Related Experiment Video
Updated: Dec 21, 2025
05:45
Quantitative Immunohistochemistry of the Cellular Microenvironment in Patient Glioblastoma Resections
Published on: July 31, 2017
10.0K
A simple model for glioma grading based on texture analysis applied to conventional brain MRI
José Gerardo Suárez-García1, Javier Miguel Hernández-López1, Eduardo Moreno-Barbosa1
1Faculty of Physics and Mathematics, Benemérita Universidad Autónoma de Puebla (BUAP), Puebla, Puebla, México.
Plos One
|May 16, 2020
Summary
This study developed a simple, low-cost MRI texture analysis model to accurately classify low-grade gliomas (LGGs) from high-grade gliomas (HGGs). The best model achieved over 91% accuracy, offering accessible glioma grading for improved patient diagnosis and prognosis.
Area of Science:
- Radiology
- Medical Imaging Analysis
- Oncology
Background:
- Accurate glioma grading is crucial for patient diagnosis, treatment planning, and prognosis.
- Distinguishing between low-grade gliomas (LGGs) and high-grade gliomas (HGGs) is essential for appropriate clinical management.
Purpose of the Study:
- To develop a low-cost, easy-to-implement classification model for differentiating LGGs from HGGs using conventional brain MRI.
- To investigate texture analysis of the necrotic and non-enhancing tumor core (NCR/NET) region on T1Gd and T2 MRI contrasts.
Main Methods:
- Texture features were extracted from the gray level size zone matrix (GLSZM).
- An under-sampling method was employed to create distinct classification models from training subsets.
- The sensitivity, specificity, and accuracy of various models were evaluated.
Main Results:
- The best classification model utilized only three texture features and achieved 91.18% accuracy, with 94.12% sensitivity and 88.24% specificity.
- HGGs exhibited more heterogeneous textures than LGGs in T1Gd images, while LGGs showed greater heterogeneity than HGGs in T2 images within the NCR/NET region.
- Glioma heterogeneity was found to be dependent on the MRI contrast used.
Conclusions:
- The developed texture analysis model is a simple, reproducible, and highly accurate classifier for gliomas.
- This accessible model can aid in glioma grading, particularly in resource-limited settings.
- Novel findings on glioma texture heterogeneity based on MRI contrast offer new insights into glioma characterization.
