Quantitative Evaluation with MRI of Glioblastoma Multiforme Growth and Treatment in Murine Models: Tumor Volume, Diffusion Mapping, and T2 Relaxometry

Glioblastoma Multiforme (GBM) is the most common form of primary brain cancer that has an extremely high mortality rate. Imaging has the capacity to serve as a crucial tool for tracking malignant growth and treatment response. Newer, advanced imaging methods such as relaxometry and diffusion have allowed for an extension of traditional anatomical imaging. In this multi-parametric study, murine models were injected with human U87 tumor cells and then scanned using MRI. Three out of six mice were treated while three remained untreated. Using relevant equations for relaxometry and diffusion, a semi-automated post processing tool was created in Matlab that could segment malignant regions of interest (ROI). Diffusion maps and T2 relaxometry maps were created. ROIs from segmentation were used to plot the distribution of T2 and ADC values as a histogram within an area of interest. Tumor volume comparison indicated that the tumor volume was significantly decreased for the treated mice, but the volume did not disappear in any of the treated mice. The distribution data showed that there was a stark change in T2 values following treatment, which was not reflected in the tumor volume, indicating that a shift in T2 values is an earlier marker of treatment response. Following initial treatment, the T2 value distribution was seen to return toward baseline levels, while the tumor volume remained small. Diffusion distributions also indicated that treatment with TMZ had a response. Mean diffusion (MD) and radial diffusion (RD) decreased with treatment, while fractional anisotropy (FA) and longitudinal diffusion (LD) increased in untreated mice. Comparison amongst these diffusion maps indicates that MD and RD show complementary information, as do FA and LD. T-test was applied to compare treated and untreated mice at each week, but there was no statistically significant difference, possibly because the sample size for this study was too small. The pattern evident in the histograms may be significant if more data is acquired to increase the sample size. This study provides new insight to the relevance of using diffusion and T2 values as potential biomarkers for treatment response