Related Experiment Video
Updated: Dec 15, 2025
06:34
Evaluation of Left Ventricular Structure and Function using 3D Echocardiography
Published on: October 28, 2020
4.4K
A Statistical Shape Model Approach for Computing Left Ventricle Volume and Ejection Fraction Using Multi-plane
Dawei Liu1, Isabelle Peck2, Shusil Dangi1
1Rochester Institute of Technology, 1 Lomb Memorial Drive, Rochester, NY 14623, USA.
Vipimage 2019 : Proceedings of the VII ECCOMAS Thematic Conference on Computational Vision and Medical Image Processing, October 16-18, 2019, Porto, Portugal. Vipimage (Conference) (2019 : Porto, Portugal)
|July 15, 2020
Summary
This study introduces a statistical shape model for estimating left ventricular ejection fraction (LVEF) from 2D ultrasound images, improving accuracy without 3D systems. The novel method accurately calculates LVEF, aiding cardiologists in clinical assessments.
Area of Science:
- Cardiology
- Medical Imaging
- Biomedical Engineering
Background:
- Accurate left ventricular ejection fraction (LVEF) assessment typically requires 3D echocardiography, which is not always accessible or preferred by clinicians.
- Current reliance on 2D ultrasound (US) often leads to visual estimation of LVEF, introducing variability.
- There is a need for methods that enable consistent LVEF estimation from readily available 2D US data, mimicking 3D volumetric accuracy.
Purpose of the Study:
- To develop and validate a statistical shape model (SSM) for estimating left ventricular (LV) volumes and LVEF from 2D US images.
- To enable accurate LVEF calculation without requiring specialized 3D ultrasound systems or extensive manual input.
- To provide a tool for consistent and reliable LVEF assessment in clinical cardiology practice.
Main Methods:
- A statistical shape model (SSM) was constructed using 13 key anchor points from LV endocardial contours in tri-plane 2D US images.
- Principal component analysis (PCA) was employed to capture LV shape variations, enabling new LV shapes to be represented as linear combinations of principal components.
- New patient LV shapes were compared to the SSM using Mahalanobis and PCA distances to determine weights for estimating LV volumes and LVEF.
Main Results:
- The SSM-based method successfully estimated LV diastolic, systolic, and stroke volumes, along with LVEF.
- LVEF estimates derived using Mahalanobis distance showed a mean difference of 6.8% compared to reference values.
- LVEF estimates derived using PCA distance demonstrated high accuracy, with a mean difference of only 1.7% compared to reference values from a clinical platform.
Conclusions:
- The proposed statistical shape model effectively estimates LV volumes and LVEF from 2D ultrasound images.
- The method offers a viable alternative for accurate LVEF assessment when 3D echocardiography is unavailable.
- This approach facilitates consistent and precise LVEF calculations, supporting improved clinical decision-making in cardiology.
