Related Experiment Video
Updated: Jun 8, 2026
04:25
Manual Segmentation of the Human Choroid Plexus Using Brain MRI
Published on: December 15, 2023
Automatic segmentation and components classification of optic pathway gliomas in MRI
Lior Weizman1, Liat Ben-Sira, Leo Joskowicz
1School of Eng. and Computer Science, Hebrew University of Jerusalem, Israel. lweizm45@cs.huji.ac.il
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|October 1, 2010
Summary
This study introduces an automated method for segmenting and classifying brain Optic Pathway Gliomas (OPGs) using MRI data. The approach accurately delineates OPGs and their components, aiding in disease progression assessment.
Area of Science:
- Medical Imaging
- Neuro-oncology
- Artificial Intelligence
Background:
- Optic Pathway Gliomas (OPGs) are challenging to segment and classify accurately using standard MRI.
- Manual segmentation is time-consuming and subject to inter-observer variability.
- Accurate OPG characterization is crucial for evaluating disease progression and treatment efficacy.
Purpose of the Study:
- To develop and validate an automated method for segmenting and classifying brain OPGs from multi-spectral MRI.
- To incorporate prior anatomical and intensity information for improved delineation.
- To classify segmented OPGs into solid, enhancing, and cystic components.
Main Methods:
- A novel automatic segmentation technique incorporating prior location, shape, and intensity information.
- Probabilistic tumor tissue modeling using training datasets to account for grey-level variations.
- Classification of segmented OPG volumes into three main components.
Main Results:
- Mean OPG boundary surface distance error of 0.73mm.
- Mean volume overlap difference of 30.6% compared to expert manual segmentation.
- High correlation between automated component classification and clinical tumor progression evaluation.
Conclusions:
- The developed method provides accurate automatic segmentation and component classification of OPGs.
- This approach supports quantitative evaluation of disease progression and treatment efficacy.
- This is the first method to offer combined automatic OPG segmentation and component classification.
