Click here to

Session: External Beam Modeling [Return to Session]

Clinical Validation of a User Defined Ethos Flattening Filter Free Beam Model in RayStation

D McClatchy*, C Foote, S Yan, J Pursley, Massachusetts General Hospital, Boston, MA


TU-F115-IePD-F3-3 (Tuesday, 7/12/2022) 1:15 PM - 1:45 PM [Eastern Time (GMT-4)]

Exhibit Hall | Forum 3

Purpose: Varian Ethos and Halcyon ring delivery system (RDS) linacs come with preconfigured and locked beam models on the Ethos and Eclipse treatment planning systems (TPSs), forcing users to rely on the vendor-supplied models. Here we report on clinical validation of a user defined RDS beam model in a third party TPS (RayStation 10A, RaySearch), which gives the user flexibility to match their clinical linac.

Methods: A RayStation 6MV FFF collapsed cone beam model was created for an Ethos RDS based on clinical commissioning data. A total of 17 VMAT plans with a range of large and small target sizes were generated for 10 previously treated genitourinary patients by certified medical dosimetrists according to clinical standards. Dose was recalculated in the Ethos TPS when the plans were imported into Ethos. Plans underwent point-based second dosimetry check (Radcalc, LAP) and 3D comparison against a preconfigured beam model (Mobius, Varian). Next, plans were delivered on a 3D diode array (ArcCheck, SunNuclear) with a centrally located micro-ion chamber (PinPoint, PTW), and trajectory log files were used to reconstruct 3D delivered dose in Mobius.

Results: The mean difference between RadCalc and RayStation point-dose calculations was 1.2%. The mean 3D gamma between RayStation dose versus plan based and trajectory log file-based Mobius dose was 99.9% and 99.8%, respectively. The mean 3D gamma at 3%/3mm between the ArcCheck measurement versus RayStation and Ethos dose was 99.0% and 99.1%, respectively, with a mean absolute difference of 0.8%. The mean 3D gamma at 1%/3mm and 2%/2mm between the ArcCheck measurement versus RayStation was 93.5% and 92.4%, respectively. The mean difference between measured micro-ion chamber point-dose and RayStation planned dose for the chamber was 0.9%.

Conclusion: A third party TPS can accurately model RDS linacs and be used to generate treatment plans for an otherwise single vendor standalone system.

Funding Support, Disclosures, and Conflict of Interest: Massachusetts General Hospital has a product evaluation agreement with Varian to evaluate Ethos in a non-standard information systems environment. SY have a research agreement with RaySearch on proton treatment planning and RayIntelligence.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email