DCE-MRI in Glioma, Infiltration Zone and Healthy Brain to Assess Angiogenesis: A Biopsy Study

PURPOSE: To explore the focal predictability of vascular growth factor expression and neovascularization using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in glioma. METHODS: 120 brain biopsies were taken in vital tumor, infiltration zone and normal brain tissue of 30 glioma patients: 17 IDH(isocitrate dehydrogenase)-wildtype glioblastoma (GBM), 1 IDH-wildtype astrocytoma °III (together prognostic group 1), 3 IDH-mutated GBM (group 2), 3 anaplastic astrocytomas IDH-mutated (group 3), 4 anaplastic oligodendrogliomas and 2 low-grade oligodendrogliomas (together prognostic group 4). A mixed linear model evaluated the predictabilities of microvessel density (MVD), vascular area ratio (VAR), mean vessel size (MVS), vascular endothelial growth factor and receptors (VEGF-A, VEGFR‑2) and vascular endothelial-protein tyrosine phosphatase (VE-PTP) expression from Tofts model kinetic and model-free curve parameters. RESULTS: All kinetic parameters were associated with VEGF‑A (all p < 0.001) expression. K(trans), k(ep) and v(e) were associated with VAR (p = 0.006, 0.004 and 0.01, respectively) and MVS (p = 0.0001, 0.02 and 0.003, respectively) but not MVD (p = 0.84, 0.74 and 0.73, respectively). Prognostic groups differed in K(trans) (p = 0.007) and v(e) (p = 0.004) values measured in the infiltration zone. Despite significant differences of VAR, MVS, VEGF‑A, VEGFR‑2, and VE-PTP in vital tumor tissue and the infiltration zone (p = 0.0001 for all), there was no significant difference between kinetic parameters measured in these zones. CONCLUSION: The DCE-MRI kinetic parameters show correlations with microvascular parameters in vital tissue and also reveal blood-brain barrier abnormalities in the infiltration zones adequate to differentiate glioma prognostic groups. SUPPLEMENTARY INFORMATION: The online version of this article (10.1007/s00062-021-01015-3) contains supplementary material, which is available to authorized users.