Automatic tablet-based monoplane quantification of stroke volume and left ventricular ejection fraction: A comparative assessment against computer-based biplane and monoplane tools

Background

Point-of-care cardiovascular left ventricle ejection fraction (LVEF) quantification is established, but automatic tablet-based stroke volume (SV) quantification with handheld ultrasound (HAND) devices is unexplored. We evaluated a tablet-based monoplane LVEF and LV volume quantification tool (AutoEF) against a computer-based tool (Tomtec) for LVEF and SV quantification.

Methods

Patients underwent HAND scans, and LVEF and SV were quantified using AutoEF and computer-based software that utilized either apical four-chamber views (Auto Strain-monoplane [AS-mono]) or both apical four-chamber and apical two-chamber views (Auto Strain-biplane [AS-bi]). Correlation and Bland–Altman analysis were used to compare AutoEF with AS-mono and AS-bi.

Results

Out of 43 participants, eight were excluded. AutoEF showed a correlation of .83 [.69:.91] with AS-mono for LVEF and .68 [.44:.82] for SV. The correlation with AS-bi was .79 [.62:.89] for LVEF and .66 [.42:.81] for SV. The bias between AutoEF and AS-mono was 4.88% [3.15:6.61] for LVEF and 17.46 mL [12.99:21.92] for SV. The limits of agreement (LOA) were [−5.50:15.26]% for LVEF and [−8.02:42.94] mL for SV. The bias between AutoEF and AS-bi was 6.63% [5.31:7.94] for LVEF and 20.62 mL [16.18:25.05] for SV, with LOA of [−1.20:14.47]% for LVEF and [−4.71:45.94] mL for SV.

Conclusion

LVEF quantification with AutoEF software was accurate and reliable, but SV quantification showed limitations, indicating non-interchangeability with neither AS-mono nor AS-bi. Further refinement of AutoEF is needed for reliable SV quantification at the point of care.