Visual versus fully automated assessment of left ventricular ejection fraction

Area of Science:

• Cardiology
• Medical Imaging
• Echocardiography

Background:

• Accurate assessment of left ventricle ejection fraction (LVEF) is crucial in cardiovascular diagnostics.
• Various echocardiographic methods exist for LVEF estimation, each with potential limitations.

Purpose of the Study:

• To compare the accuracy and reliability of three common echocardiographic techniques for LVEF assessment.
• To evaluate M-mode, automated EF (Auto-EF), and visual estimation methods.

Main Methods:

• A cohort of 268 patients underwent comprehensive echocardiography.
• Left ventricle ejection fraction was measured using M-mode, Auto-EF, and visual estimation by two independent readers.
• Bland-Altman analysis and inter-observer variability were used for statistical comparison.

Main Results:

• Automated EF was feasible in 89.5% of patients.
• Visual estimation (52% ± 12) and Auto-EF (51% ± 11) showed no significant difference (P = 0.054).
• M-mode significantly overestimated LVEF compared to both visual estimation (P < 0.0001) and Auto-EF (P < 0.0001).
• Visual EF assessment demonstrated excellent inter-observer agreement (ICC = 0.953, R = 0.911).

Conclusions:

• Two-dimensional echocardiographic methods, including Biplane Auto-EF and visual assessment, provide comparable LVEF measurements.
• M-mode echocardiography consistently overestimates LVEF and should be used with caution.
• Accurate LVEF determination relies on appropriate selection of echocardiographic techniques.