Episode 37 – Rethinking Radiology: Where AI Ends and the Vet Begins
Subscribe via your favourite podcast platform: Spotify | Apple Podcasts | Google Podcasts | Anchor | Breaker | Pocket Casts | RadioPublic
In this week’s episode:
- AI Is a Screening Tool – Not a Diagnosis
AI in veterinary practice should be treated as a preliminary aid, not a definitive diagnostic. It lacks access to clinical signs, history, and lab work – so final decisions must always rest with the vet’s training and clinical judgment. Technology supports, but doesn’t substitute, expertise. - Smarter Questions, Better Outcomes
By narrowing down the clinical question before consulting a specialist, AI can streamline workflows and reduce unnecessary diagnostics. For example, if AI highlights a potential intestinal obstruction, it helps focus the follow-up conversation and action plan – saving time and money. - Fit the Tool to the Vet, Not the Other Way Around
AI isn’t right for everyone. Its value depends on the vet’s comfort level with interpreting diagnostics. Practices should use AI where it enhances confidence – not where it creates confusion or increases reliance. - Ethical AI Demands Transparency
AI systems must be transparent about how data is used and how reliable results are. Independent validation, published accuracy metrics, and collaborative development with radiologists ensure responsible, ethical use in practice.
This episode explores how AI can responsibly enhance veterinary diagnostics by supporting clinical expertise, improving workflow efficiency, and promoting ethical, transparent use in practice.
Additional Guest Spotlights
- Recommend Resource: This week’s expert pick comes from Charlie Barton, who swears by ChatGPT, Microsoft 365, and a reliable laptop – his go-to combo for boosting productivity, brainstorming faster, and keeping everything seamlessly online.
- Next Episode Sneak Peak: Next time, we’re joined by Liz Barton – VetCT’s Head of Communications and co-founder of WellVet – for a powerful conversation on building a more sustainable, emotionally intelligent veterinary profession, tackling everything from retention and parenting in practice to dismantling hero culture and embracing flexibility in the age of AI.
Show Notes
- Out every other week on your favourite podcast platform.
- Presented by Jack Peploe: Veterinary IT Expert, Certified Ethical Hacker, CEO of Veterinary IT Services and dog Dad to the adorable Puffin.
- Eric Goldman is President of Vetology Innovations, a veterinary teleradiology company dedicated to supporting clinical diagnostic decision-making through a combination of human expertise and artificial intelligence. He leads operations, product development, marketing, and strategic partnerships, working closely with veterinary professionals to align Vetology’s service and platform with real-world clinical needs.
- Vetology is built around the goal of supporting the radiology demands of modern veterinary practices, regardless of size, specialty, or case complexity. The team includes 12 board-certified veterinary radiologists, two cardiologists, and a dental specialist. In parallel, Vetology has developed a proprietary, in-house AI platform that provides rapid screening results for canine and feline radiographs. These two services – AI and teleradiology – can be used together or independently, depending on the case and the needs of the practice.
- Eric is a strong advocate for the integration of AI as a clinical support tool – not a replacement for veterinary expertise – and is passionate about helping teams work more efficiently while maintaining the highest standard of patient care.
- Find out more about Vetology AI here!
Transcription
Jack Peploe:
Coming up on modern veterinary practice,
Eric Goldman:
One of the things that we’re careful about in Vetology is we call AI a screening result. We don’t call it a diagnostic. The reality is that software doesn’t know anything about clinical signs, doesn’t know anything about prior visits, doesn’t know anything about blood work, and sometimes the vet techs will put in the wrong breed, right? They won’t update it, they put in the wrong weight. All those things factor into the diagnostic chain, right? And so the point that what we know is computers are consistent, right? So if you take a well-positioned, well taken radiograph, complete orthogonal views, you should be able to get a consistent result. That result has to tie back to something that you’re seeing and something that,
Jack Peploe:
Welcome to the Modern Veterinary Practice Podcast. I’m your host and veterinary IT expert, Jack Peploe. In this episode, I’ll be welcoming Eric Goldman to the podcast who will talk to us about how AI can support not replace clinical expertise in veterinary medicine. We explore what AI really means for vets on the ground, the importance of using technology responsibly, and how Vetology is focusing on complimenting not competing with veterinary professionals. Whether you’re excited or skeptical about ai, this conversation will bring clarity to the practical and ethical considerations shaping its role in practice.
Eric Goldman:
My name’s Eric Goldman. I’m president of Vetology. My journey in technology, I’ve been in technology for over 25 years. The really nice part of my career is that I’ve had the opportunity to be exposed to lots of businesses and also a lot of great people along the way. I started my career in government actually putting in tax collection systems and integrated tax systems, and from there I went into trading applications. And then I ended up working for a consulting firm that did a lot of Fortune 500 customers. One of the interesting things about that particular job, it was a company called SAPN, and this was the early two thousands where a lot of technology projects were having huge overruns. And so they came out with a model where they did fixed price, fixed time technology efforts. And the interesting thing about that thing about this project was you really had to focus on the customer and what they wanted and the value that they got.
