Innovation to Impact (ft. Dr Matthew Lungren & Dr Jonathan Carlson)

Thanks for tuning to Digital Voices Podcast, where we chat digital transformation challenges and opportunities across healthcare and life sciences. And now, your host, Ed Marx.

Hey, it's Ed, and I'm so excited to have another edition of Digital Voices. Thank you for making us number 13 right now, as of February 2025, number 13 on Apple podcast for technology. It's because great guests like we have today. We have Dr. Lundgren, Matt Lundgren, and Dr. Jonathan Carlson joining us. Matt and Jonathan, welcome to Digital Voices.

It's good to be here.

Thanks for having us.

Yeah, this could be exciting. We're going to talk about innovation, Microsoft, and radiology. But before we get there, we always have a very important question that we ask everyone because your answers will make it to our Digital Voices Spotify playlist. And that is, What songs are on your playlist? So Jonathan, if we were to look, what have you been listening to lately?

Depends on who's in the car with me. If it's my youngest daughter, it's Taylor Swift. If it's my wife, it's some country song I couldn't tell you the name of. I gotta tell you what I'm listening to right now is actually the Acquired podcast, which I know is sort of blowing it all up, but that, I can't get enough of Ben and Dave right now.

That is a good podcast. We should, we should add that question to our script for now on. What about you, Matt?

I, well, so unlike Jonathan, I have 2 daughters as well, but I'm the Swiftie of the family. So I'm both. A lot of— did some Arcade Fire, believe it or not. I've still got a lot of sort of mid-2000s rock that I, that I cling to.

Yeah, no, that's, that sounds good. And Matt, sticking with you, what's your life message or mantra? Are there words that you live by or sort of guide you?

I often say there's, there are, there are no career rules. You can, you can do what you believe you're interested in or what you believe is impactful from any background.

Yeah. I like that. Jonathan, you.

2 right now. One I'm learning as a father, which is not yet, which is to say, it's not that you can't do it now, it's that you can't do it yet. It turns out that's very, very applicable. It's just a matter of attention. And the other is I can't possibly plan more than 5 years out, so don't ask me.

That's good. Yeah. You collect a lot of wisdom from a lot of super smart people like yourselves. So let's talk about each of your journeys. So Matt, we'll go back with you. Tell us about yourself. Who are you? What's your story and your life growing up?

I grew up pretty average, you know, suburban life in Arizona and always dreamed of being a writer, an English professor. Got an English degree in undergrad, made a hard pivot into medicine after a series of international experiences in refugee camps, et cetera, volunteer work. And, uh, and yeah, so I've, I'm an academic clinician. I have been, I guess now since, geez, started my first journey maybe 2003. So it's been over 20 years, um, in academic medicine. Uh, and then, and then made another pivot. So hence my mantra earlier, uh, into technology. And a lot of that was driven by the AI and healthcare research I did at Stanford as a faculty there.

Yeah, that's very cool. And I know that there's some links to some of that research that you've done on your LinkedIn profiles for people who have more interest in that. Well, let's keep down that road, Matt. So tell, tell us one story from one of these camps. So these things that changed your life in terms of caused your first pivot from English to, uh, to medicine.

Yeah, I don't tell the story often, but, uh, so I was an English major. I had a really close friend who, who was a, you know, ethnic Kosovar, uh, right during '99, and had the opportunity to help lead a camp that they had set up across the border in Albania, Shiyak, Albania. And she asked, do I want to go? And I, you know, I was an English major. So I'm like, sure, this will be great. You know, give me some inspiration, something to write about. And I went there and I was randomly assigned at the time to— this is an abandoned bread factory. And I was randomly assigned to the storage closet to keep the medical supplies organized. So when the, you know, the physicians and various volunteer organizations came by the camp, I would, I would be there to help them. Hey, do you have this? Do you have this? And I'm that guy.

Okay.

Uh, and so I kind of just parked myself most days in this little, like, it served as a clinic, but it was really a closet. And, and just got to see basically right up front, real medicine being practiced in an austere environment and met some of the most inspiring people you can possibly imagine. And I came back and said, wow, I am missing out on something that I think, uh, could, you know, really impact the world. And, uh, you know, I think I'd like to take a crack at doing that too.

