Search

Tumour hypoxia is linked to reduced treatment sensitivity and poor outcome. Oxygen-enhanced MRI (OE-MRI) is a novel technique with the potential to non-invasively image hypoxia. It involves the measurement of MR relaxation rate changes caused by changes in oxygenation within blood and tissue. This thesis is focused on the development of OE-MRI and its clinical application in tumours. Two imaging sequences were optimised and then investigated for accuracy and precision in measuring oxygen enhanced R1 changes. Following that, R1, R2* and blood flow were measured during air and oxygen breathing in healthy kidneys. This study included the implementation of model-driven registration for inversion prepared OE-MRI images. Next, a biophysical model was used to simulate expected ΔR1 in tumour tissue during hyperoxia for a range of underlying physiological conditions. Finally, OE-MRI was applied alongside BOLD and DCE-MRI in clinical studies in renal and cervical tumours. It was shown that a dynamic inversion-prepared HASTE sequence in combination with a baseline T1 map offered the most precise measure of oxygen-enhanced ΔR1. No consistent significant change was found in either T1, R2* or blood flow when using static measures, but significant changes in R1 were found by using dynamicT1-weighted signal. Image registration had variable effect on dynamic signal curves, but reduced errors in T1 calculations. Modelling showed that differences in ΔR1 are mainly caused by the interplay between blood flow, oxygen consumption and capillary length. Finally, clinical studies demonstrated the feasibility of OE-MRI in tumours. The work in this thesis represents an improvement of OE-MRI methodology and offers a new insight into how results from OE-MRI relate to both underlying tumour physiology and other imaging modalities.