And what you ended up doing was doing these very short engagements that had a very clear agenda. You would work through a solution, you would deliver it, you would then confirm that we’re adding the right value, and then you’d go on to the next piece. So it was very iterative. I mentioned it because that whole approach stuck with me for my career. Going on further from that, and then I went into the banking business. I did a mobile app development for a global banking company that had two large banks throughout the US Citizens Bank in the East Coast and chartered one bank in the Midwest. I did healthcare work, particularly Blue Cross and Blue Shield, and then I eventually was in the telecom space. So lots of different areas. I hit a point, so one of the things about me is when it becomes a Monday and you start feeling dreading like, oh, I don’t want to go into the office. That’s when I’m always like, okay, I need to start thinking about doing something different.
And I’ve always been that way. And so I kind of hit that point in the telecom space and I reached out to my network and I said, Hey, I’ve done lots of things in technology. It’s very easy for me to find a job, tell me who’s doing something really cool, something different, and also has a society component to it. I really want to figure out something more that I can contribute more. And through a mutual contact, actually going back to my first job in the government in the tax space, building tax systems connected me with Dr. Seth Wallach. It was actually someone that he went to high school with. And so I met Dr. Seth Wallak. He’s a boarded radiologist out of San Diego. He’s had a telemedicine business since around 2009. And he literally just explained to me his ideas. He’s always had a technology, Ben, and to be honest, we spent about a year every weekend just talking and him telling me his ideas. I started to do some research. At one point, I literally went down to my veterinarian down the street, and I spent a day just watching him interact with pet parents, taking the animals in the back room, having the vet techs position the animals for x-rays, getting a result, talking with them as he’s looking at the x-rays and then deciding his course of action, excuse me, with the pet parent.
And then I came back from that, sat down with Seth a little bit more, got the experience of what’s happening on the consulting side when either the radiologist, the cardiologists, the dental specialists, answer those questions. And then the last peaks was I actually paid for myself to go to VMX, which is a huge conference in the us. And I walked the hallways and literally I was blown away at the size of the industry. And I came back from that and I said, Seth, we’re all in. Let’s do this.
Jack Peploe:
That’s amazing, Eric. Well, what a background. I mean, it’s obviously, first of all, great to have you on the Modern Veterinary Practice podcast. And obviously it’s fantastic to hear about your extremely unique background going from government to Fortune 500 technology projects to multinational bank mergers. And yet what really stands out to me is your approach that technology should always serve people, not the other way around, which I love. Now at ology ai, you brought that same philosophy into veterinary medicine, which is great, focusing on how AI can support but not replace clinical expertise. And I think there’s a lot of conversation around AI and vet med right now, some excitement, some fear, and a lot of misunderstanding. So today, if it’s possible, I’d love to explore what AI actually means for veterinary professionals, how it can improve decision-making practice, and how we balance technology with the human element that makes veterinary medicine so special. Does that sound good?
Eric Goldman:
That does.
Jack Peploe:
Awesome. Alright. Well look, I suppose my first question is AI in veterinary medicine can be a polarising topic. Some see it as like a game changer, especially people like me, while others fear it will replace jobs. How do you position AI in a way that compliments rather than competes with veterinary professionals?
Eric Goldman:
Yeah, I think as you said, AI is everywhere these days and a lot of people talk about AI and diagnostics. One of the things for me is whether it’s AI or whether it’s telemedicine, that’s a transaction. It’s not about the transaction, it’s about the veterinarian on the other end with a patient trying to get information to decide the course of action period. That’s all it’s, it’s not about the underlying technology. It’s about serving that person where their skill sets are adding to that gap, giving them a set of information and then helping them drive that treatment. One of the things that we’re careful about in Vetology is we call AI a screening result. We don’t call it a diagnostic.
The reality is, is that software doesn’t know anything about clinical signs, doesn’t know anything about prior visits, doesn’t know anything about blood work, and sometimes the vet techs will put in the wrong breed, they won’t update it, they put in the wrong weight. All those things factor into the diagnostic chain. And so the point that what we know is computers are consistent. So if you take a well-positioned, well taken radiograph, complete orthogonal views, you should be able to get a consistent result. That result has to tie back to something that you’re seeing and something that, one of the things that we put specifically on the AI report is trust your training. And so it has to align with those things. If it does, great, you know what I was thinking? It’s congestive heart failure. It has those clinical signs and AI came back with positive for CHF.