That's awesome. I love that. Jonathan, tell us about your story. Where did you grow up and how did you get into what you do today?

Yeah, so I was born and raised in Beaverton, Oregon, which is right outside Portland. About mid-high school, I got super into biology and especially little things, viruses and whatnot. And the 3 things I was certain I never wanted to do were medicine, because I thought that was way too practical, business, because that's what everybody did, and computers, because that's what my dad did. Um, and then it turned out when I got to college and started actually doing a biology major, it turns out I really a disaster in the lab and kind of got sucked into computers and learned how that I could actually use computing to learn about biology. Went to graduate school, got a PhD in computer science, and then found myself at Microsoft, specifically at Microsoft Research, where I was really focusing on really basic research, but then also sort of learning about how business is fundamental to doing anything at scale in the world. And most recently really have started seeing how medicine and biology really are linked and how I really need to start paying attention to that.

Yeah. Cool. What about your daughters though? Are they, are they thinking, they're looking at you like, I want to do computer stuff or?

I try, I try not to ask that question. They're 14 and 11. And right now it's all about learning to love to learn.

Yeah.

And we'll see where the doors open later.

Yeah. That's, that's super cool. Let's talk about Microsoft. So you're both with Microsoft and, uh, we're talking healthcare. So for those of us who don't know, can you explain sort of the organization that Microsoft has where, where, where you all serve? And I know, Jonathan, you're more on the future side, but we'll bring that in as well, the health futures.

Yeah, I can start. I think, um, you know, I think what most people should know is obviously Microsoft is a very, very large company with lots and lots of different, uh, capabilities and, um, and partnerships. And I think in, as a platform, broadly speaking, that they're involved in many industries and healthcare is obviously a big one. Uh, there was an acquisition of Nuance Communications in, uh, 2019, 2020, uh, that really started to show the commitment to the healthcare vertical. And so Nuance Communications, for those in the audience who may not know, is essentially the voice-to-text capability that the vast majority of physicians in the US use to dictate their notes or to have any other sort of similar interactions. And with that kind of footprint and that lens, they really said, listen, I think we can have the horizontal and the AI capabilities and the platform capabilities meet the vertical. In other words, fit-for-purpose extensions into that to actually drive real impact and solve problems in healthcare. You know, probably the one the audience may be familiar with lately is the DAX Copilot solution, which is this ambient note generation. So the physician, rather than looking at the keyboard and the screen, can look at their patient and that transcript, that transcription can be turned into a note for the clinician to edit inside. So that's been a, that's been an absolute game changer, partially due to the capabilities of generative AI. But, but taking a step back more broadly, we have other opportunities and, you know, where Jonathan and I come, come together is what are all the other innovations happening in Microsoft Research or our strategic partners, and how could we drive that into solving problems in healthcare now? Like, how do we bring that innovation to impact pipeline? How do we shorten that timeline?

Yeah, Jonathan, how did— then do you work, you know, from a futures perspective with great clinicians like Matt?

Yeah, so we sit within Microsoft Research, which has been around for 3 decades at this point, and broadly our mandate has always been to imagine and make real the future of technology and make sure that that's developed for the good of humanity and make sure that Microsoft has a place to play in that future. And so honestly, we've been doing biology, computational biology works for about 20, 25 years at this point. And our role has kind of shifted within that. I'd say early on, our role was really to think about what is the future of computing within biology and medicine. And now with the acquisition of Nuance, the real product team, with the advances in generative AI and just the pace at which that research is progressing. We really think a lot about both what does the future look like 3, 5 years out, but also how can we translate that into products, whether it's Microsoft products or partner products right now. And increasingly we're finding motions like this last year where we publish a paper in Nature and that same quarter we ship it as a product. And that's just the pace of technology, right?

I'm curious how you both approach innovation. And I want to talk to you about some of the things that you've been doing together. Uh, in radiology specifically and with WEI. But, you know, a friend of mine works for Pininfarina. He got the job of redesigning the Zamboni. So for 50 years or whatever, the Zamboni has been exactly the same, right? All of a sudden he gets it. So I was like always curious and asking him like, hey, how do you, how do you do this? It's kind of a theoretical question for, for either one of you, but like, I, cause I think people struggle with this whole concept of innovation, but you all sort of live it. How do you go about innovating or getting new ideas around things?

