Scalar DTI features to predict AD dementia in MCI cases: comparison with brain volumetric measures Lead Investigator: Stefan Teipel Institution : DZNE E-Mail : stefan.teipel@med.uni-rostock.de Proposal ID : 548 Proposal Description: Prediction of conversion from MCI to AD dementia based on scalar diffusion indices. Features will derived for key regions of Alzheimer's disease (AD) related mictrostructural alterations as determined from the AD dementia vs. controls comparison to provide a cross-validation of region selection. Our hypothesis is that reduced FA and increased MD in intracortical projecting fiber tracts will predict conversion from MCI to AD dementia. For comparison, we will study predictive accuracy of hippocampus and cortical grey matter volume as determined from anatomical MRI scans. Due to the high number of potentially collinear predictive features from both imaging modalities, we will use regularized versions of Cox linear regression to determine accuracy of prediction of time to conversion. Based on previous findings in the European DTI Study on Dementia cohort (accessible via GAAIN), the results of this analysis will help to determine the generalizability of DTI findings from previous studies in a multicentre context and to assess the clinical value of DTI to predict conversion in MCI compared to already established imaging markers, such as hippocampus volumetry.