AI Accurately Detects LV Dysfunction Using Single-Lead Apple Watch ECG

AI Accurately Detects LV Dysfunction Using Single-Lead Apple Watch ECG

Although more work remains before this approach is widely adopted, experts see potential for use as a screening tool.

An artificial intelligence (AI) algorithm trained to interpret single-lead ECGs from the Apple Watch accurately detected signs of subclinical left ventricular systolic dysfunction in a proof-of-concept study, opening up the possibility of using measurements obtained at home to find and treat patients before they progress to more serious disease.

With an area under the receiver operating characteristic curve (AUC) of 0.89, the AI-ECG algorithm has performance that compares favorably to other screening tests commonly used in medicine, such as mammography for cancer. breast and cervical cytology for cervical cancer, researchers led by Zachi Attia, PhD (Mayo Clinic, Rochester, MN), noted in a study published online this week in natural medicine. Some of the results were presented earlier this year at Heart Rhythm 2022.

Lead author Paul Friedman, MD (Mayo Clinic), told TCTMD that this is “proof of concept” for using the AI ​​algorithm to interpret single-lead ECGs. from a smartwatch, adding that further validation and implementation work is needed before rolling out the approach more generally.

Implementation testing at other hospitals is expected to begin within the next two months, he said. “We have demonstration projects that seem incredibly exciting and encouraging, but we must continue, as with any new tool, to verify, validate and prove that in the real world it improves human life, reduces the risk of death and reduces the risk of morbidity. And that will require more study.

Other experts echoed both the call for more research and the positive outlook for the approach.

Rhodri Davies, MBBS, PhD (University College London, England), said the study was “very promising”, citing strengths of the prospective design and the use of data separate from that used to derive the algorithm initially.

If the algorithm is validated in future studies and implementation issues are resolved, the approach could find a place in clinical practice, he said. “It should not replace echocardiography or other means of correctly measuring LV ejection fraction, but as a screening tool it is potentially very useful.

Take the next step

Several research groups have developed AI algorithms, primarily applied to 12-lead ECGs obtained in clinical settings, to detect various conditions including valvular heart disease, hypertrophic cardiomyopathy, amyloid heart disease, electrolyte abnormalities, and silent arrhythmias. The Mayo Clinic team previously showed that an algorithm trained to detect signals on 12-lead ECGs increased detection of LV systolic dysfunction by more than 30% over usual care, when deployed in primary care practices in the EAGLE study.

But, until now, the approach had not been shown to work when the algorithm was adapted to interpret single-lead ECGs, such as those obtained with the Apple Watch.

It should not replace echocardiography or other means of correctly measuring LV ejection fraction, but as a screening tool it is potentially very useful. Rhodri Davies

For the current study, investigators emailed patients who had previously been treated at Mayo Clinic and had downloaded the patient’s mobile app inviting them to participate. Those who agreed were asked to download the Mayo Clinic ECG study app, which would send ECGs measured on the Apple Watch to a secure dashboard coupled with the patient’s electronic medical record. Friedman explained that the Apple Watch was chosen because the company had made all raw ECG data available through the Apple Health Kit; the company did not participate in the study.

A total of 3884 patients were recruited, of whom 2454 (63%) submitted at least one ECG during the study period. In this latter group, drawn from 46 US states and 11 countries, the average age was 53, 56% were female and 88% were Caucasian. A total of 125,610 ECGs were sent to the secure dashboard, with 92% of patients using the ECG app more than once and half using it more than five times. Most ECGs (78.5%) were classified as normal sinus rhythm by Apple Watch, 5.1% as atrial fibrillation, and 16.4% inconclusive.

To assess performance in identifying subclinical cardiac dysfunction, researchers focused on 421 patients who submitted at least one ECG within 30 days of a clinically indicated echocardiogram. Of these, 16 (3.8%) had an ejection fraction of 40% or less, with most of this group showing minimal or no symptoms of LV systolic dysfunction.

