In this episode of PathPulse, James is joined by Dr. Matthew Cecchini, a celebrated pathologist, educator, and a Top 100 Trailblazer recognized by The Pathologist magazine. Their conversation dives deep into the profound impact of digital pathology and artificial intelligence (AI) on the field, moving beyond mere clinical application to explore how these technologies are revolutionizing pathology education.
Dr. Cecchini outlines the challenges facing traditional pathology training—from its reduced role in medical school curricula to the historical barriers of glass slides. He then details his innovative work, leveraging off-the-shelf AI models and unconventional teaching methods (including creating original music!) to make pathology more accessible, engaging, and relevant for the next generation of physicians.
The core of the discussion explores two critical aspects: how to recruit and properly train future pathologists, and how to redefine the pathologist’s role in an AI-augmented world. This episode offers a compelling vision for pathology in 2035, shifting the focus from repetitive slide hunting to becoming essential managers of the diagnostic process and key educators for the entire clinical team. Dive right in!
Listen Now:
Transcript:
James Thackeray Well, welcome back to PathPulse, the digital diagnostic podcast. I’m your host, James Thackeray. Our mission is showcasing the individuals and innovations shaping the future of digital diagnostics. Today, we’re shifting our focus to a crucial area for long-term success of the pathology field—the education of future pathologists.
We started into this in our last episode, so I’m excited to kind of follow up with our guests today and kind of dive in deeper to what that could look like. This rapid advancement of digital pathology and artificial intelligence are not only transforming the clinical practice, but also immense potential to revolutionize how we teach and learn pathology.
We are delighted and welcome our guest, Dr. Matthew Cecchini, a leader in the integration of digital pathology and online resources in education. He was recognized as a Top 100 Trailblazer by The Pathologist magazine for his innovative use of social media and digital tools in pathology education.
Dr. Cecchini, welcome to the podcast. It’s a pleasure to have you on. Love to just have you introduce yourself, give us a little bit of your background, and we can kind of dive right into it.
Matt Cecchini Yeah, thanks so much. Thanks so much for the introduction and for having me here. I’m, you know, it’s my favorite thing to talk about is AI and education. So, you know, I’m sure the time’s going to go by super quick, and I’m really excited about it.
So just about me, I’m a pathologist who works in London, Ontario, Canada. I’m really interested in digital pathology and how that intersects with this AI revolution that the world is going through. And I’m really excited for a lot of the tools that we can build to really transform the field of pathology from the ground up and really make this better for everyone within the field.
The Shift in Pathology Education: Traditional vs. Digital
James Thackeray Well, that’s exciting. Maybe start by kind of just laying the groundwork. What, when it comes to education in the field of pathology, what is the traditional, and we don’t have to get into too detailed, but kind of high level, what’s the traditional methodology of training, and where do you see this going and what advancements have already, have you already implemented in this kind of new way of doing things in digital pathology?
Matt Cecchini Yeah, there’s a few forces at play here. One is that the medical school training of physicians has changed a bit. So back, you know, if you look back in time, pathology was a core course within the foundational medical curriculum, where we’re all medical students had a microscope, had to look at slides, and that’s largely gone by the wayside, right?
And so a lot of times we’re getting people into the field now…we have a hard time showcasing pathology as a field because they’re not really seeing it. But even when they do come into it, they’re not really getting a lot of that core histology understanding around that. And it’s a field that’s traditionally utilized like a microscope, right? And a lot of the data has been kind of locked behind these kind of glass slides. And that’s been a bit of a barrier to getting people involved in it.
And so digital pathology makes a lot of this easier, right? And so you can now access these slides anywhere. You can access these slides on your cell phone. Like, it reduces a lot of the barriers to kind of getting students in there, but it…pathology and histology itself can actually be very overwhelming for people, right?
Because if you haven’t seen it before, it can feel like, how do I get started with this? How do I use this? And this is something that we’re exploring in my institution. Can we use AI to start breaking down some of those barriers? And so we’re using some of these multimodal AI tools to actually give this to medical students to allow them to kind of act as a tutor, to kind of get them over that initial kind of activation energy of you know, getting comfortable with looking at a slide of what does the H&E look like?
And a lot of times people are like worried about “What if I ask a silly question and like, where do I learn these things?” And I think these things are coming together and this is going to be a real flywheel where we can kind of take these tools that enable access and then combine them with kind of AI tools that are largely freely available to kind of, you know, really accelerate getting people more interested and exposed to pathology and histology as a field.
