Joy, George; (2024) Deep structural phenotype of hypertrophic cardiomyopathy. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Joy_10191467_thesis.pdf Download (6MB) | Preview

Abstract

Hypertrophic Cardiomyopathy (HCM): HCM is characterised clinically by inappropriate left ventricular hypertrophy (LVH) and is the commonest inherited heart muscle disease worldwide. It has a prevalence estimated at 1 in 500. It is a leading cause of sudden cardiac death, arrhythmia and heart failure. The current paradigm of defining HCM by elevated wall thickness to define disease and target therapy has been inadequate; gene variant carriers without LVH (G+LVH-) undergo interval imaging surveillance, and those with LVH have heterogenous clinical outcomes ranging from asymptomatic disease to sudden death. In recent years myosin inhibitors have demonstrated resolution of left ventricular outflow obstruction, at least some positive cardiac remodelling and improvement in functional and blood biomarkers. Furthermore pre-clinical work has shown correction of sarcomeric mutations with gene editing. These developments have resulted in the need for early disease biomarkers that may respond to therapy. Whilst HCM is defined histologically by myocyte hypertrophy, it is also characterised by fibrosis, disarray and microvascular disease. Focal fibrosis is measured by cardiac MRI late gadolinium enhancement (LGE) and a high LGE burden is related to adverse clinical outcome, however this tends to occur late. Myocyte disarray and microvascular disease have been less explored. Parameters relating to these characteristics can be obtained through novel CMR techniques; diffusion tensor imaging and quantitative perfusion respectively. I have utilized these techniques to measure in-vivo these parametric changes occurring early in HCM, even in the absence of hypertrophy. Recent evidence also suggests that sarcomere variant negative overt disease (genotype negative, G-LVH+) has distinctive morphological and clinical features to G+LVH+. To explore this further, I studied the differing advanced imaging phenotypes based on presence or absence of sarcomere mutation in overt disease. Ventricular arrhythmia can occur in HCM and is a leading cause of sudden cardiac death in the young. How ventricular arrhythmia forms is poorly understood due to evidence from small invasive and pre-clinical studies. ECG Imaging is a non-invasive high spatiotemporal resolution technique of mapping ventricular activation and repolarization. I integrated this technique into advanced cardiac MRI by co-inventing a fully re-useable ECGI vest [capturECGI vest, US patent approved] to detect early electrophysiological changes in HCM and in overt disease, relate abnormalities to genotype and adverse structural change. Quantitative Perfusion CMR in physiological and pathological LVH: The presence of ischaemia in HCM has been well described but mechanisms are incompletely understood. Whilst observational studies in HCM commonly compare findings to health, the “grey-zone” patients clinically are often LVH-causing phenocopies such as Athleticism and Hypertension. I therefore explored the relationships between LVH and microvascular function and studied physiological hypertrophy (athleticism), pathological hypertrophy from excessive afterload (hypertension) and subclinical HCM. Quantitative perfusion measures pixel-wise myocardial blood flow under vasodilator stress. The technique is fully-automated, meaning that between-group differences in microvascular function can be more accurately measured. I studied 19 athletes, 10 hypertensives, 20 subclinical HCM and 14 healthy volunteers. Hypertensives and subclinical HCM compared to athletes had significantly impaired stress myocardial blood flow and perfusion reserve suggestive of microvascular disease. Compared to health, athletes had more enhanced perfusion reserve, however this finding was confounded by the requirements for higher doses of vasodilator (adenosine). Using a high-dose adenosine cohort (unmatched for age) and multivariable regression, the finding of enhanced myocardial perfusion reserve in athletes compared to health persisted. Further work is needed in larger external validation cohorts to confirm this finding. Overall, my data supports the potential use of quantitative perfusion to discriminate between physiological and pathological hypertrophy. To confirm whether quantitative perfusion has discriminatory value, larger cohorts are needed to obtain thresholds when considering potential confounders: primarily age, sex, drugs and comorbidities. Subclinical HCM and genotyped overt HCM cohorts: In order to detect the earliest manifestations of gene expression in HCM, I studied individuals with pathogenic sarcomeric mutations without hypertrophy and compared findings to healthy volunteers. Overt HCM patients (all genotyped to compare G+ vs G-LVH+) were also studied to understand the compounding effects of LVH and the effects of the presence or absence of sarcomere mutation on the advanced imaging parameters of interest. Microstructural and microvascular phenotype