Classification of breast malignancy using optimised advanced diffusion-weighted imaging : and surgical planning for breast tumour resection using MR-guided focused ultrasound
Purvis, Nina Louise
Thesis or dissertation
- © 2016 Nina Louise Purvis. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.
Intravoxel Incoherent Motion Imaging (IVIM) is a non-invasive MR-imaging technique that enables the measurement of cellularity and vascularity using diffusion-weighted (DW)-imaging. IVIM has been applied to various cancer types including breast cancer, and is becoming more popular but lacks standardisation. The quantitative parameters; diffusion, D, perfusion fraction, f, and pseudo micro capillary diffusion, D* are thought to be correlated with tumour physiognomies such as proliferation, angiogenesis and heterogeneity.
In Part 1 of this thesis, an optimised clinical b-value protocol is produced using a robust statistical method. This optimised protocol and various fitting methodologies are investigated in healthy volunteers, and then the most precise approach is applied in a clinical trial in patients following diagnosis of breast cancer, before treatment, to correlate IVIM parameters with breast cancer grade, histological type and molecular subtype with statistically significant results supporting IVIM’s potential as a non-invasive biomarker for malignancy. Monte Carlo simulations support this clinical application, where real data mean squared errors due to SNR limitations lie within simulated errors. A computed DW-imaging program is also presented to produce better quality images than acquired high b-value images as an adjunct to the optimised IVIM protocol.
In Part 2 of this thesis, MR-guided Focused Ultrasound (MRgFUS) is explored as a means to create a pre-surgical template of thermally induced palpable markers to enable a surgeon to resect occult lesions and potentially reduce positive tumour margin status and local recurrence after breast conserving surgery. A surrogate animal model with pseudo lesion is presented, as well as a clinical tool to plan spot markers around a lesion as seen on MRI.
- Hull York Medical School, The University of Hull and University of York
- Gibbs, Peter; Pickles, Martin Darren
- Sponsor (Organisation)
- Yorkshire Cancer Research Campaign
- Qualification level
- Qualification name
- 40 MB