Introduction: Low-grade gliomas are a sub-group of primary brain tumours that typically affect young adults and which present specific challenges to conventional diagnostic imaging. They demonstrate a pattern of growth whereby tumour cells infiltrate healthy brain tissue without distortion of the surrounding brain or blood-brain barrier integrity. These features limit the capacity of conventional neuro-imaging strategies to effectively delineate the tumour extent or characterise the degree of 'malignancy'. One solution is to apply multiple imaging modalities to image different aspects of the tumour behaviour, analogous to histological classification based upon changes in mitotic activity, cellular atypia, microvascular proliferation and necrosis. Published information regarding how imaging techniques that address these parameters correlate within the tumour volume is limited. This reflects the technical challenges in acquiring and processing data at an adequate spatial resolution to characterise small but heterogenous tumours. In this thesis, following a series of experiments seeking to optimise the sensitivity and reproducibility of PET analysis in gliomas, a prospective multi-modal neuro-imaging study is presented addressing this need. Methods: Retrospective [11C]-methionine PET (MET PET) data made available through a collaboration with the Max-Planck Institute for Neurological Research in Cologne was carried out first to address the optimal method of analysis of PET data in gliomas. A normal methionine uptake map was created and its use in the analysis of patient scans validated against a conventional approach. Automated methods for delineating the extent of abnormal methionine uptake and identifying the region of peak uptake were developed and evaluated to optimise the reproducibility of the approach. High-resolution MET PET and a comprehensive MRI brain tumour protocol were then acquired prospectively in 20 subjects in Manchester. Detailed analysis of the peak uptake and extent of abnormal tissue defined using PET and MRI modalities including structural, diffusion- and perfusion-weighted techniques was performed. Results: Evaluation of methionine uptake with respect to population normal data, the 'RatioMap' technique, yielded peak uptake measurements that correlated closely with a conventional approach (r = 0.97) but with improved reproducibility. The constrained 3D region-growing algorithm designed to delineate the abnormal region was shown to be reproducible and to generate volumes that correlated with tumour grade. High-resolution multi-modal data in suspected low-grade gliomas demonstrated consistent correlation between peak methionine uptake ratio and peak regional cerebral blood volume (r = 0.85) but with disparity between the location of the maximal uptake regions (mean distance = 11.2mm). Significant correlation was seen between multi-modal MRI and PET ‘tumour’ volumes (r = 0.91) but with substantially larger MRI defined abnormal volumes (ratio = 2.0) including small regions identified as abnormal by multiple MRI parameters but normal on PET imaging. Conclusion: A novel method to enhance the reproducibility of analysis of MET PET images in gliomas has been presented and validated but there remains no single imaging modality capable of fully characterising glioma extent and 'malignancy' non-invasively. Considerable correlation between PET and MRI tumour biomarkers has been demonstrated but there are significant differences between the regions identified as the 'most malignant' for biopsy targeting and the extent of potentially tumour bearing tissue. Combined use of diffusion- and perfusion-weighted MRI parameters can provide results very closely correlated to the PET findings but cannot yet completely replace the use of nuclear medicine techniques. The use of multi-modal approaches to tumour characterisation as demonstrated in this study provides the most effective currently available approach to fully characterise a suspected glioma.