2417 - A Comprehensive Workflow of Selecting Local Tolerance Limits to Detect Systematic Multi-Leaf Collimator Positional Errors by Combining Statistical Process Control and Error Curve Model

X. Yi^1,2, W. Lu¹, H. Zhang¹, and Y. Song¹; ¹Department of Oncology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, Chongqing, China, ²The Key Laboratory of Biomedical Information Engineering of Ministry of Education, School of Life Science and Technology, Xi’an Jiaotong University, Xian, Shaanxi, China

Purpose/Objective(s): This study aims to develop a comprehensive and scientifically sound methodology for determining appropriate local tolerance limits in patient-specific quality assurance (QA).

Materials/

Methods: A total of 214 volumetric modulated arc therapy/VMAT plans for cervical cancer were selected. Systematic MLC positional errors were simulated across eighteen magnitudes ranging from ±0.2 to ±5 mm. Dose verification was conducted on 808 volumetric modulated arc therapy/VMAT plans, and a retrospective review was carried out. Firstly, six commonly used QA metrics in gamma and DVH analysis were extracted from the QA results of 196 error-free volumetric modulated arc therapy/VMAT plans. These QA metrics included GP₁₀ (gamma passing rate at 3%/2mm, 10% dose threshold), GP₅₀ (gamma passing rate at 3%/2mm, 50% dose threshold), µGI₅₀ (mean gamma index at 3%/2mm, 50% dose threshold), PTV₉₅ (dose received by 95% of PTV), PTV₅ (dose received by 5% of PTV) and PTV_mean (mean dose received by PTV). Statistical process control was used to establish the corresponding tolerance limits for each metric. Subsequently, six error curve models were created based on 360 error-introduced plans to record changes of QA metrics under different magnitudes of MLC positional error. The error range of theoretical detection limits for systematic MLC positional errors was investigated to assess error sensitivity quantitatively using the error curve model. Finally, the process-based tolerance limits of the six individual QA metrics and four hybrid QA metrics were validated by using the QA results of 252 test plans. The binary classification performance (error-free/error-introduced) was assessed based on error detection rate, accuracy, precision, recall, and f1-score.
Results: The theoretical detection limits for process-based tolerance limits of GP₁₀, GP₅₀, µGI₅₀, PTV₉₅, PTV_mean, and PTV₅ were 2.19 mm, 2.71 mm, 3.52 mm, 1.93 mm, 3.20 mm, and 2.15 mm, respectively. In the validation phase, the process-based tolerance limits for PTV₉₅ effectively identified systematic MLC positional errors exceeding 0.6 mm with an error detection rate of 76.19%, displaying superior performance in binary classification among individual evaluation metrics. Regarding the hybrid metrics, the joint evaluation of process-based tolerance limits for GP₁₀ and PTV₉₅ showed a higher error detection rate of 80.16% for systematic MLC positional errors exceeding 0.6 mm.
Conclusion: The proposed workflow integrates the establishment and validation of local tolerance limits. It dose not only provide a practical tool for setting local tolerance limits based on actual clinical scenarios but also provides a quantitative method for medical physicists to understand the error sensitivity of the selected local tolerance limits.