Search

Aortic stenosis is a major cause of morbidity and mortality in the western world. Whilst symptomatic patients with severe AS are treated with valve replacement surgery, there remains uncertainty about the best treatment for asymptomatic severe AS. Current guidelines recommend wait and watch strategy and offer surgery on occurrence of symptoms or evidence of systolic dysfunction. The annual rate of sudden cardiac death in asymptomatic severe AS remains close to 1%. Aortic disease is both a disease of the valve and myocardium. The valve mechanisms and structural LV changes secondary to pressure overload have been extensively studied before, but no medical therapy directed at the valve processes has so far shown beneficial effect on cardiac physiology or disease progression. Changes in cardiac metabolism in pressure overload hypertrophy seem to play an important role in pathophysiology of AS and transition to decompensation. The healthy human heart uses fatty acids as the main energy source ~70% of adenosine triphosphate (ATP) requirements. In AS, there is substrate shift with downregulation of fatty acid oxidation (FAO) and increased reliance on glycolysis, with evidence of myocardial lipid accumulation and impairment of myocardial energetics. Peroxisome proliferator activated receptor-alpha (PPAR-a) plays a central role in FAO signalling and controlling lipid homeostasis in the heart. Downregulation of this transcriptional factor is associated with the metabolic alterations and subsequent lipotoxicity in AS, which ultimately causes reduced ATP production and heart failure. However, it is unclear whether these metabolic changes have a cause or effect relationship with the two main pathological features of AS, left ventricular (LV) pressure loading and LV hypertrophy, and their relationship with myocardial fibrosis (which is commonly seen in AS), is unclear. Furthermore, it is yet unknown whether modulating cardiac metabolism would have any beneficial effect on disease progression or outcomes in AS. This thesis set out to establish the relationship between metabolic remodelling and cardiac structure and physiology in AS. It then evaluates the role of novel metabolic modulator therapy in asymptomatic moderate-severe AS. In Chapter 3 and 4, patients across the spectrum of AS were studied with advanced CMR imaging and phosphorus-31 magnetic resonance spectroscopy (31P-MRS) and cardiopulmonary exercise testing (CPET) respectively. It is demonstrated that metabolic changes, specifically myocardial lipid accumulation (steatosis) and impaired cardiac energetics appear to occur early in the disease process. In addition, whilst impairment in cardiac energetics is related to degree of LV hypertrophy, steatosis appears to be more related to the degree of LV pressure loading from valve obstruction. Furthermore, these patients despite being asymptomatic with normal LVEF have evidence of subclinical LV dysfunction which occurs early in the disease process. Moreover, these patients though able to exercise to volitional exhaustion without developing symptoms have reduced average peak VO2 and VE/VCO2, both of which are predictors of outcome in cardiac disease. In Chapter 4 and 5, the effect of the PPARa agonist Fenofibrate was evaluated on cardiac metabolism and physiology in a randomised double-blind placebo-controlled study. Fenofibrate reduced myocardial triglyceride content significantly after 6 months' treatment with evidence of in vivo fatty acid oxidation upregulation. It also caused a modest improvement in cardiac energetics. However, this modulation did not show any measurable improvement in cardiac physiology or exercise capacity. Together, these data show that myocardial metabolic remodelling plays an important role in pathophysiology and progression of AS. The novel finding from this research is that metabolic changes occur early in the AS disease process and are associated with early subclinical systolic and diastolic dysfunction. This highlights that metabolic remodelling may play a causal role in disease progression. This research has also shown, for the first time, that metabolic alterations in AS are amenable to modulation with a PPARa agonist, with some improvement in cardiac physiology but the benefit of such modulation in advanced disease remains questionable. Whether targeting earlier disease would yield different results remains unanswered.

Heart failure (HF) remains the leading cause for hospitalisation in patients over the age of 65 years in industrialised nations. Furthermore, prognosis of HF remains poor and lags behind the improvements observed for similar chronic conditions like cancer, despite efforts for new therapies. However, these therapies are mostly available for patients with HF and a reduced ejection fraction (HFrEF), whereas prognostically relevant therapies for patients with HF and a preserved ejection fraction (HFpEF) have been absent for more than 20 years and this only changed very recently with the results of the EMPEROR-preserved trial investigating treatment with empagliflozin (10 mg once daily) in patients with HFpEF. Equally, a significant improvement of quality of life and physical limitations has been shown for patients with HFpEF following treatment with dapagliflozin in PRESERVED-HF. An increasing amount of evidence describes more subtypes of the complex and rather arbitrarily classified syndrome of HF and thus, there is a growing need for comprehensive, non-invasive imaging techniques to identify and individually phenotype patients affected. Furthermore, the same approach could be used to assess novel treatments as well as identify treatment responders, further tailoring individual patient management and aid efforts to improving prognosis. Additionally, use of a model imaging technique may lower the exponential costs of drug development by reducing the numbers of subjects needing to be enrolled and increasing reproducibility of findings. Cardiovascular magnetic resonance (CMR) fulfils many of the demands for a model imaging technique in the context of HF-phenotyping and -drug development. Firstly, CMR is the gold standard method to assess cardiac structure and function. More importantly, it allows examination of imaging and metabolic parameters alike and, with novel techniques like dynamic nuclear polarisation (i.e. hyperpolarized MR spectroscopy), even direct assessments of molecular cellular pathways. Thirdly, HF is a syndrome affecting multiple organ systems and thus, has complex and incompletely understood links in need of exploration. MR offers the possibility of multi-organ assessment in the same session and therefore, is perfectly placed to gather information on the biological interplay in HF. The underlying mechanism for the observed benefits of sodium glucose like transporter 2 inhibitors (SGLT2i) like empagliflozin in HFrEF are not yet understood. One of the leading hypotheses brought forward was that SGLT2i may induce changes in myocardial substrate selection via increased availability of ketone bodies which may serve as an additional energy source for the failing heart. Thus, I sought to investigate whether empagliflozin treatment (10mg per day) would enhance myocardial energetics in patients with HFrEF (phosphocreatine to adenosine triphosphate ratio, PCr/ATP). In this randomised controlled trial (RCT), patients did not improve measures of resting or dobutamine stress myocardial energetics (Chapter 3) and empagliflozin treatment neither altered a panel of 19 serum metabolites assessed by targeted metabolomic analysis. However, there were interesting changes to the amount of triglycerides stored in the heart, myocardial cell volume and hypertrophy as well as markers of fibrosis and quality of life (QoL). This exemplifies the usefulness of CMR in the interrogation of novel treatments and likewise emphasises that the broad range of indications for SGLT2i (diabetes, chronic kidney disease, heart failure) is possibly mirrored by a multi-organ rather than a single cardio-specific effect in patients with HFrEF. As it was unclear until very recently if patients with HFpEF would equally benefit from treatment with SGLT2i, I further integrated a HFpEF cohort into the RCT and investigated the identical variety of parameters as for the HFrEF patients (Chapter 4). Similarly to the findings in the HFrEF cohort (Chapter 3), I could not observe any changes to myocardial energetics (resting PCr/ATP as primary endpoint) or whole-body substrate usage (targeted metabolomics). Interestingly, I detected certain similarities nevertheless as myocardial triglycerides reduced in the treatment arm but not in the placebo group. Furthermore, systolic function (peak circumferential and radial strain) and measures of pulmonary function improved. This was also paralleled by a numerically improved walking distance in the six-minute walk test (6MWT) and increased QoL (assessed by the Kansas City Cardiomyopathy Questionnaire; KCCQ). The present results reinforce the need for mechanistic experiments in human subjects in-vivo, as they highlight the disparity of results from animal models that previously showed a metabolic effect following treatment with SGLT2i. Furthermore, more research is needed into effects of drugs in different HFpEF subtypes (e.g. diabetic, obese) to improve risk stratification and adequate treatment. To explore novel, non-pharmacological treatments targeting unique subgroups under the 'HFpEF-umbrella', I aimed to explore whether a lifestyle intervention of weight loss is an effective treatment for patients with obesity and HFpEF (Chapter 5). Accordingly, I enrolled patients with a clinical diagnosis of HFpEF and observed cardio-metabolic effects following 10 weeks of a very low energy diet (VLED). Assessments included exercise CMR and dobutamine stress phosphorus magnetic resonance spectroscopy (31P-MRS). While cardiac energetics at rest or during dobutamine stress did not change, prognostic serum markers (n-terminal pro-BNP; NT-proBNP) and symptom burden (New York Heart Association; NYHA) improved significantly and cardiac structure (LV-mass) as well as function (diastolic function on echo and RV function at rest and during exercise) ameliorated. The results emphasise that lifestyle treatments in select patient groups are underutilised yet cost-efficient, safe and successful. Furthermore, the obesity paradox (describing the observation that mild obesity appears to be protective in HF) may not equally apply to the group of severely obese HFpEF patients (BMI >30 kg/m²) and thus, this should be investigated in larger trials. Finally, I was interested to examine the feasibility of CMR in the context of early (phase IIa) cardio-metabolic drug development. In Chapter 6, I enrolled 22 patients with type 2 diabetes (T2D) but no underlying HF and investigated the effects of ninerafaxstat, a novel drug intended to enhance myocardial substrate metabolism via restoring metabolic flexibility in the heart. Here I show that the drug's proposed mechanism of action can indeed be successfully investigated by using [1-13C]pyruvate hyperpolarized MRS. The pyruvate dehydrogenase (PDH) flux was improved in the majority of subjects (7/9). Furthermore, this resulted in significantly improved PCr/ATP and reduced myocardial steatosis. Diastolic function and early LV-filling also improved following drug treatment for 4 or 8 weeks. Based on these results, hyperpolarized MRS is a suitable tool to investigate the mechanism of action in early drug development in humans with reduced sample sizes. The broader implication is certainly that, similar to treatments for immune modulation (e.g. canakimumab), patient selection is key for metabolic assessments in clinical trials. To summarise, this Thesis examines cardio-metabolic effects of different pharmacological and lifestyle treatments in distinct populations and further demonstrates the suitability of CMR as an investigational tool for drug development. While the results underpin the importance of substrate selection and implications on cardiac structure and function, improved patient selection is key for demonstrating these effects with statistical significance. Furthermore, refuting the 'fuel hypothesis' theory as the main contributor to SGLT2i's mode of action, I provide evidence for multi-organ effects of SGLT2i being likely responsible for the benefits observed in HF. Secondly, I add to the existing evidence that intentional weight loss is a safe and effective treatment in obese HFpEF patients.

