Primary content

Moving towards MRI alone radiation therapy treatment planning: What is the impact of MRI distortion?


P Metcalfe (Lead)

A Walker

G Liney

L Holloway


MRI provides unrivalled detail on tumour and normal tissue location, extent and biological status. Evidence for benefit of use of MRI is clear for some clinical sites and emerging in others (Metcalfe et al Technology in cancer research and treatment journal, 12, 429-446, 2013). One technical difficulty associated with use of MRI is image distortion. The aim of this project is to assess the impact of MRI distortion within the radiation therapy planning process, particularly for MRI alone planning.


We have previously demonstrated that MRI distortion varies between sequences and scanners (Walker et al Australasian Physical & Engineering Sciences in Medicine, 37, 103-113, 2014). However this work was limited due to the commercial phantoms which were available, with limitations in field of view (FOV) and in particular the phantom length. To overcome this problem we have constructed a new phantom (the Sunrise phantom). This phantom enables image distortion to be assessed for the full FOV and 512mm scan length. We have used this phantom to assess a number of MRI sequences on a 3T Siemens Skyra scanner and then used this analysis to simulate the potential impact of this distortion for MRI alone breast radiotherapy planning.


The Sunrise phantom has been successfully constructed and imaged for the required extended FOV and length. For 6 sequences considered, the greatest distortion observed was 9.8mm and distortions were demonstrated to be less than 2mm, at distances within 100mm of isocentre.


The phantom can be used to accurately assess distortion for the entire MRI FOV, as required for radiotherapy treatment planning. It will also have implications for developing technologies such as the Australian MRI-linac, currently under development at Ingham. MRI-linac technologies will require direct use of MRI data hence the distortion corrections will be essential for this equipment.

Acknowledgement:  MRI-linac project is part of a NHMRC program grant. CIs: P Keall, S Crozier and M Barton.