Analysis of tumor-influenced respiratory dynamics using motion artifact reduced thoracic 4D CT images

Respiratory motion represents a major problem in radiation therapy of thoracic tumors. Methods for compensation require comprehensive knowledge of underlying dynamics. In this study, motion of thoracic anatomical and pathological structures in lung cancer patients was analyzed using motion artifact reduced 4D CT data sets of high temporal and spatial resolution. Motion artifact reduction was achieved by applying an optical flow based 4D CT reconstruction method. Motion analysis especially focuses on inner organ and tumor mobility and the interrelation between tumor/inner organ motion and chest wall motion. Trajectories of tumor mass centers and organ specific landmarks were determined and analyzed. To study chest wall motion a non-linear registration based point tracking scheme was applied to compute trajectories of points on the chest wall skin. The interrelation of chest wall and tumor/inner organ motion was investigated using methods of multivariate statistics. Results show that, for instance, tumor motion patterns differ noticeably between the patients; a dependency between tumor motion and tumor location seems apparent. The correlation of tumor motion and motion of chest wall points depends on the patient breathing pattern (e.g. abdominal or chest wall breathing). Thus, skin regions which are suitable for prediction of tumor motion differ between the patients.