Diffusion Tensor Imaging Processing and Group Analysis

(supplement to Multimodal MRI markers for Parkinson’s disease)

PI: Norbert Schuff, PhD, Radiology UCSF

The overall goal of the PPMI study is to identify imaging markers of Parkinson’s disease for improved diagnosis of the disease, monitoring of progression, and ultimately assessment of treatment intervention. The goal of this Supplement is to determine if diffusion tensor imaging (DTI) can provide a robust marker of PD progression which could serve as a valuable adjunct to clinical assessments and other neuroimaging measures, including radioligand PET and SPECT imaging. To achieve this goal, the specific aims of this Supplement can be broadly divided into two categories: i) computation of high quality maps of diffusion tensor eigenvalues and eigenvectors which can be used efficiently by other researchers for tests of new hypotheses and other investigations, including tractography; ii) Performance of group analyses of DTI data for tests of specific hypotheses related to the PPMI.

The Supplement will process and analysis DTI data acquired at about seven imaging sites which participate in the PPMI study. All DTI sites will utilize a uniform DTI protocol for data acquisition. At each site, 10 control subjects will be scanned once and 20 patients will be scanned 4 times over a 5 years period, at baseline, 12, 24, and 48 months, totaling roughly 630 DTI scans. Processing of DTI will include: (i) co-registration of each DTI set to the respective structural MRI set for anatomical correspondence; (ii) computation of DTI spatially invariant summary measures, such as diffusion eigen-values, mean diffusivity (MD), conventional Euclidian fractional anisotropy (FA) as well as a more sensitive FA variant termed geodesic anisotropy (GA). Maps of the diffusion eigen-vectors will also be provided. Taken these computations together will provide a flexible platform for additional investigations, including tractography, by other researchers. This Supplement will further perform group analysis of the DTI during the course of the study to test specific aims.

Specific aims:
I. Region of interest analyses (ROI): We will test the hypothesis that DTI measures of the substantia nigra and nuclei of the basal ganglia, such as fractional anisotropy (FA) and mean diffusivity (MD), will be significantly altered in PD patients compared to controls at baseline, replicating previous findings. Furthermore, changes over time in these DTI measures will correlate with changes in clinical measures of PD progression as well as with changes in radioligand imaging measures, providing valuable complementary information.

II. Regionally unbiased analyses (voxel-by-voxel tests): In addition to variations in ROIs, DTI measures will show further a characteristic pattern of regional differences between PD and control subjects at baseline. Moreover, changes in these patterns over time will correlate with changes in clinical measures of PD progression and changes in radioligand imaging measures, providing valuable complementary information.

III. Exploratory tests:  DTI is a very active field of research and new methods are being developed in a fast pace. As new and more powerful DTI measures are being derived, such as geodesic fractional anisotropy (GA) and mixed tensor models, the Supplement will further compare the power of the new DTI measures to the power of conventional DTI measures in characterizing PD.