Back to search

University of Leeds

Assessing the accuracy of artificial intelligence synthetic CT generation for liver and brain MRI-only radiotherapy

Abstract

dc:description.abstract

Background: Magnetic resonance imaging (MRI) is increasingly integrated into radiotherapy because of its superior soft-tissue contrast compared with computed tomography (CT). This has prompted interest in four-dimensional (4D) MRI for motion management and MRI-only radiotherapy using synthetic CT (sCT) for dose calculation and patient positioning verification. This thesis aimed to provide clinical evidence for the technical feasibility and clinical implementation of MRI-only radiotherapy for liver and brain cancer. Methods: (i) A PRISMA-guided systematic review of the 4D MRI literature for abdominal radiotherapy was conducted. (ii) A deep-learning sCT model was developed using clinical MRI and CT data to generate liver MRI-only radiotherapy. (iii) The performance of a commercial sCT solution (Philips MRCAT) was assessed for brain MRI-only radiotherapy. For both liver and brain, dosimetric accuracy was evaluated using dose volume histogram (DVH) analysis. In addition, image-guided patient positioning was verified using the clinical XVI system. Results: (i) The systematic review, encompassing 39 studies, indicated that 4D MRI had the potential to improve abdominal radiotherapy by enabling accurate tumour definition and motion characterisation compared to 4D CT. (ii) For the liver sCT model, relative mean dose differences between CT and sCT were 0.0% for the planning target volume (PTV) and <0.5% for all organs at risk (OARs). Positioning verification revealed mean translational and rotational differences of <0.5 mm and <0.5°, respectively. (iii) For the brain MRCAT, relative mean dose differences were <0.4% for the PTV and <0.3% for OARs, with positioning accuracy maintained within ±1 mm and ±1°. Conclusion: 4D MRI shows considerable promise for motion management, but its clinical implementation remains limited by lack of robust clinical validation or standardisation. Both liver and brain sCT models demonstrated dosimetric and positioning accuracy comparable to CT, confirming the technical feasibility of MRI-only radiotherapy for the liver and its clinical applicability for the brain.

Degree

thesis:*
Name dc:type.qualificationname
Ph.D
Level dc:type.qualificationlevel
doctoral
Grantor dc:publisher.institution
University of Leeds
Year dc:date.issued
2025

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Aljaafari, Lamyaa Khalid
Advisors dc:contributor.advisor
  • Richard, Speight
  • David, Buckley
  • David, Bird
  • Bashar, Al-Qaisieh
ORCID person:identifier.orcid

Rights

Language dc:language.dcterms_iso639_2
eng

Chain of custody

source
Harvested from
White Rose University Consortium
Base URL
etheses.whiterose.ac.uk/cgi/oai2
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
related terms
citation

Aljaafari, Lamyaa Khalid. Assessing the accuracy of artificial intelligence synthetic CT generation for liver and brain MRI-only radiotherapy. doctoral thesis, University of Leeds, 2025.