Sheffield cardiac arrest survivor ‘encouraged’ by AI plans to detect signs of heart failure | News

It follows a study carried out by academics in Scotland

Artificial intelligence could be used to help detect people living at risk of heart failure, according to a study.

Researchers at Dundee University School of Medicine used artificial intelligence (AI) to scan images of the heart to identify heart failure patients from population-based electronic health records and echocardiographic heart scans.

AI deep learning was then used to examine the echocardiographic images to identify abnormalities that could increase the patient’s risk.

The experts used data provided by patients voluntarily through the Scottish Health Research Registry and Biobank (SHARE), which provides volunteer researchers with samples, health data and genomic information for clinical trials.

The trial began with a dataset of 15,000 patient records, which was narrowed down to a cohort of 578 patients.

Professor Chim Lang said: “Our research represents a breakthrough in using deep learning to automatically interpret echocardiographic images. This may allow us to streamline the identification of heart failure patients at scale in electronic medical record datasets.

“Echocardiographic heart scans that have been enhanced by AI software have helped provide more measurements – or parameters – of heart structure and function that can be used to help diagnose heart failure.

“These measurements were not routinely reported by regular heart scans in electronic health records.

“Compared to reports generated by routine heart scans, the AI-enhanced ones were more detailed and could also be processed on a larger scale than conventional images.

“This has potential clinical and research implications as it could improve the efficiency and speed of patient selection for pragmatic clinical trials, as well as improve heart failure surveillance and early diagnosis in hospital systems.”

Experts said heart failure is a very common but underdiagnosed condition, which means the heart is not able to pump blood around the body efficiently.

While symptoms can be controlled to some extent through lifestyle changes, surgery and medication, in most cases it is a serious, long-term condition that gets progressively worse over time.

Prof Lang said the latest research was “an example of how AI has the potential to deliver real-world benefits to patients”.

He said: “By evaluating large amounts of patient records, we were able to detect structural and functional abnormalities that we could not have done with traditional analysis of echocardiographic images.

“Although this is a test case, I am very excited that we were able to apply deep learning to a large-scale biobank resource.

“We hope this paves the way for other researchers to use this technology to benefit patients around the world.”

The research was published in the journal ESC Heart Failure.

The project was carried out using technology from software developer Us2.ai and was funded by Roche Diagnostics International.

Bob Reville, 55, who lives in Sheffield, suffered a cardiac arrest at Meadowhall in 2013:

“It’s still a very low survival rate, so I think if AI can start to learn more about the heart rhythms themselves and, by chance, predict who might go into cardiac arrest, then it could be a massive game changer.

“Quite often, when it is detected, it can be at a later stage, and often it is too late for many.

“So anything AI can do to help us detect heart failure at any earlier stage can only be a good thing.

“In the future, this could help save many lives.”

First and foremost, for all the latest news from across the UK every hour on Hits Radio on DAB, at hitsradio.co.uk and on the Hits Radio app.