Mayo Clinic Uses Artificial Intelligence to Diagnose Heart Failure

Published in Nature Medicine, a study conducted by Mayo Clinic researchers describes how artificial intelligence (AI) can be used to assess electrocardiogram (ECG) recordings to detect heart failure among patients outside of a clinical setting.

According to the Centers for Disease Control and Prevention (CDC), about 6.2 million adults in the US suffer from heart failure.

Acknowledging the need for increased monitoring of patients at risk for heart failure, Mayo Clinic researchers created an AI-based method to not only monitor patients more closely but also to enhance the diagnosis process.

"Currently, we diagnose ventricular dysfunction ― a weak heart pump ― through an echocardiogram, CT scan or an MRI, but these are expensive, time consuming and at times inaccessible. The ability to diagnose a weak heart pump remotely, from an ECG that a person records using a consumer device, such as a smartwatch, allows a timely identification of this potentially life-threatening disease at massive scale," says Paul Friedman, MD, senior author of the study, and chair of the Department of Cardiovascular Medicine at Mayo Clinic in Rochester, in a press release.

Often, weak heart pumps go unidentified as some patients may not experience symptoms, the press release noted. However, symptoms may emerge when the heart cannot pump enough oxygen-rich blood, leading to shortness of breath, a rapid heart rate, or swelling of the legs.

The AI method aims to improve the detection of heart issues, which could lead to better outcomes since early diagnosis is critical.

To interpret the Apple Watch's single-lead ECGs, researchers modified an existing AI algorithm to detect a weak heart pump developed for 12-lead ECGs.

To test the AI-based diagnosis method, researchers gathered 2,454 study participants, all of whom were Mayo Clinic patients from the US and 11 other countries. They were asked to download an application that allowed them to upload their Apple Watch ECGs to their EHRs. They uploaded more than 125,000 Apple Watch ECGs to their EHRs between August 2021 and February 2022. Clinicians could access this information through an AI-based dashboard within the EHR system.

After reviewing the data, researchers found that 420 patients had a traditional ECG within 30 days of logging an Apple Watch ECG in the app. Among those 420 patients, the AI method showed that 16 had a low ejection fraction, indicating a weak heart pump. The traditional ECGs confirmed this.

"These data are encouraging because they show that digital tools allow convenient, inexpensive, scalable screening for important conditions. Through technology, we can remotely gather useful information about a patient's heart in an accessible way that can meet the needs of people where they are," says Zachi Attia, PhD, the lead AI scientist in the Department of Cardiovascular Medicine at Mayo Clinic, and first author of the study, in the press release.

Similarly, a Mayo Clinic AI strategy created in September aimed to identify new cases of atrial fibrillation using ECG data. For the study, researchers had over 1,000 patients participate in usual care, and another 1,000-plus patients participate in care through the AI-based intervention method. After finding that the AI method detected atrial fibrillation in both low-and high-risk patients, researchers determined that it was an accurate and applicable tool for heart care.

Also, in May, Geisinger and Tempus created an AI model that displayed the ability to predict heart disease. After collecting data from millions of ECGs, researchers trained a deep neural network to predict which patients would develop the disease.