AUTHOR=Yu Yun , Wu Xi , Chen Jiu , Cheng Gong , Zhang Xin , Wan Cheng , Hu Jie , Miao Shumei , Yin Yuechuchu , Wang Zhongmin , Shan Tao , Jing Shenqi , Wang Wenming , Guo Jianjun , Hu Xinhua , Liu Yun 

TITLE=Characterizing Brain Tumor Regions Using Texture Analysis in Magnetic Resonance Imaging

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 15 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2021.634926

DOI=10.3389/fnins.2021.634926

ISSN=1662-453X

ABSTRACT=Purpose: To extract texture features from Magnetic Resonance Imaging (MRI) images of patients with brain tumors and use them to train a classification model for supporting early diagnosis. 
Methods: Two groups of regions (control and tumor) were selected from MRI images of 40 patients with meningioma or glioma. Texture analysis was performed on these regions to get texture features. Statistical analysis was carried out using SPSS (version 20.0), including Shapiro-Wilks test and Wilcoxon signed-rank test, which were used for testing significant differences in each feature between tumor and healthy regions. T-distributed stochastic neighbor embedding (t-SNE) was used to visualize the data distribution in order to avoid tumor selection bias. Gini impurity index in the Radom Forests (RF) was used to select top five out of all features. Based on the five features, three classification models were built respectively with three machine learning classifiers - RF, Support Vector Machine (SVM), Back Propagation Neural Network (BP).
Results: 16 out of 25 features showed significant differences between tumor and healthy areas. Through the Gini impurity index in the RF, standard deviation, first-order moment, variance, third-order absolute moment and third-order central moment were selected to build classification model. The classification model trained by SVM classifier achieved the best performance, with a sensitivity, specificity and AUC of 94.04%, 92.3%, and 0.932.
Conclusion: Texture analysis combining with an SVM classifier can help differentiate between brain tumor and healthy areas with a high speed and high accuracy, which is then useful for clinical application.