Computational Cardiac Imaging

Cardiovascular diseases are a diverse range of conditions that affect 8 million people in the UK with a cost to the UK economy of £16 billion each year.

Cardiac imaging provides a rich source of data for understanding the structure and function of the heart in health and disease but conventional approaches to analysis do not capture the complex pathophysiology of heart disease.

Our program harnesses advanced computer vision techniques for cardiac motion analysis coupled with innovative machine learning (ML) algorithms to model the genetic basis of complex cardiac traits, to provide automated approaches for personalised risk stratification and to prioritise molecular pathways for drug development.

This is achieved by analysis of human cardiac MRI data acquired in diverse populations including 100,000 participants in UK Biobank.

We will model pre-cardiomyopathic and disease-state rare variants using 3D computational models of genotype-phenotype interactions which will objectively map the expressivity of cardiomyopathy-associated variants before disease is established.

ML also has the potential to model disease heterogeneity and identify cohesive disease groups based on morphology, function, and genetic variation.

With the availability of whole exome sequencing in UK Biobank we will study the effects of predicted loss of function variants using ML phenotyping and 3D association modelling to anticipate the physiological consequences that could arise from therapeutic inhibition.

Taken together this program will develop transformative technologies for diagnosis, prediction and therapy development in cardiovascular medicine.