2292 - Motion Analysis of Abdominal Targets Using Super-Resolution of 3D Cine MRI for MR-Guided Radiation Therapy
Presenter(s)
Y. Yoon1,2, J. Sung3, J. W. Kim3, J. S. Kim2,4, T. Kim5, and J. Kim6; 1Department of Radiation Oncology, Washington University School of Medicine, St. Louis, MO, 2Department of Radiation Oncology, Yonsei Cancer Center, Heavy Ion Therapy Research Institute, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 3Department of Radiation Oncology, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea, Republic of (South), 4Oncosoft, Seoul, Korea, Republic of (South), 5Wash U School of Medicine, Department of Radiation Oncology, St. Louis, MO, 6Department of Radiation Oncology, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Republic of Korea, Seoul, Korea, Republic of (South)
Purpose/Objective(s): Magnetic resonance imaging (MRI)-Linac offers real-time target monitoring during radiotherapy. Although thoracoabdominal motion typically presents in 3D, current real-time imaging is limited to 2D views centered on the target due to the spatiotemporal resolution trade-off of MRI. This study aims to analyze motion of abdominal targets using super-resolution (SR) 3D cine MRI.
Materials/Methods: We retrospectively collected 10 abdominal cancer patients (five liver and five pancreas) treated with a 1.5 T MRI-Linac. Abdominal compression was applied only to the five liver patients. For each patient, 100 3D cine MR scans were acquired (image dimension: 80 × 80 × 34, spacing: 5.0 × 5.0 × 6 mm3, temporal resolution: 1.5 s), and their spatial resolution was enhanced with a factor of four using an in-house developed SR network. Planning target volume (PTV) and liver segmented on the planning computed tomography were deformably registered to the cine MR scans. Motion analysis was conducted using a vendor-provided tool that computed the translation of a region-of-interest in each of two-dimensional cine MR images. Unlike the current standard real-time monitoring method in MRI-guided radiation therapy, we evaluated the target motion at several off-axis locations: 5, 10, and 15 mm in all axes from the target center in the right-anterior-superior (RAS) and left-posterior-inferior (LPI) directions.
Results: All SR MRIs were compatible with the vendor-provided tool; motion analysis using 3D SR MR images were successfully implemented by extracting its 2D slices-of-interest and subjecting them to the 2D cine-MR vendor-provided tool. Average relative motion increased with the distance from the center and varied across the patients, with superior–inferior (SI) movement exceeding the other directions. At 5, 10, and 15 mm from the center, the mean ± standard deviations were 0.42 ± 0.46, 0.61 ± 0.73, and 0.72 ± 0.71 mm (RAS) and 0.42 ± 0.36, 0.41 ± 0.34, and 0.49 ± 0.44 mm (LPI), respectively. Abdominal compression reduced SI displacement, with maximum differences of 2.8 mm at 15 mm (LPI) versus 5.6 mm at 15 mm (RAS) in non-compressed patients.Conclusion: This study analyzed the abdominal target motion differences at the center and periphery using SR 3D cine MRI and vendor-provided 2D motion tracking tool. Although these differences were generally within 1 mm, they were consistently present and tended to increase in more distal locations or in the absence of abdominal compression.