Proton lattice radiotherapy for large lung tumors: effects of robust optimization and tumor motion

Purpose.The Bragg peak of proton beam enables precise dose delivery and superior dose modulation in proton lattice radiotherapy (LRT), yet it remains susceptible to uncertainties. However, detailed studies on the application of robust optimization in proton LRT are still limited. This study aims to investigate the impact of robust parameter settings and respiratory motion on intensity-modulated proton therapy (IMPT)-based LRT plans for bulky lung tumors.Methods.Eight patients with bulky lung tumors were retrospectively analyzed. Each patient underwent free-breathing 4D CT simulation and reconstructed 4D CT image sets of 10 respiratory phases. Target and organs at risk (OARs) were contoured in the RayStation treatment planning system. An in-house script was developed in RayStation to generate sphere vertices. First, a non-robustly optimized plan was generated without robust optimization. Both single-field and multi-field optimized IMPT-LRT plans were created. Then, for each optimization method, seven robust parameter settings (1 mm-9 mm) were applied to investigate the effect of setup uncertainty on LRT dose distribution. The peak to valley dose ratios (PVDR) were compared. Four-dimensional dynamic dose (4DDD) was calculated by incorporating dynamic pencil beam scanning delivery with respiratory motion to evaluate the impact of motion on PVDR and dose distributions.Results.PVDRs decreased as setup uncertainty parameters increased. Multi-field optimization resulted in slightly higher PVDR values. Most of the PVDR values were in the range of 2-4. Doses to OARs increased with larger setup uncertainties. Respiratory motion resulted larger PVDR deviations for single-field plan than multi-field plan. The 4DDD maintained the volume of high dose region inside GTV in single-field plan, while decreased in multi-field plan. The volume of high dose region decreased in 4DDD for both single-field and multi-field plans.Conclusion.Increased setup uncertainty resulted in a reduction of the PVDR. Under respiratory motion, the single-field plan maintained high-dose coverage within the target but did not preserve the PVDR. In contrast, the multi-field plan retained the PVDR at the expense of high-dose volume within the target. With appropriate vertex contouring, neither single- nor multi-field plans introduced additional high-dose volume to the target boundary, thereby ensuring safe beam delivery in proton LRT during respiratory motion. These findings support an individualized proton LRT planning strategy for bulky lung tumors, in which robustness settings, high-dose volume, and PVDR are jointly optimized to maintain target coverage while preserving the potential normal-tissue-sparing advantage of spatial dose fractionation.