September 2025—Leaders of the Digital Pathology Association met online with CAP TODAY publisher Bob McGonnagle to talk about digital pathology and much that’s related: adoption, investment, and artificial intelligence tools. “Most people are playing catch-up,” Marilyn Bui, MD, PhD, past president of the DPA and a member of the association’s board of directors, said of digital pathology. Of AI and digital pathology, DPA Foundation president Michael Rivers of Roche Tissue Diagnostics said, “Over the next several years we’re going to see a transformation in pathology.”
DPA directors Orly Ardon, PhD, MBA, of Memorial Sloan Kettering Cancer Center; Nathan Buchbinder of Proscia; and Ryan Davis of Epredia were in on the Aug. 11 discussion, which follows.
Marilyn Bui, tell us about the progress we’re making in the adoption of digital pathology across the board, and fill us in on what’s being done at Moffitt Cancer Center.
Marilyn Bui, MD, PhD, senior member and professor of pathology, and scientific director of analytic microscopy core, Moffitt Cancer Center, and chair, CAP Digital and Computational Pathology Committee: At Moffitt Cancer Center we have a two-year plan to go fully digital. We’ve finished evaluating the image management system, scanners, and monitors. There will be construction for the scanning center, and then we’ll take steps to transform the department. This is coming from the top; they infused the funding to us from the foundation and from the state and institutional contributions [https://bit.ly/MCC_digpath].
Digital pathology adoption needs intentional investment. The models I see, such as at Mayo Clinic and Moffitt Cancer Center, are top-down. It’s a smoother process because you have to engage with IT, infrastructure, et cetera. Once the institution has the green light, it’s easier for the pathology department to make this happen.
The second model is that the pathology department wants to do it, so it comes up with funding to get a scanner. And then it has to cut through the red tape with IT and others.
The third model of adoption is individual pathologists, practice owners. They fund it themselves. When they realize it works, they connect the dots and thrive from there.
The CAP Digital and Computational Pathology Committee collaborated with the House of Delegates and conducted a survey of adoption, which will be submitted to Archives of Pathology & Laboratory Medicine. We compared our survey numbers with the data from the CAP 2024 practice survey, which found that 10 percent of respondents use remote sign-out for digital pathology primary diagnosis [whole slide imaging] and that remote sign-out is much more common in independent laboratories [used by 67 percent of respondents] than in nonacademic hospitals [just under 20 percent of respondents] or academic hospitals or medical centers [24 percent of respondents] [https://bit.ly/CAP_PracticeSurvey].
The trend is digital pathology adoption is gaining significant momentum. The train has already left the station. Most people are playing catch-up. They’re no longer trying to be on the cutting edge; they have to do it.
Nathan Buchbinder, what are you seeing in terms of adoption?
Nathan Buchbinder, chief strategy officer, Proscia: We’re seeing digital pathology adoption in U.S. diagnostic practice is in the range of 15 to 20 percent. We’ve also observed two other exciting trends. The first is that a significant portion of labs that have made the shift away from the microscope are nearing 100 percent digital reads or are on track to do so over the next 12 to 18 months. It’s a big step forward. More and more labs are now saying, “This is going to be how our pathologists do their routine work. It will be a key part of what we’re planning from a technology perspective and what our strategy is around data-driven pathology.”
The 15 to 20 percent adoption rate hides the other trend, which is that 30 to 50 percent of the remaining labs are somewhere in the journey of adopting digital pathology. It doesn’t mean they’re up and running, but they see digitization not just as the future but as the present. They recognize that if they don’t make the investment today, they’re at risk of being behind in the decision-making process. If you run this roundtable again in 12 or 18 months, I’m confident we’d see that the 15 to 20 percent looks more like 25 to 30 percent. If you gave it another 12 months, I’d be shocked if we weren’t around 40 to 50 percent, which is remarkable given how the industry has taken five to 10 years to hit a degree of stability and maturity.
Ryan Davis, Epredia has historically been more well known for its histology equipment than it has for a particular play in digital pathology or AI, but what is your perspective currently and on what you’ve heard in this discussion?
Ryan Davis, director, global business, digital pathology and AI, Epredia: I agree with Dr. Bui and Nathan on their points about the adoption rate and how facilities are looking to adopt. We are seeing certain subsegments accelerate even further. Dermatology, renal, and even GI spaces are adopting at a faster rate than other areas. I’m also seeing an acceleration in the use of resources within a facility and corporate infrastructure. They want to get the most out of their pathologists and make sure their key performance indicators for turnaround time and to meet patient goals are heightened, and they see digital as a way to accelerate that.
