From East London to Global Practice: A Multi-Modal AI Tool for Early Detection of Coronary Artery Disease

Code: BC-DTP_2026_65

Title: From East London to Global Practice: A Multi-Modal AI Tool for Early Detection of Coronary Artery Disease

Primary Supervisor: Stavroula Kanoni

Email: s.kanoni@qmul.ac.uk

Institute: William Harvey Research Institute

Secondary Supervisor: Venet Osmani

Email: v.osmani@qmul.ac.uk

Institute: Digital Environment Research Institute

Lay Summary:

Coronary artery disease (CAD) is the main cause of heart attacks and is a leading cause of death and disability in the UK. People living in East London, particularly those from British South Asian backgrounds, are at especially high risk and often develop CAD at a younger age than the general population. However, the tools currently used in the NHS to estimate coronary artery disease risk were mostly developed in White European populations and do not work equally well for all communities, contributing to health inequalities. This project will use a unique set of data from 350 volunteers from the Genes & Health study, a large community-based cohort of British Bangladeshi and Pakistani people living in East London. These volunteers have provided detailed information including genetics, blood biomarkers, gene expression, proteins, clinical measurements and lifestyle data. This project will expand the dataset of biomarkers by measuring lipidomics and other CVD biomarkers (adiponectin, apoB, C-peptide, BNP, HbA1c). Using artificial intelligence (AI) methods, the student will develop an integrated model that learns early biological and clinical signals associated with coronary artery disease. Once the model has been developed in this East London dataset, it will be tested and refined in large national and international studies such as UK Biobank, Qatari Biobank, and other large datasets, and in hospital-based imaging cohorts that include coronary calcium scores. This will help ensure the tool is accurate and fair across different populations. Importantly, the project will be co-designed with the Genes & Health Community Advisory Board, to ensure that the research reflects community priorities, is culturally acceptable, and supports better prevention of heart disease in East London.

Aims:

The overarching aim of this project is to develop and validate a multimodal, interpretable AI framework for coronary artery disease prediction, initially trained in a deeply phenotyped British South Asian cohort and subsequently evaluated and adapted across broader multi-ancestry datasets.