Noninvasive assessment of glymphatic system alterations as potential biomarkers for predicting overall survival in glioma

In this retrospective study, authors aimed to evaluate the glymphatic function alterations associated with glioma and explore the prognostic value of these alterations by calculating the index for diffusivity along the perivascular space (ALPS index). The authors utilized data from the publicly available University of California San Francisco Preoperative Diffuse Glioma MRI (UCSF-PDGM) dataset, which includes 501 adult patients with histopathologically confirmed diffuse glioma, per the 2021 WHO classification, who underwent preoperative MRI, initial tumor resection, and tumor genetic testing at a single medical center from 2015 to 2021.The ALPS index was calculated from diffusivity maps for noninvasive glymphatic system (GS) analysis. The authors extracted 2288 radiomic features across four tumor regions: surrounding FLAIR abnormality, enhancing tumor, central nonenhancing and/or necrotic tumor, and whole tumor (combining necrosis, enhancement, and edema). For normally distributed variables (adjusted for age, enhancing tumor volume, and surrounding FLAIR abnormality volume), ANCOVA was utilized; nonnormally distributed data were analyzed using the Kruskal-Wallis test and Mann-Whitney U-test. Spearman's correlation coefficients were calculated to assess relationships between radiomic features and the ALPS index. Survival analysis included Kaplan-Meier curves, log-rank tests, concordance index (C-index), calibration and decision curves, and Cox regression. Ultimately, 437 patients with grade 2-4 gliomas were included in this study. The mean patient age was 57.46 ± 14.84 years, and 261 patients were male. The ALPS index correlated most strongly with shape features in the surrounding FLAIR abnormality region (MajorAxisLength, which measures elongation of the edema region, r = -0.33, p < 0.001), intensity features in the enhanced region (T2.RootMeanSquared, which quantifies variations in T2-weighted MRI signal intensity, r = 0.25, p < 0.001), and both shape and texture features in the necrotic region (Sphericity, which reflects the roundness of necrosis, r = 0.26, p < 0.001; FLAIR.glszm.GLNU, which reflects the uniformity of MRI signal distribution, r = -0.24, p < 0.001). Patients with higher-grade tumors (p < 0.001), IDH-wildtype glioma (p = 0.01), and 1p19q-intact tumors (p = 0.038) consistently exhibited reduced glymphatic function. Univariate Cox regression analysis demonstrated that a lower ALPS index was related to a shorter survival time (HR 0.297, 95% CI 0.149-0.593, p < 0.001). Subgroup analyses within histological and molecular subtypes (grade 4, IDH wildtype, and 1p19q intact) demonstrated that patients with ALPS values below the median had significantly shorter overall survival. The ALPS index combined with radiomics improved survival prediction, with the C-index increasing from 0.709 to 0.711 in the training cohort and from 0.675 to 0.693 in the validation cohort. Kaplan-Meier analysis further demonstrated more distinguishable survival curves, with the p value decreasing from 0.003 to 0.0001. Worse GS function was associated with more aggressive tumors and shorter survival times.