Scaling Innovation in Radiology: AI, Imaging Workflows, and Leadership Lessons from Ohad Arazi

You're sitting in our city, and like the you and the radiologist are like feeding off of each other because they're so into you and you're so into them because they like you are the kind of like us. They're like these tech geeks, right? That that are like they live in technology and they're so technology first and technology savvy. I find that to be hugely motivating, and I kind of always feel like we're sitting at the conference looking at each other thinking like I want your job. I want your job. Like we're we're acting like >> [laughter] >> I love that dynamic. >> Hello everyone and welcome to the Radiology Report podcast, where we are having conversations with the leaders transforming radiology today. You can find us on radiologyreportpodcast.com, Apple podcasts, Spotify, Google, or wherever you get your podcasts. >> [music] >> I'm your host, Daniel Arnold. Welcome to another episode of the Radiology Report podcast. Today, we are joined by Ohad Arazi. Ohad has enjoyed a dynamic 20-year career as an entrepreneur, investor, public company executive, and CEO in the digital health and medical technology space involved in multiple exits to strategic acquirers. Prior to joining Subtle in April 26, he most recently served as CEO and president of Clarius, an award-winning portable ultrasound and AI company, which was recognized by Fast Company as one of the world's most innovative companies in 2023. He led Clarius to profitability and scale with over 35,000 systems sold in 70 country countries. His career also included leadership roles at Zebra, Glooko, Teladoc McKesson, and Change Healthcare. Welcome to the podcast, Ohad. >> Welcome, Daniel. >> Yeah, I'm so happy to be here. >> Um you know, I love talking to founders, and I love talking to innovators in the AI and radiology space. So, you are just like right there in the intersection of all of my interests. Excited to get into all of it, but first, tell tell me about your background. How did you find your way Where Where did you grow up and how did you find your way into into our field? >> Yeah, I ended up in imaging through complete serendipity. Um, I grew up all over the world. My dad was a diplomat. So, I'm Israeli originally. And every 4 years of my life, we moved to a different country. And I grew up in some amazing cities. I lived in New York City as a kid, in Tokyo, in Seoul. Then we lived in Prague later. And I kind of always thought of myself as this human chameleon, you know, kind of wherever I go, I kind of morph into and kind of learn to fit in. And ultimately, traveling actually, I met a Vancouverite, Canadian woman, my wife, Maggie. Fell in love and we moved to Vancouver in 2006. So, I started my tech career in Israel, but we continued it here and just didn't know anything about medical imaging through complete serendipity. Ended up at a McKesson. McKesson had its medical imaging headquarters right here in Vancouver. They had acquired a Vancouver company called ALI Technologies. And then, in 2005, they acquired an Israeli company called Medcon, which was doing cardiology PACS. And they brought me on as kind of this quasi Canadian-Israeli chameleon to do the integration between the two. I always joke like, little did I know that actually integrating Israelis and Canadians was a lot easier than integrating radiologists and cardiologists, which is like I still haven't figured that out, you know, 20 years later. >> Amazing. >> that's how I got into imaging and and had a Yeah, 11-year career at McKesson. >> Two-part question on this. One, unrelated to radiology, what was the best place to grow up as a kid? Which of those cities was your Was your favorite? >> I really loved Tokyo. I lived in Tokyo from 1985 to 1990. And you know, basically like the summer we left, Japan entered a recession in you know, fall of 1990, which has been ongoing for 35 years. And so, when I lived in Japan, it was like a rocket ship, you know, it was like just so far ahead of the rest of the world coming from Israel, which at the time was before kind of the the big tech boom, was undergoing like massive inflation. It was just such a crazy dichotomy and then getting exposed to this amazing culture in Japan and growing up there. I went to an American school, but there were very few gaijins or foreigners in Japan back then and it was it was a real novelty. So I still have a very deep relationship with Japan, but I've loved every place I've lived, Daniel, honestly. I I love kind of the world. I love traveling. I get motivated by new places. Uh so I've got positive things to say about everywhere I grew up. >> There's a whole other podcast about uh uh the parallels to America cuz I've I'm just a fat and happy American who can only see America continuing to grow forever and ever and ever and sounds like you you've got a little more wisdom from your experience uh in Japan, but was McKesson at the time is this like printing film? It was at their core business? >> No, so so McKesson was a technology Yeah, McKesson now uh obviously one of the largest companies in the world, the largest healthcare company in the world, but at the time also had a big division of um of um digital health businesses, uh which it then divested in 2017 to Change Healthcare, which then IPO'd and was sold to Optum. So basically the Optum Change Healthcare PACS is originally the McKesson PACS. And so we had a big digital PACS business. We were one of the really first enterprise PACS providers. You know, when PACS started, it was really more than modality guys, right? You'd have like a GE modality and then GE would have like a workstation on which you read, you know, the GE images and you know, um each kind of modality player had its own standalone PACS. And then some film players like you suggested like introduced also some PACS systems, but very few