Artificial intelligence can meaningfully improve breast cancer screening, according to the UK's first comprehensive evaluation of AI in a routine screening service. The study, published in Nature Cancer, analysed the screening of 10,889 women in NHS Grampian using Mia, an AI tool developed by Kheiron Medical Technologies, now part of DeepHealth. AI increased breast cancer detection by 10.4 per cent, with the additional cancers found being mostly invasive and high-grade, the kinds that matter most to catch early.
The benefits extended beyond detection. The time taken to notify affected women fell from 14 days to 3 days, and the technology showed potential to reduce unnecessary recalls and biopsies. Across 17 tested deployment scenarios, the best balance came from using AI as a second reader replacing one human reader, with an additional safeguard reader retained: a configuration that preserved early detection while cutting healthcare workers' workload by more than 30 per cent.
The GEMINI project was led by the University of Aberdeen, with lead author Dr Clarisse de Vries, Lecturer in Data Science at the University of Glasgow's School of Health and Wellbeing. It was funded through the NHS AI in Health and Care Award in partnership with the NIHR. The UK National Screening Committee does not yet recommend AI in NHS breast screening, and the findings feed into the forthcoming EDITH trial, whose Scottish arm involves Aberdeen, NHS Grampian and Glasgow.