Low Perfusion Compartments in Glioblastoma Quantified by Advanced Magnetic Resonance Imaging and Correlated with Patient Survival
- Abstract:
- Background Glioblastoma exhibits profound intratumoral heterogeneity in blood perfusion, which may cause inconsistent therapy response. Particularly, low perfusion may create hypoxic microenvironment and induce resistant clones. Thus, developing validated imaging approaches that define low perfusion compartments is crucial for clinical management. Methods A total of 112 newly-diagnosed supratentorial glioblastoma patients were prospectively recruited for maximal safe resection. Preoperative MRI included anatomical, dynamic susceptibility contrast (DSC), diffusion tensor imaging (DTI) and chemical shift imaging (CSI). The apparent diffusion coefficient (ADC) and relative cerebral blood volume (rCBV) were calculated from DTI and DSC respectively. Using thresholding methods, two low perfusion compartments (ADC H -rCBV L and ADC L -rCBV L ) were identified. Volumetric analysis was performed. Lactate and macromolecule/lipid levels were determined from multivoxel spectroscopy. Progression-free survival (PFS) and overall survival (OS) were analysed using Kaplan-Meier and multivariate Cox regression analyses. Results Two compartments displayed higher lactate and macromolecule/lipid levels than normal controls (each P < 0.001), suggesting hypoxic and pro-inflammatory microenvironment. The proportional volume of ADC L -rCBV L compartment was associated with a larger infiltration area ( P < 0.001, rho = 0.42). Lower lactate in this compartment was associated with a less invasive phenotype visualized on DTI. Multivariate Cox regression showed higher lactate level in the ADC L -rCBV L compartment was associated with a worse survival (PFS: HR 2.995, P = 0.047; OS: HR 4.974, P = 0.005). Conclusions The ADC L -rCBV L compartment represent a treatment resistant sub-region associated with glioblastoma invasiveness. This approach was based on clinically available imaging modalities and could thus provide crucial pretreatment information for clinical decision making.
- Authors:
- C Li, J-L Yan, T Torheim, M McLean, N Boonzaier, Y Huang, J Yuan, BRJ Van Dijken, T Matys, F Markowetz, S Price
- Publication date:
- 1st Aug 2017
- Full text
- DOI