Background: Coronary plaque formation, progression and rupture are still the major cause for myocardial infarction. Although many ways for the identification of high-risk plaques have been developed and evaluated over the last decade, they are still not able to forecast the development and rupture of an individual plaque precisely. With the increase of new techniques such as radiomics and the fat attenuation index (FAI), new risk stratification techniques for coronary plaques could be established. Methods: The first part developed an automated coronary plaque segmentation tool for calcified and non-calcified plaque components. The second part compared statistical versus radiomics-based automated plaque classification approaches. These two parts aimed to facilitate an automated coronary plaque analysis on future big cohorts. In the last part, FAI around plaques (pFAI) was analysed and differences between individual coronary plaque categories were compared. Results: After showing excellent results on intra- and interobserver analysis for the manual segmented dataset, a model for calcified plaque components achieved very good results, even outperforming metrics from the interobserver analysis. The radiomics-based approach for non-calcified plaque detection proved the feasibility of this method by showing good initial results. While an automated plaque classification based on two first-order statistical values has already worked well, radiomics were able to even outperform this in terms of accuracy and diversification of plaque types. Measured pFAI revealed significant differences between individual plaque classes based on their degree of calcification. Discussion: Although the size of the manual labeled and segmented datasets was limited, meaningful models for an automated coronary plaque segmentation and classification were developed. Further, it was shown that FAI around different plaque types provides additional information on their inflammatory status and, therefore, eventually on their individual risk, which has to be investigated on plaque-specific outcome data in the future.

Cardiovascular disease (and particularly coronary artery disease and ischaemic heart disease) remains the leading cause of morbidity and mortality worldwide. There is increasing need for non-invasive imaging tests and biomarkers that could allow early therapeutic interventions and help risk-stratify and prognosticate to improve individual outcomes. Cardiovascular magnetic resonance (CMR) is a powerful non-invasive and radiation-free imaging tool that can comprehensively evaluate cardiac structure, function, and ischaemia, and allows advanced myocardial tissue characterisation. Stress T1-mapping is a promising contrast-free CMR application that assesses coronary vasodilatory reserve and has shown utility in the non-invasive assessment of cardiovascular disease during adenosine vasodilator stress. However, this field has not progressed beyond original proof-of-principle studies, is limited by method sensitivity and choice of pharmacological stress agent, and has not been validated against gold standard invasive coronary physiology. This Thesis addresses these issues by performing further clinical validation and method development work towards standardising the technique across a range of T1-mapping methods and stress agents. I first built on the existing adenosine stress T1-mapping proof-of-principle groundwork by performing further clinical validation. This was done in a cohort of patients with end-stage renal disease (ESRD) referred for kidney transplantation, who are most likely to benefit from a non-invasive and contrast-free imaging assessment for cardiovascular disease. I demonstrated for the first time that stress T1-mapping has good diagnostic performance for the detection of obstructive coronary artery disease (CAD) in patients with ESRD, as validated against invasive coronary physiology in a systematic study. I next explored the feasibility of stress T1-mapping using alternative pharmacological stress agents. I demonstrated that regadenoson (a 'coronary-specific', selective A2A receptor agonist vasodilator) is a viable alternative for stress T1-mapping applications by modelling regadenoson stress T1 dynamics over time in healthy volunteers. I also directly compared the performance and variability of commonly used T1-mapping methods (ShMOLLI and MOLLI variants) for stress and post-contrast T1-mapping. I showed that ShMOLLI stress T1-mapping has the strongest relationship with quantitative myocardial blood flow and has a greater effect size with less variability compared to the other T1-mapping methods tested. I then applied this in a cohort of patients with CAD and showed that regadenoson stress T1-mapping has similar clinical efficacy to adenosine for detecting obstructive CAD and differentiating between normal, remote, ischaemic, and infarcted myocardium. Finally, I demonstrated the suitability of alternative pharmacological stress agents (dobutamine and dipyridamole) for future stress T1-mapping applications, which included successful translation of the technique to an external international centre. I showed that dobutamine and dipyridamole are both viable and potentially cost-effective stress agents to use for stress T1-mapping. Overall, this Thesis reports on further development of the highly promising and novel stress T1-mapping technique for the assessment of cardiovascular disease without contrast agents, including clinical validation against invasive coronary physiology, clinical application in patients with CAD, and method comparison across a range of commonly used pharmacological stress agents and T1-mapping techniques.

The healthy human heart uses fat and glucose as its predominant metabolic substrates, approximately in a ratio of 70% to 30%. In heart failure, substrate oxidation is decreased as is ATP production. Free fatty acid uptake and oxidation is reduced, while glucose uptake and glycolysis are increased, although pyruvate oxidation falls. This thesis set out to establish the links between myocardial substrate metabolism and function in the healthy and the failing human heart, at rest and at stress. In chapter 3, an inhibitor of lipolysis, nicotinic acid, is compared to a lipid infusion in normal human subjects and in rats. While the lipid infusion increases cardiac contractility, a decline in both contractility and PCr/ATP ratio is shown with the nicotinic acid. This is revealed to result from inhibition in lipolysis combined with reduction in glucose stimulated insulin secretion, hence the heart is starved of energetic substrate, revealing metabolism's importance and close link to contractile function. In chapters 4 and 5, lipid infusion is further shown in the failing human heart to increase contractility and PCr/ATP ratio at rest as well as increasing maximum available cardiac work. It is demonstrated that the failing heart displays substrate flexibility with an ability to take up higher amounts of free fatty acids when their availability is increased and increase their uptake during rapid pacing stress. In addition, the failing heart is not shown to be oxygen starved and oxygen consumption increases during the lipid infusion without deleterious consequence. In chapter 6, in normal human subjects, vasodilator stress with glyceryl trinitrate is shown to greatly increase myocardial ATP demand while inhibiting mitochondrial ATP production, creating a mismatch between supply and demand and resulting in a fall in PCr/ATP ratio. A higher baseline PCr/ATP is shown to be protective against stress induced changes, confirming the importance of PCr/ATP as a target for interventions. Together, these data show myocardial metabolism is an important target for heart failure therapeutics and that it may be manipulated via increasing availability of free fatty acids.

Cardiovascular magnetic resonance (CMR) imaging is a powerful tool for research and clinical applications. To extract useful clinical information from the acquired CMR images, time-consuming and laborious manual delineation of cardiovascular structures is currently required. Despite promising overall performance across medical imaging applications, the current state-of-the-art automated image segmentation methods still fail in some cases, potentially jeopardising the reliability of clinical diagnosis. Thus, it is important to develop not only automation of image segmentation but also quality control of segmentation, to empower efficient and reliable CMR image data analysis. To address both segmentation and quality control problems, I have developed a novel quality control-driven (QCD) framework in this thesis. Extending upon deep ensemble learning, the framework utilises multiple convolutional neural network-based models to generate segmentation candidates, the agreement of which is exploited via additional regression models to predict segmentation quality measured by Dice similarity coefficient (DSC). The DSC prediction not only provides a quality estimate but also enables a novel approach to select a final, most optimal segmentation on-the-fly from multiple candidates, improving segmentation robustness. Following the DSC prediction, a segmentation quality classification scheme is implemented to alert human operators only when manual intervention is recommended, intended for more efficient allocation of time and labour resources for large-scale image processing pipelines. Through both quantitative and qualitative evaluation, the QCD framework has demonstrated excellent performance in both segmentation and quality control. More importantly, the framework has been successfully applied across CMR imaging techniques, anatomical structures, and large-scale datasets acquired at different sites, with high adaptability and generalisability. The QCD framework could pave the way towards large-scale automated imaging data analysis pipelines, with both efficiency and reliability, in real-world clinical applications.

This DPhil thesis outlines research undertaken in non-alcoholic fatty liver disease (NAFLD). Chapter 1 contains a literature review summarising current concepts in NAFLD including the burden of disease, clinical management including diagnosis and staging, and emerging therapeutic options to treat NAFLD. The first research chapter (Chapter 2) is an analysis of how patients with NAFLD are currently managed in clinical practice, the effectiveness of health interventions and of clinical outcomes measured by using currently available biomarkers. The outcomes and changes in liver and cardio-metabolic health markers of 165 patients with NAFLD who attended a secondary and tertiary metabolic hepatology (NAFLD) clinic at Oxford University NHS Foundation Trust over a median follow-up period of 13 months are presented. The analysis describes the burden of fatty liver disease in its various stages (steatosis, steatohepatitis (NASH), fibrosis, cirrhosis) across the Oxfordshire region. It outlines the use of non-invasive diagnostic and risk stratification tools including hepatic elastography, and how these change longitudinally at follow-up. Finally, as the quest for novel, specifically licenced therapies for NAFLD continues, this study outlines the interventions and approach used in the clinic including an analysis of changes to medical therapy including weight loss promoting agents such as glucagon-like peptide 1 (GLP-1) receptor agonists in patients with NAFLD. The two subsequent research chapters explore two important themes arising from the clinic analysis. Chapter 3 outlines novel research which aims to help meet the need for improved diagnostic and staging tools for NAFLD, and in particular the need for novel non-invasive markers to allow for better risk stratification at diagnosis and for subsequent monitoring of the disease after therapy has been instigated. Drawing on previous work which implicates dysregulated glucocorticoid metabolism as a feature of NAFLD, with specific dysregulation varying across the spectrum of NAFLD, it investigates if the wider steroid metabolome can be used as a novel non-invasive biomarker tool to diagnose and stage NAFLD. This pilot study is conducted in 275 individuals with biopsy proven disease across the spectrum of NAFLD, and explores both specific glucocorticoid metabolic pathways in these individuals as well as use of the 'global urinary steroid metabolome' coupled with machine learning techniques to determine NAFLD stage. The results show excellent promise and suggest that if validated, this technique could have future high utility in clinical practice. Following on from Chapter 1 which revealed that around 20 % of patients with type 2 diabetes attending the NAFLD clinic were commenced on GLP-1 therapy, and that the GLP1 agonist Liraglutide has been previously shown in research studies to be useful in patients with NASH, Chapter 4 investigates GLP-1 agonists as a novel therapy for NAFLD. What hitherto remains unclear is whether the benefit of GLP-1 therapy is solely due to its weight loss actions alone or whether there exist any direct GLP-1 effects on the liver. To seek to answer this question of any weight independent effects of Liraglutide in NASH, this chapter presents data from the 'Lifestyle and Liraglutide in NASH' (LiLi) clinical trial which contributes to the evidence base that may in due course inform medicines licensing agencies to licence GLP-1 agonist therapy as one the first treatments for NAFLD. 29 subjects with NASH were randomised to either Liraglutide or lifestyle intervention to induce weight loss and were followed for a minimum of 12 weeks. Both groups lost equal amounts of weight of approximately 5 % which allowed direct assessment of weight independent benefits. Data arising from metabolic tracer studies (deuterated water and 13C-glucose to detail lipid and glucose handling within the liver), biochemical and body composition data are presented. Data on liver fat and inflammation measured using magnetic resonance spectroscopy and using novel multiparametric magnetic resonance imaging (LiverMultiScanTM, Perspectum Diagnostics) pre and post weight loss intervention are also presented. This study finds that weight loss induced by Liraglutide is of equal efficacy and provides equivalent health benefits to losing weight via dietary lifestyle measures though this is only maintained whilst taking therapy. When considering that clinicians urgently require a range of tools at their disposal to tackle NAFLD and that weight loss and maintenance thereafter is difficult to achieve, this study concludes that GLP-1 agonists could play an important role as a future licenced therapy for NAFLD.