Recruiting and Engaging Future Pathologists
James Thackeray I love that. So let me dive a little deeper in some of that. So when we think about medical school, that side of it, because here we have this challenge, right? Where we have very few pathologists worldwide when you think of the aging demographic of where we’re at. And it’s really not just the US, it’s worldwide where we’re deprived of the diagnostic side of things often.
So how, when you get into the medical school side of it, yeah, that’s the key. It’s like, no, before they’ve chosen a specialty, how do you get in front of those students? So can you go a little deeper on that side of it? Because I think that’s so interesting to me.
Matt Cecchini And this is something that we’ve tried to push on and I’ve worked with one of my colleagues, Matt Quick from the University of Arkansas, on this. We’ve done a few courses at USCAP on this—about how do we recruit medical students into pathology. And people do different things. One thing that we do here in Canada here is we have all of the students during their surgery rotations, they actually have to spend a morning with us in pathology. And then we, you know, we bring them into the lab, we tour them around, we teach them about pathology, because the reality is most of them aren’t going to go into pathology, but they’re all going to use pathology resources.
So we want to explain some of the fundamentals, but we also get them excited about it, right? We show them some cool things, show them some cool cases. We get some really dynamic, engaging pathologists to talk to them about, like, why they love their job. And we bring in AI and digital pathology, and we make it very hands-on and engaging. Because a lot of times, in the medical curriculum, pathology is just presented as, “These are just some things you need to memorize for your board exams or your steps” or whatever, right? And it’s not presented as a career, right?
And so we try and break that down as when they’re in their clerkship, when they’re seeing all the different areas of medicine, we kind of give them all this exposure to it, where we can show “What does a pathologist do? How does a pathologist interact in the slide? Why is that important? What are the management decisions that are made off of that?” And that’s been really successful for us in showcasing our field. And it also just helps dispel a lot of myths, right?
Like when people expect that they could do a biopsy and get results like 15 minutes later, right? Because they’ve never been through a lab, they don’t understand that it’s got to fix, it’s got to go through a processor, right? And so we get less of those calls now because we better teach our students, like, what’s happening in the lab, right? It’s not just this black box where you submit tissue and you get reported out, right?
James Thackeray That’s so great. Talk to me a little bit more about the AI aspect of it. Because obviously, with digital pathology, we’re talking anything digital, we usually think AI is some diagnostic or prognostic algorithm that’s used in the diagnosing of, you know, or in assisting diagnosis in pathology. But a lot of what you’re talking about is AI in the education, so maybe talk a little bit more about that.
Matt Cecchini Yeah, so it’s an interesting space, right? Because like, you know, there’s not a lot of funding in this, right? So I like that I can’t use some cutting edge AI tool that’s going to cost a lot of money per thing, right? But it allows you to kind of find some innovative solutions. So we’ve actually just been using off the shelf foundation models. And it turns out they’re actually fairly good with pathology.
So like, you can take like we use a lot of things like the Gemini stream real time where it can actually see what’s on your screen. And it actually does a really good job of just explaining the basics of pathology and histology. It’s not perfect, these commercial tools are better, it’s not a medical grade, anything. But for a junior learner who’s never seen a slide before and they’re like, what is this structure? You can circle on the screen and say, that’s a hair follicle. And then it tells you why it’s a hair follicle, right? And so these freely available tools are really changing the game a lot.
And so I do a lot trying to integrate these emerging tools and technologies into the curriculum and into ways that we teach. Like my last lecture to the medical students, I don’t actually believe in PowerPoint anymore. I think PowerPoint should be banned from education. It’s a controversial opinion, but you got to get people to engage. It’s not about just people memorizing stuff anymore. It’s about how do people engage and utilize information, right?
So we actually did the whole lecture on Figma, right? So I opened up a Figma, FigJam, and then we just constructed the concepts that I wanted them to learn on the Figma. And then I had them go off and do a bunch of like deep research queries and then bring that back and analyze that data and put it in, in this real interactive way of exploring pathology education. So these are some of the things that we try and bring into this.
Because this is the world that they’re going to go into, right? The people who are in medical school, just started medical school today, they’re going to go into practice 10 years from now. And I think a lot about this: “What does medicine look like in 2035?” Given the progress that we’re seeing, it’s going to look very different to what it does now.
And so this is where I feel strongly, as an educator, I need to be pushing people to what I need to be training to where they’re going to be. They’re not going to be like me looking at all their slides under an analog microscope, typing the reports. They’re going to be using machine intelligence and everything that they do. So I think as educators, we’ve got to push them to that and not just, you know, could be constrained in our current world. Sorry. That’s a lot of words.
