Ecg AI Algorithm Can Identify Left Ventricular Systolic Dysfunction

Left ventricular (LV) systolic dysfunction was found in patients who presented to the emergency room (ED) with dyspnea using AI-analyzed electrocardiographs.

Lead investigator Demilade Adedinsewo, MD, of the Department of Cardiovascular Medicine at Mayo Clinic in Jacksonville, Florida, told Healio: "AI ECG can detect left ventricular systolic function in patients with tachypnea faster and more accurately than NT-proBNP. It may improve and speed up diagnosis in the emergency department and provide a unique opportunity to identify patients at high risk for heart disease earlier and link patients to appropriate cardiovascular care."

Patients with breathing difficulties

In the retrospective study, published in Circulation: Arrhythmias and Electrophysiology, researchers analyzed data from 1,606 patients (median age, 68; 47% of women) had difficulty breathing between May 2018 and February 2019. These patients had at least one ECG within 24 hours and 30 days of their ED presentation. Patients with a previous diagnosis of systolic, diastolic, or unexplained heart failure were excluded.

The primary outcome of this study was the discovery of new LV systolic dysfunction (defined as a left ventricular ejection fraction of 35% or less) in patients within 30 days after ED presentation. Secondary outcomes were defined as patients whose left ventricular ejection fraction (LVEF) was found to be less than 50% within 30 days of presentation. Both results are determined by the ECG evaluated by the deep Learning Network, an AI-ECG algorithm that has been developed and validated to identify LVEFs of 35% or lower without additional optimization or training.

The median time to ECG examination after ED presentation was 1 day.

In patients with dyspnea in the emergency department, the area under the receiver operating characteristic curve (AUC) of AI-ECG algorithm for identifying new left ventricular systolic dysfunction was 0.89 (95% CI, 0.86-0.91). The accuracy of the algorithm was 85.9% (95%CI, 841-87.6), the specificity was 87%, the sensitivity was 74%, the positive predictive value was 40%, and the negative predictive value was 97%.

The algorithm was also able to identify patients with LVEF lower than 50%, with an area under the receiver operating characteristic curve of 0.85(95%CI, 0.83-0.88) and an accuracy of 86%(95%CI, 842-87.7).

The algorithm was also able to identify patients with LVEF lower than 50%, with an area under the receiver operating characteristic curve of 0.85(95%CI, 0.83-0.88) and an accuracy of 86%(95%CI, 842-87.7). This also achieved 91% specificity, 63% sensitivity, 62% positive predictive value and 92% negative predictive value.

The researchers also evaluated a panel of 866 patients with available N-terminal B-type natriuretic peptide values. A level of NT-proBNP greater than 800 pg/mL indicated a new LV systolic dysfunction, with an area under the receiver operating characteristic curve of 0.8(95% CI, 0.76-0.84).

"The current study was retrospective," Adedinsewo said in an interview. "Prospective studies are needed to assess the impact of AI-ECG on long-term clinical outcomes, which is currently being evaluated by our research team."

Adedinsewo added that the technology is currently being used throughout her healthcare system. "The AI-ECG tool is currently available at all Mayo Clinic sites and is accessible through our electronic medical record system," she told Healio. "In addition, the tool was recently granted emergency use authorization through the FDA in May to screen patients with confirmed or suspected COVID-19 for left ventricular dysfunction."

The potential to advance patient care

In an accompanying editorial, Dr. KaziT.Haq of the Knight Cardiovascular Institute at Oregon Health and Science University in Portland, Oregon, and colleagues write: "Overall, the findings of Adedinsewo et al. suggest that AI using a standard 12-lead ECG improves the recognition rate of new heart failure in patients with dyspnea in the emergency department. This is a strategy that is easy to use in clinical practice and has the potential to significantly improve patient care."