Room: Exhibit Hall | Forum 5
Purpose: To implement and validate a system to monitor intra-fraction motion of paraspinal patients receiving SBRT treatment using kV projections on the Elekta XVI platform acquired during treatment.
Methods: Simulated projections from a planning CT were generated using Simple RTK and registered to acquired kV cone-beam projections using Simple ITK. A custom-built anthropomorphic phantom with a human skeleton was used to quantify the uncertainty of our monitoring pipeline as a function of gantry angle and applied displacement. Known shifts of up to 1 cm in the plane of the kV panel were applied. To date, intra-fraction projections of 3 patients receiving VMAT spine SBRT treatment were acquired with a gantry angle spacing of 1Â°. Intra-fraction motion, as assessed by our pipeline, was quantified and compared to the shifts determined by registering the reconstructed intra-fraction CBCT to the planning CT.
Results: Phantom results demonstrated that the pipeline was able to register images with a mean accuracy of 0.82 Â± 0.29 mm and 0.51 Â± 0.21 mm for displacements up to 3.2 mm and 1 cm in the X and Y axes of the panel respectively. This excluded a gantry angle range of 36Â° where a metal bar on the couch appeared in the scan. Results from the 3 patients showed average superior-inferior shifts of up to 1.2 mm with maximum displacement reaching 2.5 mm observed in the anterior-posterior direction. Larger registration errors were observed for lateral projections and were excluded from the analysis. Up to 1.9 mm of positional error in the superior-inferior direction was determined using the intra-fraction CBCT.
Conclusion: Phantom results have demonstrated that the pipeline provides sub-millimeter monitoring accuracy of vertebrae. Patient results show that displacement during treatment determined by our monitoring method were higher than that observed with the intra-fraction CBCT.