AI Platform Screens for Amyloid Cardiomyopathy Using ECG Images

Amyloid cardiomyopathy is a heart disease caused by deposits of misfolded proteins in the cardiac muscle. It is often missed until patients develop serious complications such as heart failure. Delayed diagnosis limits the opportunity to intervene before extensive cardiac damage occurs. Researchers have now developed an artificial intelligence platform that screens for the condition using images from electrocardiograms.

Yale School of Medicine (New Haven, CT, USA) scientists developed the AI to identify patients at risk for amyloid cardiomyopathy caused by misfolded transthyretin. The disease develops when transthyretin, a protein made in the liver, becomes misfolded because of genetics or age. Its accumulation stiffens the heart and impairs the electrical system that regulates heartbeat.

The platform analyzes images from electrocardiograms, which are non-invasive tests that measure the heart’s electrical activity. Physicians can access the tool from smartphones and use a photo of an electrocardiogram as the input. The model is designed to detect subtle electrocardiographic patterns associated with cardiac amyloid risk.

Researchers in the Yale Cardiovascular Data Science Lab first trained an AI model to interpret electrocardiogram results using data from thousands of de-identified patients. They then used several hundred patients in the dataset who already had amyloid cardiomyopathy to build a model that could recognize disease-associated patterns. In the study, the team tested the model across eight distinct patient cohorts in the United States and Europe.

The platform successfully identified individuals with transthyretin amyloid cardiomyopathy in those cohorts. The work was described recently in JAMA. The tool has received U.S. Food and Drug Administration (FDA) Device Designation through the Breakthrough Devices Program and is currently under FDA review. Scientists are also studying implementation of related multimodal artificial intelligence tools in the TRACE-AI Network Study, an observational study at 13 U.S. health centers.

“We are able to leverage just an image from a very simple ECG into a screening test for cardiac amyloid. For a disease that’s massively underdiagnosed, identifying those at risk is very critical,” said Rohan Khera, MD, director of the Cardiovascular Data Science Lab at Yale School of Medicine and the study’s principal investigator. 

“Our tool can really narrow down the funnel for who should be further evaluated for cardiac amyloid,” added Philip Croon, MMed, associate research scientist at Yale School of Medicine and the first author of the study.

Related Links
Yale School of Medicine