Well, I can say from at least my side, and then I think Jonathan has probably a different perspective, but from, from my side, at least, there is the tyranny of the now for us in healthcare that, you know, the burnout crisis is well described by now, but it is impacting, you know, our patients and our health systems. And, you know, we're being, we're being asked on the clinical side to do more with less, more and more, right? So, and if you kind of look back of where I started practicing to now, I, it's barely recognizable the amount of additional things that have kind of been slowly layered on in terms of the administrative burden. So I often say, That innovation is going to save time before it saves lives, at least from my point of view. And some of the solutions that I sort of touched on point to that. But where we are from a technological perspective is how do we translate what's happening in the broader acceleration of AI and other technology capabilities? How do we actually translate that into a problem that we can solve in healthcare? And then the other direction, what are some of the critical problems that we're seeing on the ground? And how do we bring that over to our innovators? our partner ecosystem and say, hey, this is something that we could actually drive real, real outcomes with if we solve this today.

That's cool. Jonathan, anything on the creative process as you deal with futures? We're going to jump into the radiology stuff here in a second, but I just find this super interesting because I think it's, like I mentioned, something people kind of struggle with. So you're given, you see a problem, what's your sort of next step? Like, how do you look at things?

You know, at a high level, it's hard. Innovation's always hard, especially when you're trying to think 3 to 5 years out. But some of this is where, why I think it's so important to be working right at the frontier of science. At some level, you can't top down this, it has to be the people that are really pushing the boundaries right now. And oftentimes the question is, where are the high leverage points where you can either create a tipping point or recognize a tipping point and push on it? And where that actually fits shifts over time. And this is where we work really closely with Matt's team in terms of, great, we're starting to see what these new opportunities are. Now, how do we actually scale it? What's the next piece that we need to do to make that more broadly applicable?

Matt, why do you think radiology At least in my opinion, and I've been in healthcare for a little while, but in my opinion, radiology has always kind of led the way when it comes to tech innovation. And if you agree, why do you think that is?

Well, I think there's lots of stereotypes about the different specialties in healthcare. I'm sure you've heard various versions of those, but the radiologist is typically someone who has a very, very high degree of technical interest. Maybe they're tinkerers, maybe they're builders, and they find themselves like always trying to innovate and bring some new technology. And one of the places that seems to be the most accepting of new approaches, mainly to diagnosis, but also treatment, is in the radiology space. And remember, radiology also includes interventional radiology, which is something that I'm, I do, right? Which is literally, here's a problem that we can't solve with surgery. We can't solve with medicine. Can you all fix it? And we use combinations of imaging and therapeutics and catheters and stents and wires and all kinds of stuff. And that's a really creative field as well. So as a whole, it's a very creative, very technical, technically forward. specialty. Um, and, you know, listen, we got, we, we, we can claim the Nobel Prize for MRI, right? We still, we still talk about that, you know? So, so there is a lot of, uh, I think a blending of science, technology, and actually practical problem solving. But, you know, to, to Jonathan's point about finding some of these tipping points, you have to be able to recognize that, uh, this technology actually can change the game. Um, but, but it's going to require, again, the full force of our scientific colleagues to, to really, to really deliver it.

You know, I, my first CMIO was a Dr. Jeff Sunshine. You may know him because there's not too many of you around, uh, interventionalist radiologist. And it was a natural thing for him to, to go into sort of the technical part of IT while still remaining in his, uh, practice. And, uh, yeah, it's super interesting. And I agree with you in terms of the typical background and maybe why they've been a little bit more progressive than some of the other specialties. Where do you think we're headed? I don't know, Jonathan, you want to take a stab at this, like with radiology, because you're working so closely together with Matt. Where are we headed? I want to kind of get, uh, you know, into AI in a second. But generally, what are you seeing or what can you share that you're working on in the next 5 years?

