Contralateral neurofluid dynamics predict survival in IDH wild-type glioblastoma: A DTI-ALPS and free water imaging study

BACKGROUND: Glioblastoma (GBM) may disrupt glymphatic function and neurofluid dynamics locally and in distant brain regions. However, the prognostic relevance of such alterations remains unclear. We investigated whether diffusion tensor image analysis along the perivascular space (DTI-ALPS) and free water (FW) imaging serve as biomarkers of glymphatic dysfunction and survival in patients with IDH wild-type GBM. METHODS: We retrospectively analyzed preoperative MRI data from 277 patients in the UPENN-GBM and 269 patients in the UCSF-PDGM cohorts. The ALPS index was quantified in tumor regions and normal-appearing white matter (NAWM) in both hemispheres, and the FW volume fraction in the contralateral NAWM. Data harmonization was performed using ComBat to adjust for intersite variability. Survival analyses were conducted using log-rank tests and Cox regression models. Optimal ALPS index and FW thresholds were derived from the UPENN-GBM dataset and validated in the UCSF-PDGM. RESULTS: The ALPS index was significantly lower in tumor regions than NAWM (P < .01). In the contralateral hemisphere of the UPENN-GBM cohort, a lower ALPS index and higher FW in NAWM were independently associated with shorter overall survival (HR = 0.75, P = .027 for ALPS index; HR = 1.34, P = .04 for FW). The identified thresholds successfully stratified survival in UPENN-GBM and were validated in UCSF-PDGM (P = .011 for ALPS; P = .038 for FW). CONCLUSIONS: Neurofluid dynamic alterations in the contralateral hemisphere, assessed using DTI-ALPS and FW imaging, were independently associated with survival in patients with IDH wild-type GBM. These findings support the use of glymphatic imaging markers for prognostic stratification and therapeutic targeting.