ML model predicts ICU cardiac arrest from ECG with 88% accuracy

The model aims to improve survival outcomes for ICU patients at risk of sudden cardiac arrest through earlier detection and intervention.

In the intensive care units of Seoul, where the heart's electrical story is already being told in real time, researchers have found a way to read that story before its ending arrives. A machine learning model trained on the subtle rhythms between heartbeats — not the beats themselves, but the silences — can now predict sudden cardiac arrest up to 24 hours in advance, with an accuracy that outpaces the most thorough clinical assessments. It is a reminder that the body often knows what is coming before we do, and that the task of medicine is, in part, to learn how to listen.

ML model predicts ICU cardiac arrest from ECG with 88% accuracy

Cobertura Relacionada

Recibe The Register en tu bandeja