Background: Imaging measurement of structure alignment has shown increasing importance in determining tissue properties. It is not known if a similar ability for characterizing neuropathology exists.
New methods: This study aimed to validate a new alignment-assessing method for measuring myelin and axonal properties using quantitative histological metrics. The new method involved analysis of the Fourier transform (FT) power spectrum in standard magnetic resonance imaging (MRI). T2-weighted MRI were collected from 10 post-mortem multiple sclerosis (MS) brain samples. Three tissue types were examined: lesions, diffusely abnormal white matter, and normal appearing white matter. MRI analysis included computing the FT power spectrum; extracting alignment histograms; and calculating dominant orientation and alignment complexity (angular entropy). Post MRI, the brain samples were processed for myelin and axonal staining, and the stained images were used to derive quantitative orientation measures using structure tensor analysis for MRI comparison.
Results: There were significant differences in orientation metrics between tissue types in both MRI and histology, and MRI measurements correlated strongly with histological indices. Moreover, the joint effect of myelin and axonal entropy explained over 95% of the variance of MRI angular entropy.
Comparison with existing method: There is no established method for characterizing myelin and axonal pathology using standard MRI. Advanced MRI methods have the potential to do this but are still in research development and are not yet routinely acquired in clinical practice.
Conclusions: Alignment measurement using clinical MRI scans may become a valuable new method for characterizing myelin and axonal properties in MS patients.
Keywords: Axons; Brain white matter; Clinical MRI; Diffusely abnormal white matter; Fourier transform power spectrum; Lesions; Multiple sclerosis; Myelin; Normal appearing white matter; Quantitative histology; Tissue alignment.
Copyright © 2018 Elsevier B.V. All rights reserved.