P03.07.B STRUCTURAL CONNECTIVITY-BASED SPREADING PATTERNS OF MULTIFOCAL AND MULTICENTRIC GLIOMAS

Le résumé fourni par la source

Abstract BACKGROUND Gliomas are hypothesized to spread along white matter tracts (WMT); coherently, it was discovered that the density of WMT overlapped by glioblastomas (“tract density index”/”TDI”) negatively correlates with patient prognosis. We aimed at studying the structural connectivity-based prognostication and distribution patterns of multifocal/multicentric gliomas, compared to random distributions of brain lesions/foci or brain metastases. MATERIAL AND METHODS We included patients with high clinical-radiological suspicion or diagnosis of multifocal/multicentric adult-type diffuse glioma at diagnosis, gathering survival and MRI data, which underwent coregistration, foci segmentation and normalisation. Control lesion distribution groups were: 1) randomly redistributed glioma foci from the study cohort, 2) artificial randomly distributed glioma-like spheres and 3) multiple brain metastases (Molab’s public dataset). Within each patient, for each single focus, we computed TDI and MRI-based indices of co-involvement of white matter tracts (tract involvement index/“TII”), functional grey and white matter networks (“GII” and “WII”) and of other lesions’ disconnectomes (“disconnectome involvement index”/“DII”). Statistics included ANCOVA correcting for lesion volume, post-hoc t-tests (to compare main cohort results with the control groups), log-rank tests and Kaplan-Meier curves; p-values were Bonferroni-corrected. The higher focus’ TDI (max TDI) of each subject was also compared to a unifocal glioblastoma cohort’s TDI. RESULTS 50 patients were included in the main group (mean age 63 years, 32 males, 23 confirmed GBMs), 50 were generated in both control glioma-like groups, 37 in the metastases group; 96 in the unifocal cohort. MRI indices were different across groups (ANCOVA p-values<0.001). TII and DII values were higher in multifocal gliomas, compared to redistributed glioma lesions (p=0.003 and p=0.036), random glioma-like spheres (p<0.001 and p=0.048) and metastases (p<0.001 and p<0.001), GII and WII were higher than in metastases (p=0.003 and p=0.009). Max TDI was higher in multifocal gliomas than unifocal GBM (33 vs 22, p<0.0001) and its median discriminated short- (higher values) and long-surviving patients (log-rank χ2=8.7, p=0.003). CONCLUSION In multifocal gliomas, higher TII and DII values suggest that different foci tend to arise more likely along the course of the same white matter fibres, compared to metastases and random foci distributions. Such discovery furtherly corroborates the hypothesis of glioma cell migration along WMT. Different glioma foci also tend to co-occur in the same functional networks, compared to metastases. Lastly, max TDI is negatively associated with overall survival in multifocal gliomas and might indicate the “original” primary foci, where surgical/radiotherapy treatments could be prioritized.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.