119 posters, 6 topics, 524 authors, 243 institutions
ePostersLive by SciGen Technologies S.A. All rights reserved.
RCR Global AI Conference 2026
29-30 June, 2026 | QEII Centre, Westminster
53
<p>Secure Data Environments for Radiology AI Research: Experiences from a Driver Project</p>
Poster Presenter
Authors
Afagh Garjani, Alan Bagnall, Mark Walsh, Hils Poole, Anna Beattie
Affiliations
School of Radiology, Health Education North East, Newcastle upon Tyne, United Kingdom, The Newcastle Upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, United Kigdom, SDEs offer considerable opportunities for developing, training, and validating AI algorithms and undertaking impact assessments., Health Innovation North East and North Cumbria, Newcastle upon Tyne, United Kingdom, Inzight, United Kingdom
Part of Topic
AI Education and research: examples of proof of concept or AI in development, technical advances, teaching approaches or pre-clinical testing
Title
Secure Data Environments for Radiology AI Research: Experiences from a Driver Project
Purpose
The RCR-NHS Global AI Conference 2025 roundtable highlighted the challenges of using Secure Data Environments (SDEs) for AI research. In response, this study aims to capture the experiences and insights of a team in developing and deploying a radiology AI project within the NHS SDE Network and to identify lessons learned.
Methods and materials
In this autoethnographic study, data included content from documents, emails, and interviews related to a driver project for refining and validating an AI algorithm for automated aortic measurements on chest computed tomography scans. The principal aim of the driver project was to develop the logistical and technical capabilities of North East and North Cumbria SDE to accommodate and handle multimodal data including Digital Imaging and Communications in Medicine (DICOM) files.
Results
The overarching theme was that internal SDE processes, rather than workflow or technical functionalities of the SDE, shaped the project pace and scope. The major challenge was obtaining sponsorship and signed data-sharing agreements. This was, however, related to the nature and aim of the specific driver project and for most projects would routinely be in place upfront. Inconsistent IG policies across stakeholder organisations and differing interpretations of regulations were identified as primary barriers to federation of data between SDEs. Federation of data and/or algorithms will be important if SDEs are to provide the most comprehensive platform for radiology AI projects and are dependent on resolving these multiple challenges.
Conclusion
SDEs can offer considerable opportunities for developing, training and validating AI algorithms along with performing impact assessments. Maximising the potential of SDEs requires aligned network-wide policies that are broad enough to accommodate diverse projects yet robust enough to ensure data security, regulatory compliance, and clarity of responsibilities.