actually came at it from an IT perspective. And McKesson based on that ALI uh acquisition was one of the first. And it was really amazing to be there kind of from the ground up really seeing Pax form and how kind of you're shifting from this more like siloed modality tool to what is truly an enterprise asset. So, it was incredible to come at it from that perspective and we had all the benefit of kind of being up in Canada, being somewhat removed from big corporate. I mean, McKesson was such a big company, you know, we were half a billion dollar business. That's not like even a rounding error on McKesson's, you know, shareholder report. And yet we had all the benefit of working for big corporate from the perspective of access and brand and just resources at our disposal to to grow. >> Yeah. Well, so from big company to startups, you've had your share of startup experiences. I remember when Zebra came out, I actually looked at some of their first vertebral compression algorithms when they were like kind of the first ones to it. And you know, now Subtle, what kind of drew you into Subtle? You you could kind of play in a lot of different places in radiology. What drew you to this opportunity? >> Yeah, I mean, first the transition from big corporate to startups, you know, I'm I'm not going back. I had an amazing big corporate career and I loved it and I learned so much. But, you know, I really love the dynamism of smaller companies. And like you said, like at Zebra, we reinvented ourselves, you know, moved from like triage tools to population health. And I think that journey generally taught me a lot about how do you find tangible proof points for use cases with AI? How do you really find value? How do you connect like this unmet need, you know, with your technology? And it's kind of not always been easy, I think, for the AI industry in imaging, right? And and we we can unpack that further later, but um when I got to know Subtle, that is really what stood out for me. Like, I really hadn't seen such a well-articulated value proposition that was so self-evident, so clear, really since Pax. You know, PACS was amazing. PACS, we would go in to a hospital that had literally three floors of film rooms. Uh basically printed film jackets being stored in the basement of the hospital, and we'd come in and say, "Oh, we can move all of that to like this hard drive or this server rack and ultimately to the cloud, right?" And um and and that's that's probably been the strongest value proposition I've seen in medical imaging to date. And then I got to know Greg and Enhao, the founders of Subtle. And they started to talk to me about this concept of you know, the upstream workflow. Like so much of what we do in imaging AI, in particular now, is really focused on the radiologist, on this interpretation workflow. And I think we're leaving a lot of you know, room on the table for innovation, for improvement in everything that happens upstream, before the image gets to the radiologist. And the core value prop of Subtle, certainly for MRI, is is acceleration. It's basically enabling you to take less time to acquire an image, which has a much better outcome for the patient. And for the provider has such a robust ROI because it adds slots to your existing equipment, right? If a If a typical MR in the US, Daniel, is seeing roughly 14 patient slots per day, per scanner per day, with a solution like ours, you're adding four to five slots per scanner per day. Using the same equipment, same capital footprint, same tech shift. And and that just increased throughput has so much pull through advantages, right? Higher revenue, shorter wait times, better experience for the patient, better quality images because you're getting substantially less motion artifacts. The patient's not as uncomfortable, and and that, you know, gets compounded, of course, with pediatrics or patients that have neurocognitive issues and are struggling to stay put. So, you know, just understanding that and saying there's so much to work with here, and this incredible technology the founders and the founding team had built. And now, kind of with the infusion of new capital with more of a market CEO, I just saw the opportunity to, you know, kind of pour rocket fuel on that, right? Is to basically now to add more commercial scale, more channel strategies, add more products. We've gone multimodal and built it much more into a platform play than this kind of amazing modality-centered innovation. >> Well, it it's really interesting to hear the tie-in from the strength of the value proposition of PACS to what drew you to settle. Um because I think one of the things that's really challenged the AI industry, probably broadly, not even outside of not just in radiology, just everyone knows the technology is incredible, but what is the value proposition and how do you tie that technology into products and services that ultimately drive um return in some way in a really clearly articulated way that the customers are coming to you uh as opposed to sort of the other way around. And so, when you look at, you know, you can picture this period of time 20 years ago where you've got closets exploding with film, that's a real operational challenge for these hospital systems and radiology practices. What are the biggest operational and business challenges facing radiologists today? >> Yeah, I mean, you know, PACS really introduced this explosion of digital data, right? But, you know, with all the innovation that's gone in, I would say that the imaging ecosystem still remains fragmented and like the way that the data's coming in is quite varied. There's this disconnect maybe between acquisition and interpretation. There's a kind of almost again two sides to workflow, like what happens upstream, what happens downstream. They're not really that well connected. We have a very diverse vendor ecosystem, especially on the acquisition