Purpose: The efficacy of SBRT in thoracic lesions depends on the effective dose deposited in the lesions. To reduce plan variability between multiple planners, a semi-automated process was validated for consistency of CI and R50 while increasing ITV mean dose and centralizing the hot spot to the ITV.
Methods: Twenty-two lesions were investigated for patients who received a five fraction SBRT regiment via VMAT. A two-step process included a python code where user selects a set of initial parameters which automates the creation of tuning ROIs, machine settings, and loads the optimizer with a consistent set of objectives. The tuning ROIs include sets of rings at certain distances from PTV, as well as overlap ROIs with nearby organs at risk. The process focuses primarily on dose falloff and PTV coverage and then, meeting nearby organ at risk dose tolerances. The clinically delivered plan, which was planned by experienced staff, was used for comparison.
Results: For the clinical plans, the CI was 1.04±0.05; R50 was 4.07±0.597; and ITV Mean relative to prescription was 117%±6.4%. For the scripted plans, the CI was 1.01±0.03; R50 was 3.61±0.351; and ITV Mean relative to prescription was 123.8%±2.8%. The scripting process was successful in meeting 335/336 OAR metrics resulting in 21 of 22 successful plans. For the clinical plans, the hot spot was located in the ITV 77.3%, while this occurred in 95.3% of scripted plans.
Conclusion: Through programmatic planning, standard deviations can be reduced compared to plans from a multiuser environment. In addition, the CI improved with a decrease of R50 without violation of critical organ dose tolerances in majority of cases. This script allowed for more planned dose deposition to PTV while also maintaining a safe dose falloff and organ tolerance preservation. This technique’s clinical implementation can improve consistency with minimal user interaction.
TH- External Beam- Photons: IMRT/VMAT dose optimization algorithms