Logo: to the web site of Uppsala University

Intro: Seizure incidence in diffuse glioma ranges between 60% and 90%. This study aimed to investigate the association between seizures and diffuse glioma in subcortical and cortical brain regions, including white matter tracts.

Methods: Adult patients with diffuse glioma at Uppsala University Hospital from 2005 to 2021 were analysed retrospectively. The relationship between tumour location in specific brain voxels and preoperative seizures was examined concerning white matter tract involvement. Tumour volumes were segmented based on T2-weighted or FLAIR MRI after spatial normalisation to standard space (MNI) and combined to create a location-specific frequency map.

Results: Of the 93 patients meeting the inclusion criteria, 70 (75%) experienced seizures. A significant decreased risk was found in tumours present within the left fronto-mesial and dorsal voxel (A3C1S1). Increased seizure risk was found in tumours located in the left supramarginal and posterior insular voxel (A4C2S3). The voxels differed in terms of type and extent of white matter networks. Additionally, there was a difference in seizure risk and voxel associations between oligodendrogliomas and astrocytoma, with specific voxels associated with seizures identified in both groups.

Conclusion: The study provides new insights into the epileptogenic potential of diffuse gliomas in relation to their spatial distribution, highlighting the need to analyse both cortical and subcortical localisation of tumours. The observed differences in seizure risks across brain regions underscore the need for personalised post-surgery treatment strategies and further research to understand the pathophysiology of brain tumour-related epilepsy, BTRE.

Background: Grade 2-3 diffuse gliomas (DGs) show extensive infiltration through white matter (WM) tracts. Along-tract analysis of WM tracts based on diffusion tensor tractography (DTI) can been performed to assess the microstructural integrity of WM tracts. The clinical implication of these DTI-related findings is still under debate, especially in tumor patients. The aim of this study was to analyze and compare diffusion-based parameters along WM tracts and variables specific to WM -tumor interactions in DGs and correlate them with preoperative neuropsychological assessment.

Methods: Fourteen patients with IDH-mutated grade 2-3 DGs were included. Tumor volumes were manually segmented on 3D-FLAIR images after spatial normalisation to MNI space. DTI was acquired using a single-shot echo-planar sequence on a 3T with 48 sampling directions. DTI data were reconstructed within the MNI space using q-space diffeomorphic reconstruction (QSDR) in DSI studio. Five bilateral sets of WM tracts were reconstructed based on the HCP-1065 template. All WM tracts were stretched to the same length of 100 indices, and for each index diffusion-based parameters fractional anisotropy (FA), radial diffusivity (RD), axial diffusivity (AD), mean diffusivity (MD) and quantitative anisotropy (QA) were sampled. Tumor-related parameters (TRP); tumor volume (Tv), maximum tumor presence (MTP) and the number of sequential indices in which a tumor is present (Te) were derived based on the along-tract analysis. Normal data were constructed by calculating the average and standard deviations of contralateral and not-affected WM tracts for each diffusion-based parameter, respectively. Affected WM tracts were individually compared to normal data using a z-test. Preoperative neuropsychological assessment was performed in all subjects and correlated to results from the along-tract analysis using correlation and logistic regression models.

Results: Abnormalities in diffusion-based parameters were detected in WM tracts. Topographical and quantitative information were presented within the same graph. AD and MD displayed the highest linear correlation with the TRPs. Abnormal QA showed a linear correlation with Tv per WM tract. Neuropsychological impairment was correlated with all the TRPs and with abnormal FA (p < 0.05) and abnormal QA (p < 0.01). Abnormal QA was the only independent variable able to predict the presence of neuropsychological impairment in the patients based on the linear regression analysis.

Conclusions: Graphical presentation of the along-tract analysis presented in this study shows that it may be a sensitive and robust method to acquire and display topographical and qualitative information regarding WM tracts in close proximity to DGs. Further studies and refinements to the methods presented herein may advance current clinical methods for evaluating displacement and infiltrations and further aid the efforts of pre-planning surgical interventions with the goal to maximise EOR and tailor oncological treatment.