Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.688350
Title: Automatic brain tumour detection and segmentation using tissue substructure features derived from MRI diffusion tensor imaging
Author: Alderson, William
ISNI:       0000 0004 5917 4849
Awarding Body: University of Bristol
Current Institution: University of Bristol
Date of Award: 2016
Availability of Full Text:
Access through EThOS:
Abstract:
The incidence of brain cancer has increased with the increase in average life span. To help combat this rise new diagnostic aids, such as Magnetic Resonance Imaging (MRI) and Computed tomography (CT) scans, have enabled clinicians to undertake non-invasive evaluation of biological tissue and to present this information in detailed 2D images and 3D volumes. Unfortunately the growing demand for clinical services is imposing great cost on health services and is highlighting the need for more highly skilled clinicians. The introduction of a novel automated computer-based brain tumour detection and segmentation process offers the potential to reduce cost, shorten waiting times and to provide tools to reduce the workload of clinicians. Magnetic Resonance Imaging (MRI) three dimensional diffusion data can be post-processed, in real time, to improve brain tumour detection and visualisation. In this thesis we present novel sub-structure features derived from MRI diffusion tensor imaging scans that go beyond conventional tractography, and show how automatic algorithms lead to tumour localisation and boundary identification. The effectiveness of this method is shown through the automatic localisation and candidate tumour boundary detection for five patients. Preliminary results from this computer based process show good correlation with the traditional visual inspection of images method by clinical experts. Further analysis of the tissue diffusion sub-structure features is expected to lead to improved non-invasive structural identification around the tumour. We compare tumours boundaries detected by our process with traditional image scans and discuss how this detection method of tumours may provide useful additional information above that which can be elicited by visual inspection of the scans.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.688350  DOI: Not available
Share: