Dr Qiang Zhang
Oxford BHF CRE Basic Science Intermediate Transition Fellow
- Start Date: 01/07/2021
- End Date: 30/09/2023
- BHF CRE Mentors: Stefan Piechnik, Vanessa Ferreira, Stefan Neubauer & Thomas Nichols
Research project title: Towards Replacing Late Gadolinium Enhancement with Artificial Intelligence Virtual Native Enhancement for Gadolinium-Free Cardiovascular Magnetic Resonance Tissue Characterisation
I am a deep learning scientist with cross-domain knowledge of cardiovascular diseases and CMR physics. My fellowship project aims to develop artificial intelligence (AI) imaging techniques to replace traditional late gadolinium enhancement (LGE) for myocardial tissue characterisation.
LGE on cardiovascular magnetic resonance (CMR) imaging is used routinely to characterize myocardial tissue and guide the treatment of heart diseases. However, LGE requires intravenous injection of a gadolinium contrast agent, which prolongs the scan, increases the cost, and is cautioned in some patients.
Using advanced AI techniques, we are developing a “virtual native enhancement” (VNE) imaging technique to replace traditional LGE for gadolinium-free myocardial tissue characterization. The new technique combines MR images that do not normally need contrast injections, and uses AI to train machines to predict what an LGE image would look like. First results on hypertrophic cardiomyopathy patient data show that the approach can produce as clear or even better-quality images than traditional LGE, providing clinicians with the same information. We are further collecting large and representative CMR datasets and developing deep learning models to address a range of cardiac pathologies, such as myocardial infarction.
AI has been a main driving force for many clinical breakthroughs. Development of AI technologies for cardiovascular research requires close collaboration between technicians and clinicians. The University of Oxford Centre for Clinical Magnetic Resonance Research (OCMR) has cross-disciplinary teams of magnetic resonance specialists, cardiologists and AI scientists to support the development and validation of VNE techniques.
Beyond this fellowship, I plan to scale up the project to address more myocardial pathologies. I also aim to expand the concept of AI-driven “virtual contrast dye” to replace more CMR imaging techniques that currently rely on contrast agents. Implementation of the prototype on MRI scanners could pave the way for wider VNE protocol dissemination and multi-centre clinical trials. With further development, this may lead to the next generation of gadolinium-free CMR scan protocols that are faster, cheaper, and more patient-friendly.
PhD, MEng, BEng, BSc
Acute Response in the Noninfarcted Myocardium Predicts Long-Term Major Adverse Cardiac Events After STEMI.
Shanmuganathan M. et al, (2023), JACC Cardiovasc Imaging, 16, 46 - 59
Artificial Intelligence for Contrast-free MRI: Scar Assessment in Myocardial Infarction Using Deep Learning-based Virtual Native Enhancement (VNE).
Zhang Q. et al, (2022), Circulation
Endogenous T1ρ cardiovascular magnetic resonance in hypertrophic cardiomyopathy.
Thompson EW. et al, (2021), J Cardiovasc Magn Reson, 23
Ensemble of Deep Convolutional Neural Networks with Monte Carlo Dropout Sampling for Automated Image Segmentation Quality Control and Robust Deep Learning Using Small Datasets
Hann E. et al, (2021), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 12722 LNCS, 280 - 293