Room: AAPM ePoster Library
Purpose: Multi-metastatic stereotactic radiosurgery involves complex VMAT planning with multiple targets, often close to Organs-At-Risk (OARs). Template-based radiosurgery assigns beam trajectories according to target location within the brain. This simplified planning approach ignores all patient-specific information. We developed a Beam Angle Optimization (BAO) algorithm using Eclipse™ Scripting API (Palo Alto, California) to automatically determine optimal table and collimator angles, and gantry trajectories using patient anatomy and target geometry, and assessed plan quality compared to a template-based approach.
Methods: A retrospective analysis was performed for 36 multi-metastatic patient plans previously treated on the Truebeam Edge™ platform. Optimized plans used the same isocenter and number of trajectories as the template-based plan. Expanding on previous work, a collision prediction algorithm was used to determine the available patient-specific treatment space. Optimal table and collimator angles and beam trajectories were determined using multi-leaf collimator coverage of the target geometry and OAR overlap avoidance, with higher priority given to OAR sparing. BAO plans were re-optimized using the Eclipse™ v13.6 photon optimizer and normalized to match clinical target dose coverage. Dose-volume criteria for OARs were compared between clinical and BAO plans using Wilcoxon signed-rank tests (a=0.05).
Results: Median maximum dose was reduced for the brainstem, right lens, right optic nerve, left optic tract, and cochlea, with statistically significant reductions in brainstem, chiasm, right optic nerve, and optic tract dose (p<0.05) when comparing BAO to template-based plans. Marginal increases in median maximum dose were observed for the left lens and left optic nerve, but were not statistically significant. Statistically significant increases in median normal brain tissue volume were seen for V10Gy and V12Gy (+2.34cc, +0.84cc), reflecting the higher priority given to OAR sparing.
Conclusion: Statistically significant dose reductions were achieved for several OARs, while maintaining target coverage, with a slight increase in normal brain tissue volume irradiated.