Great. Let’s start to put the animal on this medication. The other way to think about it is to have a more focused question to the consultant, whether that’s sending it to a specialty hospital or using a telemedicine service, whoever they desire. So one of the things, at least in the us and I will say firsthand, I don’t profess to knowing fully well the UK market just yet, is the major players, when you send in for a consult, they charge by the number of images you take. So they have a certain price, one through seven, eight through 11, and then above and so forth. At Vetology, we don’t care the number of images you take. What we’re trying to do with the AI and the telemedicine is to get a more focused question. Where radiology is inefficient is you send 12 radiographs and you say, the dog’s lethargic.
What is a radiologist doing? They’re hunting and picking through a set of images to see what’s going on. Instead, if you took a thorax and an abdomen and say, the thorax came back as normal, but the abdomen came back and it said a small intestinal obstruction. And the vet’s like, okay, do I cut? Don’t cut. You then can have a more focused question and the radiologist knows where to look and understands how that aligns with the clinical signs. And so we allow for single cavity reads, we’re focused on that question. They tend to be far less expensive than the other providers. And so the business model, I call it human and AI, go better together. It’s the idea that you get an AI report, can it help you in your treatment plan if it still, you still have further questions, do an AI reed, but have a more focused question and don’t build up the diagnostic costs.
Jack Peploe:
No, that makes sense. Now, looking at scepticism around AI, there’s obviously often resistance to new technology, especially in this industry is what I’ve noticed initially, and also especially on industries that relies so heavily on human expertise. So what have you found to be the biggest misconceptions about AI in veterinary medicine and how do you guys address them?
Eric Goldman:
Well, as I said, what I’ve seen is there’s a lot of talk, they use the term AI diagnostics, and I really feel that it’s important to set the context about what it can do. And so one of the things that I do when we get inquiries, I talk to a lot of dbms, and if the VM has no experience reading X-rays, right? They went through a programme and they had three weeks of training, and then they went on because they have to learn so many different things, and they generally have this view that you send it all out for radiology. I’m like, okay, the AI product is not right for you. If you think about, again, my CHF example, AI can calculate a vertical heart score VHS score. Just because that’s in large doesn’t mean there’s congestive heart failures taking place. You don’t want to force people into getting one metric and then saying, okay, we need to do an ultrasound, or We need to set up an appointment with a cardiologist.
All you’re doing is building up unnecessary diagnostic costs, and so the product has to be right. And I think one of the misconceptions in the market is that some companies or a lot of talk around AI in the veterinary space, they promote it as a diagnostic. I don’t believe the technology’s there yet. Maybe five years we’re having this discussion, and it’s evolved a lot further. But I’m very much a pragmatic approach as to where it is today, and I think it should be used as a screening result. I think it has to aligned with the DVM, with their level of training, with their comfort level and what they’re trying to do with the technology. And if it fits, great. If it doesn’t, don’t use it. And it’s one of the things that in our company, we don’t force people to use the ai. So yes, we’re an AI company, we’re a telemedicine practice. We’re a practice of 12 boarded radiologists, two cardiologists and a dental specialist. When I talk to the, I always try to figure out if the software is right and if the software’s not right, I’m like, don’t use it.
Jack Peploe:
No, that makes sense. Now, looking back at how radiology and diagnostics have evolved over the past decade, what do you think has been the biggest shift and where do you see things going in say, the next five years?
Eric Goldman:
Well, I think that a lot of the radiologists have been sceptical on the technology, and they’re very vocal about it, and I think they should be. One of the things that we do is, I call it democratising the tech. And what that means is that we use AI as a prelim for our radiologists, and we want other radiologists to actually use AI as a prelim for them. They don’t have to just read for ology. They should be using the technology within their own business, their own relationships, and we’ll make it available. And then they should contribute to the technology, meaning this tech should not be driven by technologists, it should be driven by radiologists. It should be driven by the doctors, and they should help it evolve, and they should help it and evolve in a way that’s going to help them in their relationship so that they work with a bunch of DVMs that normally have a comfort level sending cases to them.
It should go to them. We shouldn’t be coming in and saying, okay, well, if you use this tech, you have to use ology. That’s how the technology’s going to evolve in the right way, democratise the tech, get more doctors to contribute to it, use it from a point of view of efficiency, make them more efficient, make them workflow. The cases that we know based on that DVM skillset, where they struggle, and how do we get the cases to the radiologists that are more important and drive a workflow that becomes a priority list based on what’s important to those customers.
Jack Peploe:
No, absolutely. Now, I mean, one question I want to hit on is around data privacy and AI ethics, obviously going to vary from tool to tool, but AI does thrive on data, but obviously with that comes concerns around privacy and ethics. What sort of considerations should veterinary practices be aware of when adopting AI tools in general?
