WE‐A‐134‐09: Real‐Time Volumetric Imaging Based On PCA Lung Motion Model and Vision RT

Purpose: Real‐time volumetric imaging is highly desirable to provide instantaneous image guidance for lung radiation therapy. This study proposes a scheme to achieve this goal by using Vision RT surface camera and a principal component analysis (PCA) based lung motion model. Methods: A patient‐specific PCA‐based lung motion model is first constructed by analyzing deformable registration vector fields between a reference image and 4DCT images at each phase. In the volumetric image reconstruction stage, we only use the first 2–3 principal components, which can capture the major component of the lung motion. Specifically, we solve the reconstruction problem from an optimization approach, where the PCA coefficients are optimized to minimize the distance between the deformed patient surface and the one captured by the Vision RT camera. The optimization problem is solved by a particle swarm simulated annealing algorithm. The resulting PCA coefficients are used to generate a deformation vector field that is applied to the reference image, yielding a volumetric image corresponding to the surface measured by the Vision RT camera. We have validated our method through simulation studies on a NCAT phantom. Results: The 4DCT image at the end‐of‐exhale phase is set as reference. We have successfully reconstructed a volumetric image based on a captured surface image in an area of size 16cm×16cm. The average relative pixel intensity difference between the reference and the ground truth image is 1.3%, whereas that between the reconstructed image and the ground truth is 0.5%. We have also compared anatomical feature location, e.g. diaphragm, and satisfactory results have been observed. Conclusion: The proposed method is capable of generating a volumetric image based on the instantaneous surface image offered by the vision RT camera. This scheme will potentially offer real‐time volumetric image guidance to facilitate lung radiotherapy.