Related Experiment Video
Updated: May 16, 2026
05:33
Three-Dimensional Shape Modeling and Analysis of Brain Structures
Published on: November 14, 2019
Automated detection, 3D segmentation and analysis of high resolution spine MR images using statistical shape models
A Neubert1, J Fripp, C Engstrom
1The Australian E-Health Research Centre, CSIRO ICT Centre, Brisbane, Australia. ales.neubert@uqconnect.edu.au
Physics in Medicine and Biology
|December 4, 2012
Summary
This study introduces an automated method for segmenting spinal structures in MRI scans, enabling early detection of disc degeneration. The novel approach accurately identifies intervertebral discs and vertebral bodies, improving diagnostic capabilities for common spine disorders.
Area of Science:
- Radiology and Medical Imaging
- Biomedical Engineering
- Computational Anatomy
Background:
- High-resolution magnetic resonance (MR) imaging offers detailed anatomical insights into the spine.
- Automated assessment of intervertebral disc (IVD) and vertebral body (VB) anatomy is crucial for early detection of spinal disorders like disc degeneration.
- Existing segmentation techniques for 3D MR images have limitations, particularly for IVDs.
Purpose of the Study:
- To develop and validate an automated approach for 3D segmentation of lumbar and thoracic IVDs and VBs from MR images.
- To assess the accuracy of the automated segmentation against manual segmentations and expert radiological findings.
- To evaluate the potential of the segmented shape parameters for classifying early disc degeneration.
Main Methods:
- Utilized statistical shape analysis and registration of grey level intensity profiles for automated 3D segmentation.
- Validated the algorithm on volumetric 3T MR scans (3D T2-weighted SPACE sequence) of asymptomatic volunteers.
- Compared automated segmentations with manual segmentations and expert radiological assessments of disc degeneration.
Main Results:
- Achieved high agreement between automated and manual segmentations for IVD and VB volumes (Dice scores of 0.89 ± 0.04 and 0.91 ± 0.02, respectively).
- Demonstrated favorable comparison with existing 3D MR segmentation techniques for VBs.
- Successfully performed the first-time automatic segmentation of IVDs from 3D volumetric scans.
- Utilized shape parameters to accurately classify disc abnormalities associated with early degenerative changes (100% sensitivity, 98.3% specificity).
Conclusions:
- The proposed automated method provides accurate 3D segmentation of IVDs and VBs from high-resolution MR images.
- This technique represents a significant advancement, particularly in the automated segmentation of IVDs.
- The method shows high potential for the early detection and monitoring of spinal disorders, especially disc degeneration.
