to build from scratch the right infrastructure for kind of healthcare 2.0 way of thinking that will embrace the web, that will embrace cloud computation, that will embrace AI, that will embrace collaboration and sharing, and break down those silos, mutualize some resources on a global scale to make the system and physician work better together. Health care is a very interesting field because in other industries it would be hard for engineering engineers, for two different companies to work together, or two bankers to work together. In health care, physicans are willing to work together. And so therefore what you need is a system to enable them. And so that’s why we think more than any other industry, a web based platform for medical imaging mission critical diagnosis is essential. It’s the key and the right framework to bring knowledge, which is of course, it’s a peer to peer industry. So knowledge is extremely important health care, but it doesn’t work with a siloed system. It’s about really having multiple AIs run on every case, which is about having compute. If you want even more compute, we need to start leveraging the cloud. Start leveraging more data and not just medical imaging. I don’t know how much data is being changed or how much the data of health care could be increased every month or two months. I think it’s in the petabytes of data, isn’t it? So therefore, you need AI to bring that data to bear to get the learning out of it and health data and digesting that at the level of the case of the patient. And so that’s what we felt was needed. And that’s not to do with AI. It has to do with infrastructure. It has to do with the web. You have to do with having a browser that can process medical imaging at the real time, wherever they are. You have to deal with interoperability and workflow integration. You have to deal with data privacy. A lot of the core pieces that are not sexy but are so crazily important in order to really enable the transition of that industry in that field.
Harry Glorikian: Well, it’s interesting, right? I mean, I keep hearing, you know, AI, AI, AI, machine learning. Everybody talks about it. But at some point, you know, the user actually doesn’t care, because the user just wants the answer. So it’s almost as if it should be transparent to them in the background.
Fabien Beckers: But it’s not the same. It’s like if they tell you today, they can get a lot of transistor for your phone, like you don’t care. You just want a phone and an iPhone or whatever.
Harry Glorikian: Brand Yes. You want to pick up and make a phone call. Yes.
Fabien Beckers: You want make a phone call. You don’t want to know, like, hey, I give you the best Motorola, X, Y, Z transistor. You want. You want to make a solution. And I can tell you that from now I know AI has a buzz, and it’s true that I can really help drive productivity and accuracy. And it works. Deep learning really, really, truly can reproduce tasks that are done by radiologists accurately and very well. But the hard part is not the AI. The hard part is the platform. It took us five, six years to build a platform. It can take us weeks or months to build an algorithm. If not many more. So it’s really not the hard part. So interoperability in hospital, the workflow integration, the data privacy, those are the real Himalayas to climb. The AI, every single banker and and trader in Wall Street can download TensorFlow and and download from NIH a thousand datasets and build a model and become a medical imaging AI company. But what are they going to do when they have an AI algorithm, how are they going to get it to the to the physician at the end of the day to impact patients? And that’s last mile between the AI and the patient and the physician is, for us, we believe the much, much harder problem to solve.
Harry Glorikian: So, Fabien, what what if you had to juxtapose what were some of the challenges in the early days of the company and what are some of the business and technical challenges you face now? Year one versus year or five or six, I’m guessing.
Fabien Beckers: I think that’s a really interesting question. So I think in year one, it was really like, okay, so we had a vision on on how to to bring that data driven medicine, how to build that infrastructure. I think we far, far, far underestimated how hard it is to build it. So we thought that within, I don’t know, months or years will be done and we fully ready. And we’re five and a half to six years in and we feel like there’s still more to do and and wish every day was done more. So I think that that’s year one. And so and then also it’s covering issues about, if you come from the tech world it seems so easy to say, yeah of course we’ll do web solution, everyone is the web, why would we do a local system. But you have to understand all the complexities and intricacies of building a web solution for the healthcare environment and for hospitals. And so we had to learn a lot of those things. And it’s not about something that can do overnight. It’s a lot of iteration work when you just us one bottleneck and you have another bottleneck and another one by one, you keep addressing them over years. So I think that that was the first few years. Now we have a different problem, which is about scale. So now we’re getting health system that yave seen, And I think the mentality has shifted to now CIO, CEO, CMOs, radiologists see the value of what a web based platform can bring them as long as it protects data privacy and it’s regulated. And so now we have we are talking to health system to transition to every doctor’s office. But of course, how many can we tackle, How many can we get connected to the platform? Those are the different situations. We are live now in China, so we have three regions in the world to manage. We are in nearly 95 or hundred countries where we’re regulated today. So we have a question of scale which is very different now that are question of building from scratch, which was our case five, six years ago.
Harry Glorikian: Well, Fabien, I don’t want to take any more of your time. You’ve you’re you’re out there really making a difference in the industry as you know, thought leader and trailblazer in the area, especially from a regulatory approval perspective, getting a product out there that’s useful. I want to thank you so much for your time. And you know, everything you’ve you’ve you’ve shared with us. And thank you for joining us on the show.
