Artificial Intelligence-Powered Measurement of Left Ventricular Ejection Fraction Using a Handheld Ultrasound Device.

ABSTRACT

The aim of this study was to assess the accuracy of an algorithm for automated measurement of left ventricular ejection fraction (LVEF) available on handheld ultrasound devices (HUDs). One hundred twelve patients admitted to the cardiology department underwent assessment performed with an HUD. In each case, the four-chamber apical view was obtained, and LVEF was calculated with LVivo software. Subsequently, during the examination performed with the use of the stationary echocardiograph, the 3-D measurement of LVEF was recorded. The average LVEFs measured with LVivo and the 3-D reference method were 46 ± 14% and 48 ± 14%, respectively. The correlation between the measurements obtained with the HUD and 3-D evaluation was high (r = 0.92, 95% confidence interval 0.87-0.95, p < 0.0001). The mean difference between the LVEF obtained with LVivo and the 3-D LVEF was not significant (mean difference -0.61%, 95% confidence interval -1.89 to 0.68, p = 0.31). The LVivo software despite its limitations is capable of the accurate LVEF measurement when the acquired views are of at least good imaging quality.