Yeah, I'd also add to Matt's response too and just point out that radiology among all the disciplines has always been technology focused because the tools they use are fundamentally technological. And I mean, we also work, for example, in pathology. And one of the challenges there is that while we can have a great foundation model that's really good at detecting cancer, if the image isn't digitized to begin with, it's really hard to fit it into a workflow. Once you get to radiology, You know, think about an MRI; it's it's already digital. Now, some of what we're working on there is now how do you actually take that information off the MRI and and attach it to unbounded compute so you can unlock even more digital technological innovation, if you will. So I think you can think about it through the whole stack. How do how do you offload the compute? How do you actually build foundation models that can understand the raw signal as well as the image itself to really be able to pull out information from those images? How do you take those models now? align them with text models so that you can kind of reason with the images directly, whether that's actually generating reports or even being able to validate existing reports that have been written. So we're starting to see this convergence of large language models on the sort of human language side to models that understand images, even models that understand raw signal.

Yeah, that's very cool. What about AI? So, well, we talked a little bit already about LLM and some AI when it comes to voice and things like that. What's up next? Because Because again, radiology has been ahead of the curve and been doing AI with imaging, right? The basic stuff that everyone knows about today, very complex, but basic in that it's been around now for a few years in terms of being able to help prioritize which images to look at and which images have a high probability of requiring a more in-depth read. What else are we looking at in the next couple of years?

I mean, from our side, one of the reasons why this partnership has been so powerful, I think, is that, you know, We, like, to your point, there's a lot of applications, what we would quote unquote call narrow applications, right? That we can attack that we know are pain points, you know, not missing a nodule that could turn out to be cancer, prioritizing an emergency case, those kinds of things. But if you take a step back and kind of watch what's happened in the language space, right? Where folks were building language-based models to solve individual problems. And now we're kind of in this era of we have this general purpose technology, this tool that's incredibly performance across all kinds of language tasks. What's the equivalent of that in vision? And we see some of that in the non-medical world, right? With, you know, video captioning, uh, having conversations about, you know, photos and things like that. Well, you know, in the medical space, we know number one, there's a massive amount of information in that medical data that's sitting in your health system. And it may only get used for that one, you know, acute care moment or one doctor visit. But, but we're leaving a lot of information on the table. We've increasingly learned that, but, but mining that and intelligent kind of looking at everything from biomarker discovery all the way to being able to tell what your risk profile is for certain diseases that could come out of that pixel data. It's been kind of locked up, right? And so we see this potential of, you know, to Jonathan's point, can we actually extract the insights at a larger scale, more holistically across these modalities? And then use that to drive, you know, real, real solutions that can come from doing that. And that's really kind of been that innovation to impact pipeline that has really, that flywheel has really started to kick off.

Yeah, I like that phrase. Is that a Microsoft phrase, innovation to impact?

Well, it can be now if you like it.

And then we'll trademark that. How do healthcare organizations work with Microsoft? So, you know, aside from the big ones, right? You form these huge giant partnerships, really cool and transform things with them like Mayo or Cleveland Clinic. But for the average hospital who doesn't have that breadth and depth of an organization, and they want to do some innovative things and really take advantage of their relationship with Microsoft, how does someone go about doing that?

Some really simple things, which again, I think most people would be like, oh, right, of course. Well, so most health systems are using Microsoft, let's just say products today, like email and Word and PowerPoint. And we've shown many different use cases with different health systems that Even those productivity tools can start to lower that administration burden. That has nothing to do with healthcare-specific tools, right?

Right.

So like just being able to have a Teams meeting where the entire conversation with you and a few colleagues is transcribed, like those kinds of little workflow efficiencies do add up. And we've seen that. The second one is something that we, we have found to be incredibly powerful for a couple of reasons. It's the ability to create a HIPAA-compliant ecosystem in one health system. That allows everyone to have compliant access to some of the most innovative models on the planet. So GPT-4 and Llama and all these other models, but to do so in a way that they're not concerned about security and privacy, right? And by doing that, these health systems are realizing, hey, hold on a second, my whole workforce, no matter what you do for a living, is finding ways to use these models and driving efficiencies in their day-to-day tasks. But at the high level, they're able to look down and say, what are folks using it for across my health system? And then surface those top 3, 4, 5 use cases and say, wow, these are really popular. Let's operationalize that. The other effect that that has by setting these things up, and again, NYU is one, Stanford, many of these institutions have done this, is they've also realized they're increasing the literacy of their workforce with these new tools, which, you know, obviously we're going to argue hopefully successfully that this technology is going to change everyone's life, right? Their work life, no matter what they do for a living. And so, but, but it requires literacy and it requires getting some reps in. This, this has been really, really fruitful. And this doesn't even then go into some of the more specific, far-reaching innovation that we've really talked about. These are just basic things that we see as repeatable patterns across the industry.

