Aug 12, 2019 - reading time 8 mins
By the FHI editorial team The Philips Future Health Index editorial team is always on the look-out for great content pieces that discuss the future of health, selecting the most interesting health-related stories for you to read.
There’s little doubt that the workload for pathologists will increase over the coming years. According to the International Agency for Research on Cancer, new cancer cases are expected to rise 63% in the next two decades, and the Association of American Medical Colleges (AAMC) reports an 8.3% drop in the number of active pathologists between 2012 and 2017. Computational pathology, which uses big data integration and image analytics powered by deep machine learning to enhance diagnostic precision, has the potential to be an important part of the solution to these challenges. And, while the 2019 Future Health Index says that 47% of healthcare professionals globally are comfortable using AI for diagnosis, there is clearly a way to go before these types of technologies become mainstream. To discuss how computational pathology can help bring about a precision diagnosis future, we spoke to two of the field’s leading voices. David Snead is a consultant pathologist and leads the University Hospitals Coventry & Warwickshire (UHCW) NHS Trust Digital Pathology Center of Excellence in the UK. Nasir Rajpoot, also based in the UK, is Professor of Computational Pathology at the University of Warwick and Honorary Scientist at the Department of Pathology, UHCW NHS Trust.
How would you define computational pathology? David Snead: In my mind, it’s the application of computer algorithms or solutions generated by the computer, which provide some elements of the report that the pathologist sends back to the clinicians. It interprets data that’s presented to it and provides a computer-generated analysis of that data, which is then used to influence the report and hence how the patient is managed. Nasir Rajpoot: I agree with this and would just add that while this is an emerging branch of pathology that uses algorithms to assist with the diagnostic process, it can also help with the prognostication of disease and predicting patient responses to certain therapies. David Snead: Nasir’s right that this really is an emerging area. None of us, I don’t think, know quite how far-reaching it’s going to get – but it’s certainly going to be a game-changer. You hear the terms ‘computational pathology’ and ‘digital pathology’ discussed, and how they can enable precision medicine. Is there overlap between these terms, or what are the key differences between them? Nasir Rajpoot: Think of digital pathology as a prerequisite for computational pathology. You can’t do the latter without the former. Digital pathology allows pathologists to view and examine microscope slides on-screen digitally. It covers the digitalization of microscope slides so that they can be viewed and analyzed on a computer. Computational pathology relies on having that digital image and describes the analysis process that is performed on that data and the application of that to improve pathology workflows, efficiencies and diagnostic precision. Do you have more detail on what the next steps could be over the next two or three years? David Snead: We expect to see the first algorithms coming into practice over the next few years. There are some out there that aren’t really being used and the reason for that is they don’t fit that well into the existing workflow – so these new algorithms will need to be designed with existing workflows in mind. This means that the computer ideally performs its analytical work up front, before the pathologist gets to review the digital slides. When the early adopters take this up, we’ll start to see what can be achieved in terms of accuracy and efficiency boosts. Looking at the bigger picture, say in five or ten years’ time, what might be the improvements to both the healthcare professional experience and the outcomes for patients? David Snead: Patients will have access to better information on how likely they are to respond to certain treatments. And that’s important because the treatments we’re talking about come with risk factors and side effects. Patients are always facing this quandary of deciding whether to undergo a treatment and weighing that up against expected improvements and risks. With computational pathology and precision diagnosis, the clinician will have far more information at their disposal with which to advise the patient on these kinds of questions.
What is the potential workplace efficiency impact of computational pathology? David Snead: It will be able to automate tasks – either removing them from the workflow entirely or speeding up the time a pathologist would take to do them. For example, in the recognition of normal tissue and finding regions of interest in samples which are abnormal. This work will save some of the time spent by pathologists looking for areas of abnormality, such as the spread of breast cancer to lymph nodes, or prostate cancer in prostatic chippings. These are exciting developments from a service and capacity point of view. We've been thinking about those for some years now and we’re convinced that this is the right track. But to do this, you have to digitize your workflow. That's going to drive changes in laboratory configuration, changes in the way the services are provided across regions of the country, as well as how pathologists are trained. So there are quite significant things to get to grips with. I think it’s fair to say that most pathologists are highly skeptical that this technology is ready or nearly ready for adoption. I think they can all see it coming but are thinking it will take years to arrive and may be quite limited in its extent. We must produce the data to convince them, but pathologists are only one facet of the service. A lot of the decisions about implementing these ways of working will be made by clinical directors, chief medical officers and hospital management committees. Pathologists are an important voice, but they’re not the only one. Are there any countries outside the UK that are leading the way in this field? Nasir Rajpoot: There have been major inroads in Canada and some European countries, such as Sweden. We meet with colleagues from all over the world and UK investment into ISCF centers of excellence on AI in pathology is as good as anywhere in this discipline. The Future Health Index 2019 report indicates that the general population in China, Russia and Saudi Arabia are more likely to associate AI technologies with more accurate diagnosis. Why do you think this might be the case? David Snead: Most patients have a very limited understanding of what happens behind the scenes of their care. And pathology is, certainly in the UK and I suspect most countries, a very backroom job and is not something that patients normally interact with. The countries you mentioned probably are an exception, in that it's not unusual for patients there to hang on to their samples and scans and bring them with them when they go to consultations. These are health economies where the patient keeps his or her own records, their own data if you like, with them and go to the next doctor. So that is quite different to how most health economies work. Nasir Rajpoot: With regards to sharing data, there is some work to be done. The industry, academia and NHS need to work together with funding agencies and patient groups to come up with clear guidelines for the ethical use of medical data for research purposes and for the benefit of the population at large.
Dr David Snead, Consultant, University Hospital Coventry David is a full time NHS consultant pathologist and has been lead pulmonary and skin pathologist at the University Hospital Coventry since 1997. He has led the project to adopt digital pathology at UHCW for the last 5 years. He now leads the UHCW NHS Trust Digital Pathology Centre of Excellence (CoE), an academic venture exploring the use of this technology in routine histopathology. The CoE collaborates closely with Professor Nasir Rajpoot and colleagues at the Computer Science School University of Warwick and is currently evaluating algorithms aimed at improving pathologists’ ability to accurately grade cancers and the automation of mundane time consuming quantative tasks.
Professor Nasir Rajpoot, Professor of Computational Pathology, University of Warwick Hailing from the ancient Asian city of Multan, Nasir Rajpoot is Professor of Computational Pathology at the University of Warwick and Honorary Scientist at the Department of Pathology, University Hospitals Coventry & Warwickshire (UHCW) NHS Trust. He is the founding head of Tissue Image Analytics (TIA) lab at Warwick since 2012 and also co-Director of the recently funded £15m PathLAKE centre of excellence on AI in pathology since Jan 2019. The focus of current research in TIA lab led by Prof Rajpoot is on developing novel computational pathology algorithms with applications to computer-assisted grading of cancer and image-based markers for prediction of cancer progression and survival. He has been active in the digital pathology community for over a decade now and has delivered over 50 invited and keynote talks since 2015 at various national and international events and institutions. Prof Rajpoot was recently awarded the Wolfson Fellowship by the UK Royal Society and the Turing Fellowship by the Alan Turing Institute, the UK's national data science institute.
You are about to visit a Philips global content page
ContinueYou are about to visit the Philips USA website.
I understand