In liver disease, measurements of portal hypertension and liver function can predict clinical outcomes and help guide management, but are either invasive or lack accuracy. This thesis explores magnetic resonance imaging (MRI) techniques with gadoxetic acid administration to assess portal hypertension and liver function non-invasively. This adds to the existing capabilities of multiparametric MRI to assess liver fibrosis, fat and iron. Spleen extracellular volume fraction (ECV) and platelet count to spleen volume ratio identified portal hypertension accurately (both 0.79 area under the receiver operator curve (AUROC)). A combination of spleen T1 and spleen diameter identified oesophageal varices needing treatment (VNTs) in patients with compensated advanced chronic liver disease (cACLD) with the highest accuracy (AUROC 0.87) of all markers tested and could save 66% of screening endoscopies for VNTs without missing any VNTs. Liver T1 reduction with gadoxetic acid (%rT1) correlated well with serum-based liver function markers and clinical scores and identified patients with impaired liver function with AUROC 0.86. MRI-based markers of portal hypertension (spleen ECV; platelet count/spleen diameter ratio; spleen T1 x spleen diameter) and %rT1 were predictive of liver-related clinical outcomes in patients with cirrhosis, who are at high risk of complications, but did not add to the predictive ability of age, albumin, sodium and APRI. In conclusion, MRI provides accurate markers of portal hypertension and liver function. Of particular importance, spleen T1 and spleen diameter criteria could be used to avoid more screening endoscopies in cACLD than any other VNT exclusion criteria reported to date. Although prognostic in the high risk cohort studied here, the MRI markers of portal hypertension and liver function do not surpass the prognostic value of simpler measures.

Chronic liver diseases (CLDs) affect millions worldwide and are a major cause of mortality. Recently, multiparametric magnetic resonance (MR) methods have been proposed as a non-invasive, quantitative biomarker for assessment of CLD. CLD traverses a spectrum of severity, and the aim of this thesis was to apply MR methods across this spectrum for a quantitative assessment of disease severity. The primary variable of interest was the iron and field strength corrected T1 (cT1), generated by LiverMultiScanTM software from T1 and T2* mapping techniques. Liver cT1 was found to predict liver-related clinical outcomes in patients with CLD (p < 0.0001), including in patients with compensated disease only. Liver cT1 performed especially well in patients with non-alcoholic fatty liver disease (NAFLD), alcohol-related liver disease (ArLD) and viral hepatitis (VH) with a cut-off of liver cT1 > 825ms having a negative predictive value (NPV) of 100%. Liver cT1 quantitatively assessed changes in fibroinflammation in 15 patients with chronic hepatitis C virus (HCV) undergoing treatment by direct acting antiviral (DAA) therapy. Liver cT1 showed significant decreases from baseline to 24 weeks after end of treatment (EoT) (876 vs 806 ms, p=0.002), from baseline to 48 weeks EoT (876 vs 788 ms, p=0.0002) and from 24 weeks EoT to 48 weeks EoT (806 vs 788ms, p=0.016). Spleen cT1 had good diagnostic accuracy for presence of portal hypertension in both an adult cohort with CLD and a paediatric cohort with autoimmune liver disease (area under the receiver operating curve (AUC) of 0.85 and 0.81 respectively). In another adult cohort, spleen cT1 was predictive of future episodes of variceal bleeding (AUC: 0.74), especially in those on non-selective beta-blocker treatments (AUC: 0.97). A reference range for spleen T1 and cT1 was established using data from the United Kingdom Biobank population health study and healthy volunteers from our centre. Higher spleen T1 /cT1 was observed in women vs men and in children vs adults. This thesis provided further evidence that multiparametric MR methods combining T1 and T2* relaxometry can be used across the spectrum of CLD for stratification of disease severity.

Obesity and cardiovascular function are closely linked on a number of different levels, from the impact of the whole-body complications of obesity, to the geometric remodelling inflicted on the heart and vasculature by increased body weight, to the influence of the endocrine state of the body on intracellular signalling cascades, to the very efficiency of energy generation and transfer within the myocyte itself. Whilst obesity is one of the main contributing risks to the development of number of different cardiovascular diseases, it has a particularly strong relationship with the development of heart failure. However, there is a counterintuitive association with survival in established disease and increasing body mass index, known as the obesity paradox. This underlines the complex relationship between the two conditions, and poses the question as to whether there is a protective physiological mechanism at play within the obese myocardium. Magnetic resonance imaging is perfectly placed to interrogate this interplay in more detail, being the gold-standard technique for assessing a number of different parameters of cardiovascular structure and function, but also unveiling the potential for investigating cellular physiological parameters using non-invasive spectroscopic techniques. It is known that inflexibility in substrate selection has an impact on cardiovascular function and that this plays an important role in the development of diabetic cardiomyopathy, as well as potentially the obesity phenotype. In this thesis, I use the popular Very Low Calorie Diet in a healthy obese cohort, both to assess its efficacy in treating obesity-related cardiovascular disease, and also to demonstrate the impact of altering the availability of myocardial substrate on cardiovascular function. While after 8 weeks of dietary intervention, myocardial steatosis and diastolic dysfunction are reversed, there is an interesting early deterioration in function, associated with an increase in circulating fatty acids and deposition of ectopic lipid within the myocardium. This highlights the potential for dietary intervention as a treatment strategy in metabolic cardiac dysfunction, but also the need for caution in a population with pre-existing cardiovascular disease. While substrate selection is an important factor in determining the efficiency of energy generation, the rate of transfer of ATP from the mitochondrion to the myofibril is also a key step in regulating myocardial energy transduction. This step is catalysed by the creatine kinase enzyme, enabling PCr to act as a buffer maintaining rates of ATP delivery under fluctuating demand. The reaction rate of creatine kinase can be interrogated with sophisticated non-invasive MR spectroscopic techniques, and has been found to be reduced in heart failure. In obesity, however, there is typically preserved systolic function despite reduced PCr/ATP concentrations, which raises the question of whether creatine kinase kinetics compensate to maintain ATP delivery. I investigated this question in obese individuals without additional cardiac disease, and found not only that the reaction rate of creatine kinase was elevated, maintaining ATP delivery at rest, but that the enzyme reaction rate response to catecholamine stress was blunted, yielding no increase in energy delivery despite increased workload. This was linked to whole body exercise capacity, and reversed with intentional weight loss. These results suggest that obesity may have direct effects on the resting energetic state in the myocardium, but that this comes at a cost of reduced reserve to cope with increased demand. This may explain why exertional symptoms are so prevalent in obesity. To investigate the impact of obesity on myocardial energetics in individuals with established cardiac dysfunction, I examined volunteers either normal weight or obese with dilated cardiomyopathy, measuring myocardial PCr/ATP and creatine kinase kinetics at rest and under dobutamine stress, and interpreting these results in the context of cardiac function and exercise capacity. In obese heart failure, baseline PCr/ATP was lower compared to individuals with normal systolic function, but the creatine kinase rate constant was again elevated compared to those of normal weight. This meant that in obese heart failure, overall ATP delivery was maintained to a similar degree to those without systolic dysfunction. However, the diminished capacity to cope with increased haemodynamic demand was exacerbated, with catecholamine stress leading to a fall in PCr/ATP, blunted response in creatine kinase activity, and an overall fall in ATP delivery. This was in the context of poor exercise capacity, and may be linked to exertional symptoms in obese heart failure. The impact of intentional weight loss on cardiac function and energetic status in obese heart failure was examined in the final chapter. Volunteers were asked to adhere to a calorie controlled diet for a period of 12 months before comprehensive retesting was performed. The striking finding was an improvement in contractile function, even in volunteers with established heart failure, with a marked improvement in left ventricular ejection fraction with weight loss. Equally exciting was significant cardiac remodelling, with reduction in left ventricular end-diastolic volume and mass. The changes in myocardial energetics were more subtle, with no change in PCr/ATP, creatine kinase activity at rest, or ATP delivery. However, there was a suggestion of restoration of the stress response of enzyme activity, with individuals with successful weight loss seeing an increase in stress creatine kinase kinetics which had not previously been measured. These results raise a number of interesting points; clinically, whether intentional weight loss has the potential to become a therapeutic tool in heart failure, and on a more mechanistic level, how left ventricular contractility can improve despite no measurable change in resting energy delivery. In summary, this thesis interrogates the mechanisms underpinning the relationship between obesity and cardiac function, demonstrating the importance of substrate selection in determining cardiovascular function, establishing a potentially compensated energetic state in the obese myocardium, and providing the first stage of evidence for intentional weight loss to be used as an additional treatment strategy in obese heart failure.

