Purpose: This works investigates the feasibility of real-time motion tracking (RMT) for abdominal targets using a template registration-based algorithm between real-time 2D-cine and 3D-MRI on MR-Linac.
Methods: The RMT was evaluated using 9 sets of daily mid-position images (MidP) derived from an in-house 4D-MRI and balanced T2/T1-weighted bFFE cine images acquired with free breathing on a 1.5T MR-Linac for 8 patients with abdominal tumors. The cine images were acquired with a temporal resolution of 0.2 s for a total duration of about 10 minutes interleaved between coronal and sagittal orientations. A 2D-template was created in both orientations using data acquired over the first 20 s, followed by a rigid registration to the 3D-volume using a mask defined on the MidP. Motion tracking of the target was performed by registering the rest of the 2D-cine series with the corresponding template with a known relationship to the 3D-volume. Manually delineated contours on sampled cine frames were used as the ground truth. Tracking accuracy was defined as the standard deviation of the error (SDE) comparing the center-of-mass (CM) of the ground truth and the tracked motion.
Results: The tracking algorithm was qualitatively validated by visual assessment of 9 abdomen sets for the 8 patients. It was quantitatively validated on 4/7 Pancreatic and 2/2 Renal Fossa cases. The mean SDEs (and ranges), were 1.08 mm (0.31-1.9 mm), 1.07 mm (0.33-1.68 mm), and 0.76 mm (0.37-1.24 mm) in superior-inferior, left-right, and anterior-posterior directions, respectively. For the three remaining pancreas cases, performing tracking with ROIs including a margin encompassing the target demonstrated the robustness of the algorithm to implant-induced image artefacts in two cases and suboptimal imaging plane selection.
Conclusion: We have demonstrated the feasibility of accurate free-breathing motion tracking using the template registration for abdominal targets based on real-time orthogonal cine on MR-Linac.
Funding Support, Disclosures, and Conflict of Interest: Funding Support by Elekta