Quantification of tumor subregions in glioblastoma multiforme: A volumetric analysis.

Abstract

e14015 Background: In the last 30 years only one drug (temozolomide) has been shown to provide a significant improvement in survival of patients with newly diagnosed glioblastoma. The changes seen in overall survival are modest, virtually all patients die of their disease, and the drug primarily benefits the ~40% of patients with MGMT methylated tumors. The results of chemotherapy and targeted therapies have been extremely disappointing likely due to failure of these agents to cross the blood-brain barrier (BBB) in therapeutic concentrations. Gadolinium MRI scans have been used to identify portions of the tumor where the BBB is disrupted and where systemically administered drugs would be most likely to reach. This study was designed to estimate the proportion of a newly diagnosed glioblastoma that has sufficient BBB disruption to allow the penetration of gadolinium. Methods: MRI scans (T1, T1 post contrast, T2, and FLAIR) were analyzed from 50 patients who consecutively underwent an initial resection of their glioblastoma in 2022 at the Johns Hopkins Hospital. Volumetric tumor segmentation was performed automatically using the HD brain tumor segmentation tool followed by manual segmentation of necrosis and quality control using the 3D Slicer image computing platform. Volumes were extracted for the T2/FLAIR, the contrast enhancing and the necrotic segments. Results: The 50 patients whose pre-operative scans were evaluated had a median age of 64.5 years (range 35-85), 62% were male, and their median daily steroid dose at the time of the MRI was 10 mg (range 0-40). Their post-op pathology revealed that 30% were MGMT methylated, and all were IDH wildtype previously untreated glioblastoma. The median size of the tumor (T2/FLAIR) was 52.2 cm 3 (range 5.5-284.9), the contrast enhancing segment was 17.2 cm 3 (range 0-79.8) and the necrotic segment was 4.6 cm 3 (range 0-31.1). Overall, the tumors consisted of 73% T2/FLAIR hyperintesity (i.e. infiltration), 23% contrast enhancement, and 4% necrosis. Conclusions: The contrast enhancement volume proportion is an important surrogate for regions of the tumor where systemically administered drugs are most likely to be able to cross the BBB. This study suggests that only about one quarter of the tumor is contrast enhancing. As most chemotherapy drugs and targeted agents are much less likely than gadolinium to cross the BBB (molecular weight, surface area, charge, lipid solubility, affinity to efflux pumps, etc.) this clearly overestimates the percent of tumor likely to be treated by systemically administered drugs. These data strongly suggest that future research should focus on novel methods to deliver drugs in therapeutic concentrations to the large non-enhancing portions of the tumor if progress is to be made in the treatment of this disease.