Who can assess and diagnose cardiac function best: artificial intelligence or a sonographer?

At the military hospital in Zaporizhzhia, surgeons have been working around the clock to tend those wounded in the war – Copyright POOL/AFP Anthony Kwan

Is artificial intelligence better at assessing heart health? Research from Cedars-Sinai Medical Center suggests so. Outlining a new study, scientists have proposed that AI is superior in assessing and diagnosing cardiac function when compared with echocardiogram assessments made by human sonographers (professional who specializes in the use of ultrasonic imaging devices to produce diagnostic images).

The findings were drawn from a blinded, randomized clinical trial of AI in cardiology. The study involved the Smidt Heart Institute together with the scientists from the Division of Artificial Intelligence in Medicine at Cedars-Sinai.

At the heart of the research were novel clinical AI algorithms that were developed and tested within an established health system. This built upon earlier studies conducted at the Smidt Heart Institute and Stanford University which looked at cardiac function, specifically, left ventricular ejection fraction.

The new study assessed whether AI was more accurate in evaluating 3,495 transthoracic echocardiogram studies. This was by comparing initial assessment by AI or by a sonographer. AI was deployed to interpret echocardiograms (ECG), to automatically identify heart rhythms from an ECG and to uniquely identify an individual using the ECG as a biometric signal.

The results found that cardiologists more frequently agreed with the AI initial assessment and made corrections to only 16.8 percent of the initial assessments made by AI. In contrast, cardiologists made corrections to 27.2 percent of the initial assessments made by the sonographers.

Since this was a blinded study, the physicians were unable to tell which assessments were made by AI and which were made by sonographers.

In presenting an overview of the research, cardiologist David Ouyang says: “The results have immediate implications for patients undergoing cardiac function imaging as well as broader implications for the field of cardiac imaging.”

Ouyang  continues: “This trial offers rigorous evidence that utilizing AI in this novel way can improve the quality and effectiveness of echocardiogram imaging for many patients.”

Further examinations of the technology will be undertaken, with the potential for it be rolled out to other the clinical institutions. This will require the seamless integration of such AI into clinical software.  The aim of this type of deployment will be to save clinicians time and minimize the more tedious parts of the cardiac imaging workflow. The cardiologist will remain the final expert adjudicator.

It is additional hope that the findings will help artificial intelligence technologies to be considered for regulatory approval.

The research has been published in the science journal Nature, titled “Blinded, randomized trial of sonographer versus AI cardiac function assessment.”