side, which we haven't really normalized or standardized. Again, that's another thing PACS did. If we look downstream, right? I said that first PACS systems were modality workstations. Like Hologic like mammography, you would still read on like on a dedicated Hologic workstation until not that long ago, maybe 10 years ago, when it became integrated to PACS and kind of use of standard based approaches enable that interoperability, we don't really see that upstream. Upstream like the the protocols, the acquisition workflows are still very divergent. Um and and and and frankly the techs who are being asked to do much, you know, more and more and more are, you know, have very little tools to support them. So, I'd say that that lack of standardization and integration has really limited the efficiency of the overall radiology service. And and you said Daniel like we started to see massive adoption of, you know, AI moving kind of from research to real world deployment. That happened around the mid 2010s like 2015 like Zebras founded 2016. We're one of the first to get FDA clearance. We're the first to get two uh reimbursement codes. One of them was for vertebral compression fractures, which you noted. And so, these were all like important breakthroughs and AI really began to assist radiologists in, you know, triage, detection, structured reporting. But, I'd say that, you know, most commercial solutions in the space still today are really pretty narrow point solutions that are built for specific body parts or specific modalities or specific clinical tasks. And and even though like the conviction is strong like it makes sense and you see the demo you're like, yeah, I I could use that. The actual like value of measurable impact of that to the system, I think has not been well articulated, well measured. Um and and I think also financially like we've seen very few reimbursement codes for AI apps. And and the reason is that the payer is basically saying, listen, if this is a tool to make the radiologist more money, why should I pay for it? Right? It it does nothing for the system as a whole. And that's usually a good indication that like we still haven't cracked the full value proposition of kind of solving both this kind of variability and performance and workflow upstream and then the downstream interface into that like the metadata is not always aligned. And so, I still think we have >> I have a clarifying question for you. So, subtle and and uh we'll get into all of it, but but subtle just for a second, what does subtle do? I I know what it does, but just articulate it for everyone. And then, who do you sell to? Are you typically selling to radiologists who own their own equipment, or is it hospital systems like what is the breakdown look like when you look at imaging broadly? Bring us up to 2026. Who owns all this stuff? And cuz you brought up like who's paying for it? So, so who is paying for all this? >> Yeah, and and you know that's one of my biggest takeaways from doing this for 20 years is that um you know you really have to be very clear on like who is your user and who's your buyer. And sometimes we confound those two and we think they have the same problems and sometimes they do. Um but not always, right? So, so so subtle like what do we do? We started with you know we're really known for MRI acceleration, right? I talked about that up front. So, we solve the acquisition bottleneck by enabling any scanner regardless of vendor or vintage to acquire faster, higher quality scans and that's powered by our FDA cleared acceleration enhancement products. And that's basically uh AI solution overlaid to the existing equipment. We get a copy of the images at using DICOM format. So, we don't require like a dedicated interface to work in case base or any of the raw protocols of the modalities. We get it as if it was going to PACS. We sit as a network node, get a copy of the images, process them, and basically enable the provider to do a much shorter acquisition protocol and still get a higher quality image at the end. So, that you know improves patient comfort, reduces anxiety. It addresses motion related challenges, less callbacks. Um but what we've identified is yeah, there's so much more to do than than just speed, right? Because if we can enhance the image quality at source, kind of be that upstream PACS, so we will always enable sharper, cleaner scans at lower doses, reduce contrast. And that helps not only the acquisition side, but also the downstream side, right? It ensures that PACS, the radiologists downstream, any AI tools that receive our our data are going to get high-quality data. That's going to improve accuracy, reduce variability. It's going to support, you know, better clinical decisions. We're actually we have a study coming out now where we're measuring about a 6% improvement to the sensitivity and specificity of downstream AI apps when you measure before and after subtle image enhancement. So, not only does it help again the acquisition, but it really helps everything that comes after that in the downstream. And the beauty of it is that it doesn't really introduce any friction. There's no UI. We're a We're middleware, right? That basically enhances the image and serves it back up to the existing tools that are used by radiologists. >> So, okay, so the users are few users, right? Because you've got the radiologists, you've got the rad techs, you've got the patients themselves benefiting, but who's the buyer? >> Exactly. So, so when we started, we looked principally at areas where the buyer and the user was one. And that's basically imaging centers who their business, their revenue is basically scanning, right? And so, they're doing the technical fee for the acquisition and the professional fee for the interpretation. And that's really how we built our brand. We we work with some of the largest imaging center chains in the