Many sudden cardiac deaths could be prevented thanks to artificial intelligence.
As part of a new study published in the
European Heart Journal, on March 30, 2025
, an artificial neural network mimicking the human brain was developed by researchers from Inserm, Université Paris Cité, and AP-HP, in collaboration with American researchers.
By analyzing data from over 240,000 ambulatory electrocardiograms, this algorithm successfully identified patients at risk of severe arrhythmia that could lead to cardiac arrest within the following two weeks in more than 70% of cases.
Every year, sudden cardiac death is responsible for over 5 million deaths worldwide (1). Many of these cardiac arrests occur abruptly without identifiable warning signs and strike individuals in the general population, sometimes with no known history of heart disease.
Artificial intelligence could help better anticipate arrhythmias—unexplained heart rhythm disorders whose most severe forms can lead to fatal cardiac arrest—according to a new study led by a team of researchers from Inserm, Université Paris Cité, and AP-HP, in collaboration with American researchers.
As part of this study, an artificial neural network was developed by a team of engineers from Cardiologs (a Philips company) in collaboration with Université Paris Cité and Harvard University. Specifically, this algorithm mimics the functions of the human brain, with the goal of improving the prevention of sudden cardiac death.
The researchers analyzed millions of hours of heartbeats using data from 240,000 ambulatory electrocardiograms collected in six countries (United States, France, United Kingdom, South Africa, India, and the Czech Republic).
Thanks to artificial intelligence, they were able to identify new subtle signals indicating a risk of arrhythmia. The researchers particularly focused on the time required for electrical stimulation and relaxation of the heart's ventricles during a full cardiac contraction and relaxation cycle.
"We realized that it was possible to identify, through 24-hour electrical signal analysis, individuals likely to develop severe cardiac arrhythmia within the following two weeks. This type of arrhythmia, if left untreated, can progress to fatal cardiac arrest," explains Laurent Fiorina, the study's lead author, a researcher at the Paris Cardiovascular Research Center (Inserm/Université Paris Cité), a cardiologist at the Institut cardiovasculaire Paris Sud, and medical director in charge of artificial intelligence at Philips.
The artificial neural network is still under evaluation, but in this study, it demonstrated the ability to detect at-risk patients in 70% of cases and low-risk patients in 99.9% of cases.
In the future, this algorithm could be used to monitor at-risk patients in hospitals. With further refinement, it could also be integrated into devices such as ambulatory Holter monitors that measure blood pressure to detect hypertension risks, or even into smartwatches.
"What we are proposing here is a paradigm shift in sudden death prevention," comments Eloi Marijon, research director at Inserm within the Paris Cardiovascular Research Center (Inserm/Université Paris Cité), professor of cardiology at Université Paris Cité, and head of the cardiology department at Hôpital européen Georges-Pompidou AP-HP.
"Until now, we tried to identify at-risk patients in the medium and long term, but we were unable to predict what could happen in the minutes, hours, or days before a cardiac arrest. Today, thanks to artificial intelligence, we can predict these events in the very short term and, potentially, act before it's too late."
The researchers now aim to conduct prospective clinical studies to test the effectiveness of this model in real-world conditions.
"It is essential that this technology be evaluated in clinical trials before being integrated into medical practice," emphasizes Laurent Fiorina.
"But what we have already shown is that artificial intelligence has the potential to radically transform the prevention of severe arrhythmias."
Note:
(1)
https://www.thelancet.com/commissions/sudden-cardiac-death