Cardiovascular ageing is a key determinant of heart failure, atrial fibrillation and other major cardiovascular outcomes. Cardiac magnetic resonance (CMR) from UK Biobank provides a unique opportunity to characterise how cardiac structure and function evolve across the adult lifespan at population scale. This project will use UK Biobank cardiac MRI and linked clinical data to: (1) construct a standardised dataset of four-chamber cine CMR and core cardiovascular risk factors; (2) train conditional deep generative models to synthesise “age-progressed” cardiac images conditioned on age, sex and risk factor profiles; and (3) derive quantitative markers of “cardiac age” and evaluate their associations with established imaging traits and incident cardiovascular events. The scientific rationale is that generative models can capture non-linear trajectories of cardiac remodelling that are not easily described by conventional regression alone. By learning data-driven patterns of cardiac ageing, we aim to improve understanding of subclinical cardiac remodelling and to develop imaging-based indices that may help identify individuals at higher cardiovascular risk. All analyses will be undertaken solely for non-commercial, health-related research in the public interest.
