Aspects of MR image distortions in radiotherapy treatment planning

Abstract

Background: Registration of computed tomography (CT) and magnetic resonance (MR) images are commonly performed to define the different target regions used in radiotherapy treatment planning (RTTP). The accuracy of target definition will then depend on the spatial accuracy of the CT and MR data, and on the technique used to register the images. CT images are usually regarded as geometrically correct, while MR images are known to suffer from geometric distortion. The aim of this paper is to discuss the possible impact of MR image distortions in the radiotherapy treatment planning process. Methods: The origin, magnitude, and relative impact of the different sources of geometric distortions that affect the MR image data at different magnetic fields and for different acquisition settings are described. Techniques for distortion correction are reviewed, and their limitations are outlined. The sensitivity of image registration techniques to the presence of geometric distortions in the MR data is discussed. Finally, an overview of image registration techniques used and results obtained in clinical radiotherapy treatment planning applications is given. Results: Spatial distortions in MR images vary with field strength and with the image acquisition protocol. The spatial accuracy generally decreases with distance from the magnet isocenter. Distortion correction techniques based on phantom evaluations cannot adequately model patient-induced distortions. Conclusion: Image protocols with high gradient bandwidths should be used to reduce the spatial distortions in MR images. Correction techniques based only on phantom measurements could be sufficient at low magnetic fields, while at higher fields additional corrections of patient-related distortions might be needed. Registration techniques based on matching of landmark points located far from the magnet isocenter are especially prone to MR distortions.