A new algorithm may be able to identify people with heart disease at risk for cardiac arrest, researchers say. Photo by Myriams-Fotos/Pixabay
April 7 (UPI) -- A new artificial intelligence-based approach using scans of patients' hearts and their medical history can predict whether they will die from cardiac arrest, a study published Thursday by Nature Cardiovascular Research found.
The technology, called Survival Study of Cardiac Arrhythmia Risk, uses neural networks to build a personalized survival assessment for each patient with heart disease, the researchers said.
These risk measures provide with more than 80% accuracy their risk for a sudden cardiac death over the ensuing 10 years, and when it is most likely to happen, according to the researchers.
The name Survival Study of Cardiac Arrhythmia Risk is a reference to cardiac scarring caused by heart disease that often results in life-threatening arrhythmias, or irregular heartbeats, they said.
This scarring is key to the approach's ability to make predictions, which could allow for treatment to prevent deadly cardiac arrests, the researchers said.
"Sudden cardiac death caused by arrhythmia accounts for as many as 20% of all deaths worldwide, and we know little about why it's happening or how to tell who's at risk," study co-author Natalia Trayanova said in a press release.
"What our algorithm can do is determine who is at risk for cardiac death and when it will occur, allowing doctors to decide exactly what needs to be done" said Trayanova, a professor of biomedical engineering and medicine at Johns Hopkins University in Baltimore.
Cardiac arrest occurs when heart stops pumping blood effectively, resulting in a sudden loss of blood flow throughout the body. It often is sudden and is a medical emergency that requires immediate treatment, according to the American Heart Association.
However, research indicates many people who suffer a cardiac arrest feel unwell in the days leading up to it.
The approach developed by Trayanova and her colleagues uses contrast-enhanced heart images that visualize scar distribution from hundreds of real patients at Johns Hopkins Hospital with cardiac scarring, they said.
The images effectively train the algorithm to detect patterns and relationships between scarring and heart health not visible to the naked eye, according to researchers.
The team trained a second neural network to learn from 10 years of standard clinical patient data, including factors such as patient age, weight, race and prescription drug use, the researchers said.
With this data, algorithms' predictions were more accurate on every measure than those made by doctors, they said.
The team is now working to build algorithms to detect other heart diseases.
"There are patients who may be at low risk of sudden cardiac death getting defibrillators they might not need, and then there are high-risk patients who aren't getting the treatment they need and could die in the prime of their life," Trayanova said.
"This has the potential to significantly shape clinical decision-making regarding arrhythmia risk and represents an essential step towards bringing patient trajectory prognostication into the age of artificial intelligence," she said.