Modeling and uncertainty quantification of motion of lung tumors for image guided radiation therapy

Target localization is a key issue in the image guided radiation therapy procedures for treating tumors in thorax and abdomen. Breathing induced tumor motion necessitates larger margins during radiation therapy planning which may be harmful for healthy tissue surrounding the tumor. Large sampling time in data acquisition and latencies involved in real time imaging systems and tracking system pose a significant challenge to target localization. A framework based on pulmonary mechanics is developed to predict and precisely track the breathing induced motion of lung tumor to direct the tracking system to an estimated position instead of an observed one. A hybrid approach based on the correlation of real-time imagery data of internal markers and easy to measure external respiratory signals like flow readings etc, is proposed to support dynamic radiation therapy procedures. Issues related to reliability of proposed model predictions in the presence of parametric uncertainty are explored using Polynomial Chaos Expansion.