Room: Exhibit Hall | Forum 2
Purpose: MR system distortions are the dominant source of contouring uncertainty in patients undergoing MR simulation for treatment planning. Even after vendor corrections, residual distortions can be up to 2-3 mm, which is not brought to physicians' attention in clinical practice. Our goal was to close this gap by implementing tools within a commercial treatment planning system (TPS) to present residual distortions to physicians during the treatment planning process.
Methods: Two alternative methods were developed for presenting and quantifying distortion maps within the TPS. With the first method, the magnitude of the distortions is converted into spatial "pseudo" dose map. With the second method we correct the MR images within the volume of measured distortions and register the corrected MR onto to the original MR images. To evaluate the utility of these methods we mapped the distortions on an MR simulator (Skyra, Siemens, 3T) through phantom measurements (MagphanRT, The Phantom Laboratory) and applied the techniques to phantom and patient images.
Results: Within a volume 35x27x21cm3 around isocenter, maximum spatial distortion was 1.9 mm. Evaluation of the methods revealed that visualizing the distortion magnitude as "pseudo" dose is easily interpretable and allows evaluation of distortion through surface dose mapping and volume quantification through inherent TPS DVH tools. The second method adds evaluation of magnitude and distortion directionality.
Conclusion: While fully correcting the MR images is conceptually the ideal way of handling MR distortions, comprehensive and reliable measurements of distortions on per patient (protocol) basis is currently not feasible in routine busy clinical practice. Instead, the evaluation of the distortion magnitude as pseudo dose together with assessment of the partially corrected images gives an efficient and powerful tool to assess the areas of uncertainties due to distortion, information which might add in the clinical decision making in target and risk organ delineation.
Not Applicable / None Entered.