Related Experiment Video
Updated: May 20, 2026
08:13
In Vivo Quantitative Assessment of Myocardial Structure, Function, Perfusion and Viability Using Cardiac Micro-computed Tomography
Published on: February 16, 2016
Quantification of left ventricular volume and global function using a fast automated segmentation tool: validation in
Daniel Barbosa1, Brecht Heyde, Thomas Dietenbeck
1Laboratory on Cardiovascular Imaging & Dynamics, Department of Cardiovascular Sciences, KU Leuven, Herestraat 49, bus 7003, 3000, Leuven, Belgium. daniel.barbosa@uzleuven.be
The International Journal of Cardiovascular Imaging
|August 2, 2012
Summary
A new automated segmentation framework (BEAS) accurately quantifies left ventricular (LV) volumes and function from real-time 3D echocardiography (RT3DE). This tool significantly reduces analysis time, paving the way for routine clinical use of RT3DE.
Area of Science:
- Cardiovascular Imaging
- Medical Imaging Analysis
- Echocardiography
Background:
- Real-time 3D echocardiography (RT3DE) offers accurate left ventricular (LV) volume assessment.
- Manual LV border detection in RT3DE is time-consuming, hindering routine clinical application.
- Automated segmentation is needed to streamline RT3DE analysis.
Purpose of the Study:
- To evaluate the accuracy of a novel 3D automated segmentation framework (BEAS) for LV volumetric and functional parameter extraction in a clinical setting.
- To compare BEAS-derived parameters against expert manual measurements.
- To assess the time efficiency of the BEAS framework.
Main Methods:
- BEAS, a 3D automated segmentation framework, was developed for real-time LV morphology capture.
- 24 RT3DE exams (5 healthy volunteers, 19 patients) were analyzed.
- LV end-diastolic and end-systolic volumes (EDV, ESV) were manually contoured by 3 experts; stroke volume (SV) and ejection fraction (EF) were calculated.
- BEAS results were compared to the average expert measurements using correlation and Bland-Altman statistics.
Main Results:
- Strong correlations were found between BEAS and expert values for EDV (R=0.963), ESV (R=0.947), SV (R=0.944), and EF (R=0.853).
- Bland-Altman analysis showed minimal bias for EDV (2.59 ml), ESV (-2.11 ml), SV (4.70 ml), and EF (3.45%).
- Total analysis time using BEAS was significantly reduced to 30.7 ± 7.5 seconds.
Conclusions:
- BEAS provides fast and accurate quantification of 3D cardiac volumes and global function with minimal user interaction.
- The BEAS framework demonstrates potential for integrating 3D echocardiography into routine clinical practice.
- Automated segmentation significantly improves the efficiency of RT3DE analysis.
