Effects of breath-hold reproducibility on proton and photon lung cancer stereotactic body radiotherapy

Background and purpose: Breath-hold can mitigate respiratory motion in lung cancer radiotherapy. Reduced motion could be especially beneficial for proton therapy which is more sensitive to geometrical perturbations than photon therapy. However, failure to reproduce the breath-hold tumor position could lessen this advantage. In this study, effects of reproducibility were investigated for 3D and 4D robust optimized proton and photon therapy. Materials and methods: Fourteen patients with early stage lung cancer, imaged with a single breath-hold computed tomography, were included. Reproducibility variations were included in the treatment planning and simulated with image deformations. Further, additional larger variations were included in the evaluation.Three photon and three proton therapy treatment plans were robustly optimized for the same intended dose and compared using probabilistic evaluation. One 3D method only accounted for patient shifts. A second 3D method accounted for breath-hold variations using larger patient shifts. The final 4D method used additional planning images. Results: Protons resulted in reduced dose to organs of interest but with higher spread in target dose compared with photons. Mean heart dose was reduced, 90% probability of 6.7 Gy with photons compared with 0.3 Gy for protons for 4D planning. Decreased reproducibility affected tumor dose ranges. However, the 90% probability of median tumor dose remained stable. Conclusions: Reduction in dose to organs of interest was demonstrated with proton therapy. Using 4D compared with 3D optimization did not result in more robust plans. Larger breath-hold variability than anticipated during planning had a minor effect on the target dose.