Diagnostics at a Distance

Digital pathology is a hot topic right now. The transition to computer-based services is on everyone’s mind – to speed up service delivery, to aid in teaching and training, or to provide long-distance consultation on unique or challenging specimens. But although tales of success (and discussions of difficulties) abound, we don’t often hear the behind-the-scenes stories of implementing and running a digital telepathology laboratory. At The National Hospital for Neurology and Neurosurgery, one of the specialist hospitals of University College London Hospitals (UCLH), we have become the sole provider of diagnostic neuropathology to Brighton and Sussex University Hospital (BSUH), meaning that we offer regular long-distance services via digital technologies. We began providing routine neuropathological services on an ad hoc basis in 2013. Then, Brighton went for tender with a detailed service specification and we applied as a bidder – a contract we won in 2015. Ever since, we’ve been providing them with a full diagnostic neuropathology service.

How our system works 

Most of the specimens we examine don’t require intraoperative assessment, so they are shipped directly to our laboratory in fixative. Upon arrival, we treat them the same way we treat all other referred and local specimens – that is to say, they are booked in, cut up, processed, stained and reported. After reporting, the only difference between these and local cases is the electronic transfer of reports. Whereas our hospital’s cases and other ones are transferred directly onto the hospital information system and are immediately visible, Brighton’s go to a centralized email actioned by office administrators and forwarded to the doctors.

When we need to handle intraoperative specimens from BSUH, it’s always done digitally so that we can give rapid feedback. The surgeons usually notify us by email or telephone that they’re working on a patient who will need intraoperative service. Then, during the procedure, they submit their material directly to the local pathology department at BSUH. A trained biomedical scientist on-site performs a smear (the most common method of intraoperative microscopic assessment of neuropathological specimens), scans it, and notifies us that it’s ready to be reported. We log onto the system via the NHS N3 network (a private data network designed to ensure patient confidentiality and availability at all times) and view the specimen remotely. Our report is typed into the appropriate section of the slide management software and locked to be compliant with ISO standards. We make the diagnosis on our screen, pick up the phone, and discuss our findings directly with the surgeons (see Figure). We want our service to feel as close to having an on- site pathologist as possible – and I think we’re succeeding; the only difference right now is the additional scanning time. 

In theory, the remaining part of the tissue pathway – its processing, sectioning and staining – could also be done at BSUH, but we talked it over with their pathology department and decided that it would be preferable in terms of quality assurance, cost and turnaround to perform those tasks in our specialist center. Neuropathological specimens need an entirely different, specialized spectrum of immunohistochemical and molecular tests to those available at BSUH, and it would be expensive for them to set up, maintain and validate such a complex test portfolio for a relatively small number of samples. We routinely do these specialist tests in large quantities at our center, so we can offer them a solution that’s not only cost-effective, but provides better quality control. As most intrinsic brain tumors require additional molecular diagnostics, we can initiate these tests immediately upon examining the H&E sections, ensuring that most molecular tests are back within two to three weeks of the operation. 

For a fully integrated digital pathology remote reporting system, we could easily report sections prepared at BSUH, which would eliminate the two to three working days needed to transport specimens to our center. Until we implement that, though, I anticipate that our system will continue in its current form for some time, because it works very well – the way intraoperative sections are reported is ideal for a remote or digital pathology setup.

The good and the bad 

The high quality of modern digital pathology has made this kind of remote consulting much easier. We now have rapid, high-quality communication with surgeons across a distance of 60 miles even as they operate on a patient. This is nearly equivalent to having a pathologist available locally. That’s not to say that it’s without its drawbacks, though; the workflow, of course, is slightly more involved. We have to establish communication between the remote laboratory and the operating theater staff. We do a good job, though – many of the neurosurgery trainees in Brighton aren’t even aware that the pathologists reporting on their patients are at an entirely different center! 

The most time-consuming element of establishing a digital telepathology service isn’t necessarily the communication process, though – in our case, it was coordinating the IT departments and setting up the logistics of a fast, secure network connection. To ensure data could be transmitted securely, we needed to establish a connection accessible through a restricted number of computers in our department – which meant limiting it to a range of IP addresses. There are pros and cons to this – the advantages, of course, are the speed and security, but the downsides are a lack of access when not on site and a limited choice of hardware. Nonetheless, now that the system is fully established, it runs so smoothly that we may as well be on the ground in Brighton! 

Who would benefit most from the transition?

There are two fundamentally different scenarios in which a need for telepathology might arise. The first is a department that is staffed with a team of pathologists, but lacks specialist expertise in a few areas. In this scenario, digital pathology would help with the provision of second opinions on both a regular and an ad hoc basis. This option is well-suited to district general hospitals with a limited spectrum of expertise, but a wide range of pathological specimens, some of which may need a specialist opinion. It works especially well if the immunity chemical tests for such cases are available locally. The other potential scenario is the concentrated expertise that can be found in a center’s specialist area. These areas are what I call the “recipients” of digital pathology services – that is, they don’t need to scan any slides, but instead receive scanned slides from remote centers that require expert opinions.

There is one other possibility – a transition to digital reporting in the department where the sections are prepared. The business case for such a transition would look entirely different, and would mean a complete changeover from glass slides to computer screen- based reporting. Few departments have made this transition because of the obvious obstacles – chiefly the addition of work steps (consistent barcoding, automated scanning, automated slide and case management software, connecting to LIMS) that would be needed to enable digitization. That’s why I think telepathology is the obvious area to get started with digital pathology. In fact, it has already been available for many years, though its first incarnation was more primitive; it used to be delivered through remotely controlled microscopes. The limitation with those was the bandwidth of the Internet connections at the time – but now, our advanced software capabilities and better bandwidth make digital pathology a safe and effective reporting tool for remote pathology. 

Is a fully digital future realistic? 

Although the transition to digital telepathology seems like an obvious path, making the same move for the entire pathology department (for instance, digitizing all slides for exclusively online reporting) is a different scenario, and it’s very difficult to predict how well a rollout would go. If we look at more widely known technologies that seemed futuristic a decade or two ago there are failures and successes. Time will tell, for example, how successful electric cars are, or what the future of driverless cars really is. Under well-defined road conditions, these systems may work very well... but the devil is in the detail (like snow, rain, unexpected – but harmless – objects, or worse, harmful objects misidentified as “road signs”). Pathology is much the same. Fully digitized reporting must be properly field-tested, and in order to do so, the providers of integrated solutions must be open to discussing the pitfalls and risks. Are they? I’m not yet convinced that all of the potential pitfalls, errors, and lack of business continuity are really openly and transparently discussed and disclosed.

Generally, novel solutions will be adopted quickly if they are time- and cost-saving, especially if they allow users to simplify or omit parts of a complicated process. Digital cameras are a good example; adoption started relatively slowly because the resolution was inferior to film and the cost for a camera higher than conventional devices. But once resolution improved and price decreased, the demise of film-based cameras was very rapid. This development had the biggest impact in radiology, where digitization reduced a number of expensive steps – and now, many radiology departments are partially or entirely digital. These examples are in stark contrast to digital pathology, which requires additional equipment, IT infrastructure and software. Those are the obstacles we need to tackle now – because in my view, adoption of novel pathology technologies works better if they make people’s lives easier, cheaper and more convenient. 

Sebastian Brandner is Professor of Neuropathology at the University College London Institute of Neurology and Head of the Division of Neuropathology and honorary consultant neuropathologist at the National Hospital for Neurology and Neurosurgery, University College London Hospitals NHS Foundation Trust.