Department of Radiology, Renji Hospital, School of Medicine, Shanghai Jiao Tong University, No. 160, Pujian Rd., Shanghai, 200127, China.
MR Collaboration, Central Research Institute, United Imaging Healthcare, Shanghai, China.
Eur Radiol. 2023 Apr;33(4):2871-2880. doi: 10.1007/s00330-022-09212-5. Epub 2022 Nov 8.
The purpose of the study was to explore the performance of a three-component diffusion model in evaluating the degree of malignancy and isocitrate dehydrogenase 1 (IDH-1) gene type of gliomas.
Overall, 60 patients with gliomas were enrolled. The intermediate and perfusion-related diffusion coefficients (D and D) and fractions of strictly limited, intermediate, and perfusion-related diffusion (F, F, and F) were obtained with a three-component diffusion model. Parameters were also obtained from a diffusion kurtosis model and mono- and biexponential models. All parameters were compared between different tumor grades and IDH-1 gene types. Diagnostic performance and logistic regression analyses were performed.
High-grade gliomas (HGGs) had significantly higher F, F, and D values but significantly lower F and D values than low-grade gliomas (LGGs), and F and F differed significantly among grade II, III, and IV gliomas (p < 0.05 for all). F achieved the highest AUC of 0.872 in differentiating between LGGs and HGGs. Logistic regression analysis revealed that in each model, F, diffusion coefficient (D), apparent diffusion coefficient (ADC), mean diffusivity (MD), and mean kurtosis (MK) were associated with glioma grading. After multiple regression analysis, F remained the only differentiator. Additionally, F and F showed significant differences between IDH-1 mutated and IDH-1 wild-type gliomas (p = 0.007 and 0.01, respectively).
The three-component DWI model served as a useful biomarker for detecting microstructural features in gliomas with different grades and IDH-1 mutation statuses.
• The three-component model enables the estimation of an intermediate diffusion component. • The three-component model performed better than the other models in glioma grading and genotyping. • F was useful in evaluating the grade and genotype of gliomas.
本研究旨在探讨三成分扩散模型在评估脑胶质瘤恶性程度和异柠檬酸脱氢酶 1(IDH-1)基因突变类型方面的性能。
共纳入 60 例脑胶质瘤患者。采用三成分扩散模型获得中间和灌注相关扩散系数(D 和 D)以及严格受限、中间和灌注相关扩散分数(F、F 和 F)。还从扩散峰度模型和单指数及双指数模型获得参数。比较不同肿瘤分级和 IDH-1 基因突变类型之间的参数。进行诊断性能和逻辑回归分析。
高级别胶质瘤(HGGs)的 F、F 和 D 值明显较高,而 F 和 D 值明显较低,低级别胶质瘤(LGGs);F 在 II、III 和 IV 级胶质瘤之间存在显著差异(所有 p<0.05)。F 在区分 LGGs 和 HGGs 方面获得了最高的 AUC(0.872)。逻辑回归分析显示,在每个模型中,F、扩散系数(D)、表观扩散系数(ADC)、平均扩散率(MD)和平均峰度(MK)与胶质瘤分级相关。经过多元回归分析,F 仍然是唯一的鉴别因素。此外,F 和 F 在 IDH-1 突变型和 IDH-1 野生型胶质瘤之间存在显著差异(p=0.007 和 0.01)。
三成分 DWI 模型是一种有用的生物标志物,可用于检测不同分级和 IDH-1 突变状态的脑胶质瘤的微观结构特征。
• 三成分模型可用于估计中间扩散分量。• 三成分模型在脑胶质瘤分级和基因分型方面优于其他模型。• F 可用于评估脑胶质瘤的分级和基因型。