Voxel-based analysis of glioblastoma topography and invasiveness using the Brain-Grid system: a multicenter retrospective study.

Abstract

Precise characterization of glioblastoma (GBM) location is essential for prognosis and surgical planning. The aim of this study was to evaluate whether voxel-based analysis with the Brain-Grid (BG) system can provide prognostic insights by quantifying tumor invasiveness and spatial distribution. The authors retrospectively analyzed 116 adult patients who underwent treatment for newly diagnosed IDH-wildtype GBM. Tumors were classified on preoperative MRI using the BG system. Clinical, radiological, and molecular variables, including MGMT promoter methylation, were correlated with overall survival (OS), progression-free survival, and functional outcomes. The mean patient age was 63.5 ± 9.7 years. The mean OS was 14.4 ± 8.86 months (maximum 53 months). The mean number of infiltrated BG voxels was 5.8 ± 3.1 (median 5, IQR 3-8; maximum 14). Patients with ≥ 6 voxels had shorter OS (p = 0.012). Tumors most often involved the A2-C2-S2 and A2-C2-S3 voxels (where "A" indicates "axial," "C" indicates "coronal," and "S" indicates "sagittal"), corresponding to the basal ganglia and subventricular zone. Right-sided tumors were most common; posterior localization was associated with better survival (area under the curve = 0.71). Resection in the A2-C2-S3 voxel correlated with improved OS (p = 0.03). The median preoperative tumor volume was 38.74 cm3; higher pre- and postoperative volumes and lower extent of resection (EOR) predicted worse prognosis (p = 0.012, p = 0.001, and p = 0.002, respectively). MGMT methylation showed no significant correlation with OS (p = 0.15) or voxel distribution. Voxel-based analysis with the BG system enables quantitative and qualitative characterization of GBM topography. Invasiveness, tumor burden, and EOR were key predictors of survival, while MGMT methylation had limited spatial prognostic value. Incorporating voxel count into preoperative evaluation may enhance prognostic stratification and surgical planning in GBM.