Clinical trials rely increasingly on medical images to diagnose and stratify patients, with many studies using imaging end points as surrogate markers. Although clinical trial data is rigorously kept for future reanalysis, the same attention has not been paid to imaging data. Recently it has been recognised that well collected images from clinical trials could provide useful information about the patient phenotype and could provide further insight into prediction of response. The ability to undertake meta-analysis in these well-defined datasets could bring many benefits for patients and researchers.

The NCRI recognised this growing potential and arranged for a multidisciplinary group to meet to discuss whether or not a national archive of medical images might be of value, what this might look like and who might be responsible for this. The NCRI understood the importance of linking other routinely collected datasets to these images and how this would greatly enhance the image archive and so ensured that triallists and biologists were involved in the debate. Furthermore there was an immediate understanding that most trial patients are recruited after their routine hospital imaging. So the Royal College of Radiologists were invited to come and share their work on developing national standards and governance framework for image acquisition, IT and permissions from patients to reuse their imaging data. The final piece of the jigsaw (and arguably one of the key pieces) was the pathology data – mostly held in filing cabinets and not ready for national archiving. The innovative Leeds pathology archive was presented and this demonstrated that technology is available now.

The meeting was hugely informative and resulted in a very accessible report with a number of recommendations. The pathologists were very keen for the UK to rapidly adopt the digital techniques required to make this happen which would make a step change in how this data could be interrogated. All parties shared the vision of a national archive and recognised that a number of steps to ensure harmonisation were required. The report is available here.

These are exciting times for imaging especially for Radiology and Pathology. There is huge anticipation around the potential for AI to improve work flows and change delivery of healthcare. While the initial enthusiasm will wane as delivery of usable tools will take longer than expected the investment made by UK government and large charities will undoubtedly pay dividends long term.

Professor Fiona Gilbert Professor of Radiology at University of Cambridge School of Clinical Medicine & Honorary Consultant Radiologist at Addenbrooke’s Hospital, Cambridge. University Hospitals NHS Foundation Trust.