Yeah, I really like that. You know, a lot of times hospitals try to go it alone and it's kind of silly. And I'm a big advocate of partnering with great vendors, suppliers such as Microsoft and working together can really accelerate, you know, whatever type of innovation that you're looking to do. So that's really great. And I also recommend people read Satya's book if they have not already, Refresh. I think it kind of gives a little bit of insight too in terms of the healthcare focus and sort of his background with healthcare and why he's wired the way he is. And I think it shines through the entire organization. So I highly recommend that book. It's a really good, really good read. It's been out Well, it's been out at least 5, 6, 7 years, but it's, if you haven't read it, you need to read it. Hey, let's pivot now to leadership. So in the beginning, who are you and how did you get to the roles that you're in? We talked quite a bit about Microsoft, but I, our listeners really like to learn a lot from you around leadership. So Jonathan, this one's to you. Like, did anything early in life help prepare you for your current role? So Matt, we, we learned from you the profound background that you had in your collegiate years. Dante, what about you? Is there anything that helped you in your youth that today has helped make you successful?

I think the thing that comes to mind is curiosity. It's a thing that for me, it was always about, my parents always pushed us, be curious. I, you know, I'll add a 5th thing I said I never wanted to be when I was in high school and college, and that's a leader. But what I was always was very curious and very keen to understand deeply what other people are doing and how those connect with each other. And it turns out more and more understanding how, what, what makes people tick and what motivates them and how those can work together at some level of the fundamental core of leadership.

Yeah. And Jonathan, I hate to ask you if there's a 6th thing we haven't discovered yet that you said you probably— this could be another podcast right there. Matt, I think you're writing now, right? You're, you're doing some writing. So has the, how has the English helped you?

I, I think it's, I always think of how much it helps me. Like, I mean, if you think about some of the core fundamental things that leaders are asked to do and communication. You know, great writing, and then thinking about from the science perspective, it's always helped me in my career to be able to write clearly and you know talk about new ideas in a way that we can you know land with the audience. And obviously grants and academic papers, all this stuff comes in handy. I was maybe the first to use like a semicolon in one of my grant submissions, and it was called out saying this is over the top. And I'm like, all right, fine, I'll tone it down. But but no, I I love reading partially too to this day because of the opportunity to just really kind of walk in someone else's shoes and see the world through a different set of eyes and. Obviously that sometimes impacts how I think of the things as well. So I credit anyone who wants to do an English major, you'll never be disappointed you did it.

Hey, it's, there's something to be said about the liberal arts and not to forget it, you know, because both of your backgrounds, and that's why I'm bringing this out, is quite vast and diverse. You know, you got the English and the medicine, now the tech and Jonathan, the biology, the tech and all that comes together. So it's never a waste, you know, when people are studying whatever, as long as you're studying something that, you know, you're passionate about, it'll somehow come back and help you later on in your career and your leadership. Jonathan, what are one or two things you learned the hard way? Could have been in your youth or maybe in your career, but what's something that you, that something happened that you weren't too happy about, but, you know, you probably ended up turning it around or caused a pivot of some sort?

Yeah, I think this is an indirect answer, but one of Bill Gates's sayings has always stuck with me, which is people always overestimate what they can do in a year and underestimate what they can do in And I feel like my research career is riddled with those problems of spending way too much time on a project that I thought I could do really, really quickly and went on and on and on and on. And I just should have cut it off earlier. And then other things that it turns out, you know, you put your mind to something for a while and the pace of change can astonish you and go way beyond what you ever thought.