world, you know, RadNet, Rad Partners. These are all of our biggest customers. And in in them, in many cases, there's very strong alignment between the user and the buyer because their practice is owned maybe private equity backed, but generally they're owned by radiologists who are, you know, responsible for the top-line performer, but also for the quality output of what they do. Um and and there we would probably work with two main stakeholders. One is kind of the radiology line manager, someone that owns the P&L and knows exactly what is our yield per scanner, what is our throughput, what is our turnaround time for images. And then we'd also work with the chief of radiology that says, "Hey, I'm I I like the image quality, I like the improvement, I'm comfortable with this." And there's a lot of trust that has to be built there when you're applying AI to change an image that they're going to put their name behind when they're interpreting. Um I'd say in the last, you know, 6, 9, 12 months we've been moving much more so into the IDNs and the hospital chains. We've got some very big customers, Kaiser, New York Presbyterian, LifePoint, kind of, you know, very big deployments across large chains. And with these customers, you you sometimes see some differentiation where, you know, somebody owns kind of the business side and maybe also oversees the rad techs and the kind of acquisition throughput. And then separately, you might have a private practice radiology group that's doing the reading and they're more, you know, um kind of applying on image quality and general comfort with the workflow. But I'd say that this kind of equally serves as a user, both the rad techs, which now get to do a much shorter acquisition protocol, much simpler, more AI tools to assist with what they do, as well as the radiologists. And the buyer will typically be someone that owns a proforma and is also responsible for overall patient satisfaction like is getting measured on wait times, for example, right? It you know, did we were we able to reduce the backlog? >> So, are you ever uh hearing [snorts] from folks, "We don't want more patients right now. We can't take them. We don't want more throughput. We we can't handle higher volumes." >> I don't really see that in MR, uh quite honestly. MR just always has inherent wait times. Um and and generally like patient acquisition isn't just about throughput, it's also about like comfort of the patient. You get a lot of quality issues stem from longer exams. You really see this that, you know, like if you're running, let's say, six sequences in an exam, you know, the first one, two, or three, there's very little motion artifacts. And then by sequence four, five, six, patient's a a less comfortable, they're hot, you know, they're moving. And so there's a lot quality implications on reduced acquisition time, but I don't really see providers, customers coming back and saying, "Help us decrease the backlog, make patients wait less." >> [laughter] >> With other modalities though, there are differences, right? So like I'd say PET is another modality where, you know, we have uh we have we have a PET product. We actually just obtained FDA clearance a week ago, we just announced it this week, where we have a new PET solution. So uh PET HD or HD PET is our new new PET product. And you know, for certain instances, for like for whole body scans, it can reduce the acquisition time by 90%. Right? And so with PET, it isn't always about throughput because actually not all PET sites have a big backlog of patients. Major cancer centers do, and for them it's very important to remove to improve throughput, but a lot of that is also patient comfort, image quality, a lot of that is radiation exposure, right? Because you're basically having this radio-radioisotope flow through your body, and if you can do that for less time, it's safer for the patient, it's safer for the care team, um you know, the ability to do an image enhancement on a wide variety of of radiotracers is also very important. So there's other factors that aren't just time. And then finally, we're about to launch our CT product. Um and that one isn't about time at all because CT >> Yeah, I was going to say CT's quick. >> 15 seconds, you're in and out, right? But but with CT, there's a lot of other implications of being able to use less power, so less radiation exposure. It also allows us to better articulate the image, like do better segmentations by any downstream AI with a much lower contrast dose. And And that's usually an issue when you have like patients with high BMI that, you know, kind of the standard dose of contrast doesn't work well, a variety of pediatrics or like patients that can't tolerate dose well. So it's not only about throughput, um but I'd say for MR, that is the crux of it and it's very like rare or I haven't really found an MR site that says we don't want faster throughput. With the other modalities, it isn't only about speed but rather about just improving the overall workflow and the overall experience for the patient. >> So, I got to say you you know, you've you've come up to speed fast. As we're talking, it's uh a few days before your official start date. So, I don't even know how long you've worked at Subtle, but you talk like someone who's been in the in the seat for a while. You're you're joining a company that has accomplished quite a bit in the decade, you know, since it's founding. Um y'all have raised $85 million. You're live in 20 countries. You're live on 1,300 scanners. Um you're you're you're no longer super small startup. You've got, you know, products in market and delivering value, executing on a lot of cylinders. Um I've interviewed a ton of founders. I've actually never interviewed a hired CEO. Uh and so, what's it like taking over for a you know, venture-funded, founder-led company? What what what's that transition like? >> That's an