This thesis aims to better understand the metabolic dimension of cardiac adaptation to increased work by studying the creatine kinase (CK) system in participants with pathological hypertrophy secondary to severe aortic stenosis with preserved ejection fraction (SevAS-pEF) or secondary to severe primary mitral regurgitation (MR), participants with hearts adapted to exercise training, and participants with no hypertrophy (NH). It also aims to understand why some hearts transition to failure in the presence of pressure overload, by studying a group with severe aortic stenosis with reduced ejection fraction (SevAS-rEF). Broadly speaking there are two pathways for ATP delivery from sites of production in mitochrondria to sites of use at the myofibrils: facilitated diffusion by the CK system and simple diffusion. This thesis aims to investigate both. Closely related to this, CK also has roles in the temporal buffering of adenine nucleotides and the control of respiratory flux over varying work rates, so more broadly represents a kind of metabolic reserve. CK activity can be assessed on freeze-and-extract biochemical assay. Using 31phosphorus-magnetic resonance spectroscopy (MRS) saturation transfer to estimate the CK rate constant (kf) is the first step toward estimates of in vivo flux. Over 60 intra-operative left ventricular (LV) biopsies in participants with pre-operative 31P-MRS are collected to validate the novel TRiST sequence against biochemically-assessed maximum CK rate. Ch. 3 describes non-invasive and invasive predictors of CK total activity and derives a three-variable linear regression model in 37 triplicate datasets to predict CK total activity from total creatine, kf and LV end-systolic volume index. Ch. 4 studies invasive and non-invasive measures of the CK system in participants with sevAS-pEF, sevAS-rEF, and NH. The chapter argues for the assessment of metabolic reserve as a way of identifying - and explaining - the recently described subset of patients whose ventricles deteriorate sooner in the face of pressure overload. A stepwise fall in median CK total activity is found when arranging groups in order of increasing LV mass index. Non-invasive measures are also assessed in participants with moderate AS, healthy volunteers, and sevAS-pEF participants late post-operatively. The CK system is then studied in participants with severe primary MR and preserved systolic function pre- and post-operatively (Ch. 5), and in participants with physiological hypertrophy or normal, non-trained, hearts at rest and during dobutamine stress (Ch. 6). Ch. 7 relates CK activity with ATP diffusion distance using 3D models built from tomographic stacks of electron microscopy images in a subset of LV biopsies.

Hypertrophic cardiomyopathy (HCM), the most common inherited cardiomyopathy, is both phenotypically and clinically diverse. Sudden cardiac death is a devastating and feared complication of this disease but remains difficult to predict because of the vast heterogeneity seen in disease expression. The need to better identify individuals at risk of fatal arrhythmic events has led to the quest for more powerful and sensitive markers of risk. Myocardial fibrosis is one such marker and believed to be an important substrate capable of promoting fatal arrhythmic events. Studies have shown that myocardial fibrosis can be progressive; however, the clinical significance of progressive disease and underlying causal mechanisms remain unclear. This thesis sought to examine the underlying mechanisms and prognostic significance of fibrosis progression in HCM patients and draws insights from a set of studies that use advanced cardiac magnetic resonance techniques to probe deeper into disease pathophysiology.

Substantial evidence indicates that impaired myocardial bioenergetics play a role in the pathogenesis of various cardiovascular disorders. In particular, creatine kinase (CK) system dysfunction has been associated with heart failure, ischemia/reperfusion (I/R) injury, and doxorubicin-induced cardiotoxicity. The goal of the experiments detailed in this thesis was to determine whether cardiac CK upregulation could be of therapeutic potential in these diseases. The literature suggests that overexpressing mitochondrial creatine kinase (CK-Mt) could adversely affect mitochondria, hence experiments were undertaken to elucidate the bioenergetic phenotype of a mouse model modestly overexpressing cardiac CK-Mt. CK-Mt overexpression was found to be benign in regards to metabolism and mitochondrial respiration. The mitochondrial swelling assay showed that CK-Mt overexpression inhibits opening of the mitochondrial permeability transition pore (mPTP), which may contribute towards protection during I/R. Subsequently, CK-Mt overexpressing mice were utilized to investigate whether CK-Mt overexpression improved outcomes in an in vivo model of pressure-overload chronic heart failure. Based on echocardiography and haemodynamic analysis, CK-Mt augmentation did not significantly improve cardiac function or morphology. Biochemical assays revealed that the transgenic mice had over 90% higher CK-Mt activity and 31P-MRS study confirmed that CKMt overexpression was sufficient to restore physiological PCr/ATP ratios during heart failure, suggesting that elevated CK-Mt does not protect against pressure-overload chronic heart failure. Doxorubicin (DOX) is a potent chemotherapeutic agent whose use is limited by a dosedependent cardiotoxicity that progresses to heart failure. An in vitro model was developed in HL-1 cells to explore whether CK overexpression protected against DOX-induced cardiotoxicity. HL-1 cells were transiently transfected with muscle, brain, and mitochondrial creatine kinase (CK-M, CK-B, and CK-Mt, respectively), and a significant increase in isoenzyme-specific CK activity was confirmed. Transfected cells were challenged with 250 nM DOX for 48 hours with or without creatine, and cell death was measured by flow cytometry. Increased CK overexpression and/or creatine supplementation did not reduce DOX-induced cell death. As CK augmentation was previously found to be beneficial in various cardiac conditions, an in vitro tool was developed in HEK293 cells to investigate the transcriptional regulation of CKM and CK-Mt. Two HEK293 cell lines stably overexpressing the CK-M and CK-Mt promoter regions and luciferase reporter genes were established and validated. These cell lines were used in preliminary screens to identify potential compounds that could affect CK mRNA transcription. Sixty six compounds were screened and 12 positive hits were found. In summary, the experiments in this thesis explored the effects of overexpressing CK under physiological and pathophysiological conditions. It was found that modest overexpression of CK-Mt is safe and inhibits mPTP opening, which may attenuate cardiac damage during I/R injury. However, the current level of CK-Mt overexpression does not protect against pressureoverload heart failure nor DOX-induced cardiotoxicity. An in vitro screening model system has been developed for exploration of CK transcriptional regulators, and can be used in the future for high-throughput screening of pharmacological CK regulators that could provide potential alternative treatments for heart diseases.

Myocardial disarray is a likely focus for fatal ventricular arrhythmia in hypertrophic cardiomyopathy (HCM). Measuring the extent of this disarray could provide a better and more mechanistic marker of arrhythmic sudden cardiac death (SCD). Disarray may be inferred by using diffusion tensor cardiac magnetic resonance (DT-CMR), a novel non-invasive imaging technique which allows tissue microstructure to be assessed in vivo - something previously only possible in post-mortem specimens. DT-CMR probes the underlying myocardial architecture by characterising the three-dimensional process of water molecule diffusion as it is impeded by myocytes and interstitium. Thus, diffusion imaging surpasses the usual limits imposed by image resolution. Disarray together with other structural and ionic HCM defects are likely to cause irregularities in myocardial depolarisation and repolarisation to produce an abnormal electrocardiogram (ECG). Computational techniques such as machine learning could objectively learn from and make predictions on HCM ECG morphology, to identify distinct phenotypes within this heterogenous disease. The overall aim of this thesis was to characterise the underlying HCM substrate using in vivo DT-CMR imaging and machine learning of high-fidelity digital ECGs. Diastolic fractional anisotropy (FA), a parameter calculated from the diffusion tensor, was established as a marker of coherent myocyte alignment which demonstrated a mid-wall ring of high FA in healthy controls, corresponding to circumferentially-aligned mid-wall myocytes. In HCM, this ring was disrupted by reduced FA, despite the presence of normal helix angles which represents preserved transmural helical arrangement of myocytes. These findings are in concordance with published histology demonstrating that firstly, disarray and fibrosis invade circumferentially-aligned mid-wall myocytes at the interventricular insertion points and hypertrophied segment. And that secondly, the average myocyte orientation within the septal mid-wall remains circumferential despite disarray. FA was reduced in HCM compared to controls. Reduced FA in HCM was due to disarray and fibrosis. Markers of fibrosis, late gadolinium enhancement (LGE) and extracellular volume imaging, were both significant predictors of FA, in line with fibrosis contributing to low FA. Yet FA adjusted for fibrosis remained reduced in HCM, suggesting that low FA is likely to represent disarray after accounting for fibrosis. Reduced FA was associated with ventricular arrhythmia in HCM, even after correcting for fibrosis and hypertrophy. Phenotyping using machine learning of ECG morphology revealed four distinct HCM phenotypes. One phenotypic group with primary T wave inversion not secondary to QRS abnormalities had increased risk scores for SCD and increased extent of coexisting septal and apical hypertrophy compared to the other phenotypes. In conclusion, DT-CMR can reveal the microarchitecture of the beating left ventricle. I propose that diastolic FA is the first in vivo marker of disarray. Computational ECG phenotyping captures abnormal anatomical and ionic mechanisms and provides insights into HCM heterogeneity. Both techniques may provide independent risk factors as current clinical risk predictors do not capture the underlying pathophysiology intrinsically linked to arrhythmic risk in HCM.

