ECG, AI can predict if you'll die of heart failure - Israeli, US study

With the help of a simple ECG test and artificial intelligence, it is possible to identify candidates for heart failure and even predict mortality.

Virtual image of human heart  (photo credit: ING IMAGE/ASAP)
Virtual image of human heart
(photo credit: ING IMAGE/ASAP)

With the help of simple tools available in many of the world’s medical centers, it is possible to predict the chances of patients with inflammatory muscle diseases developing heart failure – a disease that can endanger their lives – and even predict with high percentages which of them may die. 

The new, groundbreaking study was the result of a collaboration between researchers at the Rambam Health Care Campus in Haifa and colleagues at the Mayo Clinic in the US. It was published in the Mayo Clinic Proceedings: Innovations, Quality & Outcomes under the title “Electrocardiogram-Artificial Intelligence and Immune-Mediated Necrotizing Myopathy: Predicting Left Ventricular Dysfunction and Clinical Outcomes.”

The study, led by Dr. Shahar Shelly, head of neuromuscular diseases at Rambam’s neurology department, together with cardiology researchers at Mayo, which is ranked among the best in the world. It was carried out in two stages and focused on the department‘s patients who had previously been treated in the American hospital.

How did Israeli, US researchers conduct this groundbreaking heart study?

In the first stage, the researchers took all the patients who for 20 years (2000 to 2020) received a diagnosis of inflammatory muscle diseases (89 patients). In the second stage, with the help of collecting simple electrocardiogram (ECG) charts at the time and developing an artificial intelligence algorithm, the researchers were able to identify which of them was likely to suffer from heart muscle involvement as part of the disease. 

 Heart attack (Illustrative) (credit: FLICKR) Heart attack (Illustrative) (credit: FLICKR)

The algorithm was able to predict involvement of the heart muscle with a sensitivity of 80%, and that in a group of patients with a normal-looking ECG and echocardiogram, the algorithm predicted myocardial involvement even before the diagnosis by standard means. Apart from identifying myocardial involvement in these patients, the diagnostic model studied performed well in predicting mortality, when in patients marked blue with “high risk,” mortality rates were recorded that were seven times higher compared to the others. 

Inflammatory muscle diseases are complex and life-threatening diseases that can cause significant disability and even death. It is known that many of these patients have involvement of the heart muscle, which causes a further increase in mortality and is an indication for intensive treatment. One of the heart diseases is heart failure, a disease in which the heart, usually due to a problem in the left ventricle of the heart, has difficulty meeting the metabolic demands of the various organs in the body, which causes shortness of breath, fatigue and edema.

“Further down the road, the use of this model will allow the provision of appropriate treatments at an early stage, even before the deterioration of the patients' medical condition. We are talking here about preventing serious illness and even deaths.”

Prof. Shahar Shelly

According to Shelly, the research is at the forefront of science and provides the basis for the early identification of patients at risk of heart failure: “Further down the road, the use of this model will allow the provision of appropriate treatments at an early stage, even before the deterioration of the patient's medical condition. We are talking here about preventing serious illness and even deaths.” 

From a research/technical point of view, Shelly explained that the use of artificial intelligence “as a fast and accurate research tool on the ECG saves a lot of time and costs for patients in whom an echocardiogram is not immediately available. Artificial intelligence technology continues to bring about revolutionary changes in areas such as drug development and early identification of patients at high risk for intervention. The research in collaboration with the Mayo Clinic continues, and there are many more projects in the pipeline.”