Related Experiment Video
Updated: May 28, 2026
07:34
Surgery and Sample Processing for Correlative Imaging of the Murine Pulmonary Valve
Published on: August 5, 2021
Patient-specific mitral leaflet segmentation from 4D ultrasound
Robert J Schneider1, Neil A Tenenholtz, Douglas P Perrin
1Harvard School of Engineering and Applied Sciences, Cambridge, MA, USA.
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|October 19, 2011
Summary
This study introduces an automated method for segmenting the mitral valve from four-dimensional ultrasound (4DUS) data. The technique accurately tracks valve geometry throughout the cardiac cycle, improving mechanical modeling and visualization.
Area of Science:
- Biomedical Engineering
- Medical Imaging
- Cardiovascular Research
Background:
- Accurate segmentation of the mitral valve from four-dimensional ultrasound (4DUS) is crucial for developing mechanical models and understanding valve dynamics.
- Existing segmentation methods often demand significant user input, struggle to maintain geometric consistency across the cardiac cycle, or lack detailed coaptation surface generation.
Purpose of the Study:
- To present an automated method for segmenting the mitral valve annulus and leaflets from 4DUS data.
- To enable detailed, patient-specific tracking of mitral valve geometry throughout closure for improved coaptation analysis.
Main Methods:
- The method involves minimal user interaction, requiring only the selection of two frames (start/end of closure) and a single point near the closed valve.
- Initial segmentation of the annulus and leaflets is performed in key frames, followed by geometric tracking to segment the entire cardiac cycle.
- Developed meshes were compared against expert manual tracings for both normal and diseased mitral valves.
Main Results:
- The automated segmentation method successfully tracked patient-specific mitral valve annulus and leaflets throughout closure.
- A detailed coaptation region was generated, crucial for biomechanical analysis.
- The average difference between automated and manual segmentations was 0.59 +/- 0.49 mm, comparable to ultrasound spatial resolution.
Conclusions:
- The presented 4DUS segmentation method offers an efficient and accurate approach for characterizing mitral valve geometry.
- This technique facilitates the creation and validation of patient-specific mechanical models and enhances visualization of mitral valve behavior.
