In the largest NHS breast cancer screening study ever conducted — involving 175,000 women — Google AI matched or exceeded the performance of specialist radiologists in every key measure: detecting more cancers, producing fewer false alarms, and cutting scan-reading time by nearly a third. Two papers published today in *Nature Cancer* mark a turning point in the fight against the UK's most common cancer.
The research, led by Imperial College London in partnership with Google and the universities of Cambridge and Surrey, represents years of careful testing across five NHS Trusts, including Cambridge University Hospitals, Imperial College Healthcare, the Royal Marsden, the Royal Surrey, and St George's University Hospitals.
Breast cancer affects one woman every ten minutes in the UK. Yet there is already a 29% shortfall of clinical radiologists — nearly 2,000 unfilled posts — and that gap is projected to widen to 39% by 2029. AI offers a way through.
**What the Numbers Show**
The retrospective study covered 115,973 breast cancer scans from women aged 50 to 70, screened between 2015 and 2016 and tracked for 39 months. When AI acted as the second reader — replacing the second human radiologist — the results were striking.
Cancer detection rate rose from 7.54 per 1,000 women (human reader) to **9.33 per 1,000 women** with AI. That's a meaningful improvement — more women with real cancers being identified and treated sooner. AI also identified more invasive cancers specifically, and significantly reduced false positives: the recalls that send women back for follow-up scans that turn out to be nothing, causing anxiety and system strain.
For first-time screeners — a particularly important group — AI performed even better: **39.3% fewer unnecessary recalls** and an 8.8% higher cancer detection rate. The AI also reduced the amount of time required to read a scan by **32.1%**, representing hundreds of thousands of freed-up reads.
**The Closest AI Has Come to Saving Lives in the NHS**
'This is the closest AI has ever come to helping reduce breast cancer deaths within the NHS,' said Dr Hutan Ashrafian of Imperial's Institute of Global Health Innovation, who is an author on both papers. 'The potential for the NHS to take this forward is significant, particularly in light of the National Cancer Plan for England's recognition that there are few clearer signs of the failure of the status quo than our inadequate cancer outcomes.'
The research did not stop with retrospective data. A second, prospective study followed 9,266 current cases at two screening services across 12 London sites, testing how AI performs in real-time clinical settings. The live results broadly confirmed the retrospective findings.
**Why This Matters Beyond the Numbers**
Early detection is the most powerful tool against breast cancer. Cancers caught early are more treatable, survival rates are dramatically higher, and treatment is less intensive. The central promise of AI screening isn't just efficiency — it's catching cancers that human readers miss.
The AI also excels at detecting interval cancers — cancers that develop between scheduled screening rounds. In the study, AI detected 25% of these interval cancers, flagging women who would otherwise have waited for their next scheduled scan while cancer progressed.
**What Comes Next**
The research team and NHS partners are now calling for larger real-world trials and, ultimately, integration of AI into routine NHS breast screening. The technology has been tested, the results are published in one of medicine's most prestigious journals, and the case for deployment is now evidence-based rather than theoretical.
For the 55,000 women diagnosed with breast cancer in the UK each year — and the millions who go through screening annually — this is a landmark moment. The question is no longer whether AI can help. It's how quickly we can use it. 🩺
*Sources: Imperial College London · Nature Cancer (March 10, 2026) · Google · University of Cambridge · University of Surrey · NHS Trusts*
