CAP TODAY Pathology/Laboratory Medicine/Laboratory Management
November 2023—Is digital pathology on the move? Two who know it well say it is. Esther Abels, a precision medicine and biomedical regulatory health science expert who is CEO of SolarisRTC and former president of the Digital Pathology Association, and Michael Rivers, vice president/lifecycle leader of digital pathology at Roche Tissue Diagnostics, spoke in September with CAP TODAY publisher Bob McGonnagle, who got their take on where things stand.
Abels
Where are we on an adoption curve for digital pathology?
Esther Abels: It has taken off. People are willing to implement it. Everybody is more or less starting to embrace it and to see the benefits of it. I recognize there’s not much in the field yet, and not as much as we hoped for. It’s about 10, 15 percent, but pharma and biotech companies, clinical research organizations, and big reference labs have already embraced it and are working with it. But do we see it taking off in clinic? Not so much yet, but with regard to the mindset I would say yes. I see that changing. More and more companies are working on this. I’m working with the FDA, and there is more and more involvement from the FDA, which is eager to move this field too. There is more collaboration, and last week the FDA reached out to me to ask for the direction on what they need to do to speed this up. That’s fantastic to hear.
There’s no doubt that digital pathology will also lead into computational pathology, and in doing so expand our brain power. I am positive that algorithms can be smarter than we are. Will it be better? Well, we have to ensure its use is safe and ethical.
Mike Rivers, same question, and I’d like to add that I don’t know of any pathology group of a certain size or volume that isn’t somewhere on this journey, even though many are just taking their first steps, but the goal will be to have a robust digital pathology program. Do you agree?
Michael Rivers: I absolutely agree. Where we are with this is the question we’ve all been asking ourselves for years. The tipping point has been predicted for a long time, but what I have seen is what Esther said about the mindset.
The pandemic, and coming out of the pandemic, has been a catalyst. Every lab has recognized that it has to find a way to enable remote pathology, that it is ridiculous to force pathologists to come into the lab just to read slides, and that it needs to find a way for remote diagnosis and sharing of slides.
That’s first and then, yes, there is the potential of AI to bring game-changing information for patients. That’s all pushing in the right direction, and I agree: Every lab I talk to is either in the process of digitizing or getting budgets set aside for it or planning for it. While I agree with Esther that the overall percentage of full digitization is still fairly low, that is the use case that everybody is actively pursuing, and I think we’re going to see a massive change in the coming months and years.
We know that the earliest adopters of digital pathology for full primary diagnosis were in the Scandinavian countries where there was already an acute shortage of pathologists. It’s only relatively recently in North America that we’ve been experiencing that sort of shortage of pathologists, but recruiting pathologists is a big conversation in the United States, so please comment on the role that the shortage plays in the desire to adopt digital pathology.
Esther Abels: It’s an important role because pathologists who retire continue to work, though maybe not as much and maybe only remotely, and as such digital pathology is a must because they can work a few hours from home, help the community and patients, and reduce the workloads of their colleagues.
The shortage will help drive adoption, but it is not the only driver because there’s much more than that. We know reimbursement is essential, as is interoperability.
Mike, your view of the role of the shortage?
Michael Rivers: It absolutely is a driver. I agree there are many drivers pushing this in the right direction, but there is a shortage of pathologists everywhere. The U.S. is in a pretty favorable position, but even here we’re struggling with that, and digital pathology can be an incredible tool to empower pathologists, to give them flexibility in the way they work, and then ultimately, once that digitization step has been achieved—and it’s a big step, requiring infrastructure, cost, implementation—then you can unlock the power of AI. And we’re seeing a variety of high-quality AI solutions that help with diagnosis and workflow, that automate simple quantification tasks, other simple math tasks that we ask pathologists to do. It could make them more efficient, so I do think it will be one of the mitigations to the shortage of pathologists.
Rivers
We’re seeing subspecialization in virtually every specialty of medicine, surgery in particular. And if someone is a breast surgeon only, they want a breast pathologist to read the material they excise or the cells they extract, and that has to be a strong impetus for digital pathology because there’s not enough subspecialty expertise to go around. Can you comment on that?
Michael Rivers: Easy access or fast access to second opinions is an exciting opportunity that digital pathology unlocks. It’s amazing how much shipping of glass slides is happening today, couriers traveling around towns, and digital pathology is the opportunity to get away from that and the lag time it creates in getting information to the patient. Digital pathology can bring subspecialty expertise into the equation more easily.
I’m at the tail end now of the European Congress of Pathology in Dublin, and a lot of data were presented on AI algorithms. One of the themes I continue to see is the opportunity for these algorithms to bring up the level of the general pathologist close to, if not nearly equivalent to, the subspecialty pathologists. So it has tremendous potential to raise diagnostic confidence, to raise the level of information provided, and to equip pathologists with some of that specialized knowledge.
Quest and Labcorp have been making acquisitions, but we also know that the merger and acquisition and partnership of very large systems continues to move forward. We’ve seen that with ACL and Atrium. We’ve seen Kaiser purchase Geisinger. The pure geography and the multiple hospitals and clinics that need to be served out of one system have to be sparking interest in digital pathology. Would you agree, Esther?
Esther Abels: I read that it is very much linked to it because with digital pathology you can do so much more. I’m not the expert in those labs, so I do not know all the exact locations, but I do recognize there will be a more centralized approach. That’s why I would say it does spark digital pathology, because we still need to have the slides, we still need to make sure the scanners are placed at the various locations. Or do we keep shipping slides to that central location?
Mike, from your perspective how is the consolidation of providers affecting the demand for digital pathology?
Michael Rivers: I agree with Esther that many factors are driving this. The many constituents and workflows present a challenge for digital pathology, but also a huge opportunity, and the potential benefits can be significant. As you integrate those workflows and connect the various sites, you can unlock big benefits, though I don’t want to make it sound easy. But these complex organizations will be able to extract more benefit once they become integrated. It’s up to us as manufacturers to be able to offer flexible solutions. I think scanning will continue to be done centrally just for the sake of efficiency.
Esther, we know reimbursement needs to be made, and first steps have been taken by having codes that it is hoped will hasten payment for digital pathology. In addition, let’s talk a little about the IT infrastructure of systems. We know that picture archiving and communication systems were essential to the adoption of digital radiology. Will we need to have a pathology PACS for the various images that come in? Will that be a final enabler in some ways of digital pathology, or can digital pathology stand alone outside of a PACS?
Esther Abels: PACS is very specific, and yes, digital pathology does need some system where all the images can be archived and stored and accessed. Do we need to have the same definition as a PACS? I do not think so, but I’m not an engineer, so don’t hold me to it. I believe it all comes down to interoperability—that is, the correct exchange of information between devices.
Michael Rivers: There’s not going to be a one-size-fits-all solution. There are organizations that will want a unified PACS and will want to store everything in one place, and we need to enable that. We also need interoperability within the image management software to be able to view different image types from different image sources, just even within the digital pathology world.
I would also extend this to AI. I’m seeing more and more investment by manufacturers in open platforms, environments that are able to plug in a variety of AI solutions, and this is critical. We have so many AI startups and a flood of investment into this space that is bringing great data and cool innovations. But for those to come into an efficient clinical workflow, we cannot ask pathologists in labs to log in to 10 different platforms that store images, so that’s going to be one of the key innovations over the next few years—connecting these things in an efficient way that allows the lab to still do its work but access all of these great tools.