Premium
P1‐255: PREDICTION OF PRODROMAL AD IN MCI SUBJECTS USING MULTICENTER DTI AND MRI DATA AND MULTIPLE KERNELS SVM: AN EDSD STUDY
Author(s) -
Dyrba Martin,
Ewers Michael,
Plant Claudia,
Barkhof Frederik,
Fellgiebel Andreas,
Hausner Lucrezia,
Filippi Massimo,
Kirste Thomas,
Teipel Stefan J.
Publication year - 2014
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2014.05.495
Subject(s) - fractional anisotropy , diffusion mri , support vector machine , white matter , dementia , neuroimaging , artificial intelligence , medicine , nuclear medicine , psychology , magnetic resonance imaging , pattern recognition (psychology) , radiology , computer science , neuroscience , disease
Background: Previous structural and functional MRI studies support the hypothesis that the preclinical state of Alzheimer’s disease is distinct from normal aging. The aim of this study was to use diffusion tensor imaging (DTI) measures in order to determine the existence of white matter microstructural differences in preclinical Alzheimer’s disease subjects. Additional objective was to elucidate whether high cognitive reserve also has an influence on white matter changes, and thus aid preclinical Alzheimer’s disease subjects to tolerate a more advanced stage of white matter degeneration. Methods: Thirty-eight subjects with normal cognition were included. Nineteen were in the preclinical Alzheimer’s disease stage, with decreased cerebrospinal fluid amyloid-b (Ab 42,<500 pg/ml). Nineteen matched participants with normal Ab 42 levels were included as a control group. Fractional anisotropy, radial diffusivity, axial diffusivity and mean diffusivity maps were calculated for each participant using the eigenvalues extracted from the diffusion tensors. Voxel based analyses of DTI measures maps were performed using the Tract Based Spatial Statistics. Results:We encountered increases in axial diffusivity in preclinical Alzheimer’s disease relative to controls in corpus callosum, corona radiata, internal capsule and superior longitudinal fasciculus bilaterally, and also in the left fornix, left uncinate fasciculus and left inferior fronto-occipital fasciculus. However, no differences were found in other diffusion tensor imaging indexes. Interestingly, we found significant associations between axial diffusivity values and A b 42 levels (not with total tau or phospho-tau); cognitive reserve score was also positively associated with axial diffusivity exclusively in the preclinical Alzheimer’s disease group. The axial diffusivity index presented good sensitivity (84%) and specificity (79%) for discriminating preclinical Alzheimer’s disease from controls (area under the curve: 0.87). Conclusions: The results suggest that early white matter changes in preclinical Alzheimer’s disease can be detected by DTI. The finding of axial diffusivity alteration together with preserved fractional anisotropy and radial diffusivity indexes in the preclinical Alzheimer’s disease group may indicate that A b 42 levels may be associated with subtle axonal changes, while white matter integrity is still widely preserved. In addition, cognitive reserve seems to act as a buffer for these early white matter changes, preserving cognitive functioning in patients with more advanced structural damage.