View the full publication here
Background:
Recent studies have identified that many cardiovascular diseases, cardiovascular risk factors, elevated cardiac biomarkers signaling acute myocardial injury, and other cardiac manifestations to be associated with worse prognosis in patients with COVID-19.
Transthoracic echocardiography (TTE) remains the first-line cardiac imaging modality for cardiac assessment in COVID-19 infection.
Artificial intelligence (AI) techniques have the potential for rapid and streamlined analysis of point-of-care and complete TTE examinations such as inpatients with COVID-19 infection, for which, in the current ongoing pandemic, there remains unmet clinical need. It may also facilitate risk stratification for these patients.
The aim of the study was to compare Ultromics EchoGo automated AI analytic output with standard TTE measurements of the left ventricle and assess their prognostic utility for adverse in-hospital outcomes..
50 patients with Covid-19 who underwent echocardiography were analyzed with EchoGo software, with output correlated against standard echocardiography measurements. The Acute Physiology and Chronic Health Evaluation IV (APACHE-IV) score was retrospectively calculated.
Measurements, using the apical 4-chamber and 2-chamber views, included: left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), left ventricular ejection fraction (LVEF), left ventricular global longitudinal strain (LVGLS).
Standard measurements of these left ventricle parameters were also performed using the Simpson’s biplane technique for LVEF and velocity-vector imaging technique for LVGLS.
Primary endpoint - composite of in-hospital all-cause death, acute myocardial injury, and need for mechanical ventilation.
In comparison of left ventricular measurements, there was a high correlation for LVEDV (r = 0.81) and LVESV (r = 0.87), and moderate correlation for LVEF (r = 0.53) and LVGLS (r = 0.69), all p-value < 0.001.
Adjusted for the APACHE-IV score, EchoGo LVEF and LVGLS were independently associated with with the primary endpoint, odds ratios (95% confidence intervals) of 0.92 (0.85–0.99) and 1.22 (1.03–1.45) per 1% increase, while the other EchoGo and standard TTE measurements were not.
Ultromics EchoGo LVEF and LVGLS measurements were associated with the primary endpoint, including when adjusted for the APACHE-IV score, where standard TTE measurements were not.
Ultromics EchoGo output had high correlations with left ventricular volumes and moderate correlations with LVEF and LVGLS compared with standard TTE measurements.
EchoGo demonstrated that it can play a role in the risk stratification of hospitalized COVID-19 patients.