Connected Pathology in the UK: Part 4

Digital pathology is all about improving histology workflow efficiency. But it’s only as good as the people operating it. The right knowledge and skills are paramount, so the importance of training and validation of users cannot be overstated.

Education is an important part of the Leeds-based National Pathology Imaging Co-operative (NPIC)’s remit. Along with the dizzying task of rolling out a connected digital pathology network across the National Health Service (NHS), NPIC is responsible for building knowledge around digital pathology. And that’s not only for the laboratory medicine professionals who use it, but also for the patients who benefit from it.

Bethany Williams is charged with coordinating this massive effort, and we caught up with her between training duties to chat about patient positivity, laboratory skepticism, and making digital pathology mainstream.

Can you tell me about your background and your role on the NPIC team?
 

I'm a pathologist by background, but, over the last 10 years, I've been on a very exciting sort of “choose your own adventure” career. I've diverted from the standard pathology training to get involved with all the exciting digital pathology projects that have evolved over the years at Leeds Teaching Hospitals, and now, of course, at NPIC.

It all started with some research work that I did into some of the safety aspects of digital pathology. The study looked at how we can best compare digital diagnosis with standard microscope diagnosis, to see if there are any particular pitfalls or considerations for the pathologist. From there, I moved on to a PhD in patient safety aspects of digital pathology. A lot of that work was around developing the validation protocol for pathologists. That has since been adopted by the Royal College of Pathologists, and many other organizations across the world, to help pathology professionals find a way to reach a level of not just competence, but confidence in their digital diagnosis.

From there, we’ve been looking at how we can safely and practically deploy digital pathology systems in busy clinical environments like the NHS. My current role, leading from that, is as the lead for knowledge and skills at NPIC, which covers two distinct areas. 

Firstly, there's the educational training piece. I'm still very much involved in the validation rollout across the whole of our national networks and partners, as well as providing some input for international validation efforts. I'm also in charge of our other educational offerings, such as multiprofessional opportunities for training and development in digital pathology. This includes our free-to-access webinar series, our in-person workshops, and other events as they come up. Alongside that, we're continuing to produce practical, pragmatic guideline publications for laboratory medicine professionals, laboratory managers, and anyone else with an interest in digital pathology adoption, on how to deploy these systems.

The second aspect of my work, which I find really exciting, is my work in patient and public involvement and engagement. It’s not just pathology professionals we need to bring on board, we also need to make sure we build and maintain public trust in everything that we do. And that’s part of being accountable for and transparent about what we're doing. It's crucial that we involve patients and members of the public in the decisions that we make and the partnerships that we form.

What does a typical working week look like for you?
 

I spend some of my time writing a mixture of peer-reviewed publications and guidance documents. For example, at the moment we're working on some pieces on the use of digital slides in medical education and how to effectively involve the public in digital computational pathology.

My work also involves disseminating best practice to the relevant audiences. I regularly arrange or deliver talks, webinars, and Q&A sessions on a wide variety of topics, disseminating my own work and the work of the entire NPIC team.

I also spend time on my patient and public involvement work, such as organizing meetings and events with our patient advisory group. Overall, my work is a mixture of planning and delivering all sorts of interesting and educational outputs.

In what ways is NPIC trying to achieve the aim of making digital pathology mainstream?
 

Looking back a few years, digital pathology was seen as a niche specialism – more of a “nice to have” than a necessity. Slide scanners were the preserve of particularly academic departments, or research institutions that may have had a particular individual pioneer spearheading this niche area of pathology research. But, over the years, we've seen digital pathology grow from a topic that was given an hour or so at a national pathology meeting to become the meat of pretty much every pathology meeting, whether it calls itself a digital pathology meeting or not. If you look at those meeting agendas, you'd struggle to find sessions that weren't heavily influenced by digital pathology, either as a topic or as the medium for the research.

Leeds Teaching Hospitals, and now NPIC, have been instrumental in trying to break down that barrier and make digital pathology accessible to real-world hardworking pathologists who are just looking for ways to safely and effectively deliver great outcomes for patients.

I believe digital pathology can be mainstream – largely because you don’t have to be technically minded to appreciate it. I came into this from the practical aspect of looking at how this exciting technology could be adapted safely and what the possibilities could be. My aim is to provide very pragmatic, easy-to-understand publications that look at all the different strands that you need to put together to deliver a successful digital pathology transformation programme. We’re particularly proud of the Leeds Guide to Digital Pathology, which has been popular with laboratory medicine professionals all round the world. It includes all the different aspects of digital pathology that need consideration – IT, infrastructure, project management, training, accreditation – and presents them in a practical, accessible way. I believe it’s been translated into 12 or 13 different languages now.

