Robust optimized dynamic mixed-beam arc radiotherapy for left-sided breast radiotherapy under deep inspiration breath-hold variations.

Objective.Dynamic mixed-beam arc radiotherapy (DYMBARC) combines intensity-modulated electron- and photon-arcs in order to reduce normal tissue dose compared to photon-only techniques while ensuring similar target coverage. For left-sided breast, deep inspiration breath-hold (DIBH) improves heart sparing and mitigates breathing motion. However, dose delivery uncertainties remain due to inter- and intra-breath-hold variations. We assess robustness and dosimetric plan quality of robust optimized (RO)-DYMBARC under DIBH variations with comparison to volumetric modulated arc therapy (VMAT).Approach.Ten cases were created using the extended cardiac-torso anthropomorphic phantom. NiftyReg was used to obtain deformation vector fields (DVFs) for DIBH variations. For each case, three plans were created: planning target volume (PTV)-based VMAT, RO-VMAT, and RO-DYMBARC, prescribed 42.4 Gy to the target volume (PTV or clinical target volume for RO plans). Robustness to DIBH variations was assessed by recalculating doses using the DVFs and deformable voxel geometry Monte Carlo simulations. Dosimetric endpoints for targets and organs at risk (OARs) were compared using Wilcoxon matched-pair signed-rank test (α= 0.05).Main results.RO-DYMBARC demonstrated significantly better OAR sparing compared to VMAT, with mean dose to heart and contralateral breast up to 1 Gy lower than PTV-based VMAT. Target coverage and robustness to DIBH variations were similar across all plans. For OARs, DIBH variations led to dose differences⩽ 0.1 Gy for all plans and endpoints compared to the nominal scenario.Significance.RO-DYMBARC provided superior dosimetric plan quality with similar robustness to DIBH variations compared to VMAT. The dosimetric advantage of RO-DYMBARC was maintained in the presence of DIBH variations.