Room: Exhibit Hall | Forum 7
Purpose: To evaluate robust optimization strategies with RayStationâ„¢ for lung SBRT treatment planning.
Methods: A 4DCT was acquired of a respiratory motion phantom with a 3cm-diameter GTV executing 1D sinusoidal motion with period and amplitude of 4s and 2cm, respectively. The GTV was contoured on all ten phases of the 4DCT with a 0.5cm isotropic PTV margin. Composite GTV and PTV contours were created on the average 4DCT dataset. Five separate plan optimizations were performed with otherwise identical plan settings: (1)composite PTV on average reconstruction, (2)robust optimization on 4DCT with PTV using mid-exhalation and (3)end-exhalation scans as reference, (4)robust optimization on 4DCT using 0.5cm isotropic patient setup uncertainty instead of PTV optimization objectives with mid-exhalation and (5)end-exhalation scans as reference. Dose distributions were calculated on all phases of the 4DCT, and also with 0.5cm perturbations to the isocenter position to evaluate the robustness of the plans to respiratory motion and setup uncertainty. All plans were normalized such that the mean GTV D90 over the ten 4DCT phases was 60Gy in five fractions.
Results: Plan 1 provided the greatest mean coverage of the GTV and PTV over the ten 4DCT phases, and used the fewest MU; however, plan 1 was also least conformal to the GTV, and had the greatest standard deviation in GTV coverage, indicating reduced robustness to respiratory motion. Plans 3 and 5 used more MU than plans 2 and 4 with no difference in the evaluated dose metrics. Plans 4 and 5 were more conformal to the GTV with greater dose gradient than plans 2 and 3; however, plans 2 and 3 maintained greater mean GTV coverage with perturbed isocenter positions than plans 4 and 5.
Conclusion: Robust optimization for lung SBRT has the potential to achieve target coverage goals with reduced doses to healthy tissues.