AI potentially detects 72% more heart disease patients

A recent study suggests the artificial intelligence (AI) decision support software EchoSolv could help cardiologists diagnose aortic stenosis (AS).

EchoSolv uses AI to assess echocardiographic measurements, potentially making it easier and more reliable for a cardiologist to identify AS.

A study recently completed at St. Vincent’s Hospital (Sydney and Melbourne) has shown the benefits of using EchoSolv to complement human diagnosis. According to Echo IQ, the results showed:

• 72% more patients with Severe AS were identified when EchoSolv was used to assess patient records versus current clinical practice of using human diagnosis alone.
• Women were 66% less likely to be accurately diagnosed than men through human diagnosis alone, demonstrating an unconscious bias that EchoSolv was able to address.
• When patients were under-identified using human diagnosis alone, there was a low rate of aortic valve replacement, the key treatment for Severe AS.
   

Director Cardiology and Heart Lung Program at St Vincent’s Hospital Professor Michael Feneley said, “This study clearly demonstrates not only the effectiveness of novel technologies such as EchoSolv in enhancing human diagnosis, but also its potential to reduce bias in decision-making.

“With the general aging of the population leading to the increasing prevalence of aortic stenosis, it is encouraging to see how artificial intelligence could be used to improve the identification of disease and all the increased opportunities to treat patients in a timely manner this provides.”

These results seem to reinforce the findings of an earlier proof of concept study published in the Journal of the American College of Cardiology reporting similar results.

Image credit: iStock.com/Pixelimage