A Egan¹*, D Maes¹, T Nevitt¹, E Leuro¹, A Fung¹, R Regmi¹, C Bloch^1,2, J Saini¹, P Taddei^1,2, S Bowen^1,2, M Janson³, T Wong², (1) Seattle Proton Therapy Center, Seattle, WA, (2) University of Washington School of Medicine, Seattle, WA, (3) Raysearch Laboratories, Stockholm, Sweden

• 
  A Egan

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose:
The IBA Uniform Scanning (US) modality [Ion Beam Applications, Louvain-Neuve, Belgium] was commissioned for the RayStation version 9A (RS) treatment planning system (TPS) [RaySearch Laboratories, Stockholm, Sweden] at the Seattle Proton Therapy Center (SPTC).


Methods:
The IBA US method (a.k.a beam wobbling) repeatedly raster-scans a proton pencil beam to achieve lateral uniformity while depth-wise uniformity is achieved by appropriately spaced and weighted energy layers. The beam is then patient-specifically shaped laterally by brass apertures and depth-wise with wax compensators. RaySearch provided a list of central-axis SOBP, pristine peak, and lateral profile curves of various ranges and modulation widths to be measured for model inputs. All depth-profiles were measured with a multi-layered ion chamber and lateral-profiles with an ion chamber array. Measured data was submitted to RaySearch and the returned model was validated by comparing model outputs to both input measurements and a wide array of non-input beam geometries. All model validation calculations, data export, and comparison analyses have been completely automated by in-house developed Python code in conjunction with the RS scripting interface. Final clinical commissioning included a simulation-to-delivery end-to-end test utilizing absolute film-dosimetry in an anthropomorphic phantom, development of an in-house machine-learning-based secondary MU calculation, and several other minor modifications to the existing non-RS workflow.

Results:
All model outputs were found to match measurements within clinically acceptable tolerances. Utilizing a virtual SAD method, penumbra and field width matching was found to be superior to the existing non-RS model. The in-house developed MU calculation was found to be as accurate as an existing method that does not function on RS-exported DICOM plan files. Not including measurements, Python-based automation has reduced the beam model validation process to just a few hours.

Conclusion:
The IBA US modality was successfully commissioned in the RS TPS and is currently deployed for clinical use.

Keywords

Protons, Commissioning

Taxonomy

TH- External Beam- Particle/high LET therapy: Proton therapy – computational dosimetry-deterministic

Contact Email