I would be a little more bullish on the percentage of facilities looking at digital pathology to bring it on board. Having more and more information about how labs have been successful in going digital is giving them more confidence. Many of the papers from early adopters published five to seven years ago showed several barriers to going fully digital. Today there are more options in the marketplace and technology has advanced to overcome many of those barriers.
You mentioned that Epredia is well known for its core histology equipment, and I’m also seeing more customers think about the preanalytical phases of how they get to a slide that is truly digitally ready. They are thinking about the steps in the procedure that allow them to be successful in going digital.
Mike Rivers, what are your thoughts on what you’ve heard so far?
Michael Rivers, vice president and lifecycle leader of digital pathology, Roche Tissue Diagnostics: There is tremendous momentum. We’re seeing the digital pathology segment grow much faster than the overall pathology segment.
Academic medical centers and commercial labs are seeing the value because large labs are more easily able to capture the productivity or workflow benefits that can come with digital pathology, which some may need if they’re going to justify the investment. We still have an issue with reimbursement and there’s great opportunity for organizations like the Digital Pathology Association to rally around making progress in that area.
We have tremendous potential in the AI space. There’s a lot of exciting news, research, and new tools, and a lot of work to do to bring these fully into the in vitro diagnostic realm, but that’s coming and it will provide another powerful catalyst to adoption. Over the next several years we’re going to see a transformation in pathology.
Orly Ardon, as a leader here and as someone who’s worked mightily to help with a justification and calculator for the return on digital pathology investment, are you celebrating a little as you survey the scene?
Orly Ardon, PhD, MBA, director of digital pathology operations, associate member, Department of Pathology and Laboratory Medicine, and member, Warren Alpert Center for Digital and Computational Pathology, Memorial Sloan Kettering Cancer Center: I’m very excited. We’ve been early adopters and have been practically 100 percent digital for a while. It’s a massive operation. Our hematopathology service is the first to be fully digital, so we are not distributing glass to pathologists anymore. We have fewer technical issues with more adoption by pathologists, but with time we see other issues we did not anticipate.
What we are seeing more and more is a growing digital divide. The distribution of institutions investing in and benefiting from digital pathology and AI is limited to certain geographic locations. Some regions in the U.S. are not benefiting from all this progress. The interpretation of what digital pathology is—is it fully digital, is it buying a few scanners and starting with biopsy, prospective scanning, et cetera—doesn’t matter. That you broke the barrier and introduced the digital possibilities in-house is huge.
We are going to have challenging times in health care. I hope it won’t slow anyone down because if you want to reap the benefits of digital pathology and AI, you have to make the investments now. Most institutions are aware of that. More vendors are coming out with new solutions, but we have to be cautious and think about the next steps and the next issues that we need to tackle as a community.
You’re indicating there are have and have-nots in the world of pathology. The consolidation of health systems is leading to that, and digital pathology and AI are riding along in a hugely consolidated system. Nathan Buchbinder, can you comment on that? Consolidation is part and parcel of strains in finance and labor. The thought is the bigger you get, the less vulnerable you are as a system.
Nathan Buchbinder (Proscia): The consolidation is impossible to ignore. It is a factor of both the challenging economics of laboratory medicine and the opportunity that comes when you have an extremely high-volume, high-throughput, data-rich practice that can offer more than essential pathology services.
Digital pathology is a shift that’s happening concurrently with, but not necessarily causing, consolidation. It can facilitate consolidation. But you can get meaningful benefits from digitization that favorably position even the smaller labs in this era of consolidation.
Taking a step back, there are too few pathologists to read the volume of cases coming in, and the asks on pathologists today are more complex than the asks of pathologists even five years ago. Consolidation is the economic and market response to a supply and demand disconnect and economies of scale. You can cut out some of the administrative level and centralize some physical processes and deliver the same quality of service from a centralized set of operations. Digitization facilitates that. The shift toward digital pathology, and AI in particular, is a similar response. It plugs into the economic response and enables having decentralized physical locations with a centralized operating model. It allows pathologists to practice when they’re not in the same physical location as the slide.
Digital pathology also serves well the market segment that has historically struggled with margin pressures because it opens opportunities to hire from a broader pool of talent and drive efficiency in workflow. It opens the value that comes from a data-driven practice, whether that’s by performing additional testing, layering applications of AI, or generating data and insights for patients directly or that fuel innovation, discovery, and development. These macro trends in the economy are driven or enabled by digital pathology, but there’s also a play for leaders of labs that are at risk of being consumed to look at digital pathology as a potential means of overcoming the margin pressures.
