Pre-operative Functional Magnetic Resonance Imaging: Application to Tumour Resection by Awake Craniotomy

The management of brain tumours in eloquent areas remains challenging. Surgical intervention is currently the primary method of treatment, with the surgeon attempting to maximize tumour resection while simultaneously sparing healthy cortex. Pre-operative and intra-operative brain mapping techniques are capable of localizing eloquent cortex, and are thus employed to facilitate the surgical decisions. Although direct cortical stimulation (DCS) remains the gold standard intra-operative brain mapping technique, non-invasive imaging tools like functional magnetic resonance imaging (fMRI) have become increasingly popular at the pre-operative planning stage. In this thesis, technical work has been done to improve, characterize, and provide perspective on appropriate usage of pre-operative fMRI in brain tumour surgery. The thesis encompasses three specific research studies. The first, entitled “A novel tablet computer platform for advanced language mapping during awake craniotomy procedures”, describes the development of a platform that enables behavioural testing to be standardized with respect to pre-operative fMRI, and enables sophisticated testing paradigms to be administered during DCS. The testing platform is demonstrated to provide improved language mapping capabilities over current practice. The second project, entitled “Reliability of task-based fMRI for pre-operative planning: a test-retest study in brain tumour patients and healthy controls”, shows the impact of disease and technical factors on the reproducibility of fMRI activation maps. Reliability of fMRI data is shown to be comparable in low-grade glioma (LGG) patients and healthy controls, but poorer in high-grade glioma (HGG) patients. More reliable activation patterns are demonstrated for motor tasks than for language tasks. The third project explores “Sources of variation influencing concordance between functional MRI and direct cortical stimulation in neurosurgery”. This work shows the substantial extent of false positive (FP) activations that are present in fMRI results compared to DCS, the inverse relationship between fMRI sensitivity and specificity with changing statistical threshold, and that concordance is improved when behavioural tasks are standardized across fMRI and DCS. Collectively, this thesis supports that pre-operative fMRI is a suitable adjunct to DCS for surgical decision-making. The experimental findings demonstrate that the reproducibility of fMRI data and its concordance with DCS are generally good, though directly related to the quality of fMRI activation maps. Actions can be taken to optimize fMRI data quality with good understanding of the clinical and technical factors that are influential, such that neurosurgeons can interpret activation maps more reliably.

Department dc:contributor.department

Medical Biophysics

Year dc:date.issued

2016

Author dc:creator

• Morrison, Melanie Anne

Advisor dc:contributor.advisor

• Graham, Simon J

• awake craniotomy
• direct cortical stimulation
• functional magnetic resonance imaging
• glioma
• language mapping
• reliability

Handle dc:identifier.uri

http://hdl.handle.net/1807/76580

OAI identifier oai:identifier

oai:utoronto.scholaris.ca:1807/76580