Before pathologists sign out cases from a screen instead of a microscope, the laboratory has to show that digital review is safe for its own cases, people and equipment. This guide is for pathologists, quality managers and lab directors planning the validation of digital pathology for primary diagnosis, and it focuses on study design, practical pitfalls and the documentation you will need afterward.
Why validation is required
A whole slide imaging system is a new way of producing the material a diagnosis is based on. Even if the scanner is approved or cleared by a regulator in some jurisdiction, that does not prove it performs well in your lab, with your staining, your case mix and your pathologists.
Professional bodies have addressed this directly. The College of American Pathologists publishes guidance on validating whole slide imaging for diagnostic use, and many labs outside the United States use its principles as a reference. Accreditation against ISO 15189, the international standard for medical laboratory quality and competence, also expects examination procedures to be validated or verified before use. Our article on what ISO 15189 changes day to day covers that wider context.
The core question is simple: does a pathologist reach the same diagnosis on the digital image as on the glass slide?
What you are validating: the whole system
Validation is not a test of the scanner alone. It covers everything between the glass slide and the diagnosis:
- The scanner, its settings and its scanning profiles (for example 20x versus 40x).
- Image management software and the viewer.
- Display monitors and workstation configuration.
- Network and storage, as they affect image loading.
- The workflow that links each image to the correct case.
- The pathologists who will use the system, including their training.
Change any of these significantly and you may need to revalidate. Defining the system precisely at the start makes later decisions much easier.
Designing the validation study for digital pathology
A good validation study mirrors real clinical use. It should be planned in writing, approved before it starts, and owned by a named person.
Define the intended use
Be explicit about scope. Will digital review be used for all routine H&E histology, or only specific specimen types? Will it include frozen sections, immunohistochemistry, special stains or cytology? Each distinct application may need its own validation set, because each brings different challenges.
Select cases carefully
Choose cases that represent the full spectrum of what pathologists will see, including difficult and borderline cases, not just straightforward ones. The CAP guideline has recommended a minimum of 60 cases for a single application such as routine H&E, with more advisable when the scope is broad. Include small biopsies, large resections, varied tissue types and cases with features known to be challenging on screen.
Use a washout period
Each pathologist should review the same cases on glass and on digital, with enough time between the two reads that they do not simply remember their earlier diagnosis. The CAP guideline has recommended a washout of at least two weeks. Randomize the order so that not every case is seen in the same modality first.
Measure concordance
The primary measure is intra-observer concordance: does the same pathologist reach the same diagnosis on glass and digital? Define in advance what counts as a major discordance, meaning one that could change patient management, and what counts as a minor difference with no clinical impact.
Handling discordances
Discordances will occur. Some reflect genuine differences between glass and digital review; others reflect normal diagnostic variation that would happen on two glass reads.
For every discordance, review the case and record:
- Whether the difference is major or minor.
- Whether the glass or the digital diagnosis was correct, based on consensus review.
- The likely cause: image quality, missing tissue, a feature hard to see on screen, viewer limitations, or interpretive variation.
- What action follows: rescanning protocol changes, additional training, or excluding a specimen type from digital review.
Patterns matter more than single cases. If discordances cluster around a feature such as small organisms or mitotic counting, that tells you where to apply caution or keep glass review in place.
Documentation and your quality system
Validation evidence must stand up to an assessor months or years later. Integrate it into your existing quality management system rather than keeping it in a folder on someone's desktop.
A complete validation file typically includes:
- The approved protocol, with scope, case selection criteria and acceptance criteria.
- System description: scanner model, software versions, viewer, displays and settings.
- Case list and anonymized results for each pathologist and each modality.
- Discordance analysis and the actions taken.
- Training records for participating pathologists and technologists.
- A signed conclusion stating what is approved for digital primary diagnosis and what is excluded.
- Updated standard operating procedures for scanning, image quality checks and digital sign-out.
After go-live: monitoring and revalidation
Validation is a starting point, not a one-time certificate. Ongoing monitoring keeps confidence high.
Track rescan rates, cases deferred to glass and any discordances found through peer review or second opinions. Define in advance which changes trigger revalidation, such as a new scanner model, major software upgrade, new display type, new stain or a new case type added to digital review. Smaller changes may need only a limited verification rather than a full study.
Common mistakes include validating on easy cases only, skipping the washout period because of time pressure, failing to record software versions, and expanding digital review to new specimen types without further validation. Each one weakens the evidence your accreditation and your patients rely on.
New pathologists joining the department also need training and a documented competency check before signing out digitally. A short supervised period with their own cases works well.
Key takeaways
- Validate the whole system, from scanner settings to displays and workflow, for your own intended use.
- Use a representative case set, a washout period and predefined criteria for major and minor discordance.
- Investigate every discordance and look for patterns that define where glass review remains necessary.
- Keep validation records inside your quality management system, with clear scope and software versions.
- Monitor after go-live and define in advance which changes trigger revalidation.
If you are preparing for digital sign-out and want case records, reports and slide images in one auditable place, book a demo to see how PATHOLOGY+ supports the workflow.