Despite rapid advances in artificial intelligence (AI) for digital pathology, translation from algorithm development to routine clinical use remains rare. Most studies focus on technical performance, overlooking the different implementation steps needed for safe, sustainable deployment. This study addresses that gap by detailing the real-world integration of a laboratory-developed AI tool into the daily workflow of an academic pathology department. Our approach encompasses regulatory compliance, IT integration, validation studies, workflow redesign, and user training, essential for lasting adoption under current regulation. Through this process, we developed a flexible, locally adapted system which is now used in routine care. This work offers a practical, replicable framework for other laboratories, highlighting that successful implementation takes far more than technical accuracy alone.
A practical roadmap to implementation of in-house developed computational pathology algorithms
F. Meeuwsen, M. Balkenhol, D. Geijs, L. van Eekelen, R. Lomans, R. van de Loo, K. Freund-Winkel, P. van Overdijk, S. van den Kieboom-Hageman, J. Grolleman, G. Litjens, K. Grünberg and J. van der Laak
Research Square 2026.