Methodological Considerations and Clinical Relevance of Diffusion Tensor Imaging in Acute Stroke Prognosis

Predicting motor, language, and global outcome after ischemic stroke is a major research concern. While initial impairment, age, and lesion volume have proven indicators of future outcome, the preservation of major white matter structures also play a role in these outcome domains. Diffusion tensor imaging (DTI) is sensitive to acute ischemic damage and can evaluate the integrity of important white matter bundles. Using a large cohort of patients who underwent a DTI protocol at 24 hours post-stroke and a clinical evaluation 3 months afterwards, the current thesis sought to (1) investigate spatial normalization strategies to optimally analyze imaging data, (2) establish which DTI parameter best predicts global outcome, and (3) determine if DTI can provide independent biomarkers of motor and language outcome. The major findings of the current thesis were: (1) fiber orientation distribution (FOD)-based spatial normalization performed similarly to scalar normalization for acute stroke data but yielded stronger anatomo-clinical correlations in subacute-chronic stroke patients, (2) axial diffusivity (AD) in the corona radiata highly contributes to the prediction of autonomy in patients, and (3) the AD of the corticospinal tract and arcuate fasciculus are independent markers of motor and aphasia outcome, respectively. These results support the use of AD for quantifying early brain damage in important white matter structures, such as the corona radiata, corticospinal tract, and arcuate fasciculus. These markers could be used for patient information, as surrogates of neuroprotective therapies at the hyperacute stage, or as stratification means for rehabilitative therapies.