Multiple diffusion metrics in differentiating solid glioma from brain inflammation

• DOI: 10.3389/fnins.2023.1320296
• Publication Date: 2024-01-30T04:14:22Z
• Authors: Kai Zhao; Ankang Gao; Eryuan Gao; Jinbo Qi; Ting Chen; Guohua Zhao; Gaoyang Zhao; Peipei Wang; Weijian Wang; Jie Bai; Yong Zhang; Huiting Zhang; Guang Yang; Xiaoyue Ma; Jingliang Cheng
• Keywords: Kurtosis

Background and purpose The differential diagnosis between solid glioma brain inflammation is necessary but sometimes difficult. We assessed the effectiveness of multiple diffusion metrics diffusion-weighted imaging (DWI) in differentiating from compared diagnostic performance different DWI models. Materials methods Participants diagnosed with either or a lesion on MRI were enrolled this prospective study May 2016 to April 2023. Diffusion-weighted was performed using spin-echo echo-planar sequence five b values (500, 1,000, 1,500, 2000, 2,500 s/mm 2 ) 30 directions for each value, one value 0 included. mean based tensor (DTI), kurtosis (DKI), apparent propagator (MAP), neurite orientation dispersion density (NODDI) abnormal signal area calculated. Comparisons performed. under curve (AUC) receiver operating characteristic (ROC) Results 57 patients (39 18 inflammation) finally MAP model, its metric non-Gaussianity (NG), shows greatest (AUC = 0.879) differentiation atypical manifestation. AUC DKI (MK) are comparable NG 0.855), followed by NODDI model intracellular volume fraction (ICVF) 0.825). lowest obtained DTI diffusivity (MD) 0.758). Conclusion Multiple can be used glioma. Non-Gaussianity (NG) (MAP)