A heart condition called hypertrophic cardiomyopathy thickens the heart muscle in ways that can trigger dangerous arrhythmias or sudden death, but doctors cannot yet predict which patients will worsen. This project uses artificial intelligence to analyse thousands of cardiac MRI and echocardiogram images, searching for subtle changes in wall thickness and left atrial structure that signal impending decline. The researcher will compare patients whose disease progressed against those who remained stable, then link those imaging patterns to genetic data, blood pressure, ECG markers, and other clinical factors. If successful, the work could identify which patients need early treatment with new cardiac myosin inhibitors or rhythm-control therapies for atrial fibrillation, rather than waiting for symptoms to appear. The study is a practical step toward personalised surveillance in a condition where disease course varies enormously between individuals.
View original technical description
Hypertrophic cardiomyopathy (HCM) is a leading cause of arrhythmia, sudden cardiac death (SCD) and heart failure. It is defined by inappropriate left ventricular hypertrophy (LVH) and is characterized by highly variable expression. Recently novel disease modifying agents are integrated into clinical care, meaning identifying patients more likely to progress is a rapidly emerging priority. I will study two markers of disease progression; i) the worsening of LVH through changes in maximum wall thickness (MWT) and ii) the development of atrial fibrillation (AF) through measurement of adverse left atrial (LA) structural remodelling. AI measurements in cardiac MRI (CMR) consistently exceed clinician precision and therefore can be used to detect patients with progressive LVH more sensitively. I will retrospectively analyse Echo and CMR images in patients followed-up at our inherited cardiomyopathy services (n~3000) using AI segmentation to identify two cohorts: progressive LVH and stable LVH. I will then identify baseline predictors of progression: genetics, demographics, clinical variables, blood pressure, comorbidities, CMR mapping parameters (T1, T2, ECV, quantitative perfusion) and automated ECG markers (pathological Q waves, LVH voltage criteria, repolarization changes). Our team are developing CMR AI tools to detect important LA remodelling and functional changes. We will apply this to all patients with CMR and identify automated predictors of incident AF, AF recurrence post- rhythm control strategies and further relate these to baseline characteristics above. Overall findings may inform which patients could need earlier targeting with novel cardiac myosin inhibitors and therapeutic strategies for AF in HCM.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know