Atrial fibrillation (AF) is the most common sustained arrhythmia and is associated with significant cardiovascular complications, including stroke, myocardial infarction, heart failure, and premature death. The presence of subtle left ventricular (LV) dysfunction is increasingly recognised in patients with AF, raising questions regarding the underlying pathophysiology and potential treatment strategies. I used advanced and multiparametric cardiac magnetic resonance (CMR) methods to investigate potential mechanisms that could contribute to LV dysfunction in patients with AF, controlled ventricular rate and no significant cardiovascular comorbidities (i.e., with so-called 'lone' AF). Patients were evaluated before and after catheter ablation, allowing examination of the effect of restoration of sinus rhythm and reduction in AF burden on LV structure, function, energetics, tissue characteristics, and perfusion. I demonstrated for the first time that patents with lone AF before ablation have significantly impaired ventricular energetics and a subtle reduction in LV systolic function compared to control subjects in sinus rhythm. Furthermore, there was only modest improvement (but not normalisation) in LV function following successful ablation, and myocardial energetics remained impaired despite a significant and sustained reduction in AF burden. These findings imply that lone AF may actually be the consequence (rather than the cause) of an underlying cardiomyopathy. Next, to interrogate advanced ventricular tissue characteristics (such as diffuse myocardial fibrosis) in patients with tachyarrhythmia, I developed a novel CMR method involving a systolic readout T1-mapping sequence. Methodological work in volunteers and patients with tachyarrhythmia demonstrated that this method reports clinically equivalent T1 values to the conventional diastolic readout in healthy volunteers, and was feasible in tachyarrhythmia, producing excellent quality T1 maps. When applied to the investigation of patients with AF, I demonstrated that subtle LV dysfunction in lone AF occurs in the absence of CMR evidence of diffuse myocardial fibrosis, suggesting that LV dysfunction may be reversible with appropriate and targeted therapeutic strategies initiated prior to the development of structural LV remodelling. Finally, I used quantitative perfusion imaging to determine absolute myocardial blood flow and coronary reserve in patients with AF, and determine whether microvascular coronary dysfunction could underlie impaired LV function and energetics in patients with AF. I found that myocardial perfusion is significantly reduced in patients with AF in the absence of significant epicardial coronary artery disease, both at baseline and under conditions of vasodilator stress. Lower baseline blood flow was related to reduced LV performance, and there was no significant change in perfusion after successful AF ablation. These novel findings indicate that coronary microvascular dysfunction may be an important pathophysiological mechanism in lone AF, and at least partially responsible for LV dysfunction. Overall, the findings reported in this thesis have potentially far-reaching implications for the management of patients with AF. They suggest that approaches that predominantly target rhythm control (including anti-arrhythmic medications and ablation) are insufficient to normalise the systemic and cardiometabolic phenotype in patients with AF. Further studies are needed to investigate whether novel approaches that target microvascular and energetic dysfunction in patients with AF can contribute to durable restoration of sinus rhythm and improve clinical outcomes.

Cardiovascular diseases continue to pose an unacceptable societal burden, mandating the development of new and improved methods for their diagnosis, monitoring and treatment. Current cardiovascular magnetic resonance imaging techniques provide exquisite structural and functional information, but their ability to assess the molecular processes underlying cardiac function in health and disease is limited by inherent insensitivity. Hyperpolarized magnetic resonance is a new technology which overcomes this limitation, creating molecular contrast agents with an improvement in magnetic resonance signal of up to five orders of magnitude. One key molecule, hyperpolarized [1-13C]pyruvate, shows particular promise for the assessment of cardiac energy metabolism and other fundamental biological processes with clinical relevance. This thesis describes a programme of translational research in hyperpolarized magnetic resonance conducted with the aims of identifying priority disease states for translational hyperpolarized MR studies, and of enabling first human cardiovascular studies using the technique. We identify important and new potential roles for hyperpolarized magnetic resonance in both the diagnosis, and potentially treatment, of heart disease associated with metabolic diseases and also in ischaemic heart disease. Hyperpolarized [1-13C]pyruvate also has high potential for rapid clinical translation from the laboratory to patients with cardiovascular disease, though the production and administration of clinical-grade hyperpolarized molecules poses significant challenges. We present strategies to overcome these challenges, and describe the first human cardiovascular experience with hyperpolarized magnetic resonance.

Type 2 diabetes mellitus (T2DM) is associated with an increased risk of heart failure and cardiovascular mortality even in the absence of coronary artery disease. Although the reasons for this are not clear, impaired cardiac high energy phosphate metabolism, coronary microvascular dysfunction and ectopic lipid deposition have emerged among the candidate mechanisms. Cardiac magnetic resonance imaging (CMR) and magnetic resonance spectroscopy (MRS) are powerful tools to characterise these conditions. The findings here suggest that, in diabetes, coronary microvascular dysfunction may potentially exacerbate derangement of cardiac energetics under conditions of increased workload. The relationships amongst ectopic adiposity (myocardial, epicardial and hepatic), myocardial metabolic changes and left ventricular remodelling in T2DM were determined using proton magnetic resonance spectroscopy (1H-MRS), cardiac computer tomography (CT) and multi-parametric liver MR. It was found that cardiac steatosis independently predicted left ventricular concentric remodelling and impairment of systolic strain in patients with T2DM. This work also demonstrated that ectopic adiposity is linked to insulin resistance, liver fibrosis and inflammation, and cardiac contractile dysfunction in diabetes, and that the coexistence of obesity and T2DM leads to higher epicardial fat volumes and significant non-alcoholic fatty liver disease compared to T2DM alone. These findings suggest that, since cardiac steatosis and ectopic adiposity are modifiable, strategies aimed at reducing myocardial triglyceride levels may reverse concentric remodelling and improve contractile function in the diabetic heart. The work on field strength effects on cardiac 31P-MRS has shown that 7T is feasible and reliable with reduced error and increased signal to noise compared to 3T. In summary, the work in this thesis demonstrates the powerful role of CMR and MRS in elucidating the detrimental effects that originate from the metabolic abnormalities in the diabetic heart.

The end product of the numerous cellular, electrical and mechanical processes within the normal heart is the generation of blood flow. Blood flow within the cardiac chambers exerts forces upon the myocardium, which creates a continuous remodelling process. In heart failure pathological cardiac remodelling occurs; this is a complex process involving numerous transcriptional, molecular, cellular and architectural changes within both the cardiac myocytes and surrounding extracellular structures. Intra-cardiac blood flow can now be visualised in vivo using three-directional, three-dimensional time resolved velocity encoding cardiac magnetic resonance imaging (4D flow). The interactions between intra-cardiac blood flow patterns and cardiovascular disease processes using 4D flow were determined. The relationship between atypical right atrial flow patterns and embolic stroke across a patent foramen ovale (PFO) were explored, as were the flow parameter changes in heart failure patients. It was found that in the presence of a PFO, participants with atypical flow patterns were more likely to have had an embolic event than those with classical vortical flow. Left ventricular 4D flow parameters were demonstrated to be highly repeatable and stable within a normal study population over time. In dilated (DCM) and ischaemic (IHD) cardiomyopathy patients, inefficient blood flow patterns and deranged kinetic energy profiles were found. The derangement in kinetic energy was demonstrated to correlate with worsening myocardial function, dilatation, Brain-type natriuretic peptide (BNP) levels, six minute walk test (6MWT) and patient symptoms. These findings suggest that derangements in flow parameters are novel biomarkers of disease severity in heart failure patients, which may become useful in monitoring novel heart failure therapies and predicting prognosis. An association between impairment of cardiac energetics and derangement of intra-cardiac blood flow kinetic energy was demonstrated for the first time using phosphorus magnetic resonance spectroscopy (31P MRS). The work on field strength effects on cardiac 31P-MRS has shown that 7T is feasible in a patient cohort with increased precision of metabolite quantification and increased signal to noise compared to 3T. In summary, the work in this thesis demonstrates the powerful role of 4D flow in elucidating the detrimental effects of cardiovascular disease upon intra-cardiac blood flow parameters.

Bicuspid aortic valve disease (BAV) is known to be heritable and often shows a concomitant aortopathy with increased risk of aortic dissection. This may also be present in family members. The underlying pathophysiology was largely believed to be an intrinsic aortopathy. However, recent advances in cardiovascular magnetic resonance imaging now allow in-depth haemodynamic and functional assessment of the aorta to further delineate the underlying pathophysiology. This thesis explored the association of haemodynamic and functional aortic changes with aortic dilation in BAV patients, first degree relatives and healthy volunteers. Patients with a bicuspid aortic valve were found to have an abnormal degree of helical flow in the ascending aorta with predominantly right-handed helical flow. Compared to healthy volunteers patients with a bicuspid aortic valve had larger ascending aortas, higher rotational (helical) flow, systolic flow angle and systolic wall shear stress. There were also differences between the cusp fusion patterns in BAV, with right-non coronary cusp fusion showing more severe flow and aortic abnormalities than those with a right-left coronary cusp fusion pattern: higher rotational flow, higher in-plane wall sheer stress and larger aortas respectively. BAV patients with normal flow patterns had similar aortic dimensions and wall shear stress to healthy volunteers and younger BAV patients showed abnormal flow patterns but no aortic dilation, both further supporting the importance of flow pattern in the aetiology of aortic dilation. However, aortic function measures (distensibility, aortic strain and pulse wave velocity) were similar between healthy volunteers, patients with a bicuspid aortic valve and first degree relatives with a tricuspid aortic valve. Aortic valve replacement (AVR) significantly changed the ascending aortic flow pattern in patients with a bicuspid aortic valve. The majority of patients with mechanical AVR or Ross procedure showed normal flow patterns with near normal rotational flow values post operatively and reduced in-plane wall shear stress. By contrast, all subjects with bioprosthetic AVR had residual abnormal helical flow patterns (mainly marked right-handed helical flow), with similar rotational flow values to native BAV. Wall shear stress post-bioprosthetic AVR showed a similar pattern. These findings point towards the importance of haemodynamic factors in the bicuspid aortic valve aortopathy rather than additional haemodynamic-independent mechanisms. Haemodynamic classification may allow better risk prediction and selection of patients for earlier surgical intervention. Changes in abnormal flow pattern after valve replacement could have important implications for future aortic growth, and may influence the future choice of prosthetic valves.

