Cardiovascular diseases (CVDs), such as heart attacks, heart failure, and irregular heart rhythms, remain the leading cause of death and long-term illness in the UK. Identifying people at high risk early and providing timely, personalised care can significantly reduce these impacts. However, current clinical tools often rely on fixed checklists or scores that cannot capture the full complexity of disease progression or adapt to an individual’s changing health over time.
This research project aims to develop next-generation computer models that help doctors make better decisions about CVD prevention and treatment. These models, known as digital twins, act as personalised “virtual hearts” built using real patient data. They learn from a wide range of health information-such as heart scans, blood tests, lifestyle factors, and medical records-to offer a detailed and dynamic picture of someone’s heart health. Over time, they can help predict how a person’s risk may change and what steps might be most effective to reduce it.
To achieve this, we will use advanced artificial intelligence (AI) techniques, including machine learning, deep learning, and geometric deep learning. These will allow us to analyse the 3D and 4D shape and motion of the heart, enabling more precise characterisation of different types of heart disease. Our focus will be on interpretable AI, meaning the developed tools will be transparent and explainable-so clinicians can understand and trust the predictions being made. These tools are designed to augment clinical decision-making with timely, data-driven insights.
We will integrate the world-leading health research resource of the UK Biobank including cardiac imaging, medical history, blood markers, prescriptions, and long-term follow-up for half a million participants in the UK. By using this unique dataset, we hope to build AI tools that are not only powerful and accurate, but also safe, affordable, and practical for use in real clinical settings.
The project is expected to run over three years, with the goal of developing new tools to: detect early signs of heart disease, predict who may benefit from specific treatments or monitoring, and support doctors and healthcare systems in delivering more targeted, efficient care.
This research has the potential to improve how heart disease is detected and managed for people across the UK, and eventually worldwide. It supports the NHS ambition of using digital technologies to provide earlier, more personalised, and fairer care-helping reduce the burden of CVD on individuals, families, and the healthcare system.
