Discrete-STEP Proton ARC Therapy: A Phantom-Based Planning and Dosimetric Feasibility Study
Abstract
Purpose
This study aims to evaluate the planning feasibility and dosimetric performance of a discrete-step proton arc (DSPAT) using a standardized head phantom on an IBA Proteus Plus proton therapy system.
Methods
Five distinct target geometries were defined in a standardized IBA head phantom to mimic central, lateral, and superficial clinical scenarios. For each geometry, five plans were generated in RayStation treatment planning system (TPS v12A SP1) using the commissioned pencil beam scanning (PBS) beam model. These comprised one intensity-modulated proton therapy (IMPT) reference plan and four DSPAT plans with different angular sampling strategies: coarse (15° and 30°), fine (5°), and a hybrid gantry spacing (coarse +fine). All plans were robustly optimized using identical dose prescriptions and optimization constraints. Dosimetric evaluation included target coverage (D95), conformity index (CI), homogeneity index (HI), selected organ-at-risk dose metrics, linear energy transfer (LET ≥ 6 keV/μm) within the 50% isodose of prescribe dose, and low-dose volume parameters.
Results
Comparable target coverage (D95) was observed in nominal plans for both DSPAT and IMPT plans. Dosimetric trends showed improved dose conformity and redistribution of low-dose regions, depending on the gantry spacing strategy. Most DSPAT plans exhibited reduced robustness compared to IMPT, highlighting the trade-off between conformity gains and robustness.
Conclusion
This phantom-based study demonstrates that DSPAT is technically feasible and dosimetrically comparable to IMPT. However, current treatment planning systems lacks true arc optimization for proton therapy. Employing coarse DSPAT improves planning efficiency and provides a foundation for future investigations into delivery efficiency, robustness, and biological dose considerations. Clinical adoption of PAT depends on demonstrable dosimetric gains and seamless integration into existing workflows.