Medical Physics Seminar – Monday, September 12, 2011
4D Cone Beam Computed Tomography Using Prior Image Constrained Compressed Sensing: Clinical Implementation & Evaluation
Zhihua Qi (student of Guang-Hong Chen)
Research Assistant, Department of Medical Physics, UW-School of Medicine & Public Health, Madison, WI USA
To enable 4D imaging and on-line verification of lung tumor motion inside the treatment room
prior to the treatment delivery, 4D Cone Beam CT (4DCBCT) using an on-board imager (OBI) has
been actively investigated. However, its clinical application faces considerable challenges from data
undersampling-induced streaking artifacts. The recently proposed 4DCBCT method using the
PICCS reconstruction algorithm (PICCS-4DCBCT) has shown promise in providing high SNR
4DCBCT images with the standard 1-minute data acquisition protocol. To extend the method into a
clinically feasible technique for tumor motion delineation, two questions need to be addressed: (1)
how to implement the technique to meet the clinical needs; and (2) how to evaluate the technique
under clinically relevant conditions. This work presents the clinical implementation and
performance evaluation of PICCS-4DCBCT. The results demonstrate that tumor motion
trajectories with both high temporal resolution and high accuracy can be extracted by a deformable
registration algorithm from PICCS-4DCBCT images and can be used to triage clinical actions.