Respiratory Motion Modeling on a Voxel-By-Voxel Basis Using Deformable Image Registration and 4D CT Imaging
Abstract
Purpose
To develop and test a four-dimensional (4D) motion model capable of accurately extracting respiratory motion on voxel-by-voxel basis using 4D-CT imaging and deformable image registration (DIR).
Methods
The 4D-CT images were acquired for a mobile phantom that was moved with a cyclic motion with different motion amplitudes ranging 5-35 mm. Four deformable image registration algorithms from the DIRART software were utilized to register the 4D-CT images. The 4D-CT images from different phases (10-90%) were registered to the 0%-phase CT-images. The 0%-phase was considered the reference CT image set and the displacement-vector fields (DVF) were calculated for the different motion phases relative to the 0%-phase. The DVF-values were calibrated with the motion amplitudes of the mobile phantom. The DVF were used to reconstruct the motion trajectory for each voxel in the CT-images for each phase.
Results
The DVF correlated with the motion amplitudes that were used to drive the mobile phantom. The DVF’s were to used to extract the shifts of the voxels in 3D for the CT-images from different motion phases (10-90%). The motion trajectories were reconstructed successfully for the individual voxel across the different phases in the respiratory cycle using the shifts calculated by the DIR algorithms. All four DIR algorithms produced similar motion patterns, with voxels moving away from the reference position as the phantom moved far from the reference position for the different motion patterns. Slight differences in the magnitude of the DVF’s were observed among the different algorithms, the overall motion trends remained consistent for the voxels around the diaphragm in the lung phantom.
Conclusion
The motion trajectories of the voxels were reconstructed using respiratory motion modeling with DIR on 4D-CT images. This motion modeling technique has clinical potential to perform adaptive radiation therapy considering respiratory motion and anatomical variability during treatment with radiation therapy.