awesome question, Daniel, and I I I honestly I wish it was asked more often because I think many startups get to the point where they're contemplating now a transition to say, you know, do we need more help at the CEO level to help the company achieve its full potential, kind of cross the next chasm? And I think there has to be a lot of sobriety around like this isn't easy. This is my fourth founder to hired gun CEO. For whatever reason, by the way, maybe I need to unpack that with you on a different [laughter] podcast or or like maybe maybe I need a shrink actually to like why have I not wanted to be a founder? I haven't felt ready to do that myself, but but like my comfort, my passion is like to come in when there's like a kernel of something really good and now I can believe that with like more capital and maybe my energy and kind of my market expertise, we can again kind of really fan that flame and make it bigger and stronger. I think it takes first of all a very special founder to understand that and to really want that. Sometimes maybe founders think they want that or maybe the board wants that but the founder isn't really on side and I think that I would only I only like to come in where I've had a chance to date before we get married and then I really have conviction that the founders and I are on the same page that they really want to make They really feel like this is what's best for the company and for them the realization of their vision as founders isn't necessarily being the CEO at the time of an exit or an IPO but rather is building, you know, the this kernel of something amazing and then working with someone else to help it take it to the next level. And and so I think that kind of one ingredient there is dating before you get married. And so I was supporting uh Settle's fundraise for a long time actually as an advisor starting last year. Uh actually officially into the seed in in mid-April and yeah, we're just about to announce it um in in the coming days as we closed our financing. Um I'd say another takeaway is invest a lot in trust over communication, right? Because at some level almost any decision you make as a hired gun CEO is a departure from a prior decision the founders have made. That's actually what you're paid to do, right? So like you have to be very open about that and talk about that and here's why we're doing it and and kind of communicate that very very well. Um I think also when you move from a founder-led uh company to one that now maybe is infused with more capital, starts to grow even more rapidly and brings on, you know, outside help at management. I'd say kind of two things I've I've noticed. One thing is you're moving from an environment where everybody was doing multiple jobs. They had like eight balls in the air, they're juggling them. And you know, they all want more bench strength but sometimes it's not easy to give up control. And so like doing that thoughtfully and like how do we really kind of create force multipliers, but still keep everyone comfortable because they've led, you know, they've they've been wearing so many different hats. And now as you add more executives to the team, you know, there's a little bit more of kind of like separation of church and state, defining roles, and that's just not easy and it has to be done thoughtfully. And again, communication and and trust is key. Um I'd say a second takeaway is really kind of introducing product management as a functional area. You know, founder-led companies, the founder's the product managers, right? They have this like compelling vision that's like burning in them, right? And they're like they have a hammer and they're finding nails and they did an incredible job of like finding nails they can hammer in. And then at some point I think you have to help the company reverse that. See, we've gotten really far with like a solution-oriented approach. Now, how do we transition more to a problem-led approach? How do we really focus on like where we play? Where's the really profound unmet need that our technology can solve? And how do we explain that unmet need and kind of create that connection? That's what product market fit is. And so like investing in product management is often also a very key ingredient and one of the first things I'd like to do when I come in as a hired gun CEO to a founder-led company. And then the third is I invest a lot in building out a framework for strategy to execution. Because like, you know, really being able to take the strategy from something that is kind of more visceral and more conviction-based to helping the team really understand it. And I I use I use something called the one-page plan. Like I I condense our entire strategy to a page. Yes, the font's not very big, but it does fit on a single page. And and I really make sure that everybody understands that strategy. Because that helps us to tie in strategy to execution. There's this really good uh saying by um by by Morris Chang, the the chair of of Taiwan Semiconductors, he said, "Without strategy, execution is aimless. Without execution, strategy is useless." And that kind of yin and the yang of strategy to execution is so important, and that's really again one of the first main jobs because like a founder-led company was usually just like executing like its hair is on fire, and now can we step back to just like let's just make sure we also capture what is our strategy? Does everybody understand it? And now how do we ensure that all the execution steps we're taking are tied to that strategy? So yeah, those are kind of three main >> This is a therapy session for me, too, because I don't know if you know this, but I sold my founder-led company last year. Um so not the exact same type of transition, but you know, went from being the founder in charge. Now we've got, you know, I'm the president, we have a CEO, and a lot of the the things you're describing. And