Optimizing Proton Flash In Tangential Breast Chest Wall Radiotherapy
Abstract
Purpose
Post-mastectomy chest-wall (CW) irradiation is well suited for proton FLASH due to the predominance of normal tissue within the target and the feasibility of single-energy tangential transmission beam (TB) delivery. This study investigates optimization strategies to enhance FLASH activation and modeled biological sparing for large CW targets.
Methods
Five left-sided breast cancer patients were replanned in RayStation to 32 Gy in 5 fractions with 95% CW-CTV coverage. Each plan used a single-energy (249 MeV) tangential TB plus an en face clinical beam. TB spots were placed on a 6 mm hexagonal grid, with Bragg peaks stopped in an external high-density object for inverse optimization. FLASH was evaluated using Krieger’s FLASH effectiveness model (FEM) and a modified Folkerts average dose-rate model (ADR) using 4 Gy and 40 Gy/s thresholds. A FLASH effectiveness factor of 0.67 was applied after FLASH activation. Three delivery scenarios were compared: single-TB, split-beam-TB (STB; upper and lower CW-CTV), and genetic algorithm (GA)-optimized spot sequencing. Scan speed, a key determinant of instantaneous dose rate, was varied from 10 to 20 mm/ms. Evaluation included the percentage of FLASH-activated voxels and Dmean for CW-CTV and skin.
Results
FLASH activation increased with higher scan speed (20>10 mm/ms), split-beam delivery (STB>single-TB), and ADR vs FEM, reaching 94.84% in CW-CTV and 71.96% in skin (STB, ADR at 20 mm/ms). Modeled FLASH biological sparing reduced the effective CW-CTV Dmean to normal tissue by 2.27 to 9.2 Gy, without compromising coverage. GA-based sequencing reduced scan time by 6.1%. Averaged across five patients, effective CW-CTV Dmean decreased from 29.20 Gy (TB) to 21.55 Gy (ADR-GA), and skin Dmean decreased from 20.94 to 17.42 Gy.
Conclusion
Split-TB and higher scan speed produced the greatest FLASH activation. GA-based spot sequencing achieved FLASH coverage comparable to the split-beam delivery. A larger cohort study is needed for further validation.