Fabien Beckers: Thank you so much. And it was a real pleasure. Have a great day.
Harry Glorikian: Thank you. Bye bye.
Harry Glorikian: That’s it for this week’s episode.
You can find a full transcript of this episode as well as the full archive of episodes of The Harry Glorikian Show and MoneyBall Medicine at our website.
Just go to glorikian.com and click on the tab Podcasts.
I’d like to thank our listeners for boosting The Harry Glorikian Show into the top two and a half percent of global podcasts.
To make sure you’ll never miss an episode, just open Apple Podcasts or your favorite podcast player and hit follow or subscribe.
And don’t forget to leave us a rating and review on Apple Podcasts.
We always love to hear from listeners on Twitter, where you can find me at hglorikian.
Thanks for listening, stay healthy, and be sure to tune in two weeks from now for our next interview.
FAQs about Medical Imaging
AI can help medical imaging in a variety of ways, including:
1. Image analysis: AI algorithms can be used to analyze medical images and extract useful information, such as identifying tumors or other abnormalities.
2. Computer-aided diagnosis (CAD): AI can be used to assist radiologists in interpreting medical images and making more accurate diagnoses.
3. Image enhancement: AI can be used to improve the quality of medical images, making it easier for radiologists to identify important details.
4. Automation of repetitive tasks: AI can be used to automate repetitive tasks such as image segmentation, registration and annotation, which can help to improve the efficiency of the imaging process.
5. Predictive analytics: AI can be used to analyze patient data and imaging results to predict the likelihood of certain medical conditions or the effectiveness of certain treatments.
Overall, AI has the potential to improve the accuracy, efficiency and speed of medical imaging, which can lead to better patient care.
Medical imaging is very important in healthcare as it allows doctors and medical professionals to see inside the body and diagnose a wide range of medical conditions.
Medical imaging is used to:
1. Diagnose diseases and injuries: Medical imaging allows doctors to see inside the body and identify problems such as tumors, broken bones, and internal injuries.
2. Monitor treatment progress: Medical imaging can be used to track the progress of treatment and determine if a course of treatment is working.
3. Guide medical procedures: Medical imaging can be used to guide surgical procedures and other medical interventions, such as radiation therapy.
4. Screen for diseases: Medical imaging can be used to screen for certain diseases, such as mammograms for breast cancer and colonoscopies for colon cancer.
5. Research: Medical imaging also plays an important role in medical research, as it allows scientists to see the effects of different treatments and interventions on the body.
Medical imaging is an important tool in modern medicine, as it allows doctors to see inside the body and identify problems that would otherwise be difficult or impossible to detect. With the help of AI, medical imaging will be even more accurate and efficient in the future.
Medical imaging and radiology are closely related, but they are not exactly the same thing.
Medical imaging refers to the use of various techniques to create visual representations of the interior of a body for clinical analysis and medical intervention, including X-ray, computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and nuclear medicine.
Radiology, on the other hand, is the medical specialty that deals with the use of medical imaging to diagnose and treat diseases. Radiologists are medical doctors who have completed specialized training in the interpretation of medical images, and they use medical imaging to diagnose and treat a wide range of medical conditions. They also work closely with other medical professionals, such as surgeons and oncologists, to plan and monitor treatment.
In summary, medical imaging is the techniques and technology used to create images of the inside of the body, whereas radiology is the medical specialty that deals with interpreting and using those images to diagnose and treat diseases.
Medical imaging professionals, including radiologic technologists and radiologists, typically need a combination of technical and medical knowledge and skills. Here are some of the key skills that are typically needed for a career in medical imaging:
1. Technical skills: Medical imaging professionals should be proficient in the use of medical imaging equipment, including X-ray machines, computed tomography (CT) scanners, and magnetic resonance imaging (MRI) machines. They should also be familiar with the different types of medical imaging procedures and be able to adjust the equipment settings to obtain the best possible images.
2. Anatomy and physiology knowledge: Medical imaging professionals should have a good understanding of human anatomy and physiology. They should be able to identify different structures and organs on medical images and understand how they are related to specific medical conditions.
3. Communication and teamwork: Medical imaging professionals need to be able to communicate effectively with patients, radiologists, and other healthcare professionals.
4. Attention to detail: Medical imaging professionals need to be detail-oriented and able to identify subtle differences in images.
5. Physical stamina: Medical imaging professionals are on their feet for long periods and may need to move or lift patients, so they need to be physically fit.
6. Technical Writing: Medical imaging professionals should be able to write technical reports and communicate effectively to other medical professionals about their findings.
7. Problem-solving: Medical imaging professionals need to be able to troubleshoot and solve problems related to equipment or patient care.
8. Continuing education: Medical imaging professionals should be committed to continuing education to stay up to date with the latest advancements in medical imaging technology and techniques.
9. Artificial Intelligence and Machine Learning: With the growing use of AI in medical imaging, having knowledge of AI and machine learning would be an added advantage.