Imperial College London and Institute for Cancer Research have been awarded funding to improve safety, transparency and public confidence for AI tools used in healthcare.
The Institute of Cancer Research & Imperial College London are the joint recipients of a new Medical Research Council (MRC) grant to advance how using technologies, developed in Phase 1 of the DARE UK programme, are demonstrating the possibility of linking multi-modality data – histopathology images with other routinely collected NHS data.
This will provide governed access to data from two of the UK’s cancer research institutions, the Royal Marsden NHS Foundation Trust and Imperial College Healthcare NHS Trust, and will allow future researchers to train artificial intelligence algorithms to answer more complex medical questions and make faster and better diagnoses.
An award of nearly £400,000 will fund the Federated Infrastructure for Digital Pathology Reporting and Expert Data Annotation in a Secure Environment (FIREDANSE) project. FIREDANSE will form part of the newly announced portfolio of Data and Analytics Research Environments UK (DARE UK) Phase 2 catalysts, funded through UK Research and Innovation (UKRI).
Creating trust
Led by Manual Salto-Tellez, professor of integrative pathology, Simon Doran, senior staff scientist in the Division of Radiotherapy and Imaging, both at The Institute of Cancer Research (ICR), and Erik Mayer, clinical associate professor in digital health and lead of the National Institute for Health and Care Research, they will demonstrate the feasibility of securely linking histopathology images and associated electronic record clinical data.
By enabling these data connections, the team plans to show how future researchers could train more powerful AI models to improve cancer diagnosis and treatment.
A core element of the FIREDANSE project focuses on responding to patient and public concerns about the reliability of medical AI systems. People want to know that any AI used in healthcare is closely overseen and validated by hospital consultants throughout both development and deployment. They also want reassurance that there is no significant risk of algorithms generating inaccurate or misleading results.
To address this, the team will adapt an existing web-based app developed at the ICR that allows consultants to view anonymised pathology images and clinical data securely, and to enter expert reviews in a structured and efficient way. The upgraded platform will work across both The Royal Marsden and Imperial, enabling multi-hospital expert validation of AI models. This will support not only algorithm development but also ongoing quality assurance, ensuring that AI tools perform consistently and safely in clinical settings.
“By investing in early-stage prototypes and strong public engagement, we are building the evidence needed to shape future infrastructure that is secure, flexible and worthy of public trust,” said Emily Jefferson, interim director of DARE UK.
