The optimisation of oropharyngeal radiotherapy to reduce the risk of long-term toxicities

Abstract

Abstract The University of Manchester Christina Hague, Doctor of Medicine The Optimisation of Oropharyngeal Radiotherapy to reduce the risk of long-term toxicities Purpose: The prevalence of long-term survivors with head and neck cancer is increasing. A number of strategies are required to optimise organ sparing, to improve long term quality of life. These include: establishing a correct dose threshold for organs at risk (OAR) in avoidance planning; using guidelines and auto-contouring models to standardise volumes for optimal dose delivery; adaptive re-planning to de-escalate normal tissue dose in response to tumour shrinkage and anatomical changes; exploiting the therapeutic advantages of proton beam therapy as an alternative modality and exploring the willingness of patients to travel to receive proton beam therapy in the UK's first proton trial. Aims: (1) To establish a tolerance dose for the masticatory apparatus for use in avoidance radiotherapy planning to reduce trismus. (2) To use a novel muscles of mastication atlas to standardise volumes and improve consistency and optimise dose delivery. (3) To assess the benefits of a novel CT deep learning auto-contouring model to improve clinician workload and reduce inter-observer variability. (4) To develop and evaluate a novel MR deep learning auto contouring model for adaptive re-planning. (5) To investigate the dosimetric consequences of uncertainties in set up and range with proton beam therapy in post-operative oropharyngeal and oral cavity cancers. (6) To evaluate if patients are willing to travel and stay away from home to receive proton beam therapy in the UK's first proton trial. (7) To determine if oxygen-enhanced MRI (OE-MRI) is feasible in head and neck cancer by assessing ability to detect an oxygen signal and patient tolerability. Results: (1) There was a significant association between doses >40 Gy to the ipsilateral block, lateral pterygoid and masseter and deterioration in trismus. (2) The atlas significantly reduced interobserver variability for the muscles of mastication and improved contouring consistency by trainees compared with consultants. (3) An optimised CT deep learning auto contouring model (modelCT) reduced time and inter-observer variability compared with manual contours for OAR delineation. (4) The performance of a novel MR deep learning auto-contouring model (modelMRI) to define contours was sensitive to differences in image acquisition parameters but compared with CT models, geometric accuracy with manual contours increased. (5) Multi-field optimisation is robust to inter-fraction uncertainties in set-up and range, without compromising OAR mean dose in postoperative oropharyngeal and oral cavity cancer. (6) Patients are willing to travel and stay away from home to receive proton beam therapy. (7) The OE-MRI study has received ethical approval, but is yet to begin study recruitment. Conclusions: (1) Tolerance dose to the ipsilateral block, lateral pterygoid and masseter of

Date of Award	31 Dec 2020
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Catharine West (Supervisor), Andrew Mcpartlin (Supervisor), David Thomson (Supervisor), Nicholas Slevin (Supervisor) & Marcel Van Herk (Supervisor)