The key blockers to the adoption of digital pathology
Posted 14th January 2019 by Jane Williams
In the first part of this six-part blog series, we looked at the challenges facing the pathology department. The conversation then moved onto the key blockers standing in the way of the adoption of digital pathology.
If you weren’t able to make the panel discussion, you can watch the recording here.
Peter Hamilton Efficiencies of 25% have been reported for the use of digital pathology in diagnostics. Why are we not seeing the level of adoption we would expect?
Jo Martin Cost of the system and increasing the number of steps in the process are commonly cited as reasons. If you think about a chemical reaction, the amount of energy you’ve got to put in to change the system when a system is under pressure is actually another potential blocker. There’s something around not knowing the system, in particular, the management around pathology, not really knowing the capabilities, and IT systems. The IT systems across many health services are under enormous strain. All those are potential barriers to implementation.
David Snead Cost is an important one. The systems still represent a significant investment in cellular pathology in most departments. There is certainly a significant efficiency gain to be made by going digital and the problem is in the quality of the data, how that big that gain is, and what it’s going to mean for your individual department.
To some extent, that’s unsurprising because the workflow in a cellular pathology lab is a complex beast; there’s a lot of parts to it and it’s difficult to measure it efficiently and well. You’d end up trying to present a business case with a significant capital outlay and have slightly questionable ROIs for the organization.
My colleagues are enthusiastic generally speaking about moving to digital. Most pathologists, but not all, will certainly use it, and I think once they’ve used it, they will be converted. I quite happily report all my work digitally, and I am entirely confident that any other pathologists would do the same if given the same opportunities.
Peter Hamilton It was about 18 months ago that NIHR funded a large project to help validate digital pathology in breast and colon screening. Can you tell us a little bit more about that and what it intended to achieve?
David Snead: It was a bit deflating after we’d just published our last study to receive the memo saying we weren’t allowed to use digital pathology for cancer screening samples. I’m still slightly bemused as to what was behind it, but I have to say that having talked to my colleagues that do these reports, they don’t see a particular problem with doing them digitally. We do them on glass, but they are as bemused as I am as to why that advice came out. But we are engaged and in round two of that study to look at that specific element of it, in cancer screening samples, particularly in colon and in breast.
We’ve got a great study lined up that will look at 2,000 samples and will have multiple readings from each sample. Each sample will be reported twice by four pathologists: once on the glass and once on the digital. We’ll have multiple data points to compare the difference between the two methods of reporting. I’m confident that would give everybody the evidence they need that digital pathology is as safe as glass microscopy for routine reporting.
Neil Mesher I’ve been working now as part of the Innovate program for the best part of 2 years. I’ve never seen more pull to a single piece of technology in my time in Philips than we’ve had over this particular program. There has been a perception perhaps that there is some reluctance to go down this route, digitize, to use machine learning, artificial intelligence but I think a lot of that is gone. I see a huge degree of engagement from the profession to look at different ways to do things. This is a profession that’s ready to get going.
Peter Hamilton Pathologists have a huge appetite for adopting digital technology, but are you seeing that on the ground, Jo?
Jo Martin Pathologists, given the chance, are really good at change. I know that sounds contradictory, but microbiology has gone completely molecular. They’re using mass spectrometry. We do this every day. Clinical biochemistry will adopt stuff that goes in nationally, nobody even notices. Liquid-based cytology, we put that in. We put HPV in. Give us the chance. Give us the investment. Give us the project management help to actually get the stuff in, and we can do it. Departments like David’s and the Leeds Project shows that you can actually do it; we’re just a bit pushed at the moment.
Neil Mesher We had many different institutions that wanted to participate in the program. This week, the government has announced funding for further activity, so there will be more opportunities for institutions like Coventry to take part. We should start to think about the opportunities and begin to leverage them.
We can learn something from radiology even though it was in a slightly different position. 25 to 30 years ago, the creation of MR and CT produced digital images. For 100 years, radiology has taken whatever technology is available and then produced a black and white image from it. We’re still in that kind of position today with MR and CT. We have these fantastic data capture devices, and some still look at the black and white image to read it.
As we digitize pathology, we need to think far more holistically about making joined-up data access for those image databanks. That’s something we can learn from the radiology journey.
Peter Hamilton Juan, you’ve been successful in adopting digital pathology for routine diagnostics in Granada. How did you go about presenting your case to your organization?
Juan Retamero We saw it as an opportunity to modernize service provision in our region. It is ridiculous that if I lose my suitcase in an airport, it will be fully trackable but the same doesn’t necessarily apply to cancer specimens. The same occurs with our food – it’s fully traceable to the source, but not necessarily the same with cancer diagnostic specimens. That’s something we need to modernize.
Tracking becomes the foundation upon which you are able to do digital pathology. It’s a phased approach, starting by tracking, moving into digital pathology, and then starting to contemplate computational pathology. You need to get your house in order.
Peter Hamilton Let’s talk about the key improvements that you and your pathologists have seen in Granada.
Juan Retamero You can see that digital pathology has associated its use of efficiency gains that we can quantify in the region of 20%. Of course, there might be compounding factors that we’re not aware of, but the point is observing more cases with fewer pathologists, and that’s undeniable. That occurred around the time we are able to do digital pathology. Obviously, an association is not necessarily causality but they probably go together in some respect.
When it comes to emotional reasons, pathologists prefer to work digitally. It feels like a better tool for the job at hand. Our pathologists seem to be quite happy, so we must have done something right.
It is a better working proposition. It’s a better tool for doing the job day in and day out, but it also makes you efficient, apart from being safer for patients as well. Those have been the gains that we have experienced in the past two and a half years.
Peter Hamilton There is a common misconception that by introducing that additional step of scanning, it increases the turnaround times? What’s your experience?
Juan Retamero We had a bottleneck somewhere else in the lab that we hadn’t realized. It was a case assembly and distribution, and digital pathology has resolved that bottleneck. We don’t think there are any major increases in turnaround times, so at least certainly in our institution, it was not a big problem.
David Snead The bottleneck in most laboratories is usually the pathologist’s pigeonhole. I don’t think the scanning part of digital pathology is a significant factor for us in slowing cases down. As a matter of fact, it speeds things up because any pathologist can get access to a case. A lot of really excellent work on improving turnaround times, particularly in histopathology labs, has been done around how cases are allocated to pathologists. Whether they pool the cases or pre-allocate them and push them to the pathologist makes a huge difference in the amount of time it takes to report them.
From our point of view, the blocks around putting the system in and getting it working properly were around interfacing with the laboratory information system. It’s a critical part of the whole process because it’s what drives the system. All the information is already in the laboratory system that you need as a pathologist to report the case, so it needs to flow seamlessly down with the digital slide, so the pathologist gets that information in one shot. Likewise, as you report the case, you can just use the one-system report and the report goes back to the clinicians there and then.
These things are fairly fundamental but they make a massive difference to the effectiveness of a really expensive piece of kit. Getting that right is important and it’s not straightforward. Many of my colleagues in blood sciences grapple with this problem every time they change analysers. It’s a major piece of work around interfacing new kit with a pathology information system, so it’s not particular to pathology, but it’s an important part of the process.
David Snead is Consultant Histopathologist and Clinical Service Lead, Coventry and Warwickshire Pathology services, Jo Martin is President of the Royal College of Pathologists, Juan Retamero is Pathologist at Granada Hospital and Neil Mesher is CEO of Philips UK and Ireland. With thanks to Philips for sponsoring this discussion.
Leave a Reply