James Thackeray No, this is great. And in fact, I’m going to come back to that. Like, what does pathology look like in 2035? So keep that at the back of your head because I do want to come back to that.
So, because it’s so great, what you’re doing on the education side there, how do you scale that? I mean, how are you working with other institutions or potentially working with other institutions to kind of say, “Hey, this, this is what it can look like. We want others.” And maybe it’s people like Dr. Quick and University of Arkansas, how does that look, scaling up this?
Matt Cecchini Yeah, I think it’s just sharing it. So I try and share as much as I can. In real life, we do presentations and conferences. I try and post on YouTube and LinkedIn and other social platforms. Because it’s like, a lot of this is really democratized, right? You don’t need anything, right? You don’t need any specialized tools. You don’t even need money, right? A lot of these things are largely free, right?
And so I think it’s also an important part for kind of global equity in this, right? Is that it really democratizes a lot of this education, because a lot of things in the past are locked behind paywalls or you can’t get this textbook because it costs so much money. It costs a lot of money to ship this place, right?
But I think it allows you to share a lot of information much more freely, right? And so I think it’s just being bold and pushing these things out there and rethinking our traditional ways of doing this, right? That you just like the, a lot of the barriers are going down, right? Like I can do all these things that I couldn’t do in the past, right?
Like, as an example, I was reading a paper about how music helps people learn things. So now, in all my lectures when I teach undergrads, I actually make them songs that go along with their stuff. So I take the lecture notes and I put them into an AI model I have that converts into a song.
And so I give them an album about the content that they can then listen to. Because I myself am an audio learner. And so I find that helped me. And so, in addition to their study notes, my students actually get an album about the content I’m teaching. It’s silly, right? But like a song about neoplasia, right? But it gets stuck in their head, right? And then they now remember all these terms and it’s a useful framework.
And it’s things like this, right? That we need to push the envelope on, right? We need to see beyond the, like, “Here’s your slide deck, here’s your lecture notes, please go and memorize that,” right? You need to think like just “What can these things enable?” And it’s like, they enable everything, right? I love it. It’s just creativity unlocked for me.
The Pathologist’s Role as Educator and Diagnostic Manager
James Thackeray That’s so great. And I love it. Maybe, maybe we could do a song to Ozzy Osbourne in tribute to his legacy. That would be awesome. What a cool way of doing it. And certainly something, you know, in grade school, you did stuff like that all the time to kind of remember things. So why wouldn’t we continue that?
Okay. I want to switch just a little bit. You’ve, you’ve hit on a couple of points that I think are really important. One is education of the many people who will not go into pathology and their understanding of what pathology really is.
There’s a gap, right? I believe there’s a gap in the, and maybe it’s because we’re in this field deeply enough to where people don’t recognize, always, the critical aspect that the diagnostics plays. I mean, it is what is the key to therapeutics downstream, obviously.
When you’re educating both people that would go into the field and those that surround it, what are the key elements that you kind of highlight to help them recognize the importance of the specialty?
Matt Cecchini Yeah, I think there’s a lot of misconceptions about pathology, right, that a lot of other physicians have, right. And I think it just becomes when they don’t get a lot of exposure to it. And so I, I try to just, like, unveil the curtain and show them behind the scenes what is going on, right. And I think that that’s really important, right, because a lot of times people just, there’s a disconnect, right. And like, that happens in really practical ways, right.
Like if we’re going to talk specifics in the site for a second, like so PD-L1, right? So it’s a biomarker score that we do, right? Sometimes oncologists will call me and be like, “Well, is the CPS three or four? Right. And, and, then, you know, but then I show them the slide and say, “Look, there’s, there’s literally a million cells on this slide.” And, and then I have this, this diagram where I actually calculated how long it would take me to count all the cells there. And it’s 34 hours to count every cell there. Right. And, and then I show them like, “There is no way that I can count all these cells, right? And so the number I’m giving you is based on a good estimation, but it’s an estimation of that score, right? The precision is not there where I can tell you that it’s a three versus a four or 2.5 versus a 2.7,” right?
And so that’s a very specific example, but there’s lots of these where there’s just this disconnect, right? Where they don’t see it. And so they don’t grasp what it really means, right? Which is different than radiology, right? Because radiology, they all see, right? They’re all really used to pulling up their CT scan or their PET scan.
And I wonder how much this changes as we go fully digital, right? Like, you know, does the next generation of physicians get more comfortable with pathology? Is that where we go? I’m not sure, right? Because, like I said, we’re not really training them that in medical school, right?
