ESTRO 2021 Abstract Book

S1332

ESTRO 2021

PO-1611 Characterization of Ethos therapy systems for adaptive radiotherapy: a multi-machine comparison A. van de Schoot 1 , D. Hoffmans 1 , K. van Ingen 1 , M. Simons 1 , M. Admiraal 1 , N. van Wieringen 1 , J. Wiersma 1 1 Amsterdam UMC, Radiation Oncology, Amsterdam, The Netherlands Purpose or Objective The Ethos therapy system (Varian Medical Systems) allows for CBCT-guided on-couch adaptive radiotherapy using artificial intelligence. Clinical treatment involves treatment preparation and plan adaptation using Ethos treatment management with a pre-configured beam model. Besides adequate quality assurance, accurate machine characterization and beam model verification is necessary for patient interchangeability between treatment units. Our aim was to investigate the dosimetric and mechanical characteristics of Ethos therapy systems including beam model verification and determine the variation between clinical systems. Materials and Methods Four Ethos treatment units, consisting of a 6-MV FFF beam with a dual-layer MLC and kV-CBCT system for image guidance, were recently introduced clinically. Before introduction, a comprehensive protocol was used for machine characterization and beam model verification including relative dosimetry measurements. Percentual depth dose (PDD) curves, profiles and output factors (OF) were obtained by direct measurements in a water tank and compared to vendor-provided reference data. For beam model verification, additional PDD curves, profiles and OF were collected and compared to beam model based calculations. Differences for PDD curves and profiles were quantified using 1D γ-analysis (1%/1mm). Also, point measurements at different depths were obtained for atypical field configurations including off-axis fields with non-standard source- surface distances (Figure 1A). Differences between measurements and calculations were quantified. Next to relative dosimetry, machine characterization consisted of leaf transmission verification for individual and combined MLC banks and dosimetric leaf gap (DLG) determination. The DLG was assessed by deriving the intercept of the fitted linear relation between leaf gap and measured output. Also, the isocenter accuracy for gantry and collimator was assessed by determining the diameter of the radiation isocenter using spoke-shot tests. The agreement between beam isocenter and kV-CBCT isocenter was verified by calculating MV-based and CBCT-based phantom alignment differences. Results For all machines, a good agreement was found between reference data and measurements and between calculations and measurements. Table 1 shows OF results and mean γ 1%/1mm results for selected PDD curves and profiles. Mean differences of approximately 1% were found for point measurements of atypical fields. For all machines, largest differences were found for measurements with irregular fields or large depths (Figure 1B). Furthermore, limited machine variation was found in terms of MLC performance and isocenter accuracy (Table 1).