Yeah. Well, you know, to that point, Jonathan, how do you know when to When to stop? Like, you're talking from a research point of view, and I think about it from an IT point of view, you know, like as a former CIO, it's like we're trying something, some new technology, and do we keep having it be resilient and force it through and it's going to be great and everyone's going to be so happy? Or do we stop that project because it's really never going to lead to the outcomes we wanted? Like, is there a— is it a gut feeling for you or is it, you know, talking to other people? What— how do you make that decision?

I think it's always about opportunity costs. And that's the hard thing. You always want to look at the thing you're doing right now and say, what's it going to pay off, but forcing yourself, your team to take a step back and say, yeah, but what are we not doing right now? And it's okay right now to let everything else we just want, just let it all go. There's a better opportunity right now.

Yeah. Yeah. I like that. Matt, we're going to switch it up for you. How do you recharge and stay fresh? Right? Because you're under a lot of pressure. You work for a great company, but you know, there's, there's definitely pressures and clinical and all that kind of stuff. What do you do for yourself, for your own self-care?

Well, I, so I, I was also a swimmer, uh, in, in high school and college. And so I, there's, there's something called the, uh, the master's program, which is a national program. Almost every city has a chapter. It's a bunch of ex-collegiate, few Olympians here and there that, that get together. And, you know, there's usually a coach practice every morning at 5. I'm, I'm in the pool. That's what, that's what keeps me recharged. Um, and then obviously, you know, family things and, and reading, uh, reading a lot.

Yeah. I smile because every once in a while I, I would, I do a certain, I do triathlon. And so depending what's coming up, I might like on the road join a master's thing, but I always get put in the lane of people who, who ever been in master's swimming, you know? You kind of, you go, there's fast lanes and there's slow lanes. And I always just look for the slow lane right away now. So that way my self-esteem isn't hurt.

We have a special triathlete lane. If you ever come to Stanford, you're welcome.

I'll have to look at that one. Hey, we, we talked about a lot of different things, uh, not just your very interesting backgrounds, but really your career and what was influential in your career. We talked about Microsoft, great things that you're all doing, how you innovate. How you're creative, about radiology more specifically, went into AI a bit, and then, uh, about natural curiosities and leadership, about opportunity costs as a way of looking at when do you quit, when do you start something new, when do you keep going. So we covered a lot of things, but is there something we missed or you want to double down on? Jonathan, I'll start with you, then I'll go to Matt.

Maybe this is stating the obvious, but we are in a really incredible time right now. And, you know, just think about what Matt was talking about before about what's changing in AI. Like we're moving in a place where pace is moving faster, but the models are becoming more general. And what's really powerful is they're becoming more accessible. So to me right now, I'll go back to the curiosity piece. If there's one ask I have for everybody I talk to, be curious in the space, because if there's one thing I know, the future of this technology is not going to be set by the people creating the technology. It's going to be set by the users of the technology.

Yeah. Well said, Matt.

Yeah, I think, I mean, definitely some similar sentiments. I also emphasize that for a long time, again, I've been in health AI for, I guess, over a decade and it was always like, well, just, you know, next year, you know, or next year. And, and, and really what, in addition to the astounding pace of capabilities and technology we're seeing, it's having impact today. Like DAX Copilot, something that would have been science fiction just a few years ago is in 500 health systems. Millions of, you know, draft notes have been, have been done already. Like it is, it is happening now. And so, you know, those, those folks are kind of like, oh, I have to wait for these cool new tools. They're available now and there'll be more in the future.

Yeah, my wife loves it. My wife's a clinician. She used to come, you know, pajama time, she'd come home from the clinic and it's been hours like working on notes. And then now she does it all voice and I have more time with my wife. So I appreciate it. Um, you all, you both are fabulous, uh, individuals and leaders and thank you for all that you do to sort of help transform our healthcare organizations, ultimately helping us save lives. I appreciate you taking the time. To be with me on Digital Voices.

Awesome. Thank you.

Thank you for listening to Digital Voices Podcast with Ed Marx. If you enjoyed this episode, subscribe on your preferred streaming service and leave a rating and review. And most importantly, thanks again for listening.