We don’t need to hope that future technologies will solve all our problems, existing tech can already pave the way. Read how in part five of our “Barriers to Adopting Digital Pathology” series.
Pathologists are not known for being a technologically savvy bunch, which can make the road to digital pathology a rocky one. Many assume that digital pathology simply means getting a scanner, throwing in slides, and looking at them on a screen – only to discover there are a number of other questions that need to be considered. Where will images be stored? How can they be accessed quickly without crashing the system? How are they to be moved to the appropriate locations within the network?
These questions speak to the need for a robust infrastructure, but also create another impediment for departments wanting to convert to digital pathology. Beyond additional cost, they introduce numerous layers of novelty and complexity that could, to an ordinary pathologist, seemingly require equally novel and complex solutions.
It might surprise pathologists to learn that the majority of the components of a digital pathology infrastructure could be borrowed from pre-existing, common technologies. Take the storing of digital slides, for example. Most whole-slide images are between two and four gigabytes, akin to a full-length feature film in HD. Assuming 1,000 slides produced daily, that translates to an average of 3,000 gigabytes, or three terabytes, of new data each day. Over a year, this approaches a petabyte (1,000 terabytes) of data. It sounds gargantuan, but consider the graph below.
Historical cost of computer memory and storage
This data is expressed in US dollars per terabyte (TB). It is not adjusted for inflation.
For each year, the time series shows the cheapest historical price recorded until that year. Credit: John C. McCallum (2022) | OurWorldInData.org/technological-change • CC BY
Two things become evident: the first is that the cost of data storage has been consistently falling since the fifties. The second, perhaps more important point, is that something drove this decrease in price – namely, demand. Commercial businesses, research labs, and internet sites, to name a few, all depend on mass data storage to function. This demand has not only led to better and cheaper hard drives but has spurred a revolution in data storage availability via cloud-based services. For digital pathology, the upshot is that, unlike slide scanners, which truly have had to be developed from scratch, data storage is a tried and tested technology easily converted to storing digital slides.
Another illustrative example is the methods that allow for fast viewing of whole-slide images. If a digital pathology department has even 10 pathologists working simultaneously, it could translate to a crushing amount of data running concurrently over a network. However, techniques such as “progressive” and “lazy” loading have been developed to load images only partially at any one time, making the data transfer demand significantly smaller. These techniques are widely used, with the quintessential example being Google Maps – a whole-slide image in its own right – that holds approximately 20 petabytes of data and is used simultaneously by millions.
Digital pathology infrastructure has to address numerous additional factors, including cybersecurity, computational power, and interfacing with the lab management system. But the two preceding examples demonstrate that well-established technologies already exist to solve most of these problems. Armed with this knowledge, pathologists will hopefully begin to see infrastructure not as a roadblock to going digital, but rather an opportunity to custom-build a system uniquely designed for their specific needs.
Check out the other parts of our "Barriers to Adopting Digital Pathology" series below:
Newsletters
Receive the latest pathology news, personalities, education, and career development – weekly to your inbox.
