For hundreds of years, the key tool in a pathologist’s kit has been the microscope. Through this they have peered at samples of human tissue and bodily fluids, making expert judgements as to whether disease is present.
Go to a pathology lab these days, and it’s likely you will still see microscopes in pride of place on the bench. But they’re increasingly likely to be accompanied by computers, and the lab populated by pathologists who are spending as much time looking at screens as they are looking down the scope.
This shift of focus is thanks in part to the growth of digital pathology, which involves making copies of slides via a sophisticated scanner. The image is then made available digitally, with advanced software applications allowing pathologists to view and report on it.
This is a more flexible setup than the traditional slide and microscope. In that situation, a sample can of course only be in one place at one time and, if a pathologist wants to see it, he or she needs to be in that same place. No such issue with digital pathology. Any clinician with a networked or internet-accessing device can review and assess the image.
That could provide pathologists with a much higher degree of flexibility in how and from where they work. It would become far easier to share workloads, and to consult with colleagues.
Traditionally, getting the opinion of an expert from another organisation would mean putting the slide in the post – with the attendant risk that it could get broken along the way. No such problem with a computer image.
Going digital could also aid diagnoses, with researchers exploring the application of machine learning in pathology. With cancer, for instance, the idea is that it might be possible to ‘train’ computers to recognise benign samples. This would allow pathologists to solely focus their energies on more complex cases where the result is less clear.
And when cancer is present, pathologists may no longer have to make a potentially subjective judgement on its type and aggressiveness based solely on what they see through the microscope. Instead, software packages with image analysis algorithms may be able to support in making an assessment.
Given these potential benefits, it’s perhaps little wonder that a November 2016 report from Cancer Research UK argued digitisation was key to the sustainability of pathology – a sector facing growing demand and increasing issues with recruitment and retention.
The Royal College of Pathologists, too, has recognised the possible advantages of digitising its field. But its August 2017 strategy on diagnostic digital pathology also identifies possible obstacles.
An important one: image storage and retrieval. The strategy points out there could be capacity and cost issues at play here. After all, the typical size of an scanned slide is 500 megabytes. It won’t take long before pathology departments are dealing with terabytes and petabytes of data.
There’s also the question of how to ensure images are – and remain – accessible regardless of which software solutions individual organisations might choose to use.
Could the cloud offer a viable solution to these storage and accessibility challenges? Many argue the answer is yes. UKCloud Health, for instance, has a platform which is already home to the data collected for England’s 100,000 Genomes Project.
Each genomic sequence requires more than 250 gigabytes of storage – a figure which can suddenly make the 500MB required for a pathology image seem paltry. Authorised researchers gain access to the sequencing data over the internet, using a secure virtual desktop which sits on top of the UKCloud platform.
Many point out that similar cloud-based solutions could be used to manage the data collected through digital pathology. This would allow pathologists across the NHS to easily and securely access images – and, indeed, to more easily share information with those working in other health and life science fields.
Their employers could also find the cost of cloud storage less prohibitive than more traditional methods. UKCloud Health, for instance, has a model by which clients only pay for what they use.
Not every pathology department is going to find it viable to immediately digitise its diagnostic pathology processes. But it may not be too long before many pathologists stop looking down into microscopes and start looking up into the cloud.
Claire Read is a freelance writer and editor. She has specialised in healthcare throughout her 17 year career, and has a particular interest in technology. You can find Claire on Twitter at @readthewriter.