Lessons learned from three approaches to digital pathology system validation at a large academic medical institution

Journal of Pathology Informatics · Published 2026-04-15 · DOI 10.1016/j.jpi.2026.100660

Free full text

Authors being retrieved — see the publisher record. https://doi.org/10.1016/j.jpi.2026.100660

Abstract

Context: The University of Washington Department of Pathology validated whole-slide imaging (WSI) for clinical diagnostic use in 2020 in response to the COVID-19 pandemic. Subsequently, the project was expanded to address shortcomings in the initial validation and to include the adoption of new technologies and organizational restructuring goals. Design: We describe three validation approaches: Rapid (3-week emergency deployment), Selective (single-subspecialty focused), and Big-Bang (simultaneous multi-subspecialty inclusion) methods utilized to validate digital pathology at our institution. Each method was selected based on the availability of resources, timeline constraints, and institutional needs at the time. Results: WSI for clinical use was successfully validated in each approach and eventually across all specialties. The rapid neuropathology validation achieved 100% accuracy in 3 weeks in the setting of an urgent pandemic response, however, it did not meet all updated CAP validation recommendations. The selective genitourinary service validation project achieved 96.45% concordance following all 2021 CAP WSI validation guidelines, but was time and labor-intensive, stressing a single specialty. The simultaneous multi-subspecialty validation achieved 95.16% concordance while distributing the work effort. When combined, the system-wide concordance reached 95.78%. Conclusion: This descriptive article highlights our institution's varying experiences with WSI validation. Success was dependent on matching the validation approach to the available infrastructure, timeline requirements, and clinical/organizational needs.

Abstract from DOAJ. Public domain (CC0 1.0).

Read the article at the publisher →

Publication details

Year
2026

Related articles