osimetric impact of system spe cific geometric distortion on head and neck treatment plans in MRI only based radiation therapy

Abstract: Background/ purpose: Because of the superior soft tissue contrast, in comparison with computed tomography (CT) data, magnetic resonance imaging (MRI) is of high interest to include in a larger extent in the ra diation therapy (RT) process. In an MRI only workflow, the CT scan is excluded and a synthetic CT is used for treatment plan ning. This will spare the patient extra ionizing radiation alongside with the discomfort of two separate imaging scans. However, an awareness is that geometric distortions appear in the MRI data. If uncorrected, the distortions may translate into dosimetric errors. In this study, the MRI system specific distortions were quanti fied and investigated. Dosimetric effects caused by the geometric distortions on head and neck treatment plans were also evaluated. Method: The MRI system specific distortion field was quantified by scan ning a phantom using the clinically optimized pulse sequence in a 1.5 T MRI system. The inverted system specific distortion field was calculated and applied on the patient CT and RT structure data. Both the patient CT and RT structure data were trans formed into the corresponding patient MRI frame of reference (FoR) before the inverted distortions were applied. The dosimet ric effects of the system specific distortions were investigated by creating an optimized treatment plan on the distorted patient CT data. The optimized treatment plan was copied and recalculated on the undistorted patient CT data and the absorbed dose dif ference was analyzed in clinically relevant dose volume histogram (DVH) parameters. Result: The mean geometric distortion for a radial distance of 200–250 mm from the MRI isocenter was 1.5 mm. The smallest and largest mean absorbed dose deviation was -0.19% and 1.1% for D2% CTV68 and Dmean for right parotid gland, respectively. Conclusion: In this study, the dosimetric impact of the MRI system specific distortions was found to have none, or only minor, clinical rele vance for head and neck cancer treatment plans in an MRI only workflow.