AI Capable Of Diagnosing Heart Disease In Seconds

Recent scientific tests have shown that AI powered algorithms can diagnose heart problems including heart disease in as low as four seconds. This means that artificial intelligence is capable of beating human doctors at some diagnosis’s in a fraction of the time.

For context, reading, analysing and concluding a patient’s heart status on a cardiac magnetic resonance imaging (MRI) scan usually takes a doctor around thirteen minutes. This data is based on a set of measurements taken on 300 doctors in California, USA.

Other MRI scans can produce images of other areas of the body including the brain. Source: BBC News

Other MRI scans can produce images of other areas of the body including the brain. Source: BBC News

However, artificial intelligence algorithms have yet again appeared to shake up the healthcare industry. A newly completed trial undertaken by UCL (University College London) provided evidence that a machine learning, AI program could scan the readings from MRI scans in a time far shorter than the typical thirteen minutes. What’s more, the artificial intelligence could also complete these examinations with equal accuracy and precision compared to human workers.

These scans matter. Last year, about 150,000 cardiac MRI scans were carried out on patients in the UK. The research team at University College London predicted that, if healthcare services fully embraced a more developed version of their AI prototype, the artificial intelligence could reduce clinician-days at each cardiac centre by fifty four per year.

In order to train their AI, the researchers made use of almost 600 patients’ cardiac MRI scans and fed these results to a neural network. This allowed the program to learn how the readings led to different diagnosis and when there was nothing wrong with the patient’s heart function. The AI was then tested against an expert MRI scan analyser on 110 different MRI scans. The research team found no difference in accuracy.

This AI would not only make doctors more efficient with their time, but also would allow patients to have faster feedback meaning problems can be dealt with faster leading to fewer problems in the future.

Thumbnail source: Stanford Health Care

HealthcareEdward Bristow