30th September 2019
AI identifies genes linked to heart failure
The use of artificial intelligence (AI) to analyse MRI images could pave the way to earlier identification of people at risk of heart failure and to the development of new treatments.
Researchers at the Queen Mary University of London have applied a new AI technique, which analysed the heart MRI images of 17,000 healthy volunteers from the UK Biobank – a long-term study of how both genetic predisposition and environmental factors influence the development of disease. They found that genetics account for between 22 and 39% of variation in the size and function of the heart's left ventricle, the organ's main pumping chamber. Enlargement and reduced pumping function of the left ventricle can lead to heart failure.
The research, part-funded by the Wellcome Trust and the British Heart Foundation, appears in the journal Circulation. Using their AI, the Queen Mary team identified or confirmed 14 regions in the human genome associated with the size and function of the left ventricle – each containing genes that regulate the early development of heart chambers and the contraction of heart muscle.
Lead researcher, Dr Nay Aung, said: "It is exciting that the state-of-the-art AI techniques now allow rapid and accurate measurement of the tens of thousands of heart MRI images required for genetic studies. The findings open up the possibility of earlier identification of those at risk of heart failure and of new targeted treatments. The genetic risk scores established from this study could be tested in future studies to create an integrated and personalised risk assessment tool for heart failure.
"The AI tool allowed us to analyse images in a fraction of the time it would otherwise have taken," he added. "Our academic and commercial partners are further developing these AI algorithms to analyse other aspects of cardiac structure and function. This should translate to time and cost savings for the NHS and could potentially improve the efficiency of patient care."
"Previous studies have shown that differences in the size and function of the heart are partly influenced by genes – but we have not really understood the extent of that genetic influence," said Steffen Petersen, Professor of Cardiovascular Medicine at Queen Mary, who also worked on the project. "This study has shown that several genes known to be important in heart failure also appear to regulate heart size and function in healthy people. That understanding of the genetic basis of heart structure and function in the general population improves our knowledge of how heart failure evolves. The study provides a blueprint for future genetic research involving the heart MRI images in the UK Biobank and beyond."
"High fidelity MRI measures combined with genetics is validating many known heart structural proteins, but our work also finds new genes from more heritable functional measures that are associated with ventricular remodelling and fibrosis," said Patricia Munroe, Professor of Molecular Medicine at Queen Mary. "Further genetic studies including analyses of additional heart MRI chambers are expected to provide deeper insights into heart biology."
Many more genetic markers for cardiac conditions are expected to be identified as the UK Biobank database grows. Earlier this month, the organisation announced it will begin sequencing the whole genome of 450,000 participants, following the success of a pilot programme involving 50,000 people.