We also try to deliver those messages through initiatives like our free webinar programme, and by attending as many events and responding to as many requests for help, advice, and training, as we possibly can. Although we are a limited staff, we endeavor to produce these enduring resources and make them freely accessible to as many people as possible.

I aim to address the questions that people on the ground really want answers to: Where do you put scanners in a laboratory? How long does it take to train somebody to operate a scanner? Who should be operating a scanner? Simply, the guidance that you won't find in a technical spec or a paper published in a glossy journal. It’s essential to set out the foundational things that you need to make digital pathology work in a real-life busy clinical environment. That becomes your foundation for all that really exciting developmental work down the line, when we move on to the next phase of computational pathology and artificial intelligence.

What sort of reservations, objections, or concerns do you hear from pathology labs?
 

In all the different departments I've visited across the world, there's always a clear demarcation of enthusiasm amongst the laboratory professionals. I’ve identified three different mindsets.

Firstly, you have the cheerleaders, who just want to get it installed and get going. These are fantastic people to have involved in any discussions around funding or business cases for digital pathology adoption. But they're also the people that might want to run before they can walk. They're great allies, but you also need to be a little cautious.

At the other end of the spectrum are the pure skeptics. They're the people that I spend a lot of time engaging with in a department where I'm asked to help out. They are actually great to work with because they tend to be very vocal and articulate about future concerns. Sometimes they identify problems that the NPIC team maybe wouldn't have foreseen. In that way, skeptics are great for helping you to lay out your plans and anticipating bottlenecks and areas where digital approaches might not work as well.

The hardest group to work with is that group in the middle – the people that aren't quite sure or don't like to get involved in the meetings or conversations about digitization. But I have noticed that, over the last four or five years, many people have essentially moved up a level of acceptance, and are maybe feeling more inclined to give digital a chance. I think that's partly because of the evolution of the technology. Some people’s first experiences of digital pathology would have been watching an image resolve very slowly on a flickering screen, with an unreliable broadband connection. Whereas now, if you look at the clinical systems that are available, we see high resolution images with really quick movement of pixels across the screen. We're at a stage where I think most people that sit down with an optimized system could say, “Oh, yes, I can see this could work. This could be feasible in a real-world clinical setting.”

Are you seeing a shift in perceptions of digital pathology?
 

Yes. The big thing that's shifted perceptions is our experience of the COVID-19 pandemic. We were very fortunate that, in Leeds, we had already had our digital system up and running when the lockdowns started. We were already set up to support the needs of pathologists who needed to shield or to collaborate with their colleagues to try and keep those vital diagnostic services operating under that emergency situation. And, for pathology labs without that capability during the pandemic, that might have been the push needed to make the investment.

The other area that was served by digital pathology during the pandemic was the education and training of our junior pathologists. That's the other benefit of digital pathology, of course; it really allows equitable access to high-quality, tailored learning opportunities. I think that, for a lot of those very skeptical pathologists, the educational aspect was actually their entryway to digital pathology. And once they started using it to support their junior colleagues, and to do that training aspect of the job, of course they started to appreciate the broader benefits.

The other big thing that's really shifted people's perspectives on digital pathology is the opportunities to work remotely. I think that's been a real game changer in shifting the view that pathologists have to be in a lab to sign out.

It's interesting because people often assume that the most resistance to digitization would come from our most established pathologists. But, actually, that pre-retirement group, in our experience, are really interested in working digitally, because they see that as their opportunity to start to explore that preretirement, work–life balance. It gives them the opportunity to continue to offer their valued service to a department, but maybe with more home working options. And I suppose that flexibility option also applies for our trainees and consultant pathologists who might be working part time or seeking more flexibility in their working hours and location for various reasons. It's a great tool for supporting our pathology workforce at a time when we really need to be retaining all the fantastic people in our departments and attracting more trainees and more doctors.

Digital pathology almost allows us to rebrand pathology as a very modern, open, flexible opportunity. I can't think of another area of medicine where you would have the same opportunities in terms of not just the clinical component, but also the research and educational opportunities available as a doctor. It's a really unique opportunity to have a very interesting and varied career. And digital pathology just enables that to proceed to the next level.

Do you find that pathologists can be intimidated by the validation process for digital pathology use?
 

I think some pathologists definitely are. We're asking a very busy group of professionals to do something that's going to be additive to their workload, initially. Obviously, whilst you're trying to obtain digital fluency, there is some extra work that has to be done. The research work that I did at the start of my career indicated that there were certain diagnostic features and objects that do have a slightly different appearance on digital slides versus glass slides. Part of the training is in identifying those areas where there are slight differences between the appearance of features digitally and on the light microscope, and awareness of the situations where it might be more difficult to make a digital diagnosis.

