Related Experiment Video
Updated: Mar 28, 2026
04:25
Author Spotlight: Bridging Gaps in Anatomy and Establishing a Foundation for Algorithmic Studies
Published on: December 15, 2023
4.1K
The effect of morphometric atlas selection on multi-atlas-based automatic brachial plexus segmentation.
Joris Van de Velde1,2, Johan Wouters3, Tom Vercauteren4
1Department of Anatomy, Ghent University, De Pintelaan 185, 9000, Ghent, Belgium. Joris.vandevelde@ugent.be.
Radiation Oncology (London, England)
|December 24, 2015
Summary
Selecting patient-specific atlases significantly improves autosegmentation accuracy for the brachial plexus (BP). Using six selected atlases yielded the highest accuracy, suggesting this strategy enhances medical imaging segmentation.
Area of Science:
- Medical Imaging
- Radiology
- Anatomy
Background:
- Accurate segmentation of the brachial plexus (BP) is crucial for medical procedures.
- Multi-atlas-based autosegmentation offers a promising approach for BP segmentation.
- The selection strategy of atlases can significantly impact autosegmentation accuracy.
Purpose of the Study:
- To evaluate the impact of a morphometric atlas selection strategy on BP autosegmentation accuracy.
- To determine the optimal number of atlases for accurate multi-atlas-based BP autosegmentation.
- To compare autosegmentation accuracy using morphometrically selected atlases versus randomly selected atlases.
Main Methods:
- Utilized twelve cadaver computed tomography (CT) atlases within the ADMIRE® software.
- Compared autosegmentation using morphometrically preselected atlases (based on shoulder protraction) with randomly selected atlases.
- Calculated Dice similarity coefficient (DSC), Jaccard index (JI), and Inclusion index (INI) to measure segmentation accuracy.
Main Results:
- Morphometrically selected atlases resulted in significantly higher similarity indices compared to random selection (p < 0.05).
- The highest accuracy was achieved with six selected atlases, yielding average DSC of 0.598, JI of 0.434, and INI of 0.733.
- The study demonstrated a clear advantage of strategic atlas selection over random selection for BP autosegmentation.
Conclusions:
- Morphometric atlas selection, based on patient's shoulder protraction, significantly enhances multi-atlas BP autosegmentation accuracy.
- Six selected atlases appear to be optimal for improving autosegmentation accuracy in this study.
- Further research with a larger atlas set is recommended to confirm optimal atlas numbers and clinical applicability.