And so, but I think there’s room for using AI here as well, right? To break down some of these barriers, right? Like, you can imagine AI tools help explaining some of these pathology, but this is also a great role of the pathologists, right? As we free up more and more of our time from having to do mundane things like trying to count PD-L1, you know, that we can then move towards, like, what we should be doing, right? Like we’re educating, having these discussions, right?
Like you can train a kindergartener to do PD-L1, right? You’re counting brown things and blue things, right? I didn’t need to go to medical school and residency to count things. This is where I think this hybrid of AI and humans can do better, so that we can help break down these barriers and spend the time where my medical knowledge plus pathology knowledge is most useful.
James Thackeray So great. Yeah, I appreciate that. That’s great. And the AI can help in breaking down, you know, counting the number of cells on a PD-L slide, right? So it does free you up. I like the way you verbalize that. It frees you up to educate, but also manage the entire diagnostic process.
I feel like, just having watched the specialty change and tell me, I want to get your feedback on this, but I’ll throw it out there. The pathology is changing and shifting quite a bit with new technology. You have those that are fearful that AI will replace them, but I see it more as complement them, especially if we can get to that level where the pathologist owns the entire diagnostic process.
And what I mean by that is not just the original diagnosis, but what needs to be done downstream for specific therapeutics? I think we’ve relied on oncology to just say, “Well, I want this,” which is fine, but it’s also like, this is your field, right? So talk to me a little bit about that, if you would.
Matt Cecchini Yeah, I think our specialty is actually at a critical fork in the road. And I think one side of this fork is actually very bad for us, and one side of it is very good. Right? And so the one side is where we kind of stick our heads in the sand a bit and don’t really engage with these emerging technologies and just say, “No, we don’t do that. We do the traditional light microscopy. You know, like, like I, you know, I’m a purist. This is exactly what I do.”
Because what I think is going to happen is, then, people are just going to replace us, right? Is that we’re like, there’s a need for this more stuff, right? Our current things are insufficient for what the medical world needs, right? And so if we don’t engage with it, I think they’re going to be more and more just black box solutions that just do the pathology, do the prognostication, do the things that I want to be doing, right?
And I hope that’s the fork we don’t go down. But the other fork is one where we engage with these technologies from day one, right? Where I become an AI augmented pathologist, right? Where I am the person using and building and driving this AI bus, right? And that, I think, is one which is actually better for everyone, including patients, right? Because, you know, why would you not start from everything that I have in my head, right? Why would you know, not like utilize me and then allow me build a system in a way that I can augment the AI, right?
And this is not an easy thing, right? Like, I think a lot about this interaction between human and AI and human and AI learners and type of things. Because it’s not sufficient to just say, “Here’s the AI output, what do you think?” Right? That doesn’t work. And there’s a bunch of studies where it actually shows that human physicians actually harm the AI by, know, we either have an automation bias or we don’t trust the AI.
And so this human-AI interaction is a really, really critical piece, right? And it’s one that I’m actually exploring through the education lens, because I think it’s a good way, right? Because you’re getting people who don’t have these 30 years of biases and traditions, right? And you can kind of explore these things in kind of more this AI native population that is used to this, right? And they’re used to asking chat GPT other questions, to using a perplexity search. They understand what the deep research query is, right?
And so I think, and I hope that that’s the pathway that we can go down, right? Because we haven’t solved cancer, we haven’t solved pathology, right? And so I don’t want it just to go down the road where we just try and do what we have with less. I want to do more with what we have, right? So that we can ultimately provide better care for patients.
Pathology in 2035: An AI-Augmented Future
James Thackeray That’s well said. I think those that accept that there is new technology and bring it into the practice, but also, to your point, there is a bit of a mind shift that needs to maybe change, not with all pathology, but certainly, like, “I’m not just in the basement reading, you know, the morphology. No, I need to manage this all the way through and really educate.”
You become, to your point, and everything you’re doing, an educator, but an educator to oncology, an educator to all the different specialties. And I love that you’re starting that, or at least helping others understand that way early in the process in med school and things like that.
So, okay, last question. And maybe this will give you a chance to kind of summarize things, but going back to that, what does pathology look like in 2035?
Because I do, mean, and it’s so hard because it’s come so far so rapidly. I feel like we had this period of time where it was just the same technology for a very long time. And that’s overstating it. I’m sure there are many new advancements that have come throughout the entire timeline, but it does seem like now it’s like, okay, radiology already advanced all these other specialties of advance and now pathology is kind of at that place.
