2293 - Is Separating Targets Necessary in Single-Isocenter Multi-Targets Radiosurgery Treatments?

W. Gu¹, Y. Chen², L. Ma¹, M. Bieda¹, D. Sutton¹, J. Marcel¹, and S. Nagda¹; ¹Department of Radiation Oncology, University of Pennsylvania, Philadelphia, PA, ²Department of Radiation Oncology, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, PA

Purpose/Objective(s): VMAT-based single-isocenter multiple-targets (SIMT) are sensitive to rotational uncertainties and leads to greater possibility of dose coverage loss for off-axis targets. Instead of treating all targets in one plan, it is common practice to separate them into multiple groups based on their distance. It ensures the targets are within reasonable distance and therefore increases robustness, but resulting in increased planning complexity and lower delivery efficiency. Studies have yet to evaluate the impact of separating targets for multiple scattered targets. This work aims to investigate the dosimetric and robustness impact of separating strategy. Materials/

Methods: 10 patients (95 metastases, 4-19 per patient) were retrospectively included. The PTV was 1mm margin expansion to GTV. Two sets of plans were created using HyperArc technique: (1) A single SIMT plan (non-separated) included all PTVs. The PTV to isocenter distance ranged from 2.4cm to 9.5cm (mean 5.7cm). (2) Multiple SIMT plans (separated) so in each plan PTV center to isocenter distance was <= 5cm. 3-4 plans were needed to treat all targets. The mean dose, V6Gy and V12Gy of brain were calculated. The dose coverage against setup uncertainty was evaluated by shifting/rotating dose distribution around the isocenter. 1mm translations in three directions and 1° rotations around three axes (total 12 scenarios) were simulated. The dosimetric changes of GTVs were evaluated using V100% (relative volume covered by prescribed dose) and D99% (relative minimum dose delivered to 99% of the GTV).
Results: The non-separated plan increased brain mean dose from separated plan by an average of 22.3cGy (std 15.9cGy). The brain V6Gy V12Gy were increased by 3.3cc and 0.86cc on average in the non-separated plans. Across all 12 1mm/1° translation/rotation uncertainty cases, the average GTV [V100%, D99%] is [99.8%, 103.2%] in non-separated plans, compared with [99.9%, 109.0%] in separated plans. Under worst situation, 6/95 GTVs in nonseparated plans have D99% < 95% (lowest 90.5%), while only one GTV in separated plans have D99% < 95% (lowest 94.4%).
Conclusion: Compared with separating targets based on their distances, treating all targets in one plan has similar brain dose sparing with slightly increased low dose distribution. Considering setup uncertainty, non-separated plans maintained GTV dose coverage with average V100% similar to separated plans but degraded D99%. The delivery time of non-separated plans is 3-4 times faster than separated plans.