Use this URL to cite or link to this record in EThOS:
Title: Functional imaging markers for tumour characterisation
Author: Tanner, Lydia Nathania
ISNI:       0000 0004 2721 9774
Awarding Body: Oxford University
Current Institution: University of Oxford
Date of Award: 2011
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
There is rapidly increasing interest in functional imaging for the diagnosis and monitoring of disease. In particular, Dynamic Contrast Enhanced (DCE) Mag- netic Resonance Imaging (MRI) is of enormous, and growing importance in medical imaging, fundamentally because it provides a powerful tool for assessing tumour vasculature. Furthermore it has been proposed as a means of moni- toring disease progression and for identifying potential markers for predicting patient response to therapy. In principle, DCE-MRI data can be quantified us- ing pharmacokinetic models, enabling extraction of physiologically meaningful parameters. Although the combination of DCE-MRI sequencing and pharmacokinetic mod- elling promises quantitative analysis of response, as well as providing the clinician with parametric information (such as Ktrans and kep), there are a number of fun- damental problems to be addressed. When we applied the published models to clinical colorectal cancer MRI data sets, we observe that many voxels within the tumour volume fail to produce identifiable pharmacokinetic information. We ad- dress the root causes of this problem as a necessary precursor to moving beyond calculating average pharmacokinetic values for a region of interest to observing and quantifying tumour heterogeneity. We find a high dependence of the pharmacokinetic model on uncorrected motion in the MR images, and on estimation of the pre-contrast Tl tissue relaxation time. We quantify the effects of each of these, and derive a more reliable means of generating concentration curves, necessary as a precursor to pharmacokinetic modelling. Next, we look at the standard two compartment model for DCE-MRI. While the functional form of the compartment model is known, computation of the amount of contrast agent depends critically upon knowing the arterial input function. We demonstrate a high dependence of the pharmacokinetic model on the choice of arterial input function. We quantify the effects of model and population- averaged parameter choice and find that the standard image based methods used for individual arterial input function extraction, based on identifying an artery or reference region, are unsuitable on clinical data whose temporal resolution is low. We propose an alternative method to extract the plasma concentration from the tissue of interest. We assess the new method on both simulated and clinical data; the results show a more robust estimation of the pharmacokinetic parameters. The literature suggests DCE-MRI has applications in predicting the response of tumours to neoadjuvant chemoradiotherapy. For colorectal cancer, only 70% of patients demonstrate response to therapy with 10-20% responding completely. By clustering parameters across patients to infer blood flow patterns common for different subsets of tumour, we attempt to find a suitable description of.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available