Liver disease is a major worldwide health problem with high prevalence and poor patient outcomes. In the last decade, efforts to improve liver disease severity assessment have led to the development of many non-invasive methods, including magnetic resonance (MR) techniques. The overarching aim of this thesis was to examine a multi-parametric MR technique in the evaluation of different aspects of liver disease. The MR protocol included T1 and T2* mapping techniques which were used to compute the iron corrected T1 (cT1) and the liver inflammation and fibrosis (LIF) score as measures of liver inflammation and fibrosis. In healthy volunteers, it was found that premenopausal women had a higher median liver cT1 (784ms) compared to postmenopausal women (741ms; p=0.038) or age matched men (742ms; p=0.005). Furthermore, blood volume was identified as the most significant physiological determinant of liver cT1 and a correction algorithm was developed to compute the "volume adjusted and iron corrected T1" (vacT1), which removes the confounding effect of liver congestion and improves the performance of the test as a biomarker of liver fibrosis. Against histology, liver cT1 was strongly associated with liver collagen proportionate area, particularly in patients with viral hepatitis (r=0.80; p < 0.0001). Furthermore, liver cT1/LIF had an excellent diagnostic accuracy for the assessment of multiple aspects of non-alcoholic fatty liver disease (NAFLD) with an area under the receiver operating curve of 0.89 for the identification of significant NAFLD. In the evaluation of portal hypertension, liver cT1 (r=0.49; p=0.038) was associated with the hepatic vein pressure gradient (HVPG), and spleen cT1 was identified as a new biomarker of portal pressure (r=0.66; p < 0.0001 vs HVPG). Furthermore, liver cT1/LIF was shown to be useful in predicting clinical outcomes, with a negative predictive value of 100% for a LIF < 2. Overall, this thesis demonstrated multi-parametric MR to be a robust technique that can assess the whole spectrum of liver disease.

Obesity in adults and children is associated with increased cardiovascular mortality and morbidity. This is forecast to increase markedly in the next decade as childhood obesity is a burgeoning epidemic. Excess weight is clearly associated with insulin resistance, increased circulating triglycerides, and hypertension, all of these are related to progressive heart and liver disease. Ectopic fat deposition within organs is reported to cause lipotoxicity, which may lead to dysfunction and disease, but there have been few human studies to confirm this. This doctoral thesis set out to study the early pathophysiology of obesity in adults and children using in vivo magnetic resonance (MR) imaging and spectroscopy to assess the composition and function of the heart and liver in lean and obese individuals. The central tenet of this project was to establish and validate a clinically viable method for measuring the fat content of viscera safely and accurately, and to determine a normal range for the triglyceride content of the heart and liver. The initial study demonstrated that the heart remodels in response to weight loss, with over 20% reduction in LV mass, confirming that excess weight is genuinely a modifiable risk factor. Then, using spectroscopy, it was established that the healthy myocardium has a median triglyceride content of 0.37&percnt; (IQR 0.24&percnt; - 0.47&percnt;), which increases linearly in overweight and obese adults. Obesity, in the absence of any confounders, was also associated with a 10&percnt; reduction in cardiac contractile function. In comparison, healthy liver median lipid content was 0.67&percnt; (IQR 0.44&percnt; – 0.88&percnt;), which increased in obese adults to 2.9&percnt; (IQR 1.6&percnt; - 7.6&percnt;). There was a graded association between ectopic fat deposition in the liver and dyslipidaemia in adults, characterised by increased circulating triglycerides and reduced high-density lipoprotein. This dyslipidaemia may impair reverse cholesterol transport, and thus could be expected to exacerbate weight gain. Among obese and overweight subjects, there were some with severe steatosis and evidence of coexistent hepatic inflammation and fibrosis. To verify the accuracy of these spectroscopic measures for ectopic fat, a blinded, prospective comparison of non-invasive assessment of unselected liver disease in liver biopsy patients was completed. Liver disease presents with one or more of steatosis, fibrosis and haemosiderosis, all of which are associated with adverse cardiovascular outcomes. Fifty patients were recruited, and interobserver variability among pathologists was measured for histological reference standards for fat, fibrosis and iron deposition. MR measures of each of these metrics predicted the fibrosis, steatosis and haemosiderosis scores accurately. This enabled precise tissue characterisation of all forms of liver disease, including steatohepatitis, with one non-invasive test, to allow the diagnosis and monitoring of hepatic conditions. Lastly, all these new biomarkers of early cardiac and liver disease associated with excess weight were applied to obese and lean children, to understand whether ectopic fat played a substantial role in early life. Obese children had increased ectopic fat in their hearts and livers, as well as impaired strain, evidence of dyslipidaemia, and in some cases evidence of active steatohepatitis, comparable to adults with severe disease. The thesis therefore demonstrates that in vivo magnetic resonance techniques can be used for accurate measurement of visceral lipid content. Furthermore, there is evidence of significant ectopic fat deposition in both adults and children, with evidence of organ dysfunction, which raises the possibility that cardiovascular magnetic resonance may be of value to risk stratify obese individuals based on organ involvement. Finally, the developed methods may have broader applicability and offer a promising new method for the non-invasive diagnosis of chronic liver disease in other clinical settings.

The physiological ketosis of starvation makes sound evolutionary sense, as ketone bodies have several thermodynamic advantages over other nutritional substrates, in addition to their actions to conserve protein and glucose stores. Utilising the body’s metabolic responses to ketosis by delivering a novel nutritional source of ketone bodies, the work in this thesis explored the metabolic effects of ketosis on physical performance in humans. First, the pharmacokinetics and dosing requirements for ketone containing drink preparations were characterised in a population of athletes and healthy controls (n = 45). Using endurance exercise as a model of physiologic stress, the functional impact of ketosis during sustained high intensity effort was investigated in high performance athletes (n = 22). It was shown that nutritional ketosis improved performance in 18/22 athletes, who set 14 new best performances during 30 min of rowing. Furthermore, when ketones and glucose were delivered together, cycling performance was improved by 2% (n = 8) following 1.5 hours of fatiguing effort, compared with optimal carbohydrate intake. Blood D-β-hydroxybutyrate reached 3-5 mM following ketone drinks, equivalent to several days of total fasting, but rapidly decreased during exercise. It was found that higher physical workloads correlated with larger decreases in plasma ketone concentration (n = 8), consistent with their oxidation as respiratory fuels. Nutritional ketosis significantly altered fuel metabolism during exercise in elite athletes (n = 10), decreasing peripheral lipolysis, skeletal muscle glycolytic intermediates, blood lactate, and branched chain amino acid release. In conclusion this work suggests a new hierarchy of substrate preference during physical stress, whereby mimicking the physiology of starvation, the energetic consequences of oxidising ketones may significantly enhance athletic performance. The extrapolation of these findings may have therapeutic implications for patient populations where energetic demands are high, and deleterious switches in substrate selection occur.

Both hypertrophied and failing hearts are characterised by pathological left ventricular (LV) remodelling, impaired myocardial energy status and alteration in substrate metabolism. Cardiac magnetic resonance imaging (CMR) and magnetic resonance spectroscopy (MRS) are powerful tools in the characterisation of these disease conditions. More recent techniques have allowed assessment of myocardial steatosis using 1H-MRS and tissue oxygenation using blood oxygen level dependent (BOLD) CMR. In hypertrophy and heart failure, studies on steatosis and the relationship with other parameters such as myocardial function and fibrosis, especially in humans are limited. I therefore investigated the presence of steatosis in severe aortic stenosis (AS) and dilated cardiomyopathy (DCM), and further assessed its relation to contractile function. This study found that myocardial triglyceride (TG) content is increased in both symptomatic and asymptomatic AS patients (lipid/water ratio 0.89±0.42% in symptomatic AS; 0.75±0.36% in asymptomatic AS vs. controls 0.45±0.17%, both p<0.05) and DCM patients (lipid/ratio 0.64±0.44% vs. controls 0.40±0.13%, p=0.03). Circumferential strain was lower in both AS (-16.4±2.5% in symptomatic AS; -18.9±2.9% in asymptomatic AS vs. controls 20.7±2.0%, both p<0.05) and DCM patients (-12.3±3.4% vs. controls -20.9±1.7%, p<0.001). In AS, myocardial contractility is related to the degree of steatosis, and were both reversible following aortic valve replacement (AVR), lipid/water ratio 0.92±0.41% vs. pre AVR 0.45±0.17%, p=0.04 and circumferential strain -17.2±2.0% vs. pre AVR -19.5±3.2%, p=0.04. A novel finding of this study was significant correlation of MRS-measured TG content with histological staining of TG of the myocardium, taken from endomyocardial biopsy during AVR. In DCM, myocardial TG was independently associated with LV dilatation and correlated significantly with hepatic TG, which suggests that both cardiac and hepatic steatosis might be a common feature in the failing heart. Additionally, although the hypertrophied heart is characterised by impaired perfusion, it is unknown if this is severe enough to translate into tissue deoxygenation and ischaemia. I assessed this by using adenosine vasodilator stress test and BOLD-CMR in patients with severe AS. It was found that AS patients had reduced perfusion (myocardial perfusion reserve index-MPRI 1.0±0.3 vs. controls 1.7±0.3, p<0.001), and blunted tissue oxygenation (blood-oxygen level dependent-BOLD signal intensity-SI change 4.8±9.6% vs. controls 18.2±11.6%, p=0.001) during stress. Importantly, there was a substantial improvement in perfusion and oxygenation towards normal after AVR, MPRI 1.5±0.4, p=0.005 vs. pre AVR and BOLD SI change 16.4±7.0%, p=0.014 vs. pre AVR. Overall, the work in this thesis supports the powerful role of CMR in assessing LV function and elucidating metabolic mechanisms in the hypertrophied and failing heart.