I guess the question that I have is everything you say sounds right to me. You know, I went to business school, it all makes sense, I'm nodding along. But the other thing that I hear in the back of my head is, "Ah, we're going to go so slow. You know, we're going to get slowed down as we as we scale up." And how do you keep the velocity, especially in such a dynamic and and frankly competitive field? Uh you know, you are in a very, very competitive field. Um you may be the market leader in the sort of independent space, but all the scanner companies are are are doing things. Um you know, a lot of private groups are pretty smart and can do things on their own. Like where, you know, the the vector of competition in AI is moving so fast. And so how do you keep The one s- thing founders can do is just go, right? They hear something, they they they've got it in their gut that this is the right move, and they just go. And so how do you how do you keep that speed up? Yeah, first of all, you know, when you when you talk about like a strategy to execution framework, it sounds like, "Oh, that that feels a little heavy." Actually, I think the the is to enable us to move very quickly, but to do it based on priorities. Because you know, I always say like the CEO makes a small fraction of all the decisions in the company. The executive team, very small fraction. Most decisions are made by frontline individuals, right? Someone who's writing code, who is compiling a model, who is doing an implementation on site, who's making a selling decision or supporting a customer. And the better they understand the strategy, the more direction they they understand what we're doing and what matters most to the company, the better decisions they make. And so actually kind of like moving the strategy from like this kind of tome of knowledge that is maybe shared in like board meetings or investor decks, but like most of the team doesn't see, is actually a very important step towards acceleration. Uh but but I but I I I I mean I I I I think you're right. I I it Of course like, you know, running around with your hair on fire is very effective for certain things. And and we have to keep doing that in order to beat the OEMs. At the end of the day, you know, we have to bet on the fact that we you know, software eats hardware, that we can move faster than GE, Siemens, Philips. We add a lot of value to them as well, but of course it's competition. We also compete with them. And so to me, a lot of it is to do like we keep the same dynamic. And and that's a really big part of like being a hired gun CEO is like how do you change, but also stay the same, right? Not add like layers of bureaucracy, but actually do it in a way that you're keeping a lot of the cultural velocity that exists and the real kind of innovation like like small I innovation, not big I innovation, like day-to-day innovation, like here's problem, let's find eight ways around it to get to the solution, and that is so well ingrained in startups and founded mid companies. Keep that and just do it in a way that is driven though by a common set of priorities, so that we're more focused in what we do. Um and and that's by the way a big onus when you raise capital. Because when you raise capital, one of your first steps is you become defocused. You're doing a lot of different things cuz for the first time you don't have that kind of real forcing function of not having any money. >> Yeah. >> And and that's dangerous. I think that's my number one accountability, but I've done this four times, so I think I actually this is probably where I can add the most value is how to take that next step without losing the momentum that we built. But uh you know, check in with me in a year and let's talk about >> I I look I look forward I look forward to that. So so what You brought up uh you know, the competition with the OEMs, co-opetition, you know, you got to work together. What What differentiates you? Cuz you know, my understanding is that all the OEMs sell accelerated scanning. Um maybe even embed it in in some instances. I don't know exactly the difference of of features or pricing. Kind of talk us through, you know, somebody's buying a new scanner and they've got to make these decisions. Um how do you cut through that? >> Yeah, I I I would say um I think in anywhere in imaging certainly there's a device component and there's a software component and they have to be complimentary. Again, maybe there's some co-opetition, but I mean there's amazing innovations that are happening at the scanner level and those will continue to happen and we'd love to see them continue happening. Um but you know, they're always coming at it from the hardware lens, right? Ultimately all you know, they're looking even when they have like OEM software that is embedded or overlaid onto the scanner, it's for that specific scanner and it's ultimately geared towards selling a hardware upgrade, right? It's not available on a legacy system. That's the last thing that a GE or a Philips or Siemens want, right? Because they live by the hardware referral cycles, right? They they want to convince the customers to say, "Well, your 10 or 15-year-old scanner, you have to upgrade it because to the latest and greatest uh in order to get all these efficiencies. Uh the second thing is that they always work in a multi-vendor environment. I mean, I've never been to a provider site that only has GE, that only has Siemens, right? They they're always going to be multi-vendor, multi-modal. Um I I keep going back to this, very similar to how we saw PACS evolve. Because again, the value proposition was in normalizing and harmonizing all of the interpretation workflow. And I think at some level, what we're trying to do is to normalize and harmonize all of the acquisition workflow. Part of that is acceleration, where we work with what the OEMs do. They have very good