AI flags heart failure risk five years early from routine ECG recordings

AI Algorithms Show Promise in Early Detection of Heart-Related Conditions

Recent research indicates that artificial intelligence (AI) is providing new methods for detecting heart conditions, including pulmonary hypertension and cardiovascular event risks, years before clinical diagnosis.

Detecting Pulmonary Hypertension via ECG

An AI algorithm known as the PH Early Detection Algorithm (PH-EDA) has shown the potential to identify pulmonary hypertension (PH) using data from standard, non-invasive electrocardiogram (ECG) tests. Developed by scientists at Anumana, Janssen Research and Development, the Mayo Clinic, and the Vanderbilt University Medical Center, the tool is designed to address the diagnostic delays often associated with PH.

According to the study published in the European Respiratory Journal, the algorithm demonstrated high predictive accuracy, with an area under the receiver operating characteristic curve (AUC) of 0.92 in a Mayo Clinic dataset and 0.88 in a Vanderbilt dataset. Notably, the algorithm remained effective when analyzing ECGs taken six to 18 months prior to a diagnosis, and in some cases, up to five years before a diagnosis. The U.S. Food and Drug Administration designated the PH-EDA as a breakthrough device in May 2022.

Predicting Cardiovascular Events Through Eye Scans

Separately, researchers at the University of Dundee have developed an AI technology capable of predicting a major cardiovascular event—such as a heart attack or stroke—within the next decade with 70% accuracy. The study, published in the journal Cardiovascular Diabetology, utilizes digital retinal photographs typically taken during routine eye exams.

The AI analyzes blood vessels in the back of the eye for signs of damage or narrowing, which researchers suggest act as a “window to the heart.” The technology also allows for tracking heart health over time; the study found that individuals who experienced a small increase in their AI-calculated risk score over three years had a 54% higher risk of a major cardiovascular event. While the study was trialled on patients with diabetes, researchers suggest the method could eventually be applied to the general population to help clinicians identify patients who may benefit from early lifestyle changes or medication.

Find more reporting in our Health section.