ESTRO 2025 - Abstract Book

S2677

Physics - Detectors, dose measurement and phantoms

ESTRO 2025

4377

Poster Discussion Statistical process control in patient specific QA: insights from threshold optimisation of two commercial systems Margherita Zani 1 , Livia Marrazzo 2,1 , Marta Casati 1 , Chiara Arilli 1 , Silvia Calusi 1 , Cinzia Talamonti 2,1 , Antonella Compagnucci 1 , Stefania Pallotta 2,1 1 Medical Physics Unit, Careggi University Hospital, Florence, Italy. 2 Department of Experimental and Clinical Biomedical Sciences, Univesity of Florence, Florence, Italy Purpose/Objective AAPM TG-218 [1] introduces and suggests the use of Statistical Process Control, a method for calculating action limits for Patient Specific QA (PSQA) results. This approach aims to monitor whether the process remains under control over time and to promptly identify deviations. These limits are locally defined, and therefore procedure-, equipment-, and site-specific. Aim of this study is to establish limits for two relatively new detectors/software systems and compare them with consensus values. Material/Methods -Time Interval : March 2023 - October 2024 -LINAC : 4 twin ELEKTA VersaHD -PSQA devices : 1. Integral Quality Monitor (IQM) : A large transmission ionization chamber mounted on the LINAC head, paired with software for detector output prediction. 2. iViewGT EPID and EPIbeam software : Dose distributions reconstructed from portal images, acquired without patient or couch are compared to the ones calculated from RTplans. For IQM, measured and predicted outputs are compared through the cumulative signal deviation (CSD [%]), while for EPIbeam gamma passing rate (GPR) analysis is employed (3%/2mm global, 10% threshold). For both systems, the initial thresholds were based on those established during commissioning, aligned with vendor guidelines and scientific consensus: • EPIbeam : GPR action limit = 90%; control limit = 95% • IQM : CSD action limit = ±6.3%; control limit = ±3% A total of 233 plans (EPID) and 249 plans (IQM) were analyzed (6 anatomical regions). Control limits (CL) and action limits (AL) were established following TG-218 guidelines. The Shewhart method was applied to determine CL values, while ALs were calculated using β=3 [1].