machines, and they have some of their own software. We overlay that. Uh so, on older scanners, the, you know, the improvement in um in image enhancement and speed can be 70, 80% or more. For newer scanners with AI embedded, it's maybe 30%, but there's always going to be an addition there. And then second is we standardize the way that it happens, because, you know, acceleration, even image enhancement isn't enough. You have to structure the data, correlate the metadata, um reduce motion across the board in a standardized way, standardize the way that you're doing reformats across the vendors, and then feed that data into PACS and RIS for it to be read. And so, I kind of think of us as like we're an enterprise asset that helps the modality players to, you know, get even better efficiency from the innovations that they've embedded inside the devices themselves. >> Yeah. It's I've got my popcorn out. Uh I'm I'm enjoying watching it, but it makes sense, espe- especially for the older systems, for extending the life and integrating, you know, all of the systems into a standard uh workflow, right? Which is like, I just want images to look a certain way, regardless of where I'm acquiring them from. And especially with these uh the consolidation you've seen on the radiology services side, where you might have a 100-person radiology practice reading across how many hospital systems and how many different types and you've got new equipment and old equipment and you've got so many different vendors. And so, um the trends that that make sense for unified PACS, you know, potentially apply um in what you're in what you're describing. >> Exactly. Very federated environment. I I I think our biggest competition, Daniel, is not the OEMs. Our biggest competition is the status quo. Our biggest competition is uh is a hospital in Peoria, Illinois, uh who has a 70-year-old scanner, has facing big backlogs for their uh patients, and is contemplating a $2.5 million capital upgrade to buy the latest and greatest scanner, not knowing that there's third-party overlay solutions that can standardize their inputs, that can enable them to get 60% more efficient in their acquisition, improve the quality of life for their techs, and do it at a fraction of the cost of the maintenance on that scanner. Forget even the capital expense, right? So, I think that that's actually like getting the word out there is the biggest challenge. And and and and to me, we're still there's so much to be done in that white space between uh the acquisition and the interpretation. Like you said, it's a very um kind of dynamic competitive environment. I I think downstream AI is very, very competitive, right? There's like so many companies, hundreds, right, of like different widgets. And then upstream on the hardware, there's a lot of competition, and they're all kind of trying to have better mousetraps and very, very good tools and a lot of innovation. Actually, in between, there's not a lot happening. Like we've really not applied a lot of AI, a lot of software, to make the lives of the techs better, to help address all the reformatting they do, kind of the the We're asking more and more of them. And then we we make a $2.5 million purchase on a new scanner, but the throughput and the quality of that scanner has a glass ceiling, which is the proficiency of the tech. And it could be a very good tech that like optimized the value of that capital investment. And it could be a tech that's under trained or or or is disgruntled or overworked or whatever it is and has very little tools to do their job better. And and so that area actually is an area where I think there's not a lot of competitive tension and we're really trying to kind of be that get into that white space and kind of fill the void of everything that happens between the modality and the PACS. >> You brought up subtle pet. Congrats on the FDA clearance and subtle uh HD for for CT uh as well. You you also hinted at what's happening with contrast reduction. I've read a little bit about subtle gad. What's What is that and what where is that a priority? Is that going to happen? >> Yeah, so I mean we think about the sum of what happens upstream and the patient experience. There's a speed element, there's a contrast element, there's a radiation exposure. We've had a lot of success like you you talked about subtle gad, which is an upcoming product and also we already have a product in the contrast field that's already generally available and has FDA clearance called Implify, which is a collaboration with Bracco. Um and subtle gad is a collaboration with Bayer where they're both focused on basically creating the kind of AI software um uh companion to a molecule, right? To a contrast agent because it now enables the contrast company, the biopharmaceutical, to have much better visualization with using less contrast and less dose and that's obviously harmful or can be harmful to the patient. It has a lot of complexities to it, right? So if we can use less contrast and get better vis- visualization results by really pairing and kind of closely aligning the performance of the AI to that specific molecule, it's uh it's very interesting. And and and I'll tell you this is an area where I think we're just starting to see um biopharma companies take much stronger stake in AI and kind of understanding that AI can often be the key to unlocking the value of their drug. We've seen that of course in drug discovery, but I even mean like for like drugs in the market. Uh I can tell you like we in my prior company Clarius, we had a big collaboration with Novartis, where we were kind of pairing software tools that enabled a drug manufacturer to like better visualize the condition and kind of lower the threshold, lower the barrier to make it easier for clinicians to prescribe their medication, which is life-changing if you can, you know, prove that the patient really