Glioma: Application of Whole-Tumor Texture Analysis of Diffusion-Weighted Imaging for the Evaluation of Tumor Heterogeneity

Abstract

Background and Purpose: To apply a texture analysis of apparent diffusion coefficient (ADC) maps to evaluate glioma heterogeneity, which was correlated with tumor grade. Materials and Methods: Forty patients with glioma (WHO grade II (n = 8), grade III (n = 10) and grade IV (n = 22)) underwent diffusion-weighted imaging (DWI), and the corresponding ADC maps were obtained. Regions of interest containing the lesions were drawn on every section of the ADC map containing the tumor, and volume-based data of the entire tumor were constructed. Texture and first order features including entropy, skewness and kurtosis were derived from the ADC map using in-house software. A histogram analysis of the ADC map was also performed. The texture and histogram parameters were compared between low-grade and high-grade gliomas using an unpaired student’s t-test. Additionally, a one-way analysis of variance analysis with a post-hoc test was performed to compare the parameters of each grade. Results: Entropy was observed to be significantly higher in high-grade gliomas than low-grade tumors (6.86160.539 vs. 6.26160.412, P = 0.006). The fifth percentiles of the ADC cumulative histogram also showed a significant difference between high and low grade gliomas (8366235 vs. 10306185, P=0.037). Only entropy proved to be significantly different between grades III and IV (6.29560.4963 vs. 7.11960.3165, P,0.001). The diagnostic accuracy of ADC entropy was significantly higher than that of the fifth percentile of the ADC histogram (P = 0.0034) in distinguishing high- from low-grade glioma. Conclusion: A texture analysis of the ADC map based on the entire tumor volume can be useful for evaluating glioma grade, which provides tumor heterogeneity.