Room: Exhibit Hall
Purpose: Investigating causes to unexplained failures in IMRT QA
Methods: 15 single-isocenter multi-met SRS cases were chosen. Treatment plans were left unaltered, while four verification plans were created for each case: 1) calculated with a 2.5mm dose calculation grid, our clinical standard, 2) calculated with a 1.25mm dose calculation grid, 3) calculated with a 1.0mm dose calculation grid, and 4) same as category 3 but with a couch model created specifically for the BrainLab couch of our SRS linac per TG-176 methodology. All verification plans, except those falling in category 4, were calculated with our standard treatment planning couch model, an average of the Varian IGRT and BrainLab couches. Plans were delivered to the ArcCheck and verification plans were compared in Sun Nuclearâ€™s SNC Patient using the following parameters for gamma analysis: 10% threshold, 3%, 2mm, and Van Dyk percentage difference.
Results: Categories 1, 2, 3, and 4 had average passing rates of 92Â±5%, 96Â±3%, 97Â±2%, and 98.2Â±1.3%, respectively.
Conclusion: Though we looked specifically at SRS plans, improvements were made that affected the entire IMRT QA program. The high dose gradients of these SRS cases made these plans sensitive to couch model and grid size changes, which enabled the causes of unexplained failures to be found. The BrainLab couch attenuates more than our standard model. This only affects QA since the SRS frame extends over the end of the couch for treatment with minimal attenuation. SNC Patient interpolates TPS data to a 1mm grid for gamma analysis. The plans were smoothed during analysis creating a QA plan which misrepresented the treatment plan. Now all IMRT QA verification plans are calculated with a 1.0mm dose calculation grid and SRS verification plans are calculated with the BrainLab couch, resulting in noticeable decreases in the incidence of QA failures in our clinical practice.
Quality Assurance, Intensity Modulation