KEY POINTS
- Investigators developed a script-assisted online adaptive proton workflow integrating an in-room CT-on-rails, a synchrocyclotron, RayStation, Aria, secondary Monte Carlo calculation, and post-delivery machine-log analysis.
- Validation used 50 daily CT datasets from 10 patients previously treated for recurrent rectal cancer with pelvic proton reirradiation SBRT. Prescriptions ranged from 30–40 Gy(RBE) in five fractions.
- The scheduled plan was recalculated on the daily CT and compared with a reoptimized adaptive plan using the original isocentre, beam geometry, and clinical objectives. Up to four beams, ±3% range robustness, and a 5 mm PTV margin were used.
- An adaptive plan would have been selected in 23 of 50 fractions (46%): 15 fractions (30%) because a highest-priority organ-at-risk constraint was violated and eight fractions (16%) because target coverage was inadequate.
- Scheduled plans exceeded constraints for the bladder in 16%, large bowel in 8%, and small bowel in 8% of fractions. All corresponding highest-priority constraints were met after adaptation.
- Median PTV V95% improved from 97.0% to 98.5% with adaptation (p = 0.0002). Maximum dose remained below 130% of prescription in every nominal and robustness scenario.
- The median estimated workflow duration was 61.9 minutes, with an interquartile range of 12.2 minutes. Secondary Monte Carlo calculations, measurement-based patient-specific QA, and machine-log analysis all met institutional acceptance criteria.
- No patient was treated using this online workflow. The study used one centre, one treatment platform, a small homogeneous pelvic cohort, and simulated rather than live clinical decision-making; intrafraction anatomical change during the hour-long process remains a concern.
CLINICAL TAKEAWAY
Online proton adaptation can correct clinically relevant daily bowel and bladder changes that robust planning alone cannot address. The workflow is technically close to clinical use, but a one-hour treatment slot, multidisciplinary staffing, motion management, and prospective end-to-end patient validation remain substantial barriers.
SOURCE
International Journal of Radiation Oncology, Biology, Physics