This PhD project, sponsored by GlaxoSmithKline, is well suited to students who would like to combine their knowledge of programming, computational modelling, physics and mathematics to better understand the structural changes in biological tissue in health and disease. The project would suit a physics, engineering, maths or computer science graduate with an interest in neuroscience and brain physiology. Equally it would suit a neuroscientist with a strong mathematical and computational interest.
Diffusion MRI (dMRI) is the preferred tool to study tissue-microstructure in health and disease. Notwithstanding the increasing amount of studies showcasing the sensitivity of dMRI features to diseases, it is increasingly apparent that the ultimate aim of being unambiguously specific to microstructural characteristics cannot be achieved with current methodology [1]. At Cardiff University Brain Research Imaging Centre (CUBRIC), we are developing MRI methodology and have access to an MRI system with ultra-strong gradients (one of four worldwide) to boost the performance of dMRI [2].
In addition, an emerging zeitgeist in microstructural-MRI is that combining multiple MRI-modalities will yield a more complete picture of tissue-physiology [2,3,4]. This has renewed hope of establishing biophysical models for healthy and diseased tissue. We have recently developed multi-modal MRI protocols to establish correlations between physical and chemical tissue-properties: diffusion MRI provides information on the size, shape, and orientation of tissue-compartments, while relaxometry can add complementary information on chemical composition [4], and diffusion-weighted spectroscopy on cell morphology and neuronal process complexity [5]. These rich data have the potential to improve the disentanglement of different tissue-compartments and to use approaches with fewer assumptions than commonly used biophysical models, which is important in disease-characterisation where the number and properties of tissue-constituents are unknown. So far, these efforts have been solely focussed on the healthy brain.
This project aims to develop an efficient multi-contrast MRI framework for comprehensive microstructure-characterisation in the context of oncology with ultra-strong gradients.
In the wake of the dramatic successes seen with checkpoint inhibitors in oncology, more and more targeted therapeutic options are becoming available for an ever-wider range of tumours. With the breadth of therapeutic options, it is becoming ever more relevant to identify response, or lack thereof, as early as possible. Unfortunately, conventional RECIST-like measures of response are neither very insightful, nor very sensitive for early changes. As such, there is a need to develop markers of response that provide early, in-vivo, mechanistic insights in tissue change.
Multi-contrast MRI measurements provide new windows to probe diffusion and relaxation simultaneously; relaxation provides a rich source of information complementary to diffusion e.g., on the molecular interactions of the system of spins. This simultaneous characterisation has demonstrated superior separation of tissue compartments, and improved tissue classification, e.g., for tumour gradation [6,7].
In this project, we aim to further develop and optimise integrated acquisition protocols to measure a rich set of tissue characteristics in the context of oncology. Simulations and phantom experiments will be employed to evaluate the measures, as well as a proof-of-concept study in patients in collaboration with the Velindre Cancer Centre.