Author: Chloe Doen Kim, Patricia E. Lindsay, Bern Norrlinger, Jan P. Seuntjens 👨🔬
Affiliation: Princess Margaret Cancer Centre & University of Toronto, Queen's University, Princess Margaret Cancer Centre 🌍
Purpose: To propose a method to check the consistency of VMAT delivery on linear accelerator (LINAC) with a commercial three-dimensional array device. The method is insensitive to the correction factors of the device.
Methods: Four 6-MV VMAT plans were delivered on four Elekta LINACs on two different days. An in-house code was used to parse the measurement files to extract the dose values corrected with ArcCHECK (Sun Nuclear Inc, Melbourne, FL) correction factors (method 1) and the raw counts without any further corrections except for background signal correction (method 2). The data from each method, collected for the same plan and LINAC, were compared across the two days to calculate pass rates (%) in the criteria of 10% threshold, 3% difference, and 2mm distance. The pass rates for the two methods were then compared to determine the efficacy of the raw counts on the consistency of VMAT delivery. The variations in the pass rates, as each plan was delivered across four LINACs, were also compared between the two methods.
Results: When the measurements are compared across the two days, the pass rates with the raw counts (Mean: 99.18%, SD: 1.40%) are significantly higher (p=0.01, t-test) than those with the corrected dose values (Mean: 94.29%, SD: 7.75%) for all LINACs. In addition, the pass rates of each plan delivered across LINACs showed that the method using the raw counts (96.90% - 100%) is less sensitive to the change in the LINAC than the method using the corrected dose values (76.80% - 99.47%).
Conclusion: This study shows the superior pass rates with the raw counts through a comparison of the VMAT measurements obtained at different time points and in the different LINACs. The proposed method can be used for a routine test of the consistency and reproducibility of VMAT delivery.