An AI to Predict Heart Function
Coronary heart disease is a leading cause of death globally. Clinicians rely heavily on a procedure called coronary angiography to make decisions about treatment.
However, to measure the heart’s pumping efficiency, an additional invasive procedure, known as left ventriculography, is often required.
This procedure comes with its own risks and increases the patient’s exposure to contrast dye.
In a new study, Geoff Tison, a cardiologist from the University of California, San Francisco (UCSF), and Robert Avram from the Montreal Heart Institute, decided to try a different approach.
They developed a deep neural network (DNN), a type of artificial intelligence (AI) algorithm, to predict heart function from standard angiogram videos.
This AI, named CathEF, was designed to estimate the left ventricular ejection fraction (LVEF), which is a measure of how much blood is being pumped out of the heart’s main pumping chamber.
A Revolutionary Tool
This AI tool presents a new and exciting way of using the data collected during angiograms.
Not only does it eliminate the need for an additional invasive procedure, but it also provides real-time LVEF information that can guide clinical decision-making.
To train CathEF, the researchers used over 4,000 adult angiograms from nearly 3,700 UCSF patients.
The results showed a strong correlation between the AI’s predictions and the standard measurements obtained from echocardiograms, which are usually used to assess LVEF.
The researchers then validated CathEF using real-world angiograms from the Ottawa Heart Institute.
The AI’s performance remained consistent across different patient demographics and clinical conditions, including acute coronary syndromes and varying levels of kidney function.
Future Research and Applications
Despite these promising results, further research is needed to fully understand CathEF’s potential.
The researchers are currently conducting a study to assess the impact of this AI tool on the clinical workflow for patients experiencing heart attacks.
This study will compare CathEF’s performance to the standard left ventriculogram in patients with acute coronary syndromes.
The Power of AI
This study showcases the potential of AI technology in improving diagnostic capabilities and reducing the need for invasive testing.
By accurately predicting LVEF from standard angiograms, AI tools like CathEF could revolutionize how we approach heart health, ultimately improving patient outcomes and quality of life.
If you care about heart health, please read studies about an important cause of heart disease, and how to remove plaques that cause heart attacks.
For more information about heart health, please see recent studies about How vitamin K helps protect the heart, reduce blood clots and death risk and results showing that herbal supplements could harm your heart rhythm.
The study was published in JAMA Cardiology.
Copyright © 2023 Knowridge Science Report. All rights reserved.