Automated Segmentation and Volumetry for Meningioma Using Deep Learning
Study Details
Study Description
Brief Summary
U-Net-based architectures will be applied to 500 contrast-enhanced axial MR images of different patients from a single institution after manual segmentation of meningioma, of which 50 were used for testing. Tumor volumetry after autosegmentation by trained U-Net-based architecture is final goal.
Condition or Disease | Intervention/Treatment | Phase |
---|---|---|
|
Detailed Description
U-Net-based architectures will be applied to 500 contrast-enhanced axial MR images of different patients from a single institution after manual segmentation of meningioma, of which 50 were used for testing. After preprocessing with Z-isotropification and intensity normalization of images, 3 U-Net-based networks (2D U-Net, Attention U-Net, 3D U-Net) and 3 nnU-Net-based networks (2D nnU-Net, Attention nnU-Net, 3D nnU-Net) will be trained with meningioma-segmented images. For applying to 3D networks, sagittal and coronal images will be reconstructed using axial images. After prediction, the cut-off of the probability function, which is a trade-off, will be obtained with the Gaussian Mixture Modeling algorithm using the probability density function. The voxels having a probability function higher than that will be finally predicted as meningioma. Tumor volume is calculated as the sum of the product of segmented area and thickness of axial images. For performance evaluation, dice similarity coefficient (DSC), precision, and recall will be evaluated compared with manually segmented voxels for validation datasets. The results of volumetry of each model will be compared with manual segmentation-based volume through Pearson's correlation analysis.
Study Design
Arms and Interventions
Arm | Intervention/Treatment |
---|---|
Meningioma patients
|
Other: Observation
This study does not involve any intervention to subjects.
|
Outcome Measures
Primary Outcome Measures
- Accuracy compared with ground truth [10-01-2020 until 09-30-2021]
As a primary endpoint, we will examine the ability of U-Net and nnU-Net to segment meningioma in brain MR compared with ground truth. Ground truth is defined as area on MR drawn by two neurosurgeons. Accuracy of autosegmentation of meningioma will be assessed in dice similarity coefficient, recall, and precision.
Eligibility Criteria
Criteria
Inclusion Criteria:
- Radiologically diagnosed meningioma by MRI
Exclusion Criteria:
-
under 18 years old
-
Multiple meningiomas
-
Orbital meningioma
-
Any prior treatment for intracranial meningioma before registration
Contacts and Locations
Locations
No locations specified.Sponsors and Collaborators
- Seoul National University Hospital
Investigators
None specified.Study Documents (Full-Text)
None provided.More Information
Publications
None provided.- SNUH-MNG-AI001