Research Questions:
1. What genomic signatures drive hypertrophic cardiomyopathy (HCM) and ATTR amyloidosis pathogenesis beyond established genes?
2. Can an integrated imaging-ECG deep learning model improve diagnostic accuracy for HCM/DCM compared to current clinical criteria?
3. Do combined inflammatory/myocardial injury biomarkers predict heart failure outcomes in HCM/Fabry disease?
Objectives: 1. Identify novel cardiomyopathy genes via UKB WGS burden testing excluding established pathogenic genes 2. To develop and validate a multimodal Transformer model integrating 12-lead ECG (20205) and cardiac MRI (20212) for cardiomyopathy diagnosis. 3. To construct an attention-LSTM model predicting 3-year heart failure rehospitalization (Field 42030) and all-cause mortality (Field 40000) using longitudinal: – Myocardial injury markers: TNNI3 (2743), NT-proBNP (1912) – Inflammatory cytokines: IL-6 (1418), CRP (30710) – Cardiac imaging features: MRI-ECV (20208), Echo-strain (20213)
Scientific Rationale: Three critical gaps exist: 1) the pathogenic mechanisms underlying cardiomyopathies remain incompletely understood, particularly the quantitative impact of rare genetic variants on disease expression. 2) Diagnostic delays in cardiomyopathies are prevalent, particularly among patients with comorbidities such as atrial fibrillation, coronary artery disease, or chronic kidney disease, posing significant challenges for clinical management. 3) Current prognostic biomarkers for cardiomyopathies lack disease specificity, necessitating the discovery of novel biomarkers or radiomic signatures to enable precision management and improve patient outcomes.
