Room: AAPM ePoster Library
Purpose: Magnetic Resonance Imaging (MRI) is widely expanding in Radiotherapy (RT) owing to the excellent soft tissue and the wide range of multiple contrast. However, intrinsic geometric distortion caused by magnetic field inhomogeneity and gradient nonlinearity has led to erroneous estimation the target volume delineation compared to CT. In this study, the geometric accuracy was evaluated using modular phantom with LEGO compatibility for versatile application.
Methods: The modular phantom was fabricated by Acrylonitrile butadiene styrene (ABS). This rectangular configuration phantom has stud-and-tube coupling system for LEGO Duplo compatibility. The dimension of phantoms with 12.4 mm x 6.2 mm x 7.2 mm was directly proportional to the LEGO Duplo bricks. The modular phantom was filled with water. Also, LEGO Technique bricks containing holes were construed as a grid type to evaluate the geometric distortion. Enlarge phantom was constructed by of modular phantoms assembly. The assembled modular phantom was scanned on Siemens Skyra 3.0 T (Siemens Medical Solutions, Erlangen, Germany) on head and body matrix coils. The degree of geometric distortion was calculated by the difference known size of hole and measured size of hole of LEGO Technique bricks on MR images.
Results: A highly customizable phantom was developed using modular type with LEGO coupling and implemented to evaluate the geometric distortion of MR images. No significant artifacts caused by phantom shape and LEGO Technique bricks were appeared on images. Modular phantoms and assembled modular phantom were suitable for the configuration of head and body coil. The holes and distortion were easily recognized in MR images of each coils.
Conclusion: The results show that modular phantom with LEGO compatibility is extremely versatile for geometric distortion evaluation of MRI in radiation therapy. Any size and configuration of a specific phantom can be assembled by the developed modular phantom in radiotherapy using MRI.
Not Applicable / None Entered.