Detection of Alzheimer's disease with diffusion tensor imaging and deformation-based morphometry

Abstract

Novel MRI based acquisition and analysis techniques are increasingly used as biomarkers to discriminate between Alzheimer's disease and normal aging. Evaluating the diagnostic utility of the various approaches in use is difficult though because of significant methodological differences between studies. In this research we directly compare the diagnostic utility of deformation-based morphometry (DBM) and diffusion tensor imaging (DTI) with data derived from the same group of patients with probable AD and healthy control participants. DBM was used to assess regional relative volume reductions in patients compared to controls. Distributed cortical atrophy effects were found in frontal, parietal, and temporal regions. As DTI measures, mean diffusivity and fractional anisotropy were employed to index white matter integrity. The results also point to widespread decline of white matter integrity in frontal, parieto-occipital, and temporal regions in AD. Concerning diagnostic utility, we found that the discrimination performance was best for maps of mean diffusivity and DBM. In a logistic regression model a combination of modalities reached a classification accuracy of AUC = 0.86 after leave-one-out cross-validation. We discuss the results with regard to the feasibility of current MRI based biomarkers for future applications in clinical research settings. © 2011 The authors and IOS Press. All rights reserved.