An AI system used to assess CT scans in NHS hospitals has significantly shortened the time it takes for stroke patients to reach specialist treatment, according to research published in The Lancet Digital Health.
The study found that hospitals using the Brainomix 360 Stroke platform were able to identify severe strokes more quickly and transfer patients to specialist centres sooner.
The tool, which has been rolled out to a network of over 70 hospitals across England, was used to assess scans for more than 15,000 patients during the study period.
At primary stroke centres, its use was associated with a 64-minute reduction in door-in-door-out times. Thrombectomy rates also increased, rising from 2.3% to 4.6% at participating sites, compared with smaller gains at hospitals not using the technology.
The reduction in delays is clinically important, given that every 20-minute wait for thrombectomy reduces the likelihood of full recovery by around 1%.
Patients whose scans were reviewed using AI were also more likely to receive intravenous thrombolysis and to have a favourable outcome at discharge, with no increase in in-hospital mortality.
With around 80,000 strokes occurring in England each year, the findings suggest the technology could help thousands more patients receive specialist treatment in time to improve their recovery.
Dr David Hargroves, NHS national clinical director for stroke and a co-author of the report, 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.
‘This technology supports clinicians to make rapid treatment decisions, which means more patients can receive life- and disability-saving treatments in time – giving them a better chance of returning to independent living.’
The findings form part of the national evaluation of England’s optimal stroke imaging pathway, under which AI decision-support tools were introduced across all regularly admitting stroke services in summer 2024.
The study analysed data from 452,952 patients in the national audit dataset, including more than 71,000 treated at the 26 evaluation hospitals.
Earlier this month, NICE recommended AI tools to help detect bowel cancer earlier.
