The prognostic value of advanced MR in gliomas

Kenning, Lawrence

May 2014

Thesis or dissertation

© 2014 Lawrence Kenning. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.

This work examines the prognostic value of advanced MR at selected time points during the early stages of treatment in glioma patients. In this thesis, serial imaging of glioma patients was conducted using diffusion tensor imaging (DTI), dynamic contrast enhanced (DCE) and dynamic susceptibility contrast (DSC) MRI. A methodology for the processing and registration of multiparametric MRI was developed in order to simultaneously sample whole tumour measurements of multiple MR parameters with the same volume of interest.

Differences between glioma grades were investigated using functional MR parameters and tested using Kruskal-Wallis tests. A 2-stage logistic regression model was developed to grade lesions from the preoperative MR, with the model retaining the apparent diffusion coefficient, radial diffusivity, anisotropic component of diffusion, vessel permeability and extravascular extracellular space parameters for glioma grading. A multi-echo single voxel spectroscopic sequence was independently investigated for the classification of gliomas into different grades.

From preoperative MR, progression-free survival was predicted using the multiparametric MR data. Individual parameters were investigated using Kaplan-Meier survival analysis, before Cox regression modelling was used for a multiparametric analysis. Radial diffusivity, spin–lattice relaxation rate and blood volume fraction calculated from the DTI and DCE MRI were retained in the final model.

MR parameter values were also investigated during the early stages of adjuvant treatment. Patients were scanned before and after chemoradiotherapy, with the change in MR parameters as well as the absolute values investigated for their prognostic information. Cox regression analysis was also performed for the adjuvant treatment imaging, with measures of the apparent diffusion coefficient, spin–lattice relaxation rate, vessel permeability and extravascular extracellular space, derived from the DTI and DCE datasets most predictive of progression-free survival.

In conclusion, this thesis demonstrates multiparametric MR of gliomas during the early stages of treatment contains useful prognostic information relating to grade and progression-free survival interval.

Postgraduate Medical Institute, The University of Hull
Lowry, Martin (Biochemist); Pickles, Martin Darren
Sponsor (Organisation)
Yorkshire Cancer Research Campaign
Qualification level
Qualification name
15 MB
QR Code