Nature: Watches can save people at critical moments! Mayo Hospital study: AI applied to smartwatches can diagnose heart failure!

In recent years, more and more people have begun to wear smart watches, in addition to watching the time and connecting the mobile phone to receive information, there are also sports records and health monitoring functions, it has been revealed that the heart rate detection function of Apple Watch reports physical health, and early medical treatment to retrieve a life, so that many netizens exclaimed: "It turns out that smart watches can really save people!" ”

Recently, a study from Mayo Hospital reported the ability of a smartwatch electrocardiogram to
accurately detect heart failure in a non-clinical setting.
The researchers applied artificial intelligence (AI) to Apple Watch electrocardiogram recordings to identify patients
with weak heart pumps.
Participants in the study could remotely record their smartwatch electrocardiograms
anytime, anywhere.
They regularly upload their ECGs automatically and securely to their electronic health records via a smartphone app
.
The study, titled "Prospective evaluation of smartwatch-enabled detection of left ventricular dysfunction," was published in the journal Nature Medicine
.

Image source: Nature Medicine

From August 2021 to February 2022, the researchers recruited 3,884 patients at Mayo Hospital who used the software, of whom 2,463 (63.
5%) uploaded at least one ECG
during the study.
Nine subjects were excluded due to early software application failures that prevented data collection (which were subsequently remedied), leaving 2,454 subjects
in the final cohort.
2,454 patients from 46 states and 11 countries recorded and uploaded one or more Apple Watch ECGs
.
The average age of these patients was 53 years and the majority were female and white
.
These 2,454 patients recorded more than 125,000 electrocardiograms (ECGs), of which 78.
5% were classified as sinus rhythm by the Apple Watch (n = 98,603).

The remaining Apple Watch ECGs are classified as atrial fibrillation (AF; 5.
1%) or uncertain (16.
4%)
.

The flowchart (left) is a combination chart
summarizing patient enrollment.
The map (right) depicts the geographic distribution of
registered patients within the United States.

On average, patients used the software 2.
1 times per month, with an increase in the number of uses positively correlated
with patient age.
During the study period, each patient was given an average of 51 ECGs, with an average of 7.
8 ECGs
per day.

Study participants participate in customized Mayo Clinic iPhone software (Credit: Nature Medicine)

After the Mayo Clinic ECG study software is created, the Apple Watch ECGs recorded by the patient are transferred to the
medical record in the ECG dashboard.
The dashboard is paired with data from electronic medical records (EMRs) and now has a Mobile ECG tab where the Apple Watch ECG recorded by the patient after uploading is located
.
From here, providers can view these results in real-time to assist patients with appropriate care
.

Approximately 421 unique patients were identified
within 30 days with sinus rhythms recorded by Apple Watch ECG and transthoracic echocardiography (TTE).
Of these patients, 16 (3.
8%) had ejection fractions (EFs) ≤ 40%.

A review of clinical records found that 12 of 16 patients had mild or no left ventricular systolic dysfunction (LVSD).

Assess ejection fraction (EF) using watch AI-ECG (Image: Nature Medicine)

Mayo researchers interpreted Apple Watch single-lead ECGs by modifying an earlier algorithm developed for 12-lead ECGs, which have been shown to detect weak heart pumps
.
The 12-lead algorithm for low ventricular ejection fraction has been licensed to Anumana Inc.
, an AI-powered health technology company co-created
by nference and Mayo Clinic.

Although the data is earlier, the modified AI algorithm using single-lead ECG data has an area under the curve of 0.
88 to detect weak heart pumps
.
In contrast, this measure of accuracy is as good or slightly better
than a medical treadmill diagnostic test.

"Currently, we diagnose ventricular insufficiency — weakened heart pump function — by echocardiograms, CT scans, or MRIs, but these methods are expensive, time-consuming, and sometimes inaccessible
," the researchers said.
The ability to remotely diagnose the weakened heart pump via ECG allows people to easily record using devices such as smartwatches to detect this potentially life-threatening major disease
in a timely manner.
”

People with weak heart pumps may not have symptoms, but this common heart condition affects about 2% of the population and 9% of people
over the age of 60.
When the heart can't pump enough oxygen-rich blood, symptoms may occur, including shortness of breath, difficulty breathing, rapid heartbeat, and swollen
legs.
Early diagnosis is important because, once identified, there are many treatments that can improve quality of life and reduce the risk of
heart failure and death.
The researchers added: "The project's research data are encouraging because they show that digital tools can be convenient, inexpensive and scalable to screen for important diseases
.
With this technology, we can meet people's consultation needs and collect useful information
about the patient's heart remotely.
”

In conclusion, this study applied AI capabilities to acquire ECGs from Apple Watch in non-clinical settings, demonstrating its effective role
in effectively identifying left ventricular dysfunction.
The high level of patient participation in clinical enrollment in this study indicates an opportunity to use AI in remote care and clinically validate AI-ECG models in geographically dispersed populations at a lower cost using patients' own devices, in a potentially massively scalable manner that can improve quality of life, reduce costs through early detection and cost reduction, and enhance patient engagement
.

In this way, Apple Watch and other similar smart watch products can become daily physical condition tracking devices, as long as you walk around the home can complete the test, really saving the current epidemic environment to queue up in the clinic and test nucleic acid! Not only has it become a "family doctor" around us, but the most important thing is to provide a greater incentive for the general public to have a smart watch
.

References:

#Sec4