Application of diffusion tensor imaging (DTI)/diffusion kurtosis imaging (DKI) in Dementia and multiple sclerosis (MS)

Abstract

Diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI) can be used to detect brain microstructural changes and pathophysiological procedures. This thesis explored their application in dementia and multiple sclerosis (MS). In the first study (chapter 3), we explored whether Alzheimer's disease (AD) exhibited similar abnormal patterns with MS or neuropsychiatric systemic lupus erythematosus (NPSLE). Sixteen AD plus 14 normal control (NC), 17 MS plus 25 matched NC, and 19 NPSLE plus 24 NC participants were recruited. Tract-based spatial statistics and atlas-based regions of interest were used to analyze white matter and grey matter separately. We observed that AD displayed similar abnormalities with MS in terms of DKI measurements, suggesting they may share some common pathological pathways. In the second study (chapter 4), we determined the key biomarker for tracking the dynamic alterations of MS in a longitudinal study. 18 MS and 18 NC participants were enrolled and followed at baseline, sixth and twelfth months. Both DKI and structural network measurements were calculated. We found that the DKI measurements did not show significant changes, while network measurements did. Significant association with expanded disability status scale (EDSS) score was also observed. It is speculated that network measurements are more sensitive to tracking dynamic changes. The last study explored the application of DTI and DKI in dementia. 19 NC, 11 subjective cognitive decline (SCD), 37 mild cognitive impairment (MCI), 16 AD, 11 mixed dementia (MixD), and 13 vascular dementia (VaD) participants were enrolled. Firstly, in chapter 5, we determined which kind of hippocampal measurement was sensitive in the diagnosis of AD spectrum, the volume or mean diffusivity (MD) or mean kurtosis (MK). We observed that in the dementia stage, right hippocampal volume was the best discriminator, while in the predementia stage, the best discriminators were microstructural indexes (right hippocampal MD for MCI and left hippocampal MK for SCD). In addition, the hippocampal microstructural changes from SCD to AD were not linear (MK increased for SCD while MK decreased for AD compared with NC). Secondly, in chapter 6, we further explored the association between amyloid load and microstructural changes in AD spectrum in 16 cortical GM regions. A negative relation between MD and amyloid and a positive correlation between MK and amyloid was detected in several regions for AD and MCI (such as right temporal, right parietal, and left precuneus/posterior cingulate cortex for AD, and right sensorimotor and right temporal mesial cortex for MCI). We speculated that amyloid deposition might impede water diffusion and increase microstructural complexity for AD and MCI. Thirdly, in chapter 7, graph theory was used to explore the network changes among all groups. We found structural network impairment ranking was SCD < MCI < AD < MixD < VaD, and WM hyperintensity load can account for the result. The difference in these measurements may indicate that they can be used to differentiate different types of dementia. In summary, DTI and DKI are promising diffusion techniques owing to their ability to characterize brain microstructural alterations in vivo.