The move towards higher static magnetic field strengths in MRI has allowed improved imaging quality from increased signal to noise ratio. However challenges have arisen from increased inhomogeneity in the radio frequency (RF) fields required to create MR signals and greater RF energy deposition – known as the specific absorption rate (SAR) – within imaging subjects. These factors have prompted the use of parallel transmission (PTx) MRI; in which multiple independent channels are used to control the RF electromagnetic fields. In this thesis the aim was to develop methods for controlling SAR using PTx and to assess the impact of RF safety in various scenarios. The electromagnetic behaviour of an 8-channel PTx RF coil was fully simulated which enabled the examination of differences between full simulations and a commonly modelled idealised situation. It was found that large discrepancies could result in the idealised model in certain situations. The full RF coil model was for producing SAR simulations of various adult male voxel models. These SAR models were used to perform RF shimming, in which a complex weighting is applied to each channel of a PTx system to yield improved RF conditions. This was done for two scenarios: to perform lower SAR cardiac MRI with greater RF field homogeneity in vivo for optimised imaging; and to explore methods for decoupling the transmit coil from a simulated prosthetic hip implant embedded within an adult male whilst still producing a uniform imaging field. In both scenarios, reduced SAR configurations could be found that enabled improved imaging with greater RF safety. A separate model of a 2-channel birdcage RF coil was developed to assess SAR deposition in neonates during MRI examinations. It was found that under normal operation at 3 T, local SAR constraints produced by the scanner are conservative by a factor of four.

Dynamic contrast enhanced (DCE) cardiac magnetic resonance imaging (MRI) is well-established as a non-invasive method for qualitatively detecting obstructive coronary artery disease (CAD) which can impair myocardial blood flow and may result in myocardial infarction. Mathematical modelling of cardiac DCE-MRI data can provide quantitative assessment of myocardial blood flow. Quantitative assessment of myocardial blood flow may have merit in further stratification of patients with obstructive CAD and to improve the diagnosis and prognostication of the disease in the clinical setting. This thesis investigates the development of a quantitative analysis protocol for cardiac DCE-MRI data. In the first study presented in this thesis, Fermi and distributed parameter (DP) modelling are compared in single bolus versus dual bolus analysis. For model-based myocardial blood flow quantification, the convolution of a model with the arterial input function (i.e. contrast agent concentration-time curve extracted from the left ventricular cavity) is fitted to the tissue contrast agent concentration-time curve. In contrast to dual bolus DCE-MRI protocols, single bolus protocols reduce patient discomfort and acquisition protocol duration/complexity but, are prone to arterial input function saturation caused in the left ventricular cavity by the high concentration of contrast agent during bolus passage. Saturation effects can degrade the accuracy of quantification using Fermi modelling. The analysis presented in this study showed that DP modelling is less dependent on arterial input function saturation than Fermi modelling in eight healthy volunteers. In a pilot cohort of five patients, DP modelling detected for the first time reduced myocardial blood flow in all stenotic vessels versus standard clinical assessments. In the second study, it was investigated whether first-pass DP modelling can give accurate myocardial blood flow, against ideal values generated by numerical simulations. Unlike Fermi modelling which is convolved with only the first-pass range of the arterial input function, DP modelling is convolved with the entire contrast agent concentration-time course. In noisy and/or dual bolus data, it can be particularly challenging to identify the end point of the first-pass in the arterial input function. This study demonstrated that contrary to Fermi modelling, myocardial blood flow analysis using DP modelling does not depend on the number of time points used for fitting. Furthermore, this data suggests that DP modelling can reduce the quantitative variability caused by subjectivity in selection of the first-pass range in cardiac MR data. This in turn may help to facilitate the development of more automated software algorithms for myocardial blood flow quantification. In the third study, Fermi and DP modelling were compared against invasive clinical assessments and visual MR estimates, to assess their diagnostic ability in detecting obstructive CAD. A single bolus DCE-MRI protocol was implemented in twentyfour patients. In per vessel analysis, DP modelling reached superior sensitivity and negative predictive value in detecting obstructive CAD compared to Fermi modelling and visual estimates. In per patient analysis, DP modelling reached the highest sensitivity and negative predictive value in detecting obstructive CAD. These studies show that DP modelling analysis of cardiac single bolus DCE-MRI data can provide important functional information and can establish haemodynamic biomarkers to non-invasively improve the diagnosis and prognostication of obstructive CAD.

T1 mapping of myocardial tissue is important for diagnostics of myocardial fibrosis. Cardiac magnetic resonance imaging of small animals is challenging due to high heart and respiratory rates. Pulse sequences for T1 mapping are proposed in this thesis based on inversion recovery FLASH and Intragate FLASH. The sequence IR FLASH was compared to the reference sequence RARE on a static phantom. Both sequences were applied for measuring the myocardium of a rat. For T1 quantification a software in Matlab was developed. Using this software, T1 maps of rat myocardium were calculated.

Improving imaging hardware, computational power, and algorithmic design are driving advances in interventional medical imaging. We lay the groundwork here for more effective use of machine learning and image registration in clinical electrophysiology. To achieve identification of atrial fibrosis using image data, we registered the electroanatomic map (EAM) data of atrial fibrillation (AF) patients undergoing pulmonary vein isolation (PVI) with MR (n = 16) or CT (n = 18) images. The relationship between image features and bipolar voltage was evaluated using single-parameter regression and random forest models. Random forest performed significantly better than regression, identifying fibrosis with area under the receiver operating characteristic curve (AUC) 0.746 (MR) and 0.977 (CT). This is the first evaluation of voltage prediction using image data. Next, we compared the character of native atrial fibrosis with ablation scar in MR images. Fourteen AF patients undergoing repeat PVI were recruited. EAM data from their first PVI was registered to the MR images acquired before the first PVI (‘pre-operative’) and before the second PVI ('post-operative' with respect to the first PVI). Non-ablation map points had similar characteristics in the two images, while ablation points exhibited higher intensity and more heterogeneity in post-operative images. Ablation scar is more strongly enhancing and more heterogeneous than native fibrosis. Finally, we addressed myocardial measurement in 3-D echocardiograms. The circular Hough transform was modified with a feature asymmetry filter, epicardial edges, and a search constraint. Manual and Hough measurements were compared in 5641 slices from 3-D images. The enhanced Hough algorithm was more accurate than the unmodified version (Dice coefficient 0.77 vs. 0.58). This method promises utility in segmentation-assisted cross-modality registration. By improving the information that can be extracted from medical images and the ease with which that information can be accessed, this progress will contribute to the advancing integration of imaging in electrophysiology.