Cedars-Sinai, Stanford develop AI to spot 2 difficult-to-diagnose heart conditions

Cedars-Sinai’s Smidt Heart Institute and Stanford Healthcare are doing the two-step—no dancing shoes necessary.

Researchers from both California-based medical centers have developed a two-step artificial intelligence algorithm that aims to identify signs of and distinguish between two rare, often overlooked heart conditions. They described their findings in a study published this week in JAMA Cardiology.

The algorithm is designed to flag indicators of hypertrophic cardiomyopathy, where the heart muscle thickens, and cardiac amyloidosis, also known as stiff heart syndrome, in which normal heart walls are slowly replaced by amyloid deposits.

Both conditions make it difficult for the heart to pump blood properly but may not always present with noticeable symptoms. Cardiac amyloidosis, in particular, is also more likely to affect older, Black men, a patient pool that includes many underserved by the current healthcare system, making the condition even less likely to be quickly and accurately diagnosed.

“These two heart conditions are challenging for even expert cardiologists to accurately identify, and so patients often go on for years to decades before receiving a correct diagnosis,” said the study's senior author, David Ouyang, M.D., a cardiologist in the Smidt Heart Institute. “Our AI algorithm can pinpoint disease patterns that can’t be seen by the naked eye and then use these patterns to predict the right diagnosis.”

RELATED: Using the eyes as a window, AI predicts annual heart attack risk from retinal scans

The deep learning algorithm was trained to analyze cardiac ultrasound videos, looking at the thickness of the heart’s walls and the size of its chambers to find abnormalities that may be linked to either of the two conditions.

In the study, the AI was applied to more than 34,000 cardiac ultrasounds collected by Cedars-Sinai and Stanford. It was able not only to measure the heart’s dimensions to within a few millimeters but also to classify hypertrophic cardiomyopathy and cardiac amyloidosis with remarkably high accuracy.

In fact, according to Ouyang, “The algorithm identified high-risk patients with more accuracy than the well-trained eye of a clinical expert.”

He continued, “This is because the algorithm picks up subtle cues on ultrasound videos that distinguish between heart conditions that can often look very similar to more benign conditions, as well as to each other, on initial review.”

RELATED: Cleerly's AI scans comparable to gold-standard angiography in detecting heart disease: study

With the algorithm’s potential precision now proven in the retrospective study, the Cedars-Sinai and Stanford researchers will launch clinical trials to study patients flagged by the AI.

The trial of patients suspected to have signs of hypertrophic cardiomyopathy has already begun at Cedars-Sinai. The cardiac amyloidosis study, meanwhile, is slated to kick off soon at the hospital’s Smidt Heart Institute, which hosts one of the small group of programs in the U.S. focusing on the disease.