Development and validation of a dose accumulation workflow for MR-guided adaptive radiotherapy.

Type of content
Theses / Dissertations
Publisher's DOI/URI
Thesis discipline
Medical Physics
Degree name
Master of Science
Publisher
Journal Title
Journal ISSN
Volume Title
Language
English
Date
2024
Authors
Taylor, Megan
Abstract

During radiotherapy treatment, anatomical changes and discrepancies in patient setup between fractions may result in the planned dose and actual delivered dose differing. In conventional radiotherapy, it is common practice to assume that over a full course of treatment, these differences in delivered and planned dose are negligible. This is because positioning devices, margins on target volumes, etc. define “acceptable” levels of error.

With the development of adaptive radiotherapy, images of the patient can be obtained with every fraction, allowing visualization of how the patient anatomy differs from the reference plan setup. Along with visualizing positioning changes, the Elekta Unity MR-Linac system allows for alterations to the treatment plan on each fraction to correct for these positional changes. This, in theory, should reduce the discrepancy in planned dose and delivered dose. In practice, delivered dose still differs from fraction to fraction. Without summing delivered dose per fraction, it is a qualitative assumption that the planned dose has been delivered. A quantitative method allows for reassurance and more robust adaptive planning. The aim of this project was to develop a workflow to accurately measure accumulated dose inter-fractionally – considering the affects of anatomical/positioning changes. The workflow was then to be validated and compared to planned dose.

Using data of ten patients previously treated on the MR-Linac for prostate cancer, a deformable image registration algorithm was optimised and dose mapping methods were assessed. The finalised components were combined in a dose accumulation workflow.

Uncertainties were estimated for the workflow and synthetic deformations used to validate the method. It was found that median uncertainties were generally below 1%. However, point dose uncertainties exceeded 40% in some cases where regions of high dose gradients occurred. Comparison to known deformations showed a high level of agreement in dosimetric criteria (gamma passing rates above 95%).

A comparison of planned and accumulated dose showed that there were statistically significant differences in PTV coverage suggesting dose accumulation may be a useful tool for improving understanding of planned vs delivered dose and should be considered for applications in adaptive radiotherapy.

Description
Citation
Keywords
Ngā upoko tukutuku/Māori subject headings
ANZSRC fields of research
Rights
All Rights Reserved