Respiratory gated proton therapy for liver cancer: Setup and gating validation based on fiducial marker motion extracted from 2D kV projection images acquired during cone-beam CT

Abstract Purpose To investigate the use of CBCT projection images for improved setup and gating validation in respiratory gated proton therapy for liver cancer. Methods Fifteen consecutive patients receiving proton therapy with respiratory gating guided by an external gating marker block were analyzed. The amplitude gating level was set to 50% duty-cycle around the exhale phase (50th percentile motion). Implanted markers were retrospectively segmented in the 2D projection images of five free-breathing setup CBCT scans per patient and used to estimate the 3D marker trajectories during CBCT acquisition. The trajectories and synchronized external gating block motion were used to investigate: (1) Setup accuracy of manual online CBCT marker match, (2) motion during CBCT acquisition compared to the planning 4DCT motion, and (3) internal motion during CBCT acquisition in external gate on periods. Results (1) The mean (±SD) online marker match error was 0.0 ± 0.4 mm (left–right), 1.4 ± 0.9 mm (cranio-caudal) and −0.6 ± 0.6 mm (antero-posterior). (2) The mean full marker motion during CBCT acquisition was 17.7 ± 4.4 mm (cranio-caudal). The 50th percentile motion was 5.6 ± 2.4 mm. The difference between CBCT and planning 4DCT motion was 5.8 ± 4.0 mm for the full motion and reduced to 1.0 ± 2.1 mm for the 50th percentile motion. (3) The internal motion during CBCT acquisition in external gate on periods exceeded the 4DCT planned internal 50% gating window by more than 3 mm (cranio-caudal) for 1.2 ± 2.6% of the time. Conclusions Detailed analyses of CBCT projections provided important pre-treatment setup and gating validation for liver cancer proton therapy. Online implementation may be feasible.