Cerebral blood flow and histological analysis for the accurate differentiation of infiltrating tumor and vasogenic edema in glioblastoma

Abstract

Background Glioblastoma is characterized by neovascularization and diffuse infiltration into the adjacent tissue. T2*-based dynamic susceptibility contrast (DSC) MR perfusion images provide useful measurements of the biomarkers associated with tumor perfusion. This study aimed to distinguish infiltrating tumors from vasogenic edema in glioblastomas using DSC-MR perfusion images. Methods Data were retrospectively collected from 48 patients with primary IDH-wild-type glioblastoma and 24 patients with meningiomas (Edemas-M). First, we attempted histological verification of cell density, Ki-67 index, and microvessel areas to distinguish between non-contrast-enhancing tumors (NETs) and edema (Edemas) which were obtained from stereotactically fused T2-weighted and perfusion images. This was performed for evaluating enhancing tumors (ETs), NETs, and Edemas. Second, we also performed radiological verification to distinguish NETs from Edemas. Two neurosurgeons manually assigned the regions of interests (ROIs) to ETs, NETs, and Edemas. The DSC-MR perfusion imaging-derived parameters calculated for each ROI included the cerebral blood volume (CBV), cerebral blood flow (CBF), and mean transit time (MTT). Results Cell density and microvessel area were significantly higher in NETs than those in Edemas (p<0.01 and p<0.05, respectively). Regarding radiological analysis, the mean CBF ratio for Edemas was significantly lower than that for NETs (p<0.01). The mean MTT ratio for Edemas was significantly higher than that for NETs. The receiver operating characteristic (ROC) analysis showed that CBF (area under the curve [AUC] = 0.890) could effectively distinguish between NETs and Edemas. The ROC analysis also showed that MTT (AUC = 0.946) could effectively distinguish between NETs and Edemas. Conclusions DSC-MR perfusion images may prove useful in differentiating NETs from Edemas in non-contrast T2 hyperintensity regions of glioblastoma.