New Israeli AI tool predicts heart failure with 80% accuracy

Israeli scientists have developed an artificial intelligence programme that scans ECG readings and accurately forecasts heart failure weeks in advance.

The new AI technology target people suffering from Myositis, a kind of muscle inflammation that increases the risk of cardiac failure. The research lead on this technology is physician Dr Shahar Shelly coming from the Rambam Healthcare Campus. The research was conducted in cooperation with the Cardiology Department at the US-based Mayo Clinic Medical Center.

According to his peer-reviewed research, the study took patients with immune-mediated necrotizing myopathy (IMNM) or Myositis and used an electrocardiogram (ECG)-artificial intelligence (AI) algorithm to predict when a patient will have a heart failure. The algorithm was applied to their ECG results for the study, and its predictions of who was at risk of heart failure were compared to medical data to determine who experienced heart failure.

To determine the results a validated AI algorithm was used with the 12-lead standard ECGs to detect left ventricular dysfunction (LVD).

The result was displayed as a percentage likelihood of LVD. Electrocardiograms taken before and after immunotherapy were compared. LVD-predicted likelihood scores with echocardiograms, mortality, and immunotherapy treatment response were also evaluated.

They used the ECG scans and medical information of 89 myositis patients from 2000 to 2020 to teach the AI model. The programme created a picture of minute ECG rhythms that appear to raise the risk of heart failure.

Another thing that was noted was that patients with IMNM saw better heart conditions when on immunotherapy. The research study finds it likely that early observation of cardiac involvement with proper immunotherapy escalation could reduce the risk of interstitial myocardial fibrosis.

However, the research study has pointed out some limitations to its work which is the lack of uniform follow-up with echocardiograms in all patients.