Physics-based image synthesis for MRI sequence standardisation
The last 20 years have seen a dramatic improvement in medical imaging acquisition techniques and analysis algorithmic performance, with imaging biomarkers increasingly being used as primary or secondary end-points in clinical trials and research. However, within neuroimaging applications, medical image analysis algorithms and derived imaging biomarkers are rarely used in day-to-day clinical (patient) care due to the non-quantitative nature of clinical MRI sequences and the lack of acquisition stability within and between scanners.
This project will explore the use of image synthesis to invert the signal differences within pulse-sequences (TR/TE/IR parameters) and between contrast modalities (T1, T2, FLAIR) through the joint use of imaging physics and convolutional neuronal networks. This will increase standardisation of MRI acquisition and analysis techniques and enable translation of state-of-the-art analysis algorithms to patient care.
This ambitious project provides a balanced combination of methodological research topics, where the candidate will need to understand the physics underlying the “image acquisition”, synthesise novel contrasts through “image reconstruction” processes, which will enable the development and deployment of novel “Image Analysis” tools within a ”neuroimaging” clinical context.
This project also aligns well with the research aims and vision of the UCLH BRC Imaging initiative, the NHNN Quantitative Neuroradiology Initiative, and the BRU Dementia, as it will directly impact clinical trials, clinical research and patient care.