The validation allows you a period where you are directly comparing these features, and matching your own performance on digital with your own personal performance on glass. The initial training phase includes looking at a set of curated digital slides that have some of these known pitfalls in them. Things like mitotic scoring and identifying anopheles can be tricky at first, so it's packed with lots of cases like that, that you look at digitally, compare with glass slides, and see if there's any difference.

Then you just start working on your usual workload, looking at every case digitally, in the first instance, followed by a safety check on glass on some of the key slides from that case before you sign it out. And then you collate your experience.

This approach allows pathologists to gradually gain confidence in their digital diagnoses and build up the number of cases reported digitally as part of routine practice. There is definitely extra effort required on the part of the pathologist. But we believe digital technology brings a lot of value to pathologists. And it gives our patients that extra bit of confidence that their case has been appropriately reviewed by somebody that's not just competent, but really confident in using the digital pathology images.

What's in it for the patients?
 

First and foremost, there’s the reduction in time to diagnosis. Let’s look at patients waiting for cancer test results, for example. Historically, it might have taken up to two weeks to receive glass slides from a peripheral hospital. But with one of our specialist cancer multidisciplinary teams, that transfer is now done instantaneously, as soon as the slide has been scanned.

When delicate glass slides are transported across the country, they can be lost, damaged, and delayed. The digital slide, however, is a permanent record of that patient's very valuable histopathological specimen, and can be rapidly transferred.

We also hope that digital pathology will improve the quality of our services for patients. If a pathologist can instantly share a slide with a colleague or expert in their network, they are able to check borderline-type cases with them in real time. For example: do you think this is microinvasion? Yes or no? That helps drive up quality.

There are all sorts of other little applications that benefit the patient. I could instantly compare a patient's new specimen – perhaps for a new suspected metastasis – with their previous specimens to provide a comparison to help determine if it’s a recurrence of an old tumor or something completely new. That saves time and effort of following that story and finding out what's going on for that patient.

Looking ahead, we’ll have the option of computer-assisted diagnosis to further improve the accuracy of some of those aspects of the pathology reports that are more amenable to computer analysis than human analysis. Tasks like scoring and counting are subject to variation between pathologists, or even between different times of day. But the difference between 75 percent positive for a particular marker, or 85 percent positive, might push a patient into a different treatment category. And these are tasks that can be standardized using a computer and algorithmic tool, freeing up more time for the pathologist brain to do what it does really well: drawing on all those different subtle inferences from years of histopathology training, as well as broader medical knowledge and the significance of that patient's past medical history to determine how that might affect this particular presentation of this particular case.

How are patients involved in the NPIC program?
 

We have a very active patient and public involvement and engagement program. The engagement side is around making sure that we are informing the public about what we're doing about what developments are taking place within pathology. We communicate how these could potentially benefit the public at large and patients individually. The involvement aspect is about members of the public actively influencing the NPIC program.

As well as this, we have an enthusiastic patient and public advisory group, composed of diverse members of communities across our regional footprint, who meet together regularly. We ask for their input into aspects such as information governance plans, potential clinical or research partnerships, and particular research questions that we’re developing.

In addition, many of our projects involve large amounts of patient data and images. We want to make sure that we are transparent in how that data is used, and that we really listen to the needs and interests of the public around that. This makes patient input so important when we're making decisions both about clinical service delivery and research.

We had some fascinating sessions recently, delving into artificial intelligence, involving some good debates. We have also become involved in some interesting projects looking at medical and, in particular, AI literacy. Last year, we put together a touring exhibit, created with our patient advisors, that went up in public libraries and also online, explaining aspects of medical AI and, in particular, pathology AI. It gave answers to all the key questions that our advisory group proposed. And that's all part of raising that awareness and understanding as we're making this journey forward together. These are exciting times!

What lies ahead for you and the NPIC team?
 

We have more of our in-person workshops coming out, and we'll be repeating the free webinar series that we ran last year because there was a lot of interest in that. We also have new publications coming out around patient and public involvement in digital pathology and use of digital slides in undergraduate and postgraduate medical education.

There is also one of our in-person patient advisory events coming up. Those are always exciting – we give our patient advisory group the chance to grill all the NPIC senior staff about what we're doing with digital and computational pathology.

And, of course, NPIC will be popping up at various digital pathology and pathology events for lots of different audiences. We have a very approachable team of great people that are really willing to share their expertise and their enthusiasm. So, if you see the team at any of these events, please do feel free to approach them with all your most challenging questions and put them through their paces.