KEY POINTS
- This retrospective treatment-planning study included 20 consecutive lung cancer patients previously treated with IMPT to 66 GyE in 24 fractions or 60 GyE in 30 fractions. All had respiratory motion amplitudes below 20 mm and underwent ten-phase 4D-CT imaging.
- Investigators generated 420 plans: one motion-encompassing ITV-based surrogate clinical plan, ten single-phase plans and ten three-consecutive-phase plans for each patient. Phase-targeted plans used the CTV directly and assumed intrafraction respiratory motion could be effectively frozen by gating or future ultra-high-dose-rate delivery.
- Removing the motion-encompassing target reduced irradiated target volume by a mean 25.1 cm³, or 16.8%, while all plans maintained CTV V95 ≥98% and met robustness criteria using 7-mm setup and 3% range uncertainty.
- Single-phase planning reduced mean lung dose at every respiratory phase. The largest reduction occurred at full inspiration: −0.7 GyE on average, with lung V20 decreasing by 1.2 percentage points. The reference mean lung dose was 11.7 GyE.
- Different organs favored different respiratory phases. Mean heart dose was reduced most around the 40% phase by 0.4 GyE, whereas mean esophageal dose was reduced most around the 70% phase by 0.9 GyE, with individual reductions reaching 6.0 GyE. Three-phase robust plans preserved the same overall pattern but produced smaller gains.
- Modelled clinical effects followed the dosimetry. Single-phase treatment reduced predicted radiation-pneumonitis risk by as much as 1.9 percentage points on average, treatment-related two-year mortality by 0.6 points, and dysphagia by 1.4 points at their respective most favorable phases.
- Critically, no single respiratory phase was optimal across the cohort or across endpoints. The study also did not actually deliver UHDR/FLASH treatment: residual breathing variability, reproducibility of respiratory phase, gating accuracy, compatible robustness settings and true UHDR delivery constraints remain unresolved.
CLINICAL TAKEAWAY
If proton dose could be delivered within a sufficiently short and reproducible part of the respiratory cycle, smaller motion margins alone could reduce normal-tissue exposure, even before invoking any biological FLASH effect. The study is a planning proof of concept rather than a FLASH study, and its main translational challenge is demonstrating that the assumed respiratory phase can actually be targeted reliably during treatment.