Vergleich von Diffusion Tensor Imaging und Diffusion Kurtosis Imaging basiertem Fibertracking neurochirurgisch relevanter Faserbahnsysteme

Die Traktographie stellt ein wichtiges Element der präoperativen Planung in der Neurochirurgie dar, um eine Vorhersage über den anatomischen Verlauf von relevanten Fasertrakten, deren Ausdehnung und Nähe zu Gehirntumoren treffen zu können. Hierbei ist es von vorrangiger Bedeutung, dass eine maximal...

Full description

Saved in:
Bibliographic Details
Main Author: Emde, Julia
Contributors: Nimsky, Christopher (Prof. Dr.) (Thesis advisor)
Format: Doctoral Thesis
Published: Philipps-Universität Marburg 2021
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!

Tractography is an important element of preoperative planning in neurosurgery in order to predict the course of relevant fibers, their extent and proximity to brain tumors. It is of prime importance that a maximum safe tumor resection can be ensured while preserving the surrounding structures at risk such as the corticospinal tract (CST), arcuate fascicle (AF) and the optic radiation (OR). The clinical standard method of the fiber reconstruction, diffusion tensor imaging (DTI), however, is only partially able to fulfill these aspects due to the underlying tensor model and requires some optimization. The model assumes that the direction of water diffusion in tissues is Gaussian distributed, but this is not the way diffusion is like in real complex structures such as neurons. Due to this strong simplification, fiber crossings and fanning fibers cannot be adequately represented, and the spatial extent of the fiber bundle volumes is underestimated by this technique. This last point increases the risk for postoperative deficits when using too small safety margins. The approach of Diffusion Kurtosis Imaging (DKI), which is compared with the DTI-based fibertracking in this work, also models the non-Gaussian distribution of water diffusion and thus has the potential to deal with some limitations of the DTI model. In addition, in contrast to many other alternative models DKI requires only a comparatively slightly longer data acquisition time, which should be as short as possible in terms of clinical feasibility and patient compliance. The aim of this work is to demonstrate the advantages and applicability of DKI-based fibertracking for the illustration of risk structures in preoperative planning for glioma resection and to compare them to the DTI-based fibertracking. The corticospinal tract (CST), the optic radiation (OR) and the arcuate fascicle (AF) of both hemispheres were reconstructed for N = 19 healthy volunteers and N = 16 patients with gliomas of different grades and localization using DTI- and DKI-based fibertracking. The volumes and additionally the fiber density in the cross-sectional area of the tract were statistically compared and the image data were evaluated qualitatively. The tract CST reconstructed with the help of DKI showed a significantly larger volume (volunteers: links p < 0,001, rechts p < 0,001; patients: links p < 0,001, rechts p < 0,001) and a significantly higher fiber density in both groups (volunteers: left p = 0,001, right p = 0,002; patients: left p < 0,001, right p < 0,001) in favor of the DKI reconstruction. For the OR, there was no significant difference in the right hemisphere volumes of the patient cohort (p = 0,389), while DKI-based volumes were significantly larger in both groups (volunteers: left p <0.001, right p = 0.005; patients: left p = 0.021). A significantly higher fiber density in the DKI was only found for the volunteers (left p = 0.005, right p = 0.025). The difference in patients, however, remained insignificant (left p = 0,638, right p = 0,183). There were no significant differences between DKI and DTI for volumes of the AF of the volunteers (left p = 0,101, right p = 0,044). In the patient collective, even the DTI-based reconstruction yielded significantly larger volumes (left p = 0,023, right p = 0,010) and a significantly higher fiber density in case of the left hemisphere (p = 0,010). The difference in fiber density in the other cases remained insignificant (subjects: left p <0,163, right p = 0,205; patients: right p = 0,823). The qualitative comparison confirmed the statistical results and showed a better resolution of cortical fanning fibers for the CST in many cases, especially of additional fibers to the lateral somatomotor cortex and for the OR in the DKI reconstruction. In individual cases, more fibers were seen in the vicinity of the tumor by the DKI. The observation of the AF showed the anatomically more plausible representation in the DTI reconstruction in subjects and patients, whereas many DKI-tracked fibers prematurely broke.The assumption that the DKI largely overcomes the limitations of the DTI and obtains larger tractvolumes can only partly be confirmed by the present study. The results for the reconstruction of the CST supports the hypothesis like other studies before. The result of the DKI-based AF reconstruction states a problem since the DTI based fibertracking seems to be superior for this pathway. An anatomically plausible representation of this tract failed and due to the lack of comparable studies on the DKI reconstruction, a plausible reason for this unexpected result is difficult to find. The attempt of an explanation includes the directional and diffusion properties of the AF itself and limitations of the DKI technique with respect to strong curvatures of fibers and to small angles between the examined pathway and a high number of intersecting tracts. Therefore, it would be interesting to investigate the cause of this result in additional studies for further improvement of the DKI technique. Regarding the successful results in the CST-reconstruction, a further intraoperative validation of the method would be desirable. Even if there is a restriction of the applicability of the DKI-based fibertracking to certain tracts, the great potential of this method is to be emphasized for the improvement of the preoperative representation of important structures at risk.