S1882
Physics - Dose prediction/calculation, optimisation and applications for photon and electron planning
ESTRO 2026
mouse click.
angles and static delivery segments during radiotherapy arc delivery, providing additional flexibility for dose modulation. RAD planning is an iterative and user-dependent process. The planner defines optimization objectives, runs an optimization that cannot be altered mid-run, evaluates the resulting dose distribution, adjusts objectives, and re-optimizes. This workflow is time-consuming and final plan quality is influenced by planner experience. In this work we present a fully automatic, “one-click” optimization script, AutoRAD, that operates on top of the RAD- optimizer to automate the manual and subjective optimization loop. Material/Methods: AutoRAD was tested on ten head-and-neck cancer patients previously treated with VMAT following the Danish DAHANCA guidelines. Patients received 68 Gy (34 fractions) or 66 Gy (33 fractions) to the primary tumor, 60 Gy to intermediate-risk regions, and 50 Gy to elective nodal volumes.AutoRAD was developed in Varian’s Eclipse Scripting API (ESAPI), that enables programmatic (C#) access to DVH statistics, optimization objectives, and penalty-function values for targets and OARs, as well as creation of new objectives and initiation of RAD optimization. AutoRAD was designed to mimic human planning (Fig.1): A clinically validated RapidPlan model generated initial optimization objectives.After RAD-optimization, AutoRAD extracted OAR doses and penalty-function values.For OAR’s with penalty values below a pre- defined threshold, AutoRAD tightened the corresponding objective (e.g., lowered mean-dose goals). For penalties above a pre-defined threshold the objective was loosened. AutoRAD launched a new RAD-optimization.Steps (2-4) were repeated until all OAR penalty functions were within threshold levels. AutoRAD then restored target coverage by:Assessing if all PTV V95% > 99%.Increasing PTV objective priorities.Launching a new RAD-optimization.Steps (1- 3) were repeated until all PTVs met the coverage criteria (V95% > 99%).AutoRAD-generated plans were compared with RAD plans created using the RapidPlan model alone (RAD-RP) and with the clinical VMAT plans. Comparisons focused on mean OAR doses according to DAHANCA guidelines. All plans used three arcs. RAD plans included dynamic collimator rotation and five static angles. Results: All AutoRAD plans achieved PTV's V95% > 99% and respected a 45-Gy spinal cord tolerance. AutoRAD reduced the mean (±SD) OAR dose by 1.4±1.1 Gy relative to VMAT (p<0.01) and by 1.3±1.2 Gy relative to RAD-RP (p<0.01), see Fig.2. Conclusion: AutoRAD delivers automatic RAD planning, eliminating subjective manual optimization, improving plan quality, and reducing human planning time to a single
Keywords: Automated planning, RapidArc Dynamic
Digital Poster 2359
VMAT, IMRT and RapidArc Dynamic for ultrahypofracionated left breast cancer
radiotherapy: A dosimetric comparative study. Klaudia Krzekotowska 1 , Francesco Pupillo 1 , Margherita Casiraghi 1 , Lisa Milan 1 , Maria Antonietta Piliero 1 , Alessio Minoggio 1 , Matteo Coppotelli 1 , Francesco Mose Castronovo 2 , Thomas Zilli 2 , Stefano Presilla 1 , Paola Canino 2 1 IIMSI, EOC, Bellinzona, Switzerland. 2 IOSI, EOC, Bellinzona, Switzerland Purpose/Objective: Ultrahypofractionation using a 5-fraction schedule is a well-established treatment approach for breast cancer patients; however, the optimal clinical goals and delivery technique remains to be determined. Aim of this study is to compare dosimetric results of VMAT, IMRT, and RapidArc Dynamic (RAD) techniques for left breast irradiation using a 5-fraction ultrahypofractionated schedule.
Made with FlippingBook - Share PDF online