Eric Goldman:
Yeah, I guess, again, when I think of ethics, I think of that concept of concept. They should not be using it as the be all and end all in the diagnostics, and they can’t be putting it in front of customers and saying, okay, this is a true diagnostic. I think if they’re from an ethical standpoint, if they’re using it in that way, they’re misleading the customer. I think they should be upfront that their data is being used, how it’s being used, and what the capabilities of AI can do today. So one of the things is, is we can publish our confusion matrix on our results.
We’ve done a number of studies. Two of them have got published by the American College of Veterinary Radiology Independent studies. One of the things early on as we were delivering the technology, a lot of that’s would ask me, well, how do you know it’s right? And I could sit there and go, yeah, well, our board of radiologists who are actively involved in the tech say it’s great, and here’s our confusion matrix. Or I could do independent reviews. And so for example, I’m located in Massachusetts. I went to Tufts University in Massachusetts, the coming School of Veterinary Medicine, and I had ’em do an independent study. We did one on plural pleural effusion. These were known cases that had fluid in the lungs, and we were 89% accurate. We did another one with Animal Medical Centre in New York, again, to my congestive heart failure example. It was a five month study that had an AI report come through independently.
A radiologist evaluated the case, and then at the end of five months, the head of radiology, a gentleman by the name of Dr. Anthony, went back and said, what did AI say? What did my radiologist say? What do I think? Because remember these things, radiology is an opinion, and we were 92% agreement. And so that was a study. Since then, we’ve done studies with Oregon State University, Washington State University, Purdue University. I’m always interested in reaching out to academic environments, particularly students, because I feel that this technology is going to be part of their profession. And so they need to engage it, they need to understand it. I, I’ve had them understand how we build classifiers, how you can get different results on sensitivity and specificity and be able to see happens when you want to have a particular disease. And you want to say, well, I need this to be more sensitive versus specific or the other way around, and try to adjust that with data and testing and so forth. And to me that that’s really important. And I think we need to have full disclosure on the tech, get more people involved, have these independent studies for this technology to evolve.
Jack Peploe:
No, absolutely. Eric, this has been such an incredibly insightful conversation, and I can’t believe the time I’ve warned you this was going to happen, but AI is obviously one of the most talked about and often misunderstood topics in veterinary medicine, and I think you’ve given our listeners a much clearer picture of where it fits in their world, which is amazing. So thank you. For those who sort of want to follow your work or learn more, where’s the best place that they can find you?
Eric Goldman:
So they can look@vetology.net, right? So that’s the company website. My email is eric dot goldman@vetology.net. I’m happy to reach out. I love having these conversations. I love learning about their practice. I love learning about their relationships with consultants where they struggle reading films and particularly where they want to learn. How can you use the technology to align with your personal growth. So definitely connect with me on LinkedIn, reach out to me via email, reach out to our company. There’s also info@vetology.net, which is our operations team, and they collect a bunch of information and they can share it with me, and then we can have a conversation.
Jack Peploe:
Amazing. Well, Eric, thanks again for coming on the podcast. It’s been a real pleasure and I really appreciate the perspective you bring to the conversation around AI and veterinary care. So thank you very much. You’re welcome.
Jack Peploe:
Every week we ask professionals and experts to suggest a best business resource for our listeners. This week’s recommendation is from Charlie Barton.
Charlie Barton:
Everything I could think of is being eclipsed by Chad GPT for me, which is not a terribly original answer, but I just think it’s such a good starting point for so many things and just speeds up your ability to be productive. Of course, it’s not the magic bullet for everything. So Chat GPT, Microsoft 365, get everything online and a good laptop, which for me is a MacBook. So happy days
Jack Peploe:
Coming up next week. We welcome Liz Barton, head of communications at VetCT and co-founder of Well Vet. Liz brings her wealth of experience as a vet, communicates and advocate to explore how we can build a more sustainable and emotionally intelligent future for the profession. We delve into systemic changes that could improve and retention, tackle the complexities of parenting in practice, and unpack how the rise of AI may shape the veterinary landscape.
Liz Barton:
We still need that culture shift. I think there is still that underlying hero culture, the sort of celebrating of the people that can work the longest hours and work through their lunch breaks. I still hear that, and I think we do need to continually call it out and just say, actually, the hero culture is the leader who says, right, I’m going off for my lunch break. This is my protected time. And make sure you do the same, and who says, oh, I need to go and look after my kids because things have happened at home. And build in the flexibility to allow that for other team members as well. So actually modelling the fact we are that working life is a continuum them, and we do need to respect the fact that we need to eat, sleep, socialise, and recover properly in order to be sustained in the workplace and for longer.
Jack Peploe:
That’s it for this episode. All links and recommendations we talked about are in the show notes. Don’t forget to subscribe and share the podcast if you found it useful. In the meantime, thanks for listening and see you next time.
