In dynamic image-guided radiotherapy for lung cancer, one of the major tasks is to provide the dynamic images and tumor shapes that reflect the patient's real-time anatomy as the roadmap for guiding the delivery of radiation beams. One fundamental question for these applications is how to estimate such dynamic images, as well location and shape changes of tumor using available sensors to capture the respiratory motion. This proposal focuses on solving such a lung motion tracking problem by using our newly proposed high- dimensional surface to lung motion prediction model and considering the difference of each individuals, such as gender, size, and respiratory pattern. Specifically, we wil optimize the statistical models that capture the motion distribution from training samples and the nonlinear prediction model for accurate lung motion tracking and conduct extensive evaluation for the lung motion tracking system developed to validate its feasibility in clinic practice. Our goal is to develop an efficient, effective and robust lung motion tracking system for dynamic image guidance of the radiotherapy procedures. After this clinical data validation, such a technique can also be applied to image-guided intervention.
Lung motion tracking is essential for dynamic image-guided radiotherapy for lung cancer. We propose to estimate the dynamic lung motion for patient using the high-dimensional chest surface motion. Due individual variability (patient size, gender, and different respiratory patterns), in this proposal, we will optimize the newly proposed lung motion tracking by considering practical patient variability and evaluate the performance of the proposed system with a large number of dataset collected from radiotherapy planning.
| He, Tiancheng; Xue, Zhong; Teh, Bin S et al. (2015) Reconstruction of four-dimensional computed tomography lung images by applying spatial and temporal anatomical constraints using a Bayesian model. J Med Imaging (Bellingham) 2:024004 |
