S2341
Physics - Quality assurance and auditing
ESTRO 2026
The largest variations occurring for the beam output change and beam uniformity change with a standard deviation of 0.57% and 0.4% respectively. This data is also visualised in figure (1)
grouping. These findings support site/subsite-specific TL/AL and enhance risk sensitivity by monitoring the lower tail, providing a bridge to predictive modelling that couples clinical and PSQA data. Integrating dose/ γ radiomics may further improve prediction. References: Miften M, Olch A, Mihailidis D, et al. Tolerance limits and methodologies for IMRT measurement-based verification QA: Recommendations of AAPM TG-218. Med Phys. 2018;45(4):e53–e83. doi:10.1002/mp.12810. Elekta AB. MOSAIQ® Oncology Analytics and MOSAIQ® Radiation Oncology: Product information / user documentation (product brochures). Elekta; accessed Nov 2025.Aerts HJWL, Velazquez ER, Leijenaar RTH, et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun.
2014;5:4006. doi:10.1038/ncomms5006. Keywords: PSQA, gamma analysis, ICD-10 stratification
The results in the table (1) shows the results for each test based on the 98% and 99% percentile ranges and the action and tolerance CDF ranges compared to the current Varian tolerances and action levels.
Digital Poster 1211 TrueBeam QA Stability Over Time - A Multi-Linac MPC Study Dean Wallace Mulgrave Private, Icon, Mulgrave, Australia Purpose/Objective: The study aims to review and analyse three years of Machine Performance Check (MPC) data collected from multiple TrueBeam linacs at our institution. By comparing all tests within the daily MPC Beam & Geometry enhanced couch, the aim of this work is to evaluate machine performance stability [1], inter-linac consistency, suitability of current tolerances and the utility of MPC data as a long-term QA monitoring tool. Material/Methods: Data extracted from RadMachine was processed in Python to calculate mean, standard deviation, and variability across machines and years. Outliers were removed using an adjusted interquartile range method (2.5 × IQR). Confidence intervals (95% and 99%) were computed assuming normal distribution. New tolerance limits were estimated using Gaussian cumulative distribution and percentile-based approaches (98% action, 99% tolerance) after verifying normality via skewness and kurtosis analysis. These statistical evaluations aimed to refine MPC parameter thresholds and standardize QA performance monitoring. Results: These measurements totalled 792,830 individual test results evaluating the machine performance consistency with the large dataset giving statistical confidence across the different tests and machines.
Conclusion: In conclusion the results show that for all tests a reduction in the tolerance and action levels when using both percentile and CDF evaluation that can be
Made with FlippingBook - Share PDF online