Barbara felt completely fine. No lump. No pain. No nagging feeling that something could be wrong. So, like most women, she squeezed in a routine breast screening appointment at Aberdeen Royal Infirmary between work and errands, not giving it a second thought. She certainly did not expect what came next.
The cancer inside her was so small, doctors later told her, that no human eye would have spotted it on the scan. Instead, what caught this virtually invisible tumour, potentially saving Barbara’s life, was an AI tool called Mia.
Barbara was among the very first women in the UK to have breast cancer detected through artificial intelligence. Her story sits at the centre of a landmark study happening in our own backyard: Grampian’s Evaluation of Mia in an Innovative National breast screening Initiative, known as GEMINI. The project was years in the making, built through a close partnership between the University of Aberdeen and NHS Grampian. And, when they finally published the results in March 2024, the medical world took notice.
The trial, the first of its kind in the UK, followed 10,889 patients at Aberdeen Royal Infirmary. It was led by Dr Gerald Lip, Clinical Director for breast screening at NHS Grampian, with an independent evaluation carried out by Professor Lesley Anderson, Chair in Health Data Science at the University of Aberdeen.
What the numbers showed was hard to argue with: when used alongside radiologists, Mia helped detect 12 per cent more cancers than standard practice, and these were not borderline cases. Most were invasive and high-grade, exactly the cancers it’s most important to catch early. Crucially, this improvement came without any rise in unnecessary recalls. False alarms are one of the biggest criticisms of breast screening, so the fact that Mia did not make any additional contributions was significant. The trial also modelled a potential drop in notification wait times, from 14 days to just 3.
Dr Lip put the human meaning behind the statistics plainly: “If you pick up cancer under 15mm, most women now will have a 95 per cent survival rate. Not only did Mia help us find more cancers, most of which were invasive and high grade, but we also modelled that it could reduce the time it takes to notify women from 14 days to just 3 days, reducing significant stress and anxiety for our patients.”
Professor Anderson was equally direct. “The GEMINI trial results left us amazed,” she said. “If Mia were used in breast screening, it would mean that more cancers would be detected without putting more women through additional tests.”
The project did not begin with a trial, though. It began with a problem.
When Mia was first tested on NHS Grampian data, she flagged too many women unnecessarily. Mia was too sensitive for the local population, and there were too many false alarms to be used. Before any real trial could take place, the University’s Dr Clarisse de Vries spent three years analysing 220,000 mammograms from more than 55,000 patients, calibrating Mia specifically for a Scottish setting. The work was published in Radiology: Artificial Intelligence.
“Now we know there are risks in just taking an AI tool developed elsewhere and implementing it locally,” Dr de Vries said. “You must first test the tool on the local data to ensure it will work as expected.”
It sounds obvious in hindsight, but at the time, it was a lesson that had to be learned through painstaking work. The principle has since shaped how AI diagnostic tools are approached right across the NHS.
Deservedly, recognition has followed. In February 2025, the GEMINI team won the Data Driven Innovation Award at the Digital Health and Care Awards in Edinburgh. Professor Anderson accepted on behalf of the team, noting that it “recognises the tireless efforts of our dedicated teams and the transformative potential of data innovation in improving healthcare outcomes.”
The following month brought another development. Dr Lip was appointed as the University of Aberdeen’s Lead for Artificial Intelligence in Clinical Practice, the first role of its kind in any Scottish medical school, funded by NHS Grampian. The job is not just about getting clinicians to use AI tools. It is about making sure the next generation of doctors knows when to trust them and when to push back.
Barbara’s mother had breast cancer, too. Hers was found later, and the treatment was far more invasive than anything Barbara will face. The difference between their two experiences comes down to a team in Aberdeen working quietly and methodically, year after year, on a problem most people outside medicine had never thought about.
“It’s a lifesaver,” Barbara said. “It’s a life changer.”