Health IT leaders looking to scale enterprise imaging programs should take note of digital pathology.
Whole slide images, often two to ten times the size of a radiology study, have forced pathology to confront the challenges that most imaging modalities eventually face, but at an order of magnitude that’s particularly instructive. Here are the lessons learned about what works at scale and where implementations fall short.
At Scale, Infrastructure and Workflow Are the Same Problem
It’s easy to see why early digital pathology rollouts treated infrastructure strictly as an IT concern. Before digitization, pathologists and IT teams had little reason to work together, and configuring the workflow so that pathologists could use the system often felt like a higher priority.
As programs expanded, the impacts became hard to ignore. System latency might delay a pathologist’s review, or fragmented data could complicate a multi-site consultation. The failure, in these instances, isn’t technical but a clinical roadblock that illustrates why infrastructure and workflow are one and the same.
Flexibility Matters as Much as Capacity
Any IT leader who has tried to scale by adding more storage will confirm that high-volume imaging is not a capacity problem.
This line of thinking overlooks the fact that imaging workloads aren’t predictable, and not all data ages in the same way. An image from two days ago and one from three years ago almost certainly have entirely different access patterns. Storing them together wastes money and creates friction.
Teams that have scaled successfully while keeping costs under control tend to move to cloud infrastructure, where compute is available on demand and data automatically moves to storage tiers. What’s more, the IT leaders driving these programs design systems around how pathologists actually work instead of retrofitting workflows after deployment.
Standardization Starts with DICOM
Digital pathology is known for a plethora of proprietary image formats, each of which is tied to the scanner that generated it. The resulting variability is only manageable when footprints involve one scanner, one site, and one primary workflow. With any expansion, data can’t move cleanly among systems.
DICOM reduces this complexity by standardizing how images and metadata are stored, making it easier to exchange and reducing dependencies on any single vendor. DICOMweb extends standardization further by enabling systems to connect through standard web APIs instead of custom one-off integrations.
Even still, DICOM isn’t a magic fix. Metadata must still be consistent, and new systems must be carefully integrated for DICOM data to flow properly.
Data Governance Is Where High-Volume Deployments Quietly Crumble
Whereas infrastructure problems tend to be visible, data governance challenges often hide until something breaks. Organizations that avoid these issues treat data governance like any other operational function, and critically, not one that belongs to IT alone. They share accountability across clinical, informatics, and IT leadership so that policies reflect how data is actually used, not just how it’s stored.
These organizations define lifecycle policies before they need them, automate how data moves across storage tiers, and build visibility into cost and access when the system is still small enough to course-correct. When done well, governance prevents failure while also creating the foundation for data to be reused confidently for research, quality improvement, and AI.
A Proven Blueprint for the Age of AI
The lessons we learn from digital pathology are especially relevant as AI enters the workflow. AI not only increases data volumes but also changes how that data is used, accessed, and valued. Models depend on large, well-organized datasets, consistent metadata, and infrastructure that can handle bursts of compute without slowing down other critical workflows.
While an imaging program can scale to a point with gaps in its IT foundation, those limitations become much harder to manage once AI is introduced. IT leaders who follow the digital pathology blueprint by addressing infrastructure, flexibility, standardization, and data governance will be better positioned to grow efficiently and turn imaging data into meaningful clinical and operational value, now and over time.
About Stephan Fromme
Stephan Fromme is a commercial leader in digital pathology with a focus on enterprise deployments, strategic partnerships, and clinical adoption across the globe. At Proscia, he works closely with healthcare organizations, laboratory networks, and industry partners to support the transition to digital pathology and scalable AI-enabled workflows. His work spans complex implementation programs, partner development, and collaboration initiatives aimed at advancing precision diagnostics and modern pathology operations.
