In silico modelling of radiobiological mechanisms in proton beam therapy

Abstract

Radiotherapy is premised on being able to deliver lethal doses to the tumour whilst sparing the healthy tissue surrounding it. Proton therapy is an increasingly utilised modality of radiotherapy which has a favourable dose distribution potentially lowering doses to normal tissues. A key question of radiobiology surrounds how the DNA responds to damage following irradiation of varying radiation qualities. A driver of this is to fully understand the relative biological effectiveness between photon and proton irradiation. If achieved, this would allow the leverage of the many more years of clinical experience we have with photons to be applied to proton therapy. However, the pursuit of quantifying relative biological effectiveness has been met with large uncertainties when explored experimentally, limiting clinical confidence. In turn, this has prompted a more mechanistic understanding of how the characteristics of different incident radiation damaging the DNA alters the cellular response. This mechanistic exploration challenges many areas of fundamental radiobiology as it requires `solid foundations' to build upon. In this thesis, several models were developed to tackle the issues in fundamental radiobiology modelling. Firstly, I develop a DNA Mechanistic Repair Simulator (DaMaRiS) model to describe how the two dominant DNA double-strand break repair pathways (Homologous Recombination and Non-Homologous End Joining) can function together. This model demonstrated how different repair pathways should not be thought of as purely antagonistic, as some of the steps in each pathway can be complementary, resulting in an entwined pathway. Secondly, a new methodology of modelling the DNA arrangements for radiation simulations was introduced in the form of solving Hi-C experimental data. This work, which culminated in the G-NOME solver software, demonstrated how cell types have significant differences in how their chromatin is arranged, leading to changes in how DNA damage would be distributed which is known to impact cellular fate. Finally, we purpose a computational methodology for characterising the miss-identification (miscounting) of DNA damage foci in immunofluorescence work. The developed tool, called PyFoci, identifies how the miscounting error is not constant and varies across experimental conditions, which when not accounted for, can result in perceived changes even when non-exist. These models in culmination allow for more clarity in the continued development to describe the DNA damage response. They present tools that can be adopted by others in the field for both scientific exploration and validation of their own approaches. This work continues the development of models which can offer clinical utility and promote the use of biological optimisation for patient treatments. The University of Manchester, Samuel Peter Ingram, Doctor of Philosophy. In silico modelling of radiobiological mechanisms in proton beam therapy'', 21st September 2021.

Date of Award	1 Aug 2022
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Ranald Mackay (Supervisor), Karen Kirkby (Supervisor), Mike Merchant (Supervisor) & Norman Kirkby (Supervisor)