We have not only a shortage of pathologists but also younger pathologists who are changing in many ways. I was told a story about how a group was trying to recruit young pathologists. One candidate looked particularly good—no money or location obstacles—but the candidate was unwilling to drive once a week to a small hospital to do frozen sections and talk to the surgeons. The candidate said, “If you can arrange a digital pathology solution for that small institution, I’m happy to read it, but I’m not making that commute.” What’s your reaction to that story, Marilyn? Do you agree we’re seeing changes in the younger pathologists?
Dr. Bui (Moffitt): Definitely. The newer generation is most likely not willing to do what we have done, which is to sacrifice our well-being for the job. They are looking for a healthier work-life balance. Many now consider pathology as a lifestyle medical profession because often it is not patient facing, it offers more regular and flexible hours, and it allows a long career. Having a digital pathology option is attractive to young pathologists who are balancing family responsibilities and their wellness and to pathologists who are on the cusp of retiring—now they can extend their practice even longer.
Another contemporary development is that biomarker testing has revolutionized precision medicine. Manual reading of biomarkers is no longer adequate. We’re moving to quantitative, reproducible, accurate evaluation of biomarkers so we can direct effective targeted therapy. We’re relying on quantitative image analysis and AI to standardize biomarker interpretation. That standardization relies on digital pathology, which gave birth to AI, so we have to take that step to get to AI.
Reimbursement to physicians continues to decrease, and mergers and acquisitions in pathology continue to trend up.
The economic, regulatory, and workforce benefits are driving consolidation of laboratories. When you are at a good academic center, you are somewhat insulated from this type of stress. However, if you are in a rural area and own a private practice in a community setting, venture capitalists will come in to purchase your hospital and offer you a good package to buy you out. Some merged practices may survive. For others, the goodness of the health care system will be sucked out, and when there’s no more money to be made, the venture capitalists are out. Then the rural hospitals have to close their doors, which is a bleak outcome. That’s why the CAP’s advocacy efforts on behalf of pathologists and patients are more vital now than ever before.
In the past couple of years, the digital pathology and AI market has been two halves. There’s the company side, where we’ve had dissolutions and so on, and the labs that have adopted, which are healthier. Dr. Ardon, with the news of Evident—which is today the Olympus histology and imaging line—purchasing digital pathology manufacturer Pramana, what’s your reaction to this sense of the companies perhaps being healthier in the marketplace now?
Dr. Ardon (Memorial Sloan Kettering): I’m on the laboratory user side and looking for state-of-the-art technologies and opportunities. We are seeing companies large and small develop amazing products. But it’s not enough to have a great idea and a great technology; you need a way to get it to market.
In terms of the business aspect, we are seeing a disparity between investment in AI development and the ability of institutions of all sizes to bring those technologies in or develop them in-house, not just get them to a minimal viable product but to a clinical state and then validate them so they can be used safely in the institution. Having all the infrastructure components that ensure the images are there will allow the AI to run consistently with all the legal and compliance needs that institutions have. It takes time to bring those in, and institutions need to plan for that.
The more users we have, the more technologies we’ll see, not just from startups with hardware solutions but also from established vendors. It’s an exciting time because many of the technologies we use now are becoming older and there may be promising solutions in development that will become available with the market growth.
Once we get into AI, we get caught up in a market frenzy that runs beyond the boundaries of pathology or health care in general. We’re only beginning to see a number of peer-reviewed studies of the application of AI in pathology and we’re in early innings of evaluating AI in a rigorous way. Ryan Davis, in your job what do you see as a realistic rollout of AI in pathology?
Ryan Davis (Epredia): AI will be successful in many areas; in others it will be a struggle based on the cost-versus-return ratio of the products out there and those being developed. In the past 12 months people have been thinking differently about how AI can satisfy their need and accelerate going digital. Many companies are now also thinking about how they can diversify what their software or AI can do to help pathologists most effectively.
A critical aspect of the different companies and consolidations are the regulatory bodies. For us to see technology keep pace with other industries and development time scales, we have to have an eye on what the investment is from companies for facilities to feel confident it is a viable product. We can think about developing as a marketplace many different technologies, but the cost to prove it’s of clinical benefit and a realistic adoption piece will be critical.
