This site has been optimized to work with modern browsers and does not fully support your version of Internet Explorer.

Evaluating and scaling up AI technology supporting better patient outcomes following stroke

Around 100,000 people have a stroke in the UK every year. About eight out of ten of these suffer a blockage in the brain which is called an ischaemic stroke. One third of these patients have a large vessel occlusion (LVO), a blockage in the neck or brain which contributes disproportionately to stroke-related dependence and death because of the potential to cut off blood supply to a large proportion of the brain. Early identification, rapid decision-making and prompt treatment can reduce brain damage and prevent or limit long-term disability.

Around one in eight people admitted with acute stroke are eligible for thrombolysis (IVT), a clot-busting drug which until recently was the only treatment for this type of stroke. IVT should be administered within 4.5 hours of symptom onset to achieve maximum benefit. More recently, mechanical thrombectomy (MT), manual clot removal, has revolutionised stroke care and recovery, reducing brain damage and preventing or limiting long-term disability. MT should be administered within 4.5 hours of symptom onset to achieve maximum benefit.

However, few hospitals currently provide MT services. Referral to stroke units which can perform MT is a crucial part of the treatment pathway. imaging decision support software driven by artificial intelligence (AI) can help to speed up clinical decision-making to determine patients’ suitability for treatment and transfer to their nearest thrombectomy centre.

The NHS Long Term Plan (2019) set out national targets for stroke care including increasing rates of MT from 1% to 10%. Early identification of LVO and quicker decision times can increase eligibility and MT rates. Brain scans such as CT are used to identify patients eligible for MT but specialist radiological image interpretation is not available in most hospitals. As a result, many patients who could benefit from MT are missing out.

e-Stroke by Brainomix (a company which emerged from the University of Oxford in 2010) is a CE-marked collection of tools that use artificial intelligence (AI) algorithms to support clinical decision-making, providing real-time interpretation of high-quality brain scans quickly and securely and enabling more patients to get the right treatment in the right place at the right time. More patients are receiving MT as e-Stroke extends the time window for treatment and advances the value of imaging.

The Oxford AHSN supported spread and adoption of e-Stroke and is now evaluating its impact on clinical outcomes, pathways and productivity ahead of further potential scale up and health economic evaluation after Brainomix received funding through the NHS AI in Health and Care Award in 2020.

More than 20 NHS hospitals across England are taking part in the evaluation, which aims to understand if the inclusion of e-Stroke in existing stroke pathways can increase the number of eligible patients receiving IVT and MT by reducing the time clinicians need to assess images and make treatment decisions, as well as cutting the time to treatment through improved communications.