Accuracy of tumor setup using triggered KV images at multiple gantry angles.
Li H, Jackson P, and Wen N. Accuracy of tumor setup using triggered KV images at multiple gantry angles. Med Phys 2017; 44(6):2798-2799.
Purpose: To investigate the feasibility of using triggered kV images to quantitate tumor localization uncertainty during treatment of spine and lung cancer on a SRS-based linac platform. Methods: A Rando thorax phantom was used to simulate spine SRS treatment. A T-spine vertebra was contoured as the treatment target and treated with a full arc VMAT plan. Before delivering the treatment, the phantom was positioned on the couch with pre-set offsets in six dimensions (VRT, LNG, LAT, RTN, Pitch and Roll) from the planned treatment position. CBCT and kV-kV orthogonal images were acquired to quantify the offset. Treatment was delivered while monoscopic kV image were trigged at every 30 degree gantry angle. The beam was halted after each image and a 2D/3D match between the triggered image and DRR was performed to determinate the target offset. Similar procedure was repeated for lung using CIRS lung phantom. Results: For the longitudinal component, accuracy of triggered image was comparable to CBCT; the difference was 0 ± 0.5 mm (range = -0.6 to 1.3 mm). For VRT, LAT, RTN and Pitch components, results were similar to that of CBCTwhen the offset component was perpendicular to the kV beam direction but not reliable when the offset component was parallel to kV beam. For the Roll component, all the triggered images showed zero degree offset although CBCT showed -1.7 and 0.7 degree offset for the spine and lung phantoms respectively. Conclusion: Triggered kV images can quantitate motion components perpendicular to kV beam with reasonable accuracy. For the components parallel to kV beam, not surprisingly, results are not reliable. In clinical setting, triggered kV images can be used to measure offsets perpendicular to kV beam direction, but should be used in a qualitative manner for offset components parallel to kV beam.