needs it by having the right biomarkers measured. And so I think we'll see a lot more of that. It's a really exciting field, and it's an area for us at Subtle that we've been focused on. And you mentioned PET and CT, uh those aren't intrinsically tied necessarily to a partnership with a biopharma company, but they're very much focused on decreasing radiation exposure for the patient, whether it's by using less contrast, using less energy in CT, having a shorter scan time in PET, so there's less time where this radiotracer is coursing through your body, and potentially creating harmful radiation for you and for the care team that's that's imaging you. >> So fascinating. I I I I don't think uh you know, for folks that have read about Subtle like myself, I you know, over the years, I did that wasn't an area where I understood you guys playing, but it makes a ton of sense in that intersection between um pharmaceuticals and imaging is going to only continue to grow. Uh and and so it's really exciting science and partnership at watch that uh >> It is. >> uh space. So, you know, last question, the the you've had a pretty cool career. Obviously, you couldn't uh have predicted it when you were, you know, figure taking your first job at McKesson, but you know, I think radiology is an interesting field. Some some Some people are pretty down on it. I'm pretty up on it. Um but I'm a kind of up guy, so I don't know. You can take that with a grain of salt. But like, what advice do you have for folks who are thinking about healthcare, thinking about radiology, thinking about technology, you know, how do they have an impact the way that you've had? >> First of all, I share your enthusiasm on imaging. I love imaging. I love how clinical it is. I love like the direct impact to patient care. I've worked in other fields in in medical technologies and you know, letting EMR company for a while, but you know, I I I I really really love clinical products. I also love working with radiologists, you know, Um I've attended a lot of trade shows over the year. Like my last company was a point of care ultrasound company and so we had AI and and a device, but we didn't sell them to radiology. We point of care is mostly in other clinical spaces, right? So I attended a lot of other shows, you know, like you go to um emergency medicine shows, anesthesiology shows, MSK, uh or orthopedics, um uh aesthetics and you know, those clinicians, they're happy to work with you, but like to some level like they're so focused on like hands-on the patient, like technology is a little bit of a distraction. They kind of put up with you, but uh you know, they don't need you the same way. Then you go to RSNA and you're sitting at RSNA and like the you and the radiologists are like feeding off of each other because they're so into you and you're so into them because they like you or they're kind of like us. They're like these tech geeks, right? That that are like they live in technology and they're so technology first and technology savvy. I find that to be hugely motivating and I kind of always feel like we're sitting at the conference looking at each other thinking like I want your job. I want your job. Like we're we're >> [laughter] >> like I love that dynamic. So uh I love the field. You know, like you had me at hello. Um my advice, you know, uh I'm just giving this talk about like uh minding the gap, like what your product does versus what the user needs. Like and the gap between that. I talked a bit about that. And I still think like looking back at my career, that's one of the most challenging things. It's like we tend to be as technologists, we tend to be solution-oriented. We've got like an idea and then we've got a hammer and we're looking for nails, right? And so, we really have to be um problem-oriented. We really have to like understand what is the unmet need. How do we know our users so well? There's I always use this There's a great um uh Christopher Nolan movie that I love called Inception. It's uh Leonardo DiCaprio and uh he's he's kind of like this con artist. Uh it's a science fiction movie that like goes deep into like the He has this kind of crazy technology that he can go deep into the persona, the cognition of his like targets like he goes so deep into that he plants an idea in their mind, like a con in their mind, and that like they he comes out and they think it was their idea. And and somehow I think that's always like the job of really great product management. Not not the planting a con, but like that we have to know our users so well. We have to like go into their mind that when we show them our technology, they're like, "That was my idea. I always wanted that." Whenever I see that, I'm like, "Yes, I've done my job." And and so like I think that's how you become more intuitive, more aligned, is by really understanding your user and having them firmly believe that like whatever you came up with, it was their idea, because it really was. And you just did a good job of understanding what they wanted and mining it and and helping them to helping to deliver that. So yeah, that that's my advice. Watch Inception uh and come at it from the perspective of like DiCaprio, while he's a con artist, is also a really good product manager. >> Well, I've got a long flight coming up, so I'm going to put that one on the iPad. Ohad, really wonderful to talk to you. Congratulations on the new role. We will be rooting for you and excited to see uh what the next few years bring. >> Awesome, Daniel. Thanks so much. Yeah, this was a lot of fun. >> Thanks for listening to this episode of the Radiology Report podcast. Be sure to visit us at the Radiology Report podcast.com or subscribe to the show wherever you get your podcasts. So, join us for our next episode. We are always looking for great guests. If you have someone you'd like to hear on the show, please get in touch with us online.