HCM: Diffusion tensor imaging (DTI) measures the diffusion of water in-vivo , which thereby characterizes the myocardial microstructural environment. I performed DTI and quantitative perfusion on all participants and showed that DTI changes suggestive of microstructural alteration (low fractional anisotropy, high mean diffusivity, steeper sheetlet orientation) occur in the absence of LVH in subclinical HCM. Abnormalities in DTI and quantitative perfusion also associated with each other and 12-lead ECG changes, in both subclinical and overt HCM. Furthermore in overt HCM, changes associated with genotype; G+LVH+ had a 100% prevalence of perfusion defects, and G-LVH+ unexpectedly had a more severe microstructural phenotype (steeper sheetlet orientation). Findings show the potential for these novel CMR techniques in early and disease-specific phenotype detection in the emerging era of disease modifying therapy. Detection of electrophysiological abnormalities in HCM: Electrocardiographic imaging (ECGI) is a non-invasive technique for detecting electrophysiological abnormalities. It uses body surface potentials obtained from 256 leads across the torso and heart-torso geometry obtained from imaging. Using the inverse solution of electrocardiography, 1000 unipolar electrograms (UEGs) are computed. By analysing the computed UEGS we derive ventricular activation time (AT – defined as the steepest point of the QRS upslope) and repolarization time (RT – defined as the steepest point of the T wave downslope), activation recovery interval (ARI – the difference between AT and RT), markers of spatial conduction and repolarization heterogeneity (gradients of activation GAT, GRT), fractionation and signal amplitudes. I showed that in subclinical HCM, ventricular activation is slowed, and repolarization is more spatially heterogenous. In overt disease, repolarization is prolonged and conduction heterogeneity associated with genotype (more fractionation in G+LVH+), scar and non-sustained VT. These changes are proven to be pro-arrhythmic in multiple other diseases. Findings show that EP abnormalities occur in the absence of LVH in subclinical HCM and in overt disease relate to conventional risk markers. Therefore findings support the use for ECGI in early disease detection and with further longitudinal work, may supplement current conventional risk stratification. Identification of sub-phenotypes of HCM using unsupervised learning: HCM is characterised by substantial disease heterogeneity in LV morphology and clinical outcomes. Disease is defined by arbitrary maximum wall thickness cutoffs which means the spectrum of adverse changes occurring before these cutoffs are reached goes undetected. Using unsupervised machine learning (agglomerative hierarchical cluster analysis) we discover that a third of subclinical HCM have microvascular, microstructural EP changes that are comparable to overt HCM, being either intermediate or severe. A half have an abnormal ECG meaning i) this cohort are likely to progress ii) in the absence of ECG changes and LVH, deep phenotyping is the only method of identifying such individuals. Genotype positive vs negative overt disease were balanced between intermediate and severe phenotypes suggesting large overlaps between both. Findings support the use of advanced phenotyping to identify a higher risk category of subclinical HCM that may benefit from more intensive monitoring or early initiation of novel therapy. Conclusion: Findings. In left ventricular hypertrophy, quantitative perfusion may supplement current techniques for discrimination of pathological vs physiological LVH. In HCM, abnormalities in myocardial microstructure, microvasculature and electrophysiology occur in the absence of LVH. In overt disease, abnormalities associate with important clinical characteristics (adverse structural changes, genotype). Potential Clinical Application. I have optimised new techniques to visualize in-vivo early histological and EP changes occurring in HCM. Whilst we have integrated LVH and scar in cardiac MRI for clinical care, my work has made the measurement of disarray and microvascular disease robust in HCM for future clinical and trial utility. Furthermore early electrophysiological abnormalities can now be detected and related to genotype and structural abnormalities through integration of ECGI with CMR. We found remarkable microstructural, microvascular and EP changes occurring in subclinical disease in the absence of LVH. Curious insights into genotype negative vs positive disease show both overlapping and distinctive phenotypes. These biomarkers may now be tested in longitudinal studies to track whether abnormalities relate to disease progression and ventricular arrhythmia formation. Overall findings suggest clinical techniques could supplement current care, especially for the detection of early disease but also potentially risk stratification and disease discrimination.

Download activity - last month

Export as JSONCSVXML

Download activity - last 12 months

Export as JSONXMLCSV

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item