ESTRO 2020 Abstract Book

S810 ESTRO 2020

Quantitative improvements were observed between the KB models and the reference plans, particularly for heart and spinal canal doses. As shown in Tab. 1, the models reduce the heart mean dose of about 0.5 Gy and the V20 of 1.2 Gy on average. Moreover, the maximum dose for the spinal canal is about 1.5 Gy less on average. The 3% of 270 analyzed dose-volume objectives in the clinical plans failed to reach the optimal constraints, while the respective KB plans succeeded. In 3 cases out of 18 evaluation data set plans, the KB plan failed at least one of the dose-volume constraint compared to the delivered plan, but with a mean difference less than 5%. The DVH comparisons between clinical plans and the two models for the best and the worst cases are plotted in Fig. 1. Model B results in a better OAR sparing in the 52% of the values compared to model A.

Conclusion Target volume coverage and dose to critical OARs were clinically acceptable across all modalities. While IMPT- produced improved dosimetry for contralateral parotid gland and larynx, this benefit was lost with the addition of any robustness for parotids and larynx although IMPT+3 did maintain improved sparing of the larynx compared to RP, MCO and IMPT+5. Overal for contralateral parotid sparing and CTV coverage MCO performed as well or better than protons with robustness. PO-1503 Evaluation of a generalized knowledge-based planning for VMAT irradiation of breast and lymph nodes M. Rago 1 , L. Azario 1 , M. De Spirito 1 , L. Placidi 1 1 Fondazione Policlinico Universitario A. Gemelli IRCCS, Medical Physics, Roma, Italy Purpose or Objective To evaluate the performance of two Knowledge-Based (KB) models for planning optimization (RapidPlan, Varian) of Volumetric Modulated Arc Therapy (VMAT) applied to whole-breast irradiation comprehensive of nodal station, internal mammary and/or supraclavicular regions. Material and Methods Two different KB models have been generated and trained based on 120 VMAT plans data-set (62 with Simultaneous Integrated Boost) where dose is calculated with AcurosXB Algorithm. Model A (120 VMAT plans) includes plans with whole right and left breast irradiation with lymph nodes. Model B is a subgroup of model A (100 VMAT plans) without plans with only whole breast irradiation. Only two optimization structures, generated by the planner in order to achieve a clinical acceptable standard plan, have been included in the models: a ring structure (defined around the target volumes to ensure a high dose gradient) and an optimized PTV, enlarged with a margin of 1 mm. To evaluate dosimetrically the models, a set of 18 plans, 4 of them without lymph nodes, clinically delivered in our radiotherapy department, were selected. These plans were manually performed by an expert planner to obtain the desired dose-volume objectives after several iterations. Results