Delivery of pancreatic stereotactic ablative radiotherapy on the Elekta Unity magnetic resonance - linear accelerator

Abstract

This thesis contributed to the optimisation and validation of pancreatic stereotactic ablative radiotherapy (SABR), using the Elekta Unity magnetic resonance linear accelerator (MRL) to deliver MR-guided adaptive radiotherapy (MRgART). SABR may improve patient outcomes but overlap of gastrointestinal (GI) organs at risk (OARs) with the target volume risks severe toxicity. MRgART may result in plan quality improvement due to adaption workflows compensating for inter-fraction motion. MR-based dose calculation requires definition of electron density (ED) overrides. Optimal overrides were investigated to minimise contouring complexity, particularly during adaption, whilst maintaining acceptable calculation accuracy. The dosimetric impact of how ED overrides accounted for abdominal gas was considered, followed by the impact of intra-fraction gas variation on plan dosimetry. Results demonstrated that simple ED overrides on body and bone were appropriate for patients with small volumes of abdominal gas, however gas ED overrides were required to account for larger gas volumes. An approximate gas volume threshold of ~200 cc was defined. Transient gas significantly impacted plan dosimetry in some cases, however the impact was likely clinically acceptable. Based on this data, pre-treatment clinical management of abdominal gas has been recommended. Dose accumulation of GI OAR D0.5cc was performed for an initial cohort. The dosimetric impact of intra-fraction motion for MRgART, and impact of varying GI OAR planning risk volume (PRV) margins, was assessed, based on MR images coinciding with the end of treatment (MRpost) for all fractions. Results showed that cumulative GI OAR D0.5cc remained within planning tolerances for all patients, despite scaled tolerances being exceeded for several fractions. Initial toxicity data confirms safe delivery (no grade >2 toxicity). Simulation of alternative PRV margins suggests that OARs would have exceeded tolerances in several cases if no PRVs had been applied, however only the duodenum for 1 patient would have exceeded tolerance if 3 mm PRVs had been used. 3 mm PRV margins may offer a clinically appropriate risk-benefit balance between target volume coverage and OAR dose control. The dosimetric impact of respiratory motion, and associated interplay effects, had not been reported for MRgART pancreatic SABR and may be significant due complex plan modulation. A method was devised to predict dose delivery throughout the breathing cycle for individual patients, generate dose distributions for all phases, and sum distributions based on deformable registration. Results showed that the dosimetric impact of respiration was generally small and clinically acceptable. However, systematic under-reporting of stomach dose, and over-reporting of large bowel dose, on the reference plans was observed due to OAR location relative to the pancreas. This suggests a systematic bias in OAR dose reporting for these patients, which has not previously been demonstrated.

Date of Award	6 Jan 2025
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Robert Chuter (Supervisor)