Abstract
BACKGROUND: The focus of neuro-oncology research has changed from histopathologic grading to molecular characteristics, and medical imaging routinely follows this change. This study aimed to demonstrate the feasibility of using continuous-time random walk (CTRW) and fractional order calculus (FROC) models, together with histogram analysis, in identifying the states of molecular biomarkers of diffuse gliomas in adults. METHODS: A total of 111 diffuse glioma patients undergoing multi-b-value diffusion-weighted imaging (DWI) were included. The histogram parameters of CTRW, FROC, and mono-exponential models were compared between diffuse gliomas with different molecular states [isocitrate dehydrogenase 1 (IDH1), X-linked alpha-thalassemia/mental retardation syndrome (ATRX), O6-methylguanine-DNA methyltransferase (MGMT), and telomere reverse transcriptase (TERT)] using independent samples t-test or Mann-Whitney U test. The diagnostic performance of each diffusion parameter was evaluated using receiver operating characteristic (ROC) curve. RESULTS: Statistically significant differences (P<0.05) were found between IDH1-mutant and wildtype gliomas in all diffusion parameters except for the kurtosis D(CTRW), 90(th) percentile α(CTRW), kurtosis β(FROC), 90(th) percentile µ(FROC), and kurtosis apparent diffusion coefficient (ADC). Moreover, the areas under the curve (AUCs) of the 10(th) percentile β(CTRW), as well as the 10(th) percentile, mean, and median β(FROC) were significantly higher than all ADC histogram parameters following the DeLong test (P<0.05) in IDH1 genotyping. The 90th percentile ADC (AUC =0.797) provided the highest diagnostic efficiency among individual parameters in ATRX genotyping of IDH1-mutant gliomas. The median β(CTRW) (AUC =0.758) and 10(th) percentile β(CTRW) (AUC =0.869) provided the highest differential efficiency for MGMT and TERT, respectively. CONCLUSIONS: The CTRW and FROC models demonstrate good diagnostic performance in predicting different molecular subtypes in diffuse gliomas, and provide new imaging biomarkers for probing tumor structural heterogeneity at a subvoxel level.