Automatic Tablet-Based Monoplane Quantification of Stroke Volume and
Left Ventricular Ejection Fraction: A Comparative Assessment against
Computer-Based Biplane and Monoplane Tools
Abstract
Purpose: Point-of-care cardiovascular left ventricle ejection
fraction (LVEF) quantification is established, but automatic
tablet-based stroke volume (SV) quantification with handheld ultrasound
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 handheld ultrasound scans, and LVEF and SV were quantified
using AutoEF and computer-based software that utilized either apical
four-chamber views (AS-mono) or both apical four-chamber and apical
two-chamber views (AS-bi). Correlation and Bland-Altman analysis were
used to compare AutoEF with AS-mono and AS-bi. Results: Out of
43 participants, 8 were excluded. AutoEF showed a correlation of 0.83
[0.69:0.91] with AS-mono for LVEF and 0.68 [0.44:0.82] for SV.
The correlation with AS-bi was 0.79 [0.62:0.89] for LVEF and 0.66
[0.42:0.81] 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.