AI-powered ECGs detect subtle changes in the heart brought on by COVID-19
AI-powered electrocardiograms (ECGs) can be trained to detect subtle, subclinical electrical changes in the heart associated with COVID-19, according to a new study published in Mayo Clinic Proceedings. This could provide clinicians with a fast, affordable screening tool that boosts patient care.
Researchers trained their advanced algorithm using clinical data and ECG results from more than 26,000 patients from four different continents. Nearly 8,000 more ECGs were then used to test the AI model’s accuracy. Nearly one-third of the training and test cohorts had a confirmed COVID-19 diagnosis.
Overall, the authors wrote, their AI-powered ECG achieved an area under the curve (AUC) for detecting acute COVID-19 of 0.767. Its sensitivity was 98%, specificity was 10%, positive predictive value was 37% and negative predictive value was 91%. When adding data from more than 50,000 additional healthy control patients, the AI model’s AUC, specificity and negative predictive value improved to 0.780, 12.1% and 99.2%, respectively.
“Accuracy is one of the biggest hurdles in determining the value of any test for COVID-19,” Zachi Attia, PhD, an engineer in Mayo Clinic’s department of cardiovascular medicine, said in a statement. “Not only do we need to know the sensitivity and specificity of the test, but also the prevalence of the disease. Adding the extra control ECG data was critical to demonstrating how a variable prevalence of the disease—as we have encountered with regions having widely different rates of disease at different stages of the pandemic disease—would impact how the test would perform.”
“This study demonstrates the presence of a biological signal in the EKG consistent with COVID-19 infection, but it included many ill patients,” added senior author Paul Friedman, MD, chair of Mayo Clinic’s department of cardiovascular medicine. “While it is a hopeful signal, we must prospectively test this in asymptomatic people using smartphone-based electrodes to confirm that it can be practically used in the fight against the pandemic. Studies are underway now to address that question.”
Click here for a PDF of the team’s full analysis.
Michael has more than 18 years of experience as a professional writer and editor. He has written at length about cardiology, radiology, artificial intelligence and other key healthcare topics.