So as you educate others on what it looks like in the future, not knowing all things. What does that look like?
Matt Cecchini So I’ll answer this question, but I want to give a few caveats first. So one is I’m very much an AI optimist, okay? So I’m going to paint a picture for you with what I think it could be, while making a few assumptions. One, I’m an AI optimist and I believe that these tools are going to keep scaling exponentially over the next few years. And so this may not happen, right? We may hit a wall with these tools and this may not happen, but let’s paint a picture of what this looks like.
Right. So I honestly think we’re going to hit artificial general intelligence within 1-2 years, if we have not already, right. Where there’s AI tools that are essentially as smart as most humans. Right. And so, you know, along those timelines, in the 2035, we’re talking about worlds where we have artificial super intelligence, right. Where there’s like commercial AI tools that are smarter than the collective of humanity. Right.
And you need to take a step back and say, well, what does anything look like in a world where we have access to that level of technology? And it results in a bunch of things that can be scary. The value of intelligence essentially goes to zero, because you now have unlimited access to intelligence that’s smarter than you infinitely. And so what then becomes my role?
My role becomes harnessing that intelligence, right? My role becomes integrating that intelligence. My role becomes directing that intelligence, right? And so my role becomes more one of an operator, right? Where I have a problem in front of me and I am utilizing variance, varying amounts of machine intelligence to do my job, right? So let’s imagine this through a case, right?
So there’s a case that’s coming through you know, I probably have the whole patient history summarized to me automatically by a tool that reads the entire EMR. I then have that go through the case and give me an auto summary of all the slides, and it’s flagged some areas where, you know, there’s some uncertainty, and then it’s focusing my attention there with my expertise on these areas of uncertainty. And then my job is crafting a narrative back to the clinical team of how that best guides it.
And maybe I’m realizing, “Okay, this patient’s a candidate for advanced therapy. Okay, I’m gonna call in some advanced tools and get prognostic markers and predict this.” And then I’m turning this all around really quickly and giving this kind of informed aspect. And so that job is very different than my job today, right? I expect in 2035, I will be spending less than 10%, if not less than that, looking at microscopic slides.
Because most of the stuff that I do is highly repetitive. It’s like looking for tumor in lymph nodes, right? There’s already AI models that are better than me at that, right? And so, in the world where things keep scaling and keep doing better, right? Then it’s like, “Well, I don’t need to do that anymore, right?
But this is where we have these two forks again, right? So there’s one fork where we just automate and replace most pathologists and there’s a few pathologists left in the world chain to a desk somewhere clicking buttons, right? Like that’s the bad version, right?
The good version of this is where I don’t have to spend six hours of my day hunting for lymph nodes, but I can spend six hours of my day communicating to oncologists, developing integrated reports, making sure the reports fit within the context, integrating the molecular results into my findings, right? And really doing the transformative stuff that I can do as a physician, Not the like playing where’s Waldo finding tumor lymph nodes.
That’s important, but I see my role moving more towards as a manager, manager of a team of AI that’s helping me through a case, right?
And I also do molecular pathology, and that’s kind of how my role is molecular pathology, right? I oversee a bunch of processes that are going on, right? There’s someone who’s running the sequencer, then someone doing initial analysis, someone’s looking at variance, and I’m overseeing that whole process, and they’re bringing in my expertise in certain areas to make decisions or to correlate that or put it together, but I’m kind of, I’m not doing all the little mundane little pieces of this puzzle, right?
And so this is where I see the transition being, but we need to consistently think about how we can do more with this, not how we can do the same, right? Because if we’re just focused on doing the same with less, then I think that that’s bad for our specialty, right? But how can I do more?
And this is the thing I sell to the students, right? Of like, “This is the future of the specialty right? It’s not, you’re still gonna need all these foundational skills on how to do this, but your role is gonna be different and exciting and all these different things to it.” But anyway, sorry, that’s kind of, that’s my two cents of this.
James Thackeray That’s perfect. And it’s actually a perfect summary. I have a lot of things that I try to add only because it’s in my head and I shouldn’t because it’s a perfect summary of where I think it’s going as well. I couldn’t agree more with the trajectory and the role of the pathologist as it changed.
Well, thank you so much for being on the podcast, Dr. Cicchini, and thank you for the work that you’re doing on the education side because it’s so critical for the specialty, but to your earlier point, it’s so critical for pathology to be better understood amongst other specialties in the medical field.
So really appreciate your time. Look forward to doing this again at some point and seeing how this progresses over time.
Matt Cecchini Thanks so much for the invitation.
