NHS AI tool cuts stroke transfer times by an hour, doubles clot-removal treatment

What is the background?

Around 80,000 people have a stroke in England every year. Speed of treatment matters enormously: every 20-minute delay in thrombectomy, a procedure that physically removes a blood clot from the brain, cuts a patient’s chance of full recovery by about 1%. Many patients first arrive at a local hospital without thrombectomy capability, so doctors must quickly decide whether to transfer them to a specialist stroke centre. In 2020, Brainomix and Health Innovation Oxford and Thames Valley won an NIHR NHS AI in Health and Care Award to scale up an AI imaging tool, Brainomix 360 Stroke, built to speed up that decision.

What happened?

A study published in The Lancet Digital Health on 2 December 2025 found the tool cut the time to transfer patients from a local hospital to specialist treatment by 64 minutes at primary stroke centres. Hospitals using it also doubled their thrombectomy rate, from 2.3% of stroke patients to 4.6%. Researchers analysed 452,952 stroke admissions recorded across all 107 NHS hospitals in England between 2019 and 2023, comparing the 26 hospitals that had adopted Brainomix 360 Stroke with hospitals that had not yet done so. NHS England said the tool is now deployed in more than 70 hospitals nationwide, with over 15,000 patient scans reviewed since January 2022.

How did it happen?

Brainomix 360 Stroke analyses a patient’s CT brain scan within minutes, automatically flagging signs of a major clot. That speed lets doctors at a local hospital decide sooner whether a patient needs transfer for thrombectomy, in which specialists remove the clot using a thin tube threaded through an artery. The same scans also help identify candidates for clot-dissolving drug treatment: researchers found AI-reviewed patients were significantly more likely to receive that too. Royal Berkshire Hospital in Reading served as the programme’s first “champion site” under the clinical lead of Dr Kiruba Nagaratnam, before the NIHR award funded expansion to four further stroke networks and 25 stroke units. Jean Hines, 83, experienced the result directly: after a scan at Royal Berkshire flagged a major clot, she reached the John Radcliffe Hospital in Oxford within 25 minutes for a successful thrombectomy. “I feel incredibly lucky,” she said. “Everything happened so quickly and I know that made all the difference.”

Why does it matter?

Dr David Hargroves, NHS national clinical director for stroke and a co-author of the study, said: “This landmark study confirms what we have already been seeing in daily practice: that stroke AI imaging is helping us deliver faster decision-making and better care for our patients.” Professor Gary Ford, chief executive of Health Innovation Oxford and Thames Valley, called it “an AI solution that is meeting a clear need and delivering impact.” Faster decisions and higher treatment rates matter because thrombectomy substantially reduces long-term disability after a major stroke, and every minute of delay narrows the window in which that benefit is possible.

What were the key results?

• Door-in-door-out transfer time cut by 64 minutes at primary stroke centres using the AI tool
• Thrombectomy rate doubled at evaluation sites, from 2.3% to 4.6% of stroke patients
• Comparison hospitals without the tool saw a smaller rise, from 1.6% to 2.6%
• Study covered 452,952 stroke admissions across all 107 NHS hospitals in England, 2019–2023
• AI-assisted patients were statistically more likely to receive thrombectomy (odds ratio 1.57, p<0.0001) and clot-dissolving drug treatment (odds ratio 1.99, p<0.0001)
• Brainomix 360 Stroke now deployed in more than 70 NHS hospitals, reviewing over 15,000 scans since January 2022