New AI tool detects often overlooked heart disease

Doctors and scientists at the Smidt Heart Institute in Cedars-Sinai have created an artificial intelligence (AI) tool that can effectively identify and differentiate between two life-threatening heart diseases that are often overlooked: hypertrophic cardiomyopathy and cardiac amyloidosis degeneration

"Both heart diseases are a challenge even for cardiologists, so patients often go years or even decades before they are properly diagnosed," said David Ouyang, MD.

The new, two-step algorithm was used on more than 34,000 cardiac ultrasound videos from Cedars-Sinai and the Echocardiography Laboratory at Stanford Medical Center

"The algorithm identified high-risk patients more accurately than the eye of a trained clinical expert," said Ouyang Sidi

Without comprehensive testing, cardiologists have found it difficult to distinguish between similar developing diseases and changes in the shape and size of the heart that are sometimes considered part of normal aging

Cardiac amyloidosis, commonly referred to as "cardiac sclerosis syndrome," is a disorder caused by the deposition of an abnormal protein (amyloid) in heart tissue

The disease tends to affect older adults, black men, or patients with cancer or diseases that cause inflammation

Hypertrophic cardiomyopathy is a disease that causes the heart muscle to thicken and harden

Cardiac amyloidosis and hypertrophic cardiomyopathy, although two distinct diseases, look similar on echocardiography, the most commonly used cardiac imaging diagnostic method

Importantly, each heart disease can also mimic the appearance of a normal heart in the early stages of the disease, which gradually changes in size and shape with age

"One of the most important aspects of this AI technology is the ability to not only distinguish abnormal from normal, but also distinguish between these abnormal conditions, because each heart disease is treated and managed very differently," Ouyang said

Ouyang Sidi said they hope the technology can be used to identify patients early in their disease course

The researchers plan to soon launch a clinical trial in patients with suspected cardiac amyloidosis flagged by an artificial intelligence algorithm

A clinical trial in patients flagged by the algorithm as suspected hypertrophic cardiomyopathy has just begun at Cedars-Sinai Hospital

"In a relatively short period of time, the application of artificial intelligence in cardiology has grown rapidly and significantly," said Susan Cheng, MD, MPH, who is a member of the Smit Heart Institute's Department of Cardiology Health Director of the Institute on Aging and co-senior author of the study

article title

High-throughput accurate phenotyping of left ventricular hypertrophy using cardiovascular deep learning