Dokument

Summary:
Diffusion tensor imaging (DTI) has become a standard procedure in clinical routine as well as research as it enables the reconstruction and visualization of fiber tracts in the human brain. Due to the simplified assumption the tensor model – a Gaussian distribution of the diffusion – it typically fails to provide neither accurate spatial mapping nor quantification of crossing or kissing fibers. A clinically feasible development might be diffusion kurtosis imaging (DKI), an extension of DTI also integrating non-Gaussian distribution diffusion processes and thereby shall overcome some of its limitations. The potential DKI will be evaluated in case of the detection of the interhemispheric asymmetry of the white matter in healthy volunteers (n = 20), as well as the analysis of tumor-related impairments of fiber tracts and their correlation with neurological deficits in patients (n = 13) diagnosed with glioma. In order to analyze interhemispheric asymmetry across the whole brain, especially of nine large fiber tracts, tract-based spatial statistics (TBSS) analysis was performed using DTI- and DKI-based parameters, a laterality index was calculated for asymmetries and DTI- and DKI-based results were compared. With regard to fractional anisotropy as marker of integrity, asymmetry was found for all nine fiber tracts based on DTI and seven tracts based on DKI. For mean diffusivity, asymmetries were found for three (DTI) and two (DKI) fiber tracts. Regarding mean kurtosis, asymmetry was found in one tract. The interhemispheric asymmetry thereby varied in anatomical location as well as in cluster size. Only small parts of the tracts were affected. A comparison of DTI and DKI showed significantly higher fractional anisotropy and mean diffusivity based on DKI compared to DTI. Gender and handedness did not seem to have any influence. For the assessment of tumor-related changes of fiber tracts in patients diagnosed with glioma, especially in relation to pre-existing and postoperative neurological deficits (hemiparesis, aphasia), templates for the corticospinal tract and the arcuate fasciculus were created based on DTI- and DKI-derived parameters, respectively. The corticospinal tract and the arcuate fasciculus were reconstructed for each patient and the associated parametric maps were projected onto the templates. Based on this, alterations along the tracts could be identified and quantified. Alterations were found on fiber tracts regardless of the spatial proximity to the lesion. There was a correlation between alterations based on fractional anisotropy, mean diffusivity and mean kurtosis. Increased mean diffusivity was associated with alteration in mean kurtosis, a decreased fractional anisotropy was found concurrent with a likewise decreased mean kurtosis. In the case of pre-existing neurological deficits (hemiparesis, aphasia) with regard to the changes along the fiber tracts (corticospinal tract, left arcuate fasciculus), most often increased mean diffusivity and altered mean kurtosis was found. Applying this pattern for prediction of corresponding postoperative neurological deficits a sensitivity of 75.0% and a specificity of 87.5% was achieved. DKI seems to more precisely estimated and depict the underlying microstructure in comparison to DTI. Thereby, in pathological cases especially the mean kurtosis seems to be of special interest. A combination of DTI- and DKI based parameters, particularly with regard to its clinical usability and value, offers great potential in clinical routine.

Das Dokument ist im Internet frei zugänglich - Hinweise zu den Nutzungsrechten