Room: AAPM ePoster Library
Purpose: To determine patient specific collision zones which will be integrated into automated trajectory calculation for 4-Pi treatment planning. This system assures the creation of safe treatment trajectories which will not lead to collisions. Initial 4-Pi cranial studies utilized a single manually measured collision zone for trajectory planning, which is insufficient for clinical applications.
Methods: collision detection system utilized a point cloud of the patient obtained from the treatment planning CT. The couch, immobilization equipment and gantry point clouds were automatically aligned to the patient. The patient contour was extended to account for the remainder of the body. Each combination of couch and gantry angles were tested for collision using an OcTree detection method. Treatment arcs were generated for eight patients using both the specific and general collision zones applied for 4-Pi treatment planning. These arc trajectories were imported into the Eclipse treatment planning system (v13.6) and inverse-optimized with VMAT (PRO v13.6) according to clinical standards. The mean and maximum dose to several OARs, conformity index of the PTVs and trajectories were compared. By overlaying the trajectories from the patient’s specific collision map onto the general collision zone map, potential collisions may be identified.
Results: difference in the maximum and mean dose to the OARS when comparing the general and patient-specific collision zones and conformity index values, were statistically insignificant (p > 0.7). At least one possible collision was identified for the remaining seven patients if the general arc trajectories were delivered.
Conclusion: patient specific collision zones for 4-Pi SRS/SRT is required to assure the calculation of treatment trajectories which are safe to deliver. Patient specific collision zones did not affect the treatment outcomes when examining the dose to OARs and are essential clinical implementation for 4-Pi SRS/SRT and for further extension to 4-Pi SBRT.
Funding Support, Disclosures, and Conflict of Interest: This research is supported in part by BrainLab and the Atlantic Canada Innovation Fund. J Lincoln reports funding from the Canadian Institute for Health Research.