Abstract:Background: Measuring non-parametric intravoxel mean diffusivity distributions (MDDs) using magnetic resonance imaging (MRI) is a sensitive method for detecting intracellular diffusivity changes during physiological alterations. Histological and molecular glioma classifications are essential for prognosis and treatment, with distinct water diffusion dynamics among subtypes. Purpose: We developed a data-driven approach using a fully connected network (FCN) to enhance the speed and stability of calculating MDDs across varying SNRs, enable tumor microstructural mapping, and test its reliability in identifying MIB-1 labeling index (LI) levels and molecular status of gliomas. Methods: An FCN was trained to learn the mapping between the simulated diffusion decay curves and the ground truth MDDs. We performed 5 000 000 simulation curves with various diffusivity components and random SNR ∈ [ 30 , 300 ] $ \in [ {30,\ 300} ]$ . Eighty percent of simulation curves were used for the FCN training, 10% for validation, and the others were external tests for the FCN performance evaluation. In vivo data were collected to evaluate its clinical reliability. One hundred one patients (44 years ± $ \pm $ 14, 67 men) with gliomas and six healthy controls underwent a 3.0 T MRI examination with a spin echo-echo planar imaging (SE-EPI) diffusion-weighted imaging (DWI) sequence. The trained FCN was employed to calculate MDDs of each brain voxel by voxel. We used the Fuzzy C-means algorithm to cluster the MDDs of tumor voxels, facilitating the characterization of distinct glioma tissues. Quantitative assessments were conducted through sectional integrals of the MDDs, demarcated by six bands to derive signal fractions ( f n , n = 1 - 6 ${{f}_n},\ n = 1 -6$ ) and diffusivities of the maximum peaks ( D p e a k ${{D}_{peak}}$ ). Cosine similarity scores (CSS) were used for MDD similarity. ANOVA and Mann-Whitney U test were used for difference analysis. Logistic regression and area under the receiver operator characteristic curve (AUC) were used for classification evaluation. Results: The simulation results showed that the FCN-based MDD approach (FCN-MDD) achieved higher CSS than non-negative least squares-based MDD (NNLS-MDD). For in vivo data, the spectra of ET and NET obtained by FCN-MDD are more distinguishable than NNLS-MDD. Fraction maps delineate the characteristics of different tumor tissues (enhancing and non-enhancing tumor, edema, and necrosis). f 3 , f 4 , D p e a k ${{f}_3},\ {{f}_4},{{D}_{peak}}$ showed a positive and negative correlation with MIB-1 respectively ( r = 0.568 , r = - 0.521 , r = - 0.654 $r = 0.568,\ r = - 0.521,\ r = - 0.654$ , all p < 0.001 $p < 0.001$ ). The AUC of D p e a k ${{D}_{peak}}$ for predicting MIB-1 LI levels was 0.900 (95% CI, 0.826-0.974), versus 0.781 (0.677-0.886) of ADC. The highest AUC of isocitrate dehydrogenase (IDH) mutation status, assessed by a logistic regression model ( f 1 + f 3 ${{f}_1} + {{f}_3}$ ) was 0.873 (95% CI, 0.802-0.944). Conclusion: The proposed FCN-MDD method was more robust to variations in SNR and less reliant on empirically set regularization values than the NNLS-MDD method. FCN-MDD also enabled qualitative and quantitative evaluation of the composition of gliomas.

Accurate low and high grade glioma classification using free water eliminated diffusion tensor metrics and ensemble machine learning

Differentiation Between Low-Grade and High-Grade Glioma Using Combined Diffusion Tensor Imaging Metrics

Incorporating Multiple Magnetic Resonance Diffusion Models to Differentiate Low- and High-Grade Adult Gliomas: a Machine Learning Approach.

Nimg-71. T2 Relaxivity And Diffusion Tensor Metrics In The Prediction Of Progression-Free Survival In Non-Recurrent High Grade Glioma

Machine-learning based classification of glioblastoma using delta-radiomic features derived from dynamic susceptibility contrast enhanced magnetic resonance images: Introduction.

Application of MR diffusion tensor imaging and diffusion tensor tractography in diagnosis and treatment of glioma

Application of Diffusion Tensor Imaging Quantitative Parameters in Grading of Cerebral Glioma on a 3.0T Scanner

Preliminary Study of Diffusion Tensor Imaging in Differentiating Diagnosis of High-grade Gliomas and Metastatic Tumors

Machine-learning based classification of glioblastoma using dynamic susceptibility enhanced MR image.

Qualitative and Visual along-Tract Analysis of Diffusion-Based Parameters in Patients with Diffuse Gliomas

A data-driven intravoxel mean diffusivities distribution approach for molecular classifications and MIB-1 prediction of gliomas

Diffusion Tensor Imaging with Fluorescein Sodium Staining in the Resection of High-Grade Gliomas in Functional Brain Areas

Application of Diffusion Tensor Imaging on Glioma Grading

The Application of Diffusion Tensor Imaging in Glioma Grading

Differences in the architecture of low-grade and high-grade gliomas evaluated using fiber density index and fractional anisotropy.

The role of MR diffusion tenser imaging in evaluating the relationship between the grade of gliomas and the state of adjacent fiber tract

Texture feature analysis of MRI-ADC images to differentiate glioma grades using machine learning techniques

Integrating diffusion kurtosis imaging, dynamic susceptibility-weighted contrast-enhanced MRI, and short echo time chemical shift imaging for grading gliomas

Axial diffusivity and tensor shape as early markers to assess cerebral white matter damage caused by brain tumors using quantitative diffusion tensor tractography.

Differentiation between high-grade gliomas and solitary brain metastases based on multidiffusion MRI model quantitative analysis

Intergrating conventional MRI, texture analysis of dynamic contrast-enhanced MRI, and susceptibility weighted imaging for glioma grading