A fundamental challenge for personalised radiotherapy (RT) is enabling regular in-treatment RT plan adaption to account for motion of organs. This motion may be associated with everything from respiration and peristalsis to bladder/bowel filling, and even coughing. In each case, its effect is to diminish the accuracy with which RT can be delivered to target regions (tumours), meaning that significant treatment margins must be incorporated in the plans, with the certainty that normal tissues also will then receive elevated levels of radiation exposure. If tissue motions in individuals during treatment can be characterised reliably, RT plans may be adapted correspondingly, potentially allowing dose escalation in target regions, while simultaneously reducing toxicity in sensitive nearby tissues. To this end, reliable and regularly updated per-patient models of these motions are required, along with robust and efficient techniques for driving such models with intra-treatment patient measurements – most importantly, images.
The project’s overarching aim, therefore, is formulation of image-driven modelling approaches that enable real-time estimation of tissue motion states from sparse and low quality images acquired during treatments. We envisage a hybrid modelling approach for this purpose that integrates biomechanical models of the relevant organs with artificial intelligence (AI) techniques that can learn mappings between model configurations and corresponding images. This will ensure motion predictions are based on reliable physical and physiological principles, while enabling robust and efficient exploitation of information derived from images.
The work will thus involve: - Development of patient-specific computational (biomechanical) models of organ motion, based on finite element methods or other suitable techniques. - Formulation of innovative AI methods that link model predictions with image-derived observations. - Development of automated methods and tools for generating individual patient motion models from routine patient clinical data (images etc).
The work will be based within CISTIB, at the University of Leeds, but will involve close collaboration with researchers from the Leeds Cancer Centre, based at St James’s hospital. As can be seen, the project will require a genuinely cross-disciplinary approach that spans (at least) traditional disciplines of computational biomechanics, data science, and medical imaging.
It presents an opportunity for a motivated and able candidate to develop expertise in all three of these areas, while also gaining experience of close working with clinical collaborators and of the challenges associated with translating technical solutions into the clinic.
Funding will be awarded on a competitive basis. A full standard studentship consists of academic fees (up to £22,750 in Session 2019/20), together with a maintenance grant (up to £15,009 in Session 2019/20) paid at standard Research Council rates.