Room: ePoster Forums
Purpose: To determine the necessity of the enormously complex and costly gantry for carbon ion therapy, beam orientation optimization (BOO) is used to create and compare gantry-based plans vs. plans using two fixed beamlines and a robotic couch.
Methods: BOO was performed for the two different scenarios. The fixed beam plans included 42 beams combining two fixed gantry angles at 45Â° and 90Â°, respectively, and couch angles ranging from -60Â° to 60Â° with 6Â° interval. The gantry beam plans started with 1162 non-coplanar beams that uniformly distributed across the 4Ï€ steradians combining the gantry and couch rotational degrees-of-freedom. The number of beams was further reduced to 772 by eliminating beams with couch angles between 60Â° to 120Â° and -60Â° to -120Â° to exclude geometrically undesired beams. The carbon ion pencil beam dose was calculated using matRad. For both scenarios, the BOO and physical dose optimization was formulated to include a quadratic dose fidelity penalty and an L2,1/2-norm group sparsity term to select 2 beams from the candidate beam pool. The Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) was used to solve the optimization problem. The dosimetry comparison was performed on a complex head-and-neck (H&N) patient with three prescription levels.
Results: The selected two beams from the two fixed gantry angles achieved comparable dosimetry as beams selected from a significantly larger beam pool using the gantry-based plan. The gantry and static-beamline plans took 2 hours and 6 minutes to generate, respectively.
Conclusion: We showed that the dosimetry of carbon ion therapy using fixed beamlines is not disadvantageous in comparison to fully-rotating gantry delivery. This work indicates the potential to significantly simplify gantry design for carbon ion therapy thus overcoming a major hurdle in availing this technology. The future study will incorporate biological effect of carbon ions.
Funding Support, Disclosures, and Conflict of Interest: This research is supported by NIH Grants Nos. NIH R01CA230278, NIH R44CA183390 and NIH R01CA188300
Not Applicable / None Entered.