Room: AAPM ePoster Library
Purpose: MLC beam model parameter values from our institution, recent literature, and the vendor were evaluated through comparison with ionization chamber measurements for a suite of VMAT test plans.
Methods: Analysis was performed in RayStation 7 for flattened 6, 10, and 15 MV beams from a TrueBeam equipped with Millennium 120 MLC. Model parameter values were collected from three sources: our institution’s clinical model, the template machine from the vendor, and average values from a recently published IROC poll of RayStation users. The following MLC-specific parameters were considered: Transmission, Position Calibration, Leaf Tip Width, and Tongue-and-Groove Width. Dose to a cylindrical Solid Water® phantom was calculated for seven VMAT test plans, four geometrically-based per TG-119 recommendations (small and large cylinders, c-shape, and off-axis target) and three anatomically-based (unilateral neck, chest wall, and lung). For each plan, average dose calculated to an ionization chamber ROI was compared to measurements using calibrated A1SL chambers at six positions within the high dose region. Corrections were made for accelerator output measured same-day using a water tank.
Results: For a given energy, the percent difference between calculated and measured values averaged over all plans was found to be <1.5% for our institution’s model, <1.6% for the IROC poll, and <2.2% for the vendor template. Results for the maximum dose deviation at a chamber position were +3.0% for our model, +4.0% for the IROC poll, and +5.3% for the vendor.
Conclusion: Our institution’s MLC parameter values in RayStation, fine-tuned using VMAT test plans, achieved the best agreement with measurements. To this end, published RayStation parameters from a recent IROC poll may serve as a helpful starting place when creating a beam model. Furthermore, clinical users can better contextualize end-to-end results from their own beam model and identify any outlying parameter values through comparison with available datasets.