Friedman said the AI-ECG algorithm “worked very well” for detecting cardiac dysfunction, noting that the AUC of 0.89 is similar to what is seen with a treadmill test (AUC 0.85) . “An ECG recorded from a consumer device, when we apply this AI analysis to it, can detect what could potentially be a life-threatening silent heart disease,” he said.

Potential uses

Friedman called this study the first test of this approach, which needs to be followed by further research. But if the results are validated, there are some potential roles for the technology in clinical practice. First, it could be used to screen for subclinical LV dysfunction in high-risk individuals, such as the elderly or those with diabetes, and then initiate appropriate treatments when detected.

Another possibility is for patients undergoing chemotherapy for cancer, who could monitor treatment-related heart damage using ECGs obtained on a smartwatch rather than undergoing periodic echocardiograms, he said.

Partho Sengupta, MD (Robert Wood Johnson University Hospital, Rutgers Robert Wood Johnson Medical School, New Brunswick, NJ), said the demonstration of the accuracy of the AI-ECG approach to detecting cardiac dysfunction with single-cell ECGs lead – with performance similar to when the algorithm was applied to 12-lead ECGs – this is a “quite spectacular development” as it brings the approach into focus.

He raised some questions about the size of the study, the relative lack of racial/ethnic diversity among participants, and whether all patients were truly asymptomatic, and he highlighted the need for further research conducted in other centers and countries. . With these caveats, he called it a “fantastic technological innovation”. . . . It democratizes the ability to get that information.

It is extremely promising. Partho Sengupta

Sengupta also saw a role for this approach as a screening tool if the study results are confirmed in future research. Patients with subclinical LV dysfunction could be treated early, preventing progression to symptomatic heart failure, or examined more closely to discover the reason for the reduced ejection fraction, which could then be treated, he said. suggested.

He stressed that any potential impact on clinical outcomes remains to be proven in randomized trials. Future studies should also assess whether all patients with heart failure, regardless of ejection fraction, can be identified using the AI-ECG algorithm, he said.

Implementation challenges

There are challenges when it comes to implementing such an approach more widely in clinical practice, experts agreed. Davies emphasized the need for the infrastructure to collect the ECG data and for the right clinician to verify the information and act on it. There is no point in alerting a patient that their watch has detected signs of LV dysfunction if there is no connection to a cardiologist or other physician who can take a history and perform the tests. imaging needed to confirm the result, he said. .

Also, as something like this is introduced, “you have to be careful who you apply it to”, Davies said, noting that the algorithm should work better in older, high-risk patients than in younger, healthier individuals. False positives should also be minimized, he noted.

Friedman agreed that there is a risk of false positives if testing is applied to the wrong kinds of individuals, saying there is a need to have some operational processes in place to ensure it is offered to the correct populations.

The technical aspect of screening is already built in, as the secure dashboard has already been built in, Friedman said, adding that while most hospitals don’t have a similar platform yet, it’s likely that more will get in the next two years.

When optimally integrated into a health system’s electronic medical records system, this strategy should not add much additional time for a clinician, especially as hospitals gain experience with it. , pointed out Friedman. “Once in this situation, I don’t think it will have a significant impact on workflows. It must be consistent to be effective. Doctors are overwhelmed. We can’t make them busier.

Another question regarding implementation, Sengupta pointed out, is whether individuals will be prone to “alarm fatigue” if they constantly receive alerts about potential issues with their smartwatches. And this is an aspect that needs to be addressed by studying AI-ECG screening and its impact on clinical outcomes in randomized trials.

Despite the lingering questions, Sengupta said he’s excited about the possibilities, especially if the AI-ECG is combined with other types of wearable monitors in the future.

“It’s extremely promising,” he said. “It’s an important technological step that has been taken.”

#Accurately #Detects #Dysfunction #SingleLead #Apple #Watch #ECG

Leave a Comment

Your email address will not be published. Required fields are marked *