Advancing magnetic resonance imaging (MRI) guided adaptive radiotherapy techniques for prostate cancer

Abstract

Advances in radiotherapy technology have increased precision of treatment delivery, improved cure rates and decreased side effects. The advent of MR-Linacs, which are hybrids of an MRI scanner and linear accelerator, holds potential to further transform the management of cancers. MRI-guided adaptive radiotherapy (MRIgART) facilitates superior soft tissue definition within the pelvis and enables daily adaptive replanning allowing for possibilities such as dose escalation, tumour tracking and further sparing of normal tissue. Building upon previous research, this thesis delves into multiple aspects of this innovative treatment in the non-metastatic prostate cancer setting.

This thesis begins with an exploration of aspects of the MR-Linac workflow. I investigate the hypothesis that high levels of inter- and intraobserver variability exists in prostate delineation during online treatment and compare this to ‘normal’ variability which is evaluated in an offline setting on MR-Linac acquired MR images. The factors that might contribute to inter- and intraobserver variability are also investigated. Further work concentrates on whether autocontours, propagated by rigid and deformable image registration, using the Monaco treatment planning system can be utilised, instead of clinician-drawn contours, for daily treatment on the Elekta MR-Linac.

I then examine clinical outcomes of MRIgART for localised prostate cancer with moderately hypofractionated radiotherapy. I report toxicity, early biochemical response and patient-reported outcomes from the MOMENTUM study, a multicentre international registry study for patients who received treatment on the Unity MR-Linac.

With moderately hypofractionated radiotherapy established as the standard of care, the prostate cancer community has turned its focus towards ultrahypofractionated radiotherapy. The potential benefits of the MR-Linac can be harnessed to deliver this. I explore whether intrafractional changes in anatomy contribute to significant variations in dose delivered to organ-at-risk structures during 5-fraction radiotherapy.

Finally, I turn my attention to the postoperative setting. Significant variability exists amongst the multiple guidelines currently in use for delineating the prostate bed using CT-guided radiotherapy. I assess baseline interobserver variability in prostate bed contouring on MRI images obtained from the MR-Linac and using this data, create consensus guidelines for contouring the prostate bed clinical target volume (CTV) for MRIgART. Lastly, I investigate the impact of these guidelines on interobserver variability. Using these guidelines, further research looking at the benefits of MRIgART in the postoperative setting may be carried out.

The findings presented in this thesis lay the groundwork for realising the full potential of MRIgART for treating prostate cancer.