gms | German Medical Science

72. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC)
Joint Meeting mit der Polnischen Gesellschaft für Neurochirurgie

Deutsche Gesellschaft für Neurochirurgie (DGNC) e. V.

06.06. - 09.06.2021

Comparison of diffusion signal models for fibre tractography in eloquent glioma surgery – determination of accuracy under awake craniotomy conditions

Diffusionstensorbildgebung oder hochaufgelöste Diffusionsmodelle zur Faserbahntraktographie – Evaluation der Genauigkeit bei der Chirurgie eloquenter Gliome unter Wachkraniotomiebedingungen

Meeting Abstract

  • presenting/speaker Daniela Becker - Heidelberg University Hospital, Neurosurgery, Heidelberg, Deutschland
  • Moritz Scherer - Heidelberg University Hospital, Neurosurgery, Heidelberg, Deutschland
  • Peter Neher - Deutsches Krebsforschungszentrum, Medical Image Computing, Heidelberg, Deutschland
  • Christine Jungk - Heidelberg University Hospital, Neurosurgery, Heidelberg, Deutschland
  • Jessica Jesser - Heidelberg University Hospital, Neuroradiology, Heidelberg, Deutschland
  • Regina Brinster - Ruprecht-Karls-University Heidelberg, Medical Biometry and Informatics, Heidelberg, Deutschland
  • Andreas W. Unterberg - Heidelberg University Hospital, Neurosurgery, Heidelberg, Deutschland

Deutsche Gesellschaft für Neurochirurgie. 72. Jahrestagung der Deutschen Gesellschaft für Neurochirurgie (DGNC), Joint Meeting mit der Polnischen Gesellschaft für Neurochirurgie. sine loco [digital], 06.-09.06.2021. Düsseldorf: German Medical Science GMS Publishing House; 2021. DocP230

doi: 10.3205/21dgnc511, urn:nbn:de:0183-21dgnc5112

Published: June 4, 2021

© 2021 Becker et al.
This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License. See license information at http://creativecommons.org/licenses/by/4.0/.


Outline

Text

Objective: Besides intraoperative electrostimulation methods (IOM), fiber tractography (FT) became an important non-invasive tool to ensure maximal safe tumor resection in eloquent glioma surgery. Intraoperatively applied FT is still predominantly based on Diffusion Tensor Imaging (DTI). However, multiple reconstruction schemes of high angular resolution diffusion imaging (HARDI) data for high resolution fiber tractography (HRFT) are becoming a focus of interest. However, clinical reasonability has to be considered. We evaluate the accuracy of HRFT based on q-ball imaging (QBI) and Constrained Spherical Deconvolution (CSD), compared with DTI-FT for the major language-associated pathways.

Methods: Ten patients with eloquent gliomas underwent surgery under awake craniotomy conditions. Points(AP) were acquired according to IOM results under repetitive neuropsychological examination. The distances of APs to fiber bundle margin and centerline (offsets) were calculated. Probabilistic HRFT models QBI-, and CSD were compared with conventional DTI-FT. The data was derived from a diffusion sensitized dataset of the clinical routine. Pairwise comparison of the methods was performed using the Wilcoxon signed rank test.

Results: Among the measured 183 min-offset values, 46% were located closer than 1cm to the estimated fiber bundle (CSD: 38%; DTI: 39% and QBI: 62%). However, only 10 min-offsets were further than 3cm away (5%). For QBI, lowest mean min-offsets (SLF: 7.7 ± 7.9mm; IFOF: 12.7 ± 8.3mm; ILF/MLF: 17.7 ± 6.7mm) were found. A significant advantage was objectified for QBI-FT compared with CSD and DTI (p<0.001). Comparing CSD-, and DTI-FT, there was no significant difference for the total fiber object (p=0.105), only for SLF (p<0.001) (illustrated with Figure 1 [Fig. 1]).

Conclusion: Comparing HRFT techniques QBI and CSD for FT with DTI, QBI delivers significantly better results with lowest offsets and a good correlation to the IOM results. Besides, QBI-FT is still feasible for neurosurgical pre- and intraoperative applications. We consider a combined approach of QBI-FT and IOM under awake craniotomy for best preservation of neurological function in the presented setting. The implementation of well selected and sophisticated HRFT models into neuronavigation systems should be pursued.