Robust carbon-ion optimization improves dose stability but does not always preserve LET

In a constrained phantom target, robust optimization reduced worst-case dose loss but increased minimum LET loss from 11.7% to 16.7%.

KEY POINTS

  • This computational phantom study from Gunma University, Japan, investigated how robust optimization and beam configuration affect dose-averaged linear energy transfer (LETd) stability in carbon-ion radiotherapy. Investigators used a 30 × 30 × 30 cm³ water-equivalent phantom containing either a cylindrical target or a C-shaped target surrounding a central OAR. Plans prescribed 50 Gy(RBE) in 10 fractions, using RayStation 12A and the modified microdosimetric kinetic model (mMKM).
  • PTV-based planning was compared with robust optimization incorporating setup, density, or combined uncertainties. Geometric margins of 1, 3, and 5 mm were evaluated against corresponding robust scenarios with translation errors of 1–5 mm, density uncertainties of ±1–5%, and combined perturbations. Seven beam configurations were also examined, ranging from two opposing lateral fields to four-field oblique arrangements, with LETd and RBE-weighted dose assessed independently.
  • Density uncertainty had a substantially greater effect on LET distributions than translational errors alone. The most sensitive endpoint was near-maximum LETd2%, particularly for the C-shaped target, where the average interquartile variation reached approximately 40%. For the central OAR, LETd2% variability reached as high as 165% under density perturbations. Robust optimization incorporating both density and translation generally reduced the dispersion of LET outcomes compared with PTV-based optimization.
  • LET and RBE-weighted dose responded differently to the same uncertainties. Under density perturbations, mean LETd could increase while mean RBE-weighted clinical dose decreased, reflecting redistribution of high-LET regions despite loss of physical dose. This divergence demonstrates why a treatment plan meeting conventional dose-robustness criteria may still experience important changes in its underlying LET distribution.
  • Robust optimization improved worst-case target dose coverage without consistently improving worst-case LET coverage. For the cylindrical target, combined robust optimization reduced the worst-case loss of LETd98% from 14.3% to 8.8%, while reducing the worst-case loss of RBE-weighted D98% from 32.8% to 23.5%. For the more constrained C-shaped target, however, the worst-case loss of LETd98% increased from 11.7% to 16.7%, despite improved dose robustness, with the worst-case D98% loss decreasing from 77.6% to 65.8%.
  • Beam configuration rankings depended strongly on the optimization objective. Two opposing lateral fields achieved the highest ranking for minimizing changes in LET metrics under perturbation. In contrast, four-field oblique delivery with robust optimization ranked first under the authors' biologically favorable criterion, balancing greater target LET with reduced OAR LET exposure. No configuration simultaneously optimized both criteria, and the rankings were derived from idealized geometries rather than actual patient anatomy.
  • Spot-weight redistribution provided a potential mechanism for improved LET stability. In the cylindrical target, combined robust optimization increased the average number of energy layers from 36.3 to 38.0 and reduced the fraction of spot weight assigned to distal energy layers from 60.8% to 57.4%. This redistribution was not reproduced in the C-shaped geometry. The study did not include explicit LET objectives during optimization, patient-specific anatomy, anatomical motion, alternative RBE models, or clinical outcomes, limiting direct translation into treatment-planning recommendations.

CLINICAL TAKEAWAY

Conventional robust optimization can improve RBE-weighted dose stability in carbon-ion radiotherapy without guaranteeing preservation of the intended LET distribution, particularly for anatomically constrained targets. The findings support evaluating LET robustness separately from dose robustness, but beam-angle rankings and optimization strategies require validation in patient-specific plans before clinical implementation.

SOURCE

Journal of Applied Clinical Medical Physics

Browse more research Suggest a correction