Free-water diffusion tensor imaging improves the accuracy and sensitivity of white matter analysis in Alzheimer’s disease

Abstract

Magnetic resonance imaging (MRI) based diffusion tensor imaging (DTI) can assess white matter (WM) integrity through several metrics, such as fractional anisotropy (FA), axial/radial diffusivities (AxD/RD), and mode of anisotropy (MA). Standard DTI is susceptible to the effects of extracellular free water (FW), which can be removed using an advanced free-water DTI (FW-DTI) model. The purpose of this study was to compare standard and FW-DTI metrics in the context of Alzheimer’s disease (AD). Data were obtained from the Open Access Series of Imaging Studies (OASIS-3) database and included both healthy controls (HC) and mild-to-moderate AD. With both standard and FW-DTI, decreased FA was found in AD, mainly in the corpus callosum and fornix, consistent with neurodegenerative mechanisms. Widespread higher AxD and RD were observed with standard DTI; however, the FW index, indicative of AD-associated neurodegeneration, was significantly elevated in these regions in AD, highlighting the potential impact of free water contributions on standard DTI in neurodegenerative pathologies. Using FW-DTI, improved consistency was observed in FA, AxD, and RD, and the complementary FW index was higher in the AD group as expected. With both standard and FW-DTI, higher values of MA coupled with higher values of FA in AD were found in the anterior thalamic radiation and cortico-spinal tract, most likely arising from a loss of crossing fibers. In conclusion, FW-DTI better reflects the underlying pathology of AD and improves the accuracy of DTI metrics related to WM integrity in Alzheimer’s disease.

Amyloid-PET of the white matter: relationship to free water, fiber integrity, and cognition in patients with dementia and small vessel disease

“…In diffusion tensor imaging (DTI) modeling of dMRI data, fractional anisotropy (FA) reflects an ellipsoid shape of diffusivity within a given voxel. Higher FA values are thought to arise from higher diffusivity along one preferred direction (i.e., along intact myelinated fibers, thus representing fiber integrity) or from a loss of crossing fibers [11][12][13] . On the other hand, mean diffusivity (MD) represents the magnitude of diffusion (average of eigenvalues) regardless of the directionality.…”

“…It has been hypothesized that this WM-cortex PET correlation may be partly due to an additive off-target signal within the cortex, which may explain the improved stability of longitudinal PET in AD target regions using white matter as a reference region 53 . Apart from the WM-cortex correlation in PET signal, the relationship between lower 18 F-AV45 SUVR and higher FW-adjusted FA may also be explained by a loss of crossing fibers typically resulting in higher FA and FW-adjusted FA 11,13,54 . For example, Douaud et al 12 showed that the only region where DTI metrics were significantly different between healthy controls and MCI was an area of crossing fibers in the centrum semiovale; here, the motor-related projection fibers that crossed the degenerating fibers were selectively preserved, resulting in an increased FA in MCI.…”

A preliminary study to evaluate white matter structural correlates of delayed visuospatial memory and one-week motor skill retention in nondemented older adults

“…One limitation of this study regards the diffusion-weighted image acquisition protocol. Recent work has shown that free-water correction improves the accuracy and sensitivity of white matter analyses (59,60) by fitting a bi-tensor model to each voxel to account for partial volume effects (i.e., voxels that contain brain tissue and free-water such as cerebrospinal fluid); however, it is advised to apply this technique to single-shell diffusion-weighted images that were acquired with b-values less than 1000 s/mm 2 (61). Our data were acquired with b-value = 2500 s/mm 2 , therefore we were unable to apply free-water correction due to our imaging acquisition.…”

Free-water diffusion tensor imaging detects occult periependymal abnormality in the AQP4-IgG-seropositive neuromyelitis optica spectrum disorder

“…Free-water corrected DTI maps were calculated using an in-house MATLAB script (Fig. S1 ) 20 , 29 . Finally, we acquired the mean of the diffusion measures of each periependymal lining shell.…”

“…Further, it is important to correct for partial volume effect in the voxels surrounding ventricles because contamination of CSF has a large effect on DTI parameters. To overcome the effect of extracellular free-water on DTI metrics, we adopted a free-water correction algorithm that quantify and remove the contribution of extracellular free-water 19 , 20 .…”