Predicting respiratory motion for real-time tumour tracking in radiotherapy

Abstract

Radiation therapy is a local treatment aimed at killing cells in and around a tumor. Accurate predictions of lung tumor motion help to improve the precision of radiation treatment by controlling the position of a patient during radiation treatment. Our goal is to develop an algorithmic solution for predicting the position of a target in 3D in real time. In addition to prediction accuracy and low fluctuation of the prediction signal (jitter) we aim for minimum calibration time each patient at the beginning of the procedure. Our solution is based on a model form from the family of exponential smoothing. Performance is evaluated on clinical datasets capturing different behavior (quiet, talking, laughing), and validated in real-time on a prototype with respiratory motion imitation. Proposed solution (ExSmi) achieves good accuracy of prediction (error 4-9 mm/s) with tolerable jitter values (5-7 mm/s). The solution performs well to be prototyped and deployed in applications of radiotherapy