Feasibility and Accuracy of Assessing Left Ventricular Systolic Function by Measuring Mitral Annular Excursion using a Hand-Held Echocardiography Device with an Automated Atrioventricular-Plane Tracking Algorithm

Abstract

Background: Hand-carried ultrasound (HCU) and automated echocardiographic (echo) software may allow point-of-care, focussed assessment of left ventricular systolic function (LVSF). This could potentially allow for assessment and monitoring by non-specialised staff to reduce the global burden of heart failure. Aim: Determine accuracy and feasibility of assessing LVSF via mitral annular excursion (MAE) using HCU with integrated first-in-man, automated atrio-ventricular (AV) plane tracking software (AutoMAE). Methods: Forty-five consecutive patients in sinus rhythm undergoing clinical echo (including MAE via Motion-Mode ‘M-MAE’, and left ventricular ejection fraction (LVEF) measurements; Vivid E9), also had AutoMAE performed (AV plane App on Vscan ExtendTM) within 24 hours. AutoMAE (mean from septal and lateral annular measures from an apical four chamber view) was compared to M-MAE using mean difference (absolute and relative) and Bland-Altman, as well as to LVEF using receiver operating characteristic (ROC; ≤55% LVEF). AutoMAE reproducibility was also assessed. Results: The 45 patients (56% male) were 45 ± 14 yrs with LVEF 59 ± 7%. AutoMAE had 91% feasibility. Intra (n = 39) and inter-observer (n = 11) comparisons: 0.9 ± 0.8 mm; 1.0 ± 1.0 mm absolute and 9 ± 7%; 10 ± 9% relative mean differences respectively. AutoMAE was comparable to M-MAE, although with systematic underestimation: 3.0 ± 1.7 mm absolute and 25 ± 14% relative mean differences; Bland-Altman −2.6 (−7.0, +1.8) mm. AutoMAE ROC found an AUC of 0.77 at ≤14.1 mm (Sensitivity 80%; Specificity 68%). Conclusion: MAE measurement using automated software is highly feasible. While there was consistent underestimation of AutoMAE compared with M-MAE, ROC analysis suggests this may allow reliable detection of a reduced LVEF. Further investigations are required into AutoMAE accuracy when used by non-specialised staff.