Mapping cardiac microstructure in aortic stenosis and diabetes using diffusion MRI and machine learning

About the Project

Heart failure remains one of the leading causes of hospitalisation and death worldwide. Two of its most important causes, type 2 diabetes and aortic stenosis, both lead to thickening and stiffening of the heart muscle, but the microscopic processes underlying these changes are not fully understood. Conventional cardiac MRI techniques can visualise heart anatomy and function, yet they are unable to capture the fine organisation of muscle fibres and tissue microstructure where early disease begins.

Cardiac diffusion tensor imaging (cDTI) is a new MRI technique that measures how water molecules move within heart tissue, providing information about the orientation and integrity of muscle fibres. This approach can detect subtle microstructural changes associated with hypertrophy, fibrosis, and metabolic dysfunction, potentially before conventional imaging reveals any abnormality. Early studies have demonstrated characteristic patterns of remodelling in the diseased heart (1).

This project aims to advance the clinical translation of cDTI in cardiometabolic and valvular heart disease. The work will begin with optimisation and validation of the technique in healthy volunteers, using the same harmonised imaging protocol developed by the Society for Cardiovascular Magnetic Resonance (SCMR) through its international multicentre SIGNET collaboration (2). Leicester is a participating site in this study, providing a unique opportunity to benchmark local data against global standards.

Following validation, cDTI will be applied in patient cohorts with type 2 diabetes and aortic stenosis to identify disease-specific microstructural signatures. The study will integrate quantitative diffusion imaging with conventional cardiac MRI measures, blood biomarkers, and machine-learning approaches to explore links between tissue microstructure and disease severity.

By developing and applying state-of-the-art diffusion MRI methods, this research will help to establish cDTI as a sensitive, quantitative imaging biomarker for detecting early cardiac remodelling and monitoring therapeutic response.

The project provides comprehensive multidisciplinary training in advanced cardiac MRI, diffusion tensor imaging, and quantitative data analysis. The student will gain hands-on experience in MRI acquisition, reconstruction, and post-processing, including exposure to motion-compensated diffusion methods and AI-based analysis. Training will also cover clinical research design, governance, and patient recruitment in collaboration with cardiology teams. Engagement with the international SCMR SIGNET consortium will offer opportunities for collaboration and networking with leading experts. The experience will equip the student with transferable technical, analytical, and research skills for careers in academic, clinical, or industry-based cardiovascular imaging.