Dr. Ardon (Memorial Sloan Kettering): What we are seeing is not necessarily AI for pathologists, but AI that can help with preanalytical and postanalytical issues with digital pathology. These may be the first to get adopted because they are relatively low risk to institutions and will make a huge impact on operations. We may see easier adoption of those AI tools and at the same time look at what can benefit pathologists and how we can bring the benefits of investment to patient care.
Not every AI tool will have the same impact. We have to be strategic in what is needed, what will be helpful to the laboratory, pathologist, and patient. Having input from pathologists is important because anyone can develop an AI tool these days. But what will make that impact? This is where we have to concentrate.
Mike Rivers, what are your thoughts around AI and its adoption?
Michael Rivers (Roche): I’m excited about the investment I’m seeing from pharma in general, not just from Roche. It is fueling a move toward precision diagnostics. The diagnostic tests that pathologists are being asked to evaluate are more and more complex, and some are getting into the realm of being beyond human interpretability alone. I think we’re close to being able to establish through evidence and the painstaking development of data and regulatory review that there are insights you can gain from AI that can drive a treatment decision and a better outcome for the patient.
Pharma particularly in the area of cancer drugs is the all-star here and is unrelenting in the pursuit of solutions to optimize therapy.
Michael Rivers (Roche): Yes, and there are late-stage trials in progress and exciting devices in development, breakthrough device designations that have been granted. I think we’ll see some of these come into the market as in vitro diagnostic-labeled solutions in the next 12 to 24 months.
Dr. Bui, what would you like to add about AI?
Dr. Bui (Moffitt): There are two big categories for AI tools. One is the focus on precision—finding the tumor; classifying a tumor; quantifying ER, PR, HER2; and predictive and prognostic capability. A recently FDA-approved multimodal AI-based prognostic and predictive software for prostate cancer is an excellent example. This AI tool analyzes digital pathology images of a patient’s biopsy slides to predict the outcome and determine the most appropriate treatment.
The second category is AI algorithms for workflow improvement. For example, a certain high-volume scanner has great QA for whole slide imaging. With the current workflow the histotechnologist has to do everything manually; that’s a bottleneck. With this type of scanner, you cut down the effort and time required for slide QAs. Visiopharm has Qualitopix with which you can monitor the quality of your IHC stain, so your histology stain is almost like a chemical stain now. You can map it if the quality is good; if it’s not, you can troubleshoot. With intentional effort in product development, the histology instrument company collaborates with the scanner company, so when the stainers are done with H&E, IHC, you can pull out the rack and put it directly into the scanner. That saves the histotechnologist time.
Another example of workflow improvement is DigitCell’s TWOD system, which eliminates the need to manually retrieve tissue from multiple formalin jars one by one by combining three tissue cores into one, simplifying specimen handling. Additionally, a barcoding and color-coded system tracks specimens throughout the process. Digital pathology assists gross image capture and specimen measuring and description, and preserves specimen integrity. After the processing, the whole slide images are ready for review and AI analysis. This increases the lab output by up to 40 percent, reduces consumables by 33 percent, and improves overall productivity.
There are also AI-assisted workload distribution solutions. If you have five breast pathologists on service, their workflow may not be even. Using AI they can set the parameters and evenly distribute the cases.
Many tools are coming out, and you can break them down into precision and workflow tools.
Nathan Buchbinder, your thoughts on what’s been said?
Nathan Buchbinder (Proscia): Take a step back, outside the realm of pathology, and think of how big AI has been over the past two years in transforming how we think about the world around us. Day-to-day tasks, how you run errands, how you do a Google search, if you’re even using Google to search anymore. The vision of what the future holds and how real that vision is has changed because the technology is at a point where we’re not talking about science fiction; it’s a matter of application and intent.
Pathology is not any different. There’s a massive opportunity, whether it’s in the domain of precision medicine or workflows, to rethink what the future of pathology looks like. I’m not suggesting pathology will see that kind of transformation overnight. Medicine in general is a tough nut to crack for new technology introduction, but when it cracks and when technology is adopted, it can be truly trans-
formative.
Pathology has the benefit above almost every other specialty in medicine of being at the cusp of digitization in the first place. There’s a real convergence happening today between the potential of AI and the fact that labs are beginning to go digital en masse. As they move toward full digitization, it’s exciting to see how realizable AI’s promise becomes when labs can build the foundation for it from the start. We used to say pathology is the last ology to go digital, but as a result it will be the first ology to see truly native AI interwoven into routine practice.
Dr. Marilyn Bui’s views are hers only and not those of the Digital Pathology Association, the CAP, or Moffitt Cancer Center.