Inflammatory arthropathies and systemic rheumatic diseases (IASRDs) commonly involve the cardiovascular system, and are associated with high morbidity and mortality. Mechanisms of cardiovascular involvement in these clinical entities are not fully understood. Cardiovascular magnetic resonance (CMR) is the single imaging modality capable of assessing non-invasively cardiac function, strain, ischaemia, altered vascular function, perfusion, oedema/inflammation and fibrosis. Furthermore, magnetic resonance spectroscopy (MRS) can give further insights into the status of myocardial energetics and lipidosis. The pathophysiological mechanisms and phenotype of cardiovascular disease (CVD) in IASRDs need further clarification. CMR is an ideal tool for the early identification and monitoring of cardiac manifestations in patients with IASRDs. Hence, the aims of this D.Phil project were to (i) utilise CMR and MRS to study disease mechanisms in patients with rheumatoid arthritis (RA), systemic lupus erythematosus (SLE) and systemic sclerosis (SSc), and (ii) to assess the role of anti-cytokine therapies in abrogation of cardiovascular complications and effects on cardiovascular function in patients with RA, ankylosing spondylitis (AS) and psoriatic arthritis (PsA). First, we used CMR to assess myocardial structure and function in RA, SLE and SSc patients with no known cardiovascular symptoms. Patients and controls were stratified by the presence of traditional cardiovascular risk factors (CVRFs). Our data demonstrated no differences in overall left ventricular (LV) systolic function, size and mass between patients and matched controls. There were, however, impairments in regional function and myocardial deformation, which is most severe in RA, SLE and SSc patients with CVRFs. We also found evidence of impaired vascular function in RA, SLE and SSc, using pulse wave velocity (PWV) and aortic distensibility, and again, showed that patients with CVRFs demonstrated the most severe aberrations, while patients without CVRFs and controls with CVRFs had an intermediate phenotype. Abnormalities in vascular and regional function were most severe in patients with SSc. Also, we showed that impaired vascular function correlated with abnormal systolic myocardial strain and diastolic strain rate in all groups of IASRDs studied. These data have implications for the clinical care of patients with RA, SLE and SSc and show that there is extensive cardiovascular involvement in asymptomatic patients. These results also suggest that early identification and stratification of CVD in IASRDs, with non-invasive techniques like CMR, may permit earlier intervention, thus potentially reducing the effect of CVD on morbidity and mortality in IASRDs. Lastly, these data highlight the importance of early detection and aggressive management of co-existent traditional CVRFs, as they confer incremental risk of CVD in patients with IASRDs. Second, we used CMR for comprehensive phenotyping of tissue characteristics in patients with RA, SLE and SSc. Our data confirmed that subclinical myocardial changes are common in patients with IASRDs (even with apparently normal hearts), which can be detected using multiparametric CMR. In addition to focal areas of fibrosis (detected by late gadolinium enhancement [LGE]), there were also areas of focal myocardial oedema or inflammation (detected by T2-weighted imaging). Further, using more sensitive techniques such as native T1 mapping and extracellular volume (ECV) quantification, we were able to demonstrate even more areas of myocardial involvement in IASRD patients than conventional CMR techniques can reveal, with patients showing significantly larger areas of T1 abnormality and expanded ECV, which likely represent a combination of low grade inflammation and diffuse myocardial fibrosis that are well-described disease processes in this cohort. We also found that T1 and ECV measures were associated with subtle myocardial systolic and diastolic dysfunction. The results of this study suggest that CMR, particularly T1 and ECV quantification, can be used for early detection of subclinical myocardial involvement in IASRD patients, potentially serving as an early screening tool before overt LV dysfunction or irreversible myocardial damage occurs. Third, we utilised CMR to study myocardial perfusion in patients with RA, SLE and SSc. We found that myocardial perfusion was impaired in asymptomatic IASRD with no overt heart disease. Non-segmental, subendocardial perfusion defects consistent with microvascular dysfunction were present in 47%, 31% and 41% of RA, SLE and SSc patients, respectively. Furthermore, there was a significant correlation between MPRI and systolic and diastolic regional function in all groups of patients. In RA and SSc, there was also a correlation between myocardial perfusion reserve index (MPRI) and disease activity. SSc patients had the greatest burden of ECV expansion, and in this group ECV also correlated with MPRI. These data led us to hypothesise that myocardial ischaemia likely leads to impaired myocardial relaxation and diffuse fibrosis, which predate overt dysfunction in contractility. Fourth, we investigated the effect of TNF-alpha inhibitors on myocardial and vascular function and structure in RA, AS and PsA patients. We confirmed that anti-TNF therapy was associated with improvements in serum inflammatory parameters like CRP and ESR, as well as with improved clinical measures of disease activity. Anti-TNF therapy, however, was not related to a change in left ventricular size, mass and global systolic function. We found that inhibition of TNF-alpha activity does result in better myocardial strain and strain rate, likely reflecting an improvement in myocardial inflammatory burden. Moreover, these findings were also supported by improvements in T2 weighted measures, native T1 values and ECV calculations. There was, however, no significant change in myocardial perfusion following anti-TNF therapy. These results support the hypothesis that during episodes of disease activity, myocardial oedema is present in patients with IASRDS and that by reducing the systemic inflammatory response, improvements in myocardial and vascular function can be achieved. Finally, we used CMR and MRS in asymptomatic RA and SLE patients (with normal hearts on echocardiography) to investigate cardiac metabolic status in this cohort. We found that myocardial energetics were impaired in patients, despite preserved overall ejection fraction. Interestingly, abnormal myocardial energetics were associated with presence of LGE, decreased myocardial perfusion, expanded ECV, volume fraction of T1 >990ms (which represents >2 standard deviations above the mean T1 value at 1.5T) and left atrial size. We did not find any difference in myocardial and hepatic lipid content between patients and controls. These data clearly demonstrate that abnormalities in cardiac energetics are present in IASRD patients even before the development of overt cardiac dysfunction, and may be driven by microvascular function and fibrosis.

Cardiovascular magnetic resonance (CMR) is superior to other cardiac imaging modalities in its ability to perform multiparametric myocardial tissue characterisation. Commonly applied techniques include T2-weighted (T2W) imaging for detecting oedema and late gadolinium contrast enhancement imaging for detecting fibrosis, allowing characterisation of most cardiac conditions encountered in the clinical setting. While these techniques have proven useful, conventional MR images acquired by these techniques are displayed on an arbitrary scale and only semi-quantitative with some inherent deficiencies. Quantitative techniques, such as T1-mapping, allow direct quantification of tissue properties and may circumvent many of the issues of conventional MR imaging, with potential to offer more insight into diseased myocardium. Although T1-mapping has been well-established for brain T1-mapping, cardiac T1-mapping has been slow to mature due to technical and practical challenges. The aim of this thesis was to develop a novel method for cardiac T1-mapping that is robust and clinically applicable so that cardiac T1-mapping can transition from a research tool into clinical practice. Further, cardiac T1-mapping using this new technique was validated in normal and disease states, including its ability to detect regional and global myocardial oedema and acute myocarditis. This work developed a novel MR sequence for cardiac T1-mapping, named Shortened Modified Look-Locker Inversion Recovery (ShMOLLI). The conventional method (MOLLI) required 17 heart-beats (a 17 second breath-hold for a patient with a heart rate of 60 bpm) for the acquisition of one slice of T1-map. This is difficult to achieve for patients, especially those who are acutely ill and if whole heart coverage (~8 slices) were desired. One of the goals in the design of ShMOLLI was to bring the breath-hold time down to below 10 seconds so that patients may tolerate the protocol. The design of ShMOLLI was successful in breaching the 10-second breath-hold time without compromising the accuracy of T1 measurements compared to the conventional method; further, unlike MOLLI, ShMOLLI did not exhibit increasing error with increasing T1 or heart rates, making it advantageous in measuring long T1 values and at fast heart rates. These features of ShMOLLI make it a very attractive method ready for clinical application. The ShMOLLI method was then further applied to disease states, starting with models of acute myocardial oedema. ShMOLLI T1-mapping was performed in patients with acute stress-induced cardiomyopathy (model of global oedema) and in patients with regional oedema without infarction, and compared to T2W imaging. T1-mapping demonstrated an excellent diagnostic performance compared to the T2W methods in detecting acute myocardial oedema, whether it was global or regional. ShMOLLI T1-mapping was well-tolerated by acute cardiac patients and showed promise in its application to other conditions in which oedema would be a major component in the pathophysiology. ShMOLLI T1-mapping was next applied to patients presenting with acute myocarditis, in which oedema and inflammation are major components. Compared to the conventional MR methods of T2W and LGE imaging, T1-mapping again showed superior diagnostic performance to T2W imaging and equivalent performance to LGE imaging in detecting acute myocarditis. Further, T1-mapping was able to detect a larger extent of injury compared to both of these methods, and provided topographic information such as the non-ischemic patterns of injury similar to that seen on LGE imaging, but without the need for exogenous contrast agents. In conclusion, this thesis reports the development of a novel method for cardiac T1-mapping, its validation and clinical application to detect myocardial oedema and characterise acute myocarditis without the need for exogenous contrast agents. ShMOLLI T1-mapping holds promise in quantitative tissue characterisation in a wide range of myocardial diseases.

There is a large body of clinical and experimental evidence linking heart disease with impairment of myocardial energetics, particularly the creatine kinase (CK) system. The goal of the experiments described in this thesis was to develop and study models of increased CK phosphotransfer, by overexpressing the CK isoenzymes and/or augmenting intracellular creatine stores. Pilot experiments were performed in cultured cells, which were used to (a) study the effects of CK overexpression in vitro, and (b) validate constructs prior to generation of transgenic mice. Expression was verified at the protein level for all constructs in HL-1 and HEK293 cells, and enzymatic activity was confirmed. Mitochondrial CK (CKmt) was expressed in the mitochondria, as expected, and CKmt overexpression was associated with a significant reduction in cell death in a model of ischemia/reperfusion injury (68.1 ± 7.1% of control, p≤0.05). Transgenic mice overexpressing CKmt in the heart were generated by a targeted approach, using PhiC31 integration at the ROSA26 locus. Transgene expression was confirmed in vitro in embryonic stem cells, and in vivo at the mRNA and protein levels. There was only a modest increase in CKmt activity; therefore, homozygous transgenic mice were generated to increase expression levels, and had 27% higher CKmt activity than wild-types (p≤0.01). Mitochondrial localization of CKmt was confirmed by electron microscopy. Citrate synthase activity, a marker of mitochondrial volume, was ~10% lower in transgenic mice (p≤0.05). Baseline phenotyping studies found that CKmt-overexpressing mice have normal cardiac structure and function. These mice are currently being backcrossed onto a pure C57BL/6 background for further studies in models of heart disease. In addition to CKmt, transgenic mice overexpressing the cytosolic CK isoenzymes, CK-M and CK-B, were generated. Due to the modest level of expression observed at ROSA26, random-integration transgenesis was used, and multiple lines were generated for each construct (carrying 2 or 6 transgene copies in the CK-M line; 2, 3, or ~30 in CK-B). Transgene expression was validated at the mRNA, protein, and activity levels. These lines are currently being expanded for further validation and phenotyping studies. Previous experiments in our group have demonstrated that increasing intracellular creatine (Cr) reduces ischemia/reperfusion injury, and a series of in vitro experiments was performed to determine whether this effect may be mediated by inhibition of the mitochondrial permeability transition pore (mPTP). The mPTP plays a significant role in ischemia/reperfusion, and there is evidence linking the CK system to regulation of the mPTP. Therefore, a model was developed to test whether Cr affects mPTP opening in cardiac-derived HL-1 cells, as this mechanism may contribute to the protective effect observed in vivo. Cr incubation conditions were determined empirically, and 24-hour incubation with 5mM or 10mM Cr was found to significantly delay mPTP opening, to a similar degree to the established mPTP inhibitor, cyclosporin A. This provides evidence that Cr may exert protective effects in the heart by a variety of mechanisms, in addition to its traditional role in energy metabolism. In summary, the experiments conducted in this thesis have produced a range of tools for studying augmentation of the creatine kinase system as a therapeutic target in heart disease. The results of in vitro assays indicate that mitochondrial CK may be a particularly promising target, and that inhibition of the mitochondrial permeability transition pore may contribute to the cardioprotective effect of creatine. Finally, the transgenic models generated and validated over the course of this project will allow for a wide range of future studies into the potential benefits of CK overexpression in the mammalian heart.

