AI machine learning for cardiometabolic disease and cardiovascular population health
- 16 Sep, 2026 to 1 Dec, 2026
- Project No: BDI009
- DPhil Project 2027
- Big Data Institute (BDI)
Background
Cardiovascular disease remains the leading cause of death worldwide. While traditionally associated with conditions such as diabetes, hypertension, and obesity, it is now increasingly recognised as a systemic, multi-organ disorder comprising diverse cardiometabolic disease subtypes. The complex interactions between the heart and other organs cannot be fully characterised using conventional clinical measurements alone.
Recent advances in deep learning and population-scale imaging provide unprecedented opportunities to investigate cardiometabolic disease using large, multimodal datasets. Building on the team's pilot work in AI-driven multi-organ cardiometabolic phenotyping [1], this DPhil project will develop novel deep learning methods using the latest Biobank imaging resources to improve understanding of disease mechanisms and enable precision risk prediction.
The project may also build on the group’s pioneering work in AI-based Virtual Native Enhancement (VNE) technology [2-4], and investigate myocardial scarring through AI enhanced images in large population cohorts. In addition, it will explore the generalisability and clinical translation of AI models through cross-cohort validation across large clinical and population studies, including the HCM Registry (hcmregistry.org), OxAMI (oxami.org.uk), and China Kadoorie Biobank, with Oxford collaborating teams.
References and pilot works:
- [1] Kway Y, et al. Precision phenotyping of multi-organ fat. ESC Late-Breaking Science. Aug 2026.
- [2] Zhang Q, et al. Circulation. 2021;144:589–599. doi.org/10.1161/CIRCULATIONAHA.121.054432
- [3] Zhang Q, et al. Circulation. 2022;146:1492–1503. doi.org/10.1161/CIRCULATIONAHA.122.060137
- [4] Zhang Q, et al. J AM Coll Cardiol. 2026 (in press). www.jacc.org/doi/10.1016/j.jacc.2026.05.029
Research experience, research methods and skills training
The team brought together expertise in AI, health statistics, cardiac imaging and medical big data, with a track record of high-impact research at the interface of AI and medicine. You are expected to present abstracts at leading conferences, and publish main research works at machine learning and medicine journals.
Fieldwork, secondments, industry placements and training
You will be primarily based at the Big Data Institute. You will have access to additional facilities, mentorship and networking at the Division of Cardiovascular Medicine at John Radcliffe Hospital. This DPhil offers a unique opportunity to contribute to the interdisciplinary collaborations between the two departments, and develop deep learning methods at the forefront of cardiovascular imaging and population health.
Prospective student
The ideal candidate will have a strong background in machine learning, medical imaging, health statistics, or population health.