KEY POINTS
- This retrospective simulation study included 18 head and neck cancer patients with substantial anatomical change and 576 daily CBCT scans. The patients had received VMAT, while proton plans and accumulated doses were generated retrospectively.
- The authors created an actual proton replan for every possible daily adaptation point, assuming that only one offline replan could be performed during treatment.
- Fraction 15, near the end of the third treatment week, was the collective optimal timing. Four of eight evaluated structures achieved at least 95% of their structure-specific maximum benefit, while six were within 0.1 Gy(RBE) of their optimal result.
- Maximum-dose reductions were greatest for serial organs: 2.93 Gy(RBE) for the brainstem and 4.35 Gy(RBE) for the spinal cord. High- and low-risk CTV D98 improved by 1.00 and 1.99 Gy(RBE), respectively.
- Optimal timing varied by structure, from fraction 9 for the brainstem to fraction 23 for the larynx and constrictor muscles. The substantial interpatient variability supports individualized, trigger-based monitoring rather than automatic replanning at a fixed fraction.
CLINICAL TAKEAWAY
For centres limited to one offline adaptation, the end of the third treatment week may be a practical window for intensified imaging review and replanning readiness. It should not be treated as a universal replanning rule: the cohort was small and selected, the patients were originally treated with photons, and the proton treatments were simulated rather than delivered clinically.