S2004
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
photon transport. Conclusion:
Digital Poster 4577 Dosimetric comparison of Eclipse AAA and Acuros XB for TBI-VMAT: a phantom validation study Susana Gonçalves 1,2 , Joana Lencart 1,3 , Anabela Gregório Dias 1,3 1 Medical Physics, Radiobiology and Radiation Protection Group, IPO Porto Research Center (CI- IPOP), Porto Comprehensive Cancer Center (Porto.CCC) & Rise@CI-IPOP (Health Research Network), Porto, Portugal. 2 Escola Internacional de Doutoramento, Universidad de Vigo, Vigo, Spain. 3 Medical Physics Department, Portuguese Oncology Institute of Porto (IPO-Porto), Porto, Portugal
AXB produced dose distributions broadly consistent with AAA but with measurable differences in low density regions and in high-dose subvolumes. The reduced lung Dmean and increased MU values obtained with AXB align with known differences between convolution-superposition (AAA) and grid- based Boltzmann transport algorithms (AXB). These differences reflect both the more realistic photon transport model of AXB and the distinct dose reporting conventions of the two algorithms. These findings support the use of AXB as a more accurate reference for TBI-VMAT dosimetry, highlighting the importance of algorithm choice for clinical dose evaluation and plan validation. Keywords: Total Body Irradiation, AAA, Acuros XB Navigating uncertainty: a treatment planning Multi-Criteria Optimization framework to include robustness in the decision-making process Remo Cristoforetti 1,2 , Tobias Becher 1,2 , Philipp Süss 3 , Niklas Wahl 1,4 1 Department of Medical Physics in Radiation Oncology, DKFZ, Heidelberg, Germany. 2 Faculty of Physics and AstronomyInstitute for Radiation Oncology, Heidelberg University, Heidelberg, Germany. 3 Fraunhofer Institute for Industrial Mathematics, ITWM, Kaiserslautern, Germany. 4 Heidelberg Institute for Radiation Oncology, HIRO, Heidelberg, Germany Purpose/Objective: Radiotherapy is a highly personalized treatment modality. During treatment planning, robust optimization (RO) is performed to maintain plan quality under uncertainty, followed by probabilistic evaluation to estimate the confidence levels for achieving clinical goals. However, conventional RO relies on rigid operators and population-based parameters, limiting treatment personalization. This work proposes a novel planning framework to include robustness as a criterion in the decision-making process. Material/Methods: The recently proposed scenario-free approach [1] defines efficient variance reduction objectives (VRO) that can encode and minimize multiple uncertainty sources, models and error magnitudes. Within a multi- Digital Poster Highlight 4591 criteria optimization (MCO) framework, VROs are traded against standard dose objectives and among themselves.VROs are constructed for increasing error magnitudes, and a Pareto front approximation is generated through a sandwiching algorithm. Interactive navigation of the resulting surface enables
Purpose/Objective: Volumetric modulated arc therapy (VMAT) is
increasingly adopted for total body irradiation (TBI) due to its ability to improve dose homogeneity while respecting organ-at-risk constraints. However, the accuracy of dose calculation algorithms remains a critical factor in TBI planning, particularly given the large irradiation fields and heterogeneities involved. The Anisotropic Analytical Algorithm (AAA) and Acuros XB (AXB), both available in Eclipse TPS, use fundamentally different dose calculation approaches, with AXB providing a more physically accurate model of radiation transport in heterogeneous media. This study aimed to compare AAA and AXB dose calculations for a TBI-VMAT plan on an anthropomorphic phantom, evaluating their impact on target coverage, hotspots, lung dose, and monitor units (MU). Material/Methods: A previously optimized TBI-VMAT plan and calculated with AAA in Eclipse (version 16.1.2) was recalculated using AXB, maintaining identical beam geometry, optimization objectives, and calculation grid. The plan was generated on an anthropomorphic phantom representing a pediatric patient following institutional TBI-VMAT protocol parameters. Dose–volume histogram (DVH) metrics were compared between algorithms. For the PTV, D98%, V110%, and D1cm3 were evaluated. For the lungs, the mean dose (Dmean) was recorded. The number of MU per arc was compared across all 12 arcs of the plan. Results: AAA and AXB produced identical PTV D98% (96.2%). AXB yielded a slightly higher hotspot volume, with V110% increasing from 0.15% to 0.28%, and D1cm3 increasing from 116.8% (AAA) to 118.4%. The mean lung dose differed substantially, decreasing from 7.93 Gy with AAA to 6.72 Gy with AXB, consistent with AXB’s more accurate modeling of dose deposition in low-density media. AXB required consistently higher MU across all 12 arcs, with increases ranging from 2.0% to 3.3%, reflecting its more rigorous modelling of
Made with FlippingBook - Share PDF online