Copenhagen, Denmark

ESTRO 2022

Session Item

Optimisation and algorithms for photon and electron treatment planning
Poster (digital)
Commissioning of a new treatment planning system for CyberKnife robotic stereotactic radiotherapy
Maud Jaccard, Switzerland


Commissioning of a new treatment planning system for CyberKnife robotic stereotactic radiotherapy

Maud Jaccard1, Nicolas Perichon2, Marie Fargier-Voiron1, Oscar Matzinger1, Shelley Bulling1

1Swiss Medical Network, Radiation Therapy, Genolier, Switzerland; 2Swiss medical Network, Radiation Therapy, Genolier, Switzerland

Show Affiliations
Purpose or Objective
Treatment planning with RayStation (RaySearch Laboratories AB, Stockholm, Sweden) for the CyberKnife (CK, Accuray Incorporated, Sunnyvale, CA, USA) has been available since July 2021 (RayStation 11A). We present the beam data measurements, and describe the commissioning and validation process for this new vendor-independent TPS for the fixed, iris and MLC collimators of CK S7. The CK TPS (Precision, Accuray) was commissioned in parallel. Our commissioning data will be among the first to be part of the reference standardized CK S7 beam data set.
Material and Methods

Beam data acquisition for RayStation was performed at fixed SSD, and included percent depth dose (PDD), profiles at 15, 50, 100 and 300 mm depth, a single diagonal for the largest field, and output factors (OF). A microDiamond detector 60019 and a Beamscan water tank (PTW, Freiburg, Germany) were used for all measurements. Output factors (OF) were double-checked with a microSilicon detector (60023, PTW) and not adjusted by Monte Carlo (MC) correction factors. Beam modelling was performed for collapsed cone (CC) and MC algorithms for all collimators, except for the iris 5-mm aperture that cannot be modelled in RayStation. Validation tests for the quality of beam modeling were performed following the AAPM Practice Guideline 5.a. This included: a verification of the dose calculated in calibration conditions, comparisons of calculated dose distributions with commissioning data, comparisons of calculated dose distributions with 2D detector array measurements (Octavius 729 2D-Array, PTW), measurements for various MLC-shaped fields and fields at oblique incidence, and comparison of calculated dose in an heterogeneous medium with dose measurements in bone and lung inserts of a thorax phantom (CIRS, VA, USA). 


Beam data acquisition and processing for RayStation took approximately seven days (vs. 11 days for Precision). Calculated dose in reference conditions (only available through scripting in RayStation) was below 0.3% compared to calibration. Comparison between commissioning data and clinical dose calculated in a homogeneous phantom with CC showed that more than 97% of PDD and profile points respected gamma criteria of 1%/1mm (global), except for the iris aperture of 7.5 mm, the MLC smallest field, and at the depth 300 mm. The agreement increased when modeling the beam with MC algorithm, which also improved the beam modeling in the profile tails of the largest fields. MLC fields measured with the 2D detector at perpendicular and oblique incidence all agreed with TPS calculation within gamma 1.5%/1.5mm criteria. Differences between calculated and measured doses in the heterogeneous phantom were below 2% in the lung and bone inserts, for both CC and MC algorithms. 


RayStation 11A TPS for CK was successfully commissioned for clinical use in our department. Preliminary results show promising planning capabilities for stereotactic treatment planning with the CK S7.