2079 - Isocenter Optimization for Robust VMAT-Based Single-Isocenter Multiple-Targets SRS
Presenter(s)
E. Bilokon1, Y. Chen2, L. Yin3, M. R. Bieda4, K. K. Teo3, and W. Gu4; 1Department of Radiation Oncology, University of Pennsylvania, Philladelphia, PA, 2Department of Radiation Oncology, Sidney Kimmel Medical College at Thomas Jefferson University, Philadelphia, PA, 3Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, 4Department of Radiation Oncology, University of Pennsylvania, Philadelphia, PA
Purpose/Objective(s): Volumetric modulated arc therapy (VMAT)-based single-isocenter multiple-target (SIMT) plans are sensitive to rotational and translational uncertainties occurring in setup as well as intra-fractional motion, which can lead to dose coverage loss, particularly for off-axis targets. The extent of coverage loss depends on factors such as margin, gross tumor volume (GTV) size, and the distance between the GTV and isocenter. Importantly, isocenter positioning plays a key role in plan robustness. This study aims to develop an optimized isocenter placement strategy to enhance robustness of clinical SIMT plans against rotational and translational errors.
Materials/Methods: Eight patients (4–15 metastases per patient) treated using the HyperArc technique with an isocenter placed at the center of mass (CoM) were retrospectively analyzed. The planning target volume (PTV) was 1 mm uniform expansion to the GTV. The dose distributions from the original clinical plans served as the reference dose envelope for isocenter optimization. An optimization algorithm was developed to reposition the isocenter by minimizing an L2 cost function that penalized underdosing across 200 error scenarios incorporating random rotational setup errors up to 1° and random translational errors up to 1 mm. Genetic algorithm was used to solve the optimization problem. New HyperArc plans were reoptimized using the repositioned isocenter with the same objectives, and compared against the original CoM-based plans. GTV V100% values were evaluated across identical error scenarios to assess improvements in coverage.
Results: The optimized isocenter was, on average, 15.7 mm (standard deviation: 5.4 mm) away from the original CoM. Under simulated random uncertainties (rotational up to 1° and translational up to 1 mm), the mean GTV V100% improved from 96.4% (standard deviation: 7.9%) to 97.2% (standard deviation: 6.4%). The improvement in dose coverage was statistically significant (paired t-test, p < 0.001).
Conclusion: An optimization framework was developed to improve isocenter placement in SIMT VMAT plans. The optimized isocenter positioning resulted in improved GTV dose coverage compared with conventional CoM-based placement, enhancing plan robustness against rotational and translational setup uncertainties. This method offers a clinically applicable approach to improving the robustness of SIMT VMAT treatments.