A scalable digital pathology platform supporting patient-centered artificial intelligence research

Background: National pathology imaging cooperative, NPIC is a digital pathology and artificial intelligence program providing a scalable platform for clinical deployment and research. The program has deployed digital pathology across 20 hospitals and scanned over 2.5 million digitized images in the national health service, UK. The program aims to deploy digital pathology and support artificial intelligence research across the full lifecycle from development to adoption. DesignThe research is undertaken in collaboration with patient groups, academia, clinicians, and industry. The team has established processes and built a platform to support artificial intelligence research at scale, providing secure access to high-quality data and image standards. The platform will enable access to a large repository of de-identified digital slides for research use. The multi-scanner facility houses over 15 digital pathology scanners from different vendors supporting small and large-scale digitization and creating data sets, including the largest multimodal dataset in partnership with Genomics England. Patient participation and active user involvement underpin the program's research initiatives. ResultsThe NPIC program includes an in-house team supporting partnerships and new opportunities. The program ensures transparency and evidence-based assessment of artificial intelligence tools in the interest of the public. The research findings contribute to the evolving regulations and establishing international quality and image standards, and sharing best practices. The program has established a secure and trusted infrastructure to support artificial intelligence research in digital pathology, underpinned by standardized processes that align with ethical, data governance, and patient and public engagement frameworks. Researchers can access de-identified digital slides through validated workflows developed in collaboration with clinical experts, ensuring the quality and integrity of training and evaluation datasets. The platform supports a wide range of collaborators from start-ups to multinational companies. The portfolio supports over 20 research projects with global reach and has the potential to scale and share over a million digitized images for research and generate 3-4 petabytes of data. Conclusion NPIC has built a national platform for clinical digital pathology and artificial intelligence research. Supported by a dedicated team, standardized research processes, governance, and scalable infrastructure, the NPIC program enables broad collaboration and fosters strategic data partnerships. User needs, patient involvement, and engagement are central to achieving the program's research goals and advancing artificial intelligence research in digital pathology.

Learning Objectives: 

1. Enabling access to a national digital pathology platform for research
2. Identify approaches for involving patients and healthcare users in digital pathology research
3. Apply best practices for shaping artificial intelligence research in digital pathology