The way forward for digital pathology: Beyond radiology’s blueprint
Radiology may have blazed the digital trail, but pathology needs its own map. As NHS infrastructure strains under the weight of Whole Slide Imaging and data fragmentation, Liam Canavan argues that pushing pathology to follow radiology’s lead risks stalling progress, and missing a rare opportunity to build smarter from the ground up.
While The Advanced Research Projects Agency for Health’s (ARPA-H) INDEX programme and the Digital Pathology Association (DPA) have laid out a comprehensive plan for digital development in pathology, they also admit that infrastructure challenges across the NHS are holding things back. The question is, how much responsibility do they bear for these very challenges?
The criticism that pathology is lagging, often based on flawed comparisons to fields like radiology and cardiology, misses the mark. Trying to force pathology onto others’ developmental paths will ultimately hinder its digitalisation roadmap and undermine efforts to solve its significant infrastructure problems.
A solitary genesis
Pathology’s distinct infrastructure demands a customised approach:
1. Glass slides
Unlike cardiology and radiology, pathology’s centuries-long reliance on glass slides means it faces a significantly longer and more challenging path to digitalisation.
2. Larger images
Pathology images dwarf those in other disciplines. A single high-resolution pathology image, for instance, can be 80 times larger than a radiology image, a stark difference that highlights the massive gap between current IT infrastructure and what’s truly needed.
| Single Whole Slide Image (WSI) | Single CT scan | Single frame echocardiography |
| 10 Gb > per slide | 30 Mb > per image | 1 Mb > per image |
Staining Whole Slide Images (WSIs) is also necessary for proper diagnosis, leading to significantly larger patient datasets.
| WSI with 10 stainings | Whole body CT | Echocardiography study |
| 100 Gb > per patient | 60 Gb > per patient | 250 Mb > per patient |
3. Differing practice
Pathology’s transition is uniquely intricate, largely because its Laboratory Information Systems (LIS) oversee the entire process, from specimen handling to result verification. This stands in sharp contrast to radiology’s more segmented approach, where Radiology Information Systems (RIS) manage workflows and reports, and Picture Archiving and Communication Systems (PACS) handle images separately.
The technical challenge
Despite recognising that inadequate digital pathology infrastructure severely hinders AI, ARPA-H’s INDEX programme and the DPA fall short of providing concrete solutions to address these challenges.
Storage
Digital storage is overwhelmingly dominated by pathology. According to industry analysis, a large hospital could require 1 petabyte (PB) of storage for pathology alone each year. This demands storage solutions that are not only compliant and scalable but also high performing.
Fragmentation
Fragmented data and inconsistent file formats create significant roadblocks, making data exchange difficult, slowing workflows, and complicating communication across departments.
Networks
WSI files’ immense size necessitates robust network bandwidth and high-speed connectivity. Yet, systems often struggle with such vast data volumes, causing slow transfers and frequent service interruptions.
Laying the groundwork
The promise of digital pathology is undeniable, but we must soberly assess our infrastructure’s readiness to support it on a large scale. Continuing to follow radiology and cardiology’s trajectory will be futile if our current healthcare network infrastructure continues to undermine progress.
A shared problem
Advancements hinge on interdepartmental collaboration to share infrastructure and budgets to ease the storage burden. Integrating with existing DICOM (Digital Imaging and Communications in Medicine) systems is crucial for standardised image formats and interoperability, future-proofing pathology data within a unified system for universal benefit.
A new way of working
Adopting digital workflows means altering ingrained lab practices, making their buy-in crucial. Key obstacles such as a lack of real-time magnification control also need to be overcome. This requires pathology diagnostic processes to be completely redesigned to ensure accuracy and mitigate these disadvantages.
A new way of transferring data
Digital pathology, driven by Whole Slide Images (WSI), demands a strong network: high-throughput scanners alone require 1 Gb/s to databases and 10 Gb/s to storage. Data compression is crucial for managing this. While lossy compression offers smaller files, the risk of losing key diagnostic information pushes us toward lossless compression, though even that isn’t a full answer.
A sole trajectory
The digital infrastructure needed for comprehensive WSI is a distinct challenge for pathology. Paradoxically, the intense focus on AI right now is impeding progress. Though slow, methodical infrastructure development may lack the excitement of AI breakthroughs, empowering pathology to build its digital imaging foundation from the ground up will lead to a more robust and advantageous ecosystem for all, ultimately underpinning future AI innovations.
About Loadbalancer
Loadbalancer‘s application delivery solutions are crucial for over 1,000 healthcare networks globally, managing more than 200 million studies annually. Their significant reach is evident through partnerships with a quarter of UK NHS trusts, the NIH and VISNS Network in the USA, Saudi Arabia’s MOH, and most Fortune 100 Enterprise Imaging companies. Their solutions are also validated and integrated with leading vendors like Philips, Fujifilm, and GE HealthCare.
