Related Experiment Video
Updated: Jul 31, 2026
09:17
Digital Spatial Profiling for Characterization of the Microenvironment in Adult-Type Diffusely Infiltrating Glioma
Published on: September 13, 2022
Nuclear spatial and spectral features based evolutionary method for meningioma subtypes classification in
Kiran Fatima1, Hammad Majeed1, Humayun Irshad2
1Department of Computer Science, National University of Computer and Emerging Sciences, A. K. Brohi Road, H-11/4, Islamabad, Pakistan.
Microscopy Research and Technique
|April 6, 2017
Summary
This study presents an optimized computational framework for classifying benign meningioma subtypes. The method achieved 94.88% accuracy, aiding neuropathologists in diagnosing these brain tumor lesions.
Area of Science:
- Neuropathology
- Computational Pathology
- Bioinformatics
Background:
- Meningioma subtype classification is complex due to high intra-class variability and low inter-class variation.
- Accurate classification is crucial for diagnosis and prognosis.
- Computational tools can assist pathologists in meningioma characterization.
Purpose of the Study:
- To develop an optimized evolutionary framework for classifying benign meningioma into four subtypes.
- To investigate the role of RGB color channels and various phenotypes in tumor discrimination.
- To improve the accuracy of meningioma subtype classification using computational methods.
Main Methods:
- Proposed an optimized evolutionary framework using Genetic Algorithm (GA) and Support Vector Machine (SVM).
- Extracted structural, statistical, and spectral phenotypes from tissue samples.
- Investigated the utility of RGB color channels for feature discrimination.
- Tuned classifier parameters and selected optimal phenotype combinations.
Main Results:
- Achieved a classification accuracy of 94.88% on a meningioma histology dataset.
- Demonstrated the framework's robustness in discriminating four benign meningioma subtypes.
- Highlighted the importance of selected phenotypes and RGB channels for accurate classification.
Conclusions:
- The developed computational framework effectively classifies benign meningioma subtypes.
- This approach shows potential to aid neuropathologists in diagnosis and classification of meningioma lesions.
- Optimized evolutionary algorithms combined with phenotype analysis offer a promising avenue for computational neuropathology.
