IC‐P‐072: PREDICTION OF PRODROMAL AD IN MCI SUBJECTS USING MULTICENTER DTI AND MRI DATA AND MULTIPLE KERNELS SVM: AN EDSD STUDY

Alzheimer's disease (AD) patients show significant changes in white matter (WM) structural integrity. Diffusion tensor imaging (DTI) is a neuroimaging technique that allows in vivo assessment of WM fiber tract integrity and could therefore improve the detection of AD at the predementia stage. We used a Support Vector Machine (SVM) classifier to automatically detect AD-specific structural WM changes in patients with amnestic mild cognitive impairment (MCI). Within the framework of the European DTI study on Dementia (EDSD) we collected multicenter DTI and structural MRI data from 146 subjects with MCI, 159 patients with clinically probable AD and 227 healthy controls (HC). MCI subjects with available CSF information were stratified into amyloid-β42 negative (MCI-Aβ42-) and positive (MCI-Aβ42+) subjects using previously established thresholds. To automatically determine group membership, the SVM was applied to DTI-derived fractional anisotropy (FA), mean diffusivity (MD), and mode of anisotropy (MO) maps, as well as volumetric gray matter (GM) and white matter (WM) maps derived from the structural MRIs. We used linear regression to estimate and reduce the effect of the covariates age, education, gender, and center, and validated our approach using a tenfold cross-validation. For multimodal analysis we used the multiple kernels SVM (MK-SVM) algorithm. We achieved accuracies of up to 77% for MD and 62% for GM volume for the discrimination of MCI-Aβ42+ and HC. With the MK-SVM we obtained 73% accuracy. For the group separation between MCI-Aβ42+ and MCI-Aβ42- we obtained accuracies of up to 69% for MO, 52% for GM volume, and 68% for the MK-SVM.