AI can now diagnose heart disease in just four seconds, as study shows machines now 'as good' as doctors

Henry Bodkin
If widely used, the technology could safe thousands of days of doctors' time - WARNING: Use of this copyright image is subject to the terms of use of BBC Pictures' Digital Picture

Robots can diagnose heart problems in as little as four seconds, tests have shown, as a review of artificial intelligence (AI) found machines are now as good at spotting illness as doctors.

Analysing a patient’s heart function on a cardiac MRI scan currently takes doctors around 13 minutes.

But a new trial by University College London (UCL) showed an AI programme could read the scans in a fraction of the time with equal accuracy.

There are approximately 150,000 such scans performed in the UK each year, and researchers estimate that fully utilising AI to read them could save 54 clinician-days at each cardiac centre per year.

It is hoped that AI - where computer systems are able to learn from data to identify new patterns with minimal human intervention - will transform medicine by helping doctors spot diseases such heart disease and cancer quicker and earlier.

However, most scans are still read by specially trained doctors.

In the new study, Published in the journal Circulation: Cardiovascular Imaging, researchers trained a neural network to read the cardiac MRI scans using results from nearly 600 patients.

The team then tested the system’s precision against an expert and trainee on 110 separate patients from multiple centers.

They found no significant difference in accuracy.

It comes as the first ever systematic review of AI in medicine, published in the Lancet Digital Health, found that machines are now equally as good at diagnosis as doctors in a number of fields.

Dr Charlotte Manisty, who led the UCL research, said: "Cardiovascular MRI offers unparalleled image quality for assessing heart structure and function.

“However, current manual analysis remains basic and outdated.

“Automated machine learning techniques offer the potential to change this and radically improve efficiency, and we look forward to further research that could validate its superiority to human analysis."

She added: “Our dataset of patients with a range of heart diseases who received scans enabled us to demonstrate that the greatest sources of measurement error arise from human factors.

“This indicates that automated techniques are at least as good as humans, with the potential soon to be 'super-human'--transforming clinical and research measurement precision."

The Lancet study, conducted by doctors at University Hospitals Birmingham NHS Foundation Trust, reviewed 14 studies which compared the performance of AI and health professionals.

Professor Alastair Denniston said: “Within those handful of high-quality studies, we found that deep learning could indeed detect diseases ranging from cancers to eye diseases as accurately as health professionals.

“But it’s important to note that AI did not substantially out-perform human diagnosis.”