Introduction: Model discovery methods identify dynamical relationships between state variables, enabling improved understanding of complex biological systems. In Dynamic Contrast Enhanced (DCE) magnetic resonance imaging (MRI), localized convolutional function regression (LCFR) has been used to quantify spatially varying PDE transport parameters in vivo [1]. While these measurements have been validated, there is currently no direct means to modulate these parameters for tumor control, nor a clear causal link between transport dynamics and patient outcomes. To address this gap and assess the clinical utility of these methods, we studied patients with high-grade gliomas (HGGs) treated with IL13Rα-targeted chimeric antigen receptor (CAR) T-cell therapy. This cohort was selected because both HGG and CAR-T cell proliferation are sensitive to flow [2,3]. Methods:We retrospectively analyzed 39 patients with HGGs treated at City of Hope National Medical Center using IL13Rα-targeted CAR-T cells, all of whom received DCE-MRI and diffusion-weighted MRI before and after an initial CAR-T infusion. LCFR was applied to DCE-MRI data to quantify intratumoral distributions of fluid velocity (u), diffusion (D), perfusion (Ktrans), plasma volume fraction (vp), and the apparent diffusion coefficient (ADC) of water within T1-contrast-enhancing tumor regions. These distributions were compared to clinical outcomes to probe explanatory relationships. Results: High heterogeneity of LCFR-measured fluid speed (std(|u|), Cox proportional hazards ratio 0.08, p < 0.05) and increases in speed heterogeneity following CAR-T infusion (Δstd(|u|), Cox proportional hazards ratio 0.01, p < 0.05) protected against decreased progression-free survival. Additionally, patients with a low lower-quartile ADC exhibited significantly reduced progression-free survival (log-rank p < 0.05). Conclusion: These findings demonstrate that LCFR-derived parameters are clinically relevant and may aid in prognosis following further validation. They also motivate testable hypotheses that fluid flow heterogeneity facilitates CAR-T cell distribution and proliferation, while increased tumor density may limit CAR-T cell penetration.