SU-E-T-624: Comparison of PTV+PRV-Based Optimization and Robust Optimization in Intensity-Modulated Proton Therapy

Area of Science:

• Medical Physics
• Radiation Oncology
• Cancer Treatment

Background:

• Intensity-modulated proton therapy (IMPT) planning aims to optimize dose delivery while accounting for uncertainties.
• Traditional planning target volume (PTV)-based optimization incorporates setup uncertainties, implicitly adding margins to targets.
• Robust optimization (RO) explicitly considers both setup and range uncertainties, adding margins to targets and organs-at-risk (OARs).

Purpose of the Study:

• To investigate whether the superiority of robustly optimized IMPT plans is due to the exclusion of organs-at-risk (OARs) margins in traditional PTV-based optimization.
• To compare the robustness and optimality of PTV-based optimization with and without OAR margins (PTV+PRV) against fully robustly optimized plans.

Main Methods:

• Compared PTV+PRV-based optimization and robust optimization for head and neck and rhabdomyosarcoma cases.
• Generated 9 dose distributions per case to account for setup and range uncertainties.
• Assessed plan robustness and optimality using worst-case dose distributions and specific dose-volume histogram (DVH) metrics (e.g., D1cc, Dmean, D1%, D5%, D95%).

Main Results:

• Robust optimization yielded superior plan robustness and optimality compared to PTV+PRV-based optimization for head and neck cases.
• PTV+PRV-based optimization showed improved OAR sparing compared to standard PTV-based optimization but was still inferior to RO.
• Target dose coverage and homogeneity were comparable between PTV+PRV and PTV-based optimization.

Conclusions:

• The PTV+PRV method offers partial improvement in plan optimality but remains inferior to the robust optimization method.
• Robust optimization is the preferred approach for achieving both robust and optimal IMPT plans, effectively managing uncertainties.
• Findings were consistent across both head and neck and rhabdomyosarcoma cancer types.