Quantification and correction of distortion in diffusion-weighted MRI at 1.5 and 3 T in a muscle-invasive bladder cancer phantom for radiotherapy planning

Objective: Limited visibility of post-resection muscle-invasive bladder cancer (MIBC) on CT hinders radiotherapy dose escalation of the residual tumour. Diffusion-weighted MRI (DW-MRI) visualises areas of high tumour burden and is increasingly used within diagnosis and as a biomarker for cancer. DW-MRI could, therefore, facilitate dose escalation, potentially via dose-painting and/or accommodating response. However, the distortion inherent in DW-MRI could limit geometric accuracy. Therefore, this study aims to quantify DW-MRI distortion via imaging of a bladder phantom. Methods: A phantom was designed to mimic MIBC and imaged using CT, DW-MRI and T2W-MRI. Fiducial marker locations were compared across modalities and publicly available software was assessed for correction of magnetic susceptibility-related distortion. Results: Fiducial marker locations on CT and T2W-MRI agreed within 1.2 mm at 3 T and 1.8 mm at 1.5 T. The greatest discrepancy between CT and apparent diffusion coefficient (ADC) maps was 6.3 mm at 3 T, reducing to 1.8 mm when corrected for distortion. At 1.5 T, these values were 3.9 mm and 1.7 mm, respectively. Conclusions: Geometric distortion in DW-MRI of a model bladder was initially >6 mm at 3 T and >3 mm at 1.5 T; however, established correction methods reduced this to <2 mm in both cases. Advances in knowledge: A phantom designed to mimic MIBC has been produced and used to show distortion in DW-MRI can be sufficiently mitigated for incorporation into the radiotherapy pathway. Further investigation is therefore warranted to enable individually adaptive image-guided radiotherapy of MIBC based upon DW-MRI.