Abstract
Purpose: In the simultaneous motion estimation and image reconstruction (SMEIR) algorithm for 4D cone-beam CT (4D-CBCT), the motion model is obtained by enforcing a global smoothness regularization term on the motion fields. The objective of this work is to enhance the performance of the SMEIR for 4D-CBCT by using a multi-organ meshing model to explicitly consider the discontinuity in the motion fields between different organs during the respiration. Methods: In the SMEIR algorithm, the deformable motion between a reference phase 4D-CBCT and other phases is obtained by matching the 4D-CBCT projections at each phase with the corresponding forward projections of the deformed reference phase. In this work, the sliding motion between lung and thoracic cage is considered by a multi-organ meshing model. Specifically, lung and thoracic cage are first segmented from a reference phase 4D-CBCT. Multi-organ tetrahedral meshes are then created from the segmented images to control different motions for different organs during the respiration. In the multi-organ meshes, continuous motion is enforced within each organ and along the normal direction of the organ interface while discontinuous motion is allowed along the tangential direction of the organ interface. The updated motion model is then used in the motion-compensated image reconstruction step in the SMEIR algorithm. Results: In 4D NCAT digital phantom, normalized cross correlations between reconstructed image and the ground truth image for homogeneous and multi-organ mesh-based methods are 0.9893 and 0.9920, respectively; the maximum motion errors along the interface between the lung and thoracic cage for homogeneous and multi-organ mesh-based methods are 6.12 mm and 0.95 mm, respectively. Conclusion: The multi-organ mesh-based 4D-CBCT reconstruction method can estimate the motion fields accurately by considering the sliding motion between different organs. Both image quality and motion estimating accuracy are improved by the proposed method as compared to traditional homogeneous mesh-based approach.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have