Prognostic Value of Optimized Dynamic Contrast-Enhanced Magnetic Resonance Imaging of High-Grade Gliomas

Abstract

This thesis investigates several issues in DCE-MRI of high-grade gliomas for the purpose of optimizing the sequence for increased reproducibility of the estimated kinetic parameters. The effect of reducing the temporal resolution and total acquisition time was systematically investigated using simulations and clinical data. Furthermore, the necessity of a baseline T1 measurement for accurate estimation of the evolution of the kinetic parameters was assessed by simulations and clinical data using different measurement strategies. Finally, the evolution of DCE-MRI and DSC-MRI in glioblastoma patients during radiochemotherapy treatment and their ability to predict overall survival was compared to structural images. Our results show that a temporal resolution of 10 s and a total scan time of 6 min are adequate for accurate parameter estimation. Additionally, a T1 measurement is not needed to assess changes in median tumor values but histogram distribution is affected by a set T1 error. Lastly, DCE-MRI was more robust than structural MRI and DSC-MRI for early overall survival prediction but combining several modalities increases the predictive value of early MRI.