When the heart's electric system malfunctions, it causes irregular beating, leading to cardiac arrest. Sudden cardiac arrest could happen to anyone and anywhere; that is why it is dubbed the "silent killer." However, researchers from Johns Hopkins University developed a new form of artificial intelligence (AI) that may accurately predict if and when a person will die from cardiac arrest.
The AI was trained using raw images of a patient's diseased heart and background. Healthline reported that the AI-based approach examines scarring in heart muscles not visible to the naked eye to make predictions with accuracy. The team believes that the program stands to revolutionize clinical decision-making and increase the survival rate of sudden and lethal heart arrhythmia.
New Form of AI Detects Invisible Patterns of MRI to Accurately Predict If and When a Person Dies From Cardiac Arrest
First-Of-Its-Kind Survival Predictor Detects MRI Patterns Invisible to the Naked Eye
Even with 20% of all deaths worldwide due to heart arrhythmia, there is no way to know why it is happening and who is at risk, SciTech Daily reported. With this in mind, researchers sought to make a program that will accurately determine who is at risk or cardiac and when it occurs to help doctors decide the clinical action to follow, says senior author Natalia Trayanova.
They are the first to use neural networks to build a personalized survival assessment for heart disease patients, measuring risk to provide high accuracy of their chances of a sudden cardiac arrest in the next 10 years and when it is most likely to happen.
The new AI technology called Survival Study of Cardiac Arrhythmia Risk (SSCAR) uses contrast-enhanced cardiac images that visualize scar distribution from hundreds of real patients at Johns Hopkins Hospital. After training the AI with the data, they found that it can detect patterns and relationships not visible to the naked eye.
ALSO READ: Changes in Blood Vessels in the Retina Linked to Risk of Heart Disease, Study Suggests
Importance of the New AI to Healthcare
Dan Popescu, the first author of the study and former Johns Hopkins doctoral student, said that the images from the AI carry critical information that doctors have not been able to access.
Then they trained a second neural network that learned from 10 years worth of standard clinical patient data, identifying 22 factors, including age, weight, race, and prescription drug use, Science Daily reported. The AI's predictions were more accurate on every measure than doctors.
When validated in tests, it showed that the platform could be adopted anywhere, a potential that will significantly change decision-making in terms of heart arrhythmia risk and presents an essential step toward using AI in conducting prognosis. Trayanova describes their innovation as the epitome of a trend of merging AI, engineering, and medicine in the future of healthcare.
Dangers of Cardiac Arrest
According to the American Heart Association, about 356,000 cardiac arrests happen outside of a hospital in the US every year. Cardiac arrests occur due to the abrupt loss of heart function that may have been diagnosed with heart disease. It can come suddenly without warning and is often fatal.
Moreover, cardiac arrest is different from a heart attack. Whereas the former is caused when the heart's electrical system malfunctions, the latter is caused by a blockage that stops blood flow to the heart. Death from heart attack refers to the death of heart muscle because of the loss of blood supply, creating circulation problems.
Cardiac arrests may happen because of a kind of irregular heartbeat called arrhythmia. Johns Hopkins researchers are hopeful that their AI will help improve the survival rates of cardiac arrest.
RELATED ARTICLE: Sudden Cardiac Arrest Is Actually Not That Sudden -- A New Study
Check out more news and information on Heart Health in Science Times.
