New UK NICE guidance endorses AI-derived software in stroke assessments

A new guidance document from the UK National Institute of Health and Care Excellence (NICE) has endorsed National Health Service (NHS) deployment of two artificial intelligence (AI)-derived software technologies to support the review and reporting of computed tomography (CT) brain scans in patients with suspected stroke.

Based on existing data, RapidAI (RapidAI; formerly iSchemaView) and e-Stroke (now known as Brainomix 360; Brainomix) can be used in the UK healthcare system “while more evidence is generated”, but only once they have gained appropriate Digital Technology Assessment Criteria (DTAC) approval.

According to the latest diagnostic guidance from NICE on stroke and transient ischaemic attack (DG57), “the software should only be used with healthcare professional review, and centres should maintain existing scan reporting protocols to reduce the risk of incorrect results”. In addition, the document notes that “centres should ensure that images shared between different stroke centres can be remotely reviewed to help with decision making by healthcare professionals at a different site”.

“We welcome this latest guidance from NICE, which reflects the clinical value that Brainomix 360 brings to stroke care—which we have been seeing since we first implemented the software in March 2020,” said Kiruba Nagaratnam (Royal Berkshire Hospital, Reading, UK), as quoted in a recent Brainomix press release. “Our recently published study showed that the implementation of Brainomix 360 not only improved our speed of treatment but, most notably, improved patient outcomes, with the number of patients achieving functional independence rising from 16% to 48%.”

“AI will never replace the clinical expertise that our doctors and consultants have, but harnessing this latest technology is allowing us to make very quick decisions based on the experiences of thousands of other stroke patients,” stated Jenny Vernel (Hereford Hospitals NHS Trust, Hereford, UK), speaking in December 2023 after her hospital became the first in the UK West Midlands to roll out RapidAI software for stroke assessments. “The system is linked to the comprehensive stroke centre at University Hospitals Birmingham NHS Foundation Trust [Birmingham, UK], meaning that clinical teams and experts at both sites can make faster and more informed clinical decisions. This is essential when treating stroke patients.”

NICE’s recent guidance—published online on 23 January—also lists a number of technologies that “can only be used in research” currently, including Accipio (MaxQ AI), Aidoc (Aidoc Medical), BioMind (BioMind AI), BrainScan CT (Brainscan AI), Cercare Perfusion (Cercare Medical), CINA Head (Avicenna AI), CT Perfusion 4D (GE Healthcare), icobrain CT (Icometrix), Neuro Solution (Nanox AI), qER (Qure AI), and Viz (Viz AI).

The guidance states that “more research is needed” on all of these AI-derived software to support review and reporting of CT brain scans in patients with suspected stroke, and that access to these technologies “should be through company or research funding” i.e. non-core NHS funding.

Additionally, NICE advises that further evidence generation and more research are needed on: the impact of the addition of AI-derived software on a healthcare professional’s ability to identify people for whom thrombolysis and thrombectomy is suitable; how often the software is unable to analyse CT brain scans, with reasons for this; the impact of using the software on time to thrombolysis or thrombectomy; and the impact of using the software on how many people have thrombolysis or thrombectomy.

Alongside this guidance, NICE has released an evidence generation plan providing further information on prioritised evidence gaps and outcomes, ongoing studies, and potential real-world data sources. NICE also details that the UK Department of Health and Social Care (DHSC) and NHS England have launched several initiatives, such as the AI Diagnostic Fund and the AI Deployment Platform, that will help generate more evidence on the use of AI technologies.

“Clinical evidence on the software is limited in quality,” the NICE guidance further states. “There is no evidence on their diagnostic accuracy when used alongside healthcare professional review that met review inclusion criteria. Some studies on two technologies (e-Stroke and RapidAI) in clinical practice suggest that people had faster or greater access to treatment after using the software, but it is unclear to what extent this is an effect of the software. In the economic model, a small increase in the number of people having thrombectomies because of AI-derived software would likely make the software cost effective.”

According to NICE, e-Stroke and RapidAI are already “widely used” in the UK NHS, and can continue to be deployed while further evidence is generated to help better determine their cost effectiveness—but the other aforementioned technologies, for which there is “no evidence on how they impact time or access to treatment”, should only be used in research.

Latest news
Related news