Both hypertrophic (HCM) and dilated cardiomyopathy (DCM), though differing in their aetiologies, share features of impaired resting energetics. The aim of this thesis was to determine if cardiac high energy phosphate metabolism, measured as the phosphocreatine (PCr)/ATP ratio using 31Phosphorus magnetic resonance spectroscopy (31P MRS), is further impaired during exercise in these pathologies. This would provide a possible explanation for the high incidence of exercise related death in HCM and DCM as well as the blunted inotropic response to exercise in DCM. Furthermore, this thesis investigates the role of stress perfusion and stress tissue oxygenation in HCM (as these are hypothesized to exacerbate the primary defect in energetics) and exercise training in DCM (which is hypothesized to improve function though the mechanisms are uncertain). This work developed a novel protocol for measuring 31P MRS in a clinically acceptable time frame. The traditional acquisition is at least 20 minutes (as much as 40 minutes in subjects with lower pulse rates). This is a particularly long time to allow for exercise in the magnet particularly in the symptomatic DCM cohort. Hence this work meticulously developed a shorter 8 minute protocol. Its validity, reproducibility and application to exercise were confirmed. The post processing of the MRS data was further improved for calculating blood contamination and tested with both simulated and patient data, including normal, hypertrophied and thinned myocardium. Applying this new method, this thesis is the first to report a further decrease in exercise energetics in HCM. The relationship between perfusion, tissue de-oxygenation and energetic compromise during exercise was then explored in HCM. Athletes, with physiological hypertrophy, were used as an additional control group in these experiments. These results demonstrated a strikingly blunted oxygenation response of the HCM heart to stress even in the pre-hypertrophy HCM mutation carriers. However, as a group, the data did not show a correlation between the blunted oxygenation response and the percentage change in PCr/ATP during exercise. None-the-less, these results can potentially be useful for distinguishing between hypertrophy in the athletes and pathological hypertrophy in HCM and for distinguishing HCM mutation carriers’ pre hypertrophy and the normal heart. In the DCM cohort, this thesis explored the impact of exercise training on cardiac metabolism and function. The results showed no change in cardiac energetics and left ventricular ejection fraction during 8 minutes of exercise. In addition, an eight week home exercise programme did not alter resting or exercise cardiac PCr/ATP, but improved cardiac function during rest and exercise, and increased exercise tolerance and quality of life scores. In conclusion, this thesis reports further insights into cardiac exercise energetics in HCM and DCM and its relationship to perfusion and oxygenation in HCM and to exercise training in DCM. Therapies that decrease the energy cost of cardiac work during exercise may prove beneficial targets to explore further in these conditions.

Marfan syndrome (MFS) is the commonest inherited disorder of connective tissue and is associated with a high risk of potentially life-threatening complications, including aortic aneurysm, dissection and rupture, and, perhaps, ventricular disease. This work describes the prospective application of advanced cardiovascular magnetic resonance techniques to the aorta and heart of consecutive, unselected subjects with MFS, and to appropriately matched control populations. Comprehensive 3D visualisation of blood flow in the entire thoracic aorta of subjects with MFS was achieved using a time-resolved phase-contrast magnetic resonance technique with 3-directional velocity encoding (4D flow), demonstrating a high prevalence of major flow disturbance (87 ± 12%), compared to controls (28 ±18%), localising to those regions of the aorta known to be most prone to aortic dissection (sinuses of Valsalva and proximal descending aorta). Wall shear stress, recently identified as a potentially important determinant of aneurysm progression and rupture, was interrogated in these datasets at the sinuses of Valsalva (SOV), ascending aorta (AA), arch, and proximal (PDA) and distal descending aorta (DDA), using the 4D flow datasets, and was shown to be significantly reduced at each location (SOV -15%; AA -12%; Arch -17%; PDA -18%; and DDA -14%, p<0.05 for each), in subjects with MFS compared to healthy subjects. 4D flow datasets were used to generate relative pressure maps in healthy subjects and in subjects with several aortic diseases. A novel method for the separate evaluation of the components of relative pressure was applied, revealing marked differences in the relative contribution of the components of pressure (unsteady > convective >> viscous), and characteristic differences between subjects in overall relative pressure, and its components. Left ventricular volumes and function were evaluated in subjects with MFS, and did not differ significantly from healthy subjects in the absence of significant valvular regurgitation and / or shunt. Left ventricular end-diastolic volume varied markedly with degree of regurgitation (r=0.75, p=0.0001). The mechanistic implications of these findings, and the potential role of these techniques in the evaluation of cardiovascular disease, are discussed.

The management of the asymptomatic patients with severe aortic stenosis (AS) is challenging; clinicians have to balance the risks of early surgery against the risk that irreversible myocardial damage may occur with a conservative management strategy. It has become increasingly apparent that prognosis in asymptomatic AS depends not only on the degree of valvular stenosis, but also on the myocardial response to pressure overload and understanding the mechanisms of myocardial decompensation may help to guide management in the future. The degree of myocardial fibrosis, microvascular dysfunction, hypertrophy and left ventricular (LV) geometry may all play important roles. However, current guidelines for management of asymptomatic AS limit assessment of the myocardium to the measurement of ejection fraction with echocardiography. More advanced techniques may provide greater information that could be clinically useful. This thesis seeks to further our understanding of the mechanisms of the myocardial response to AS, using Cardiac Magnetic Resonance (CMR) in patients with moderate and severe AS. Myocardial perfusion in AS is examined in chapter 3. The results show that CMR first pass perfusion can be carried out safely and is well tolerated by AS patients. Microvascular dysfunction in these patients was associated with age, exercise time and markers of diastolic dysfunction. Myocardial strain is examined in chapter 4, utilizing a new software tool to look at strain throughout the left ventricle, and also to explore the relationship between strain and myocardial fibrosis. The results show that there are significant variations in circumferential strain measurements, depending on slice position in the LV, and also that there was no relationship found between strain and the degree of LV fibrosis. In chapter 5, the potential of CMR T1 mapping to identify fibrosis is examined using a new shortened non-contrast sequence (ShMOLLI - Shortened Modified Look-Locker Inversion) developed in our unit. CMR T1 values were validated against histological quantification of myocardial fibrosis in a large group of moderate and asymptomatic AS. A good correlation was found between ShMOLLI derived T1 values, with T1 values increasing with the severity of AS. The clinical value of measuring myocardial perfusion and LV global strain is examined in chapter 6 by linking these to prognosis. Measurement of circumferential strain could predict prognosis in asymptomatic AS, but myocardial perfusion showed poor ability to predict events. In conclusion, this thesis offers further insights into the changes that occur in the myocardium of patients with asymptomatic moderate and severe AS, using established and new CMR techniques. The clinical value of measuring these CMR parameters to aid risk stratification is shown, and the future potential for monitoring new therapies in these patients is discussed in the final chapter.

Changes in myocardial energetics have been implicated in the pathophysiology of heart failure (HF). However, the precise contribution of creatine (Cr) / phosphocreatine (PCr) / creatine kinase (CK) energy buffer and transfer remains unclear. The aim of this thesis was to study the effects on murine cardiac function of both impairment and enhancement of creatine metabolism. In order to longitudinally follow the cause and effect relationship of myocardial creatine concentration, a non-invasive method of quantification was required. Cardiac Cr levels measured in vivo by 1H-MRS were therefore compared with gold-standard invasive HPLC and found to correlate over a wide-range (r2=0.91). 1H-MRS was reproducible for measuring Cr levels in the heart, brain, and skeletal muscle. The cardiac phenotype of a novel model of creatine depletion, the AGAT-/- mouse, was characterized using in vivo MRI, 1H-MRS and LV catheterisation, under conditions of gradually reducing Cr concentrations; zero Cr; and attempted phenotype rescue with dietary Cr. For the first time in the heart, the rate of Cr turnover was quantified (~3 % per day) and demonstrated that cardiac function was preserved even when creatine levels reduced by ~70-90%. Total absence of myocardial Cr induced impairment of inotropic and lusitropic cardiac function and reduced inotropic reserve. Cardiac dysfunction was only partially rescued by replenishment of the Cr pool, suggesting this to be a consequence of long-term adaptations to chronic low Cr. Finally, we tested the hypothesis that combined elevation of myocardial creatine and ribose would be beneficial in a mouse model of chronic HF by increasing cardiac energy availability. Despite an increase in myocardial ribose concentration, this did not prevent loss of total adenine nucleotides (TAN), and there was no improvement in post-infarct LV remodeling or function. Future studies are needed to explore alternative approaches for maintaining TAN in combination with total creatine.