ESTRO 2025 - Abstract Book

S2620

Physics - Detectors, dose measurement and phantoms

ESTRO 2025

Conclusion: High resolution patient specific QA measurements using 3D-printed phantom and Gafchromic film has been developed and gamma pass rate was on average over 90%. Localization of gamma fails facilitates the analysis of steep dose gradients, indicating that hippocampal-sparing was less than planned for. The methodology of 3D printing a phantom with film verification is feasible for relative dosimetry but rather laborious.

Keywords: Gafchromic film, proton therapy

3111

Digital Poster Characterization of a commercial EPID based pretreatment verification system for radiotherapy beams Marta Casati 1 , Cinzia Talamonti 2 , Margherita Zani 1 , Ilaria Cupparo 2 , Silvia Calusi 1 , Livia Marrazzo 2 , Chiara Arilli 1 , Antonella Compagnucci 1 , Stefania Pallotta 2 1 Medical Physics Unit, Careggi University Hospital, Florence, Italy. 2 Department of Experimental and Clinical Biomedical Sciences, University of Florence, Florence, Italy Purpose/Objective: The aim of the study is the characterization of Dosisoft EPIbeam radiotherapy pre-treatment verification system, based on EPID (Electronic Portal Imaging Device), by investigation of panel extraction procedure repeatability and system sensitivity to delivery errors, comparison with a reference pre-treatment verification device. EPIbeam was tested on a large sample of clinical plans, to characterize performances in the clinical setting. Material/Methods: 4 ELEKTA VersaHD LINACs are equipped with iViewGT EPID and EPIbeam software, which calculates and compares two dose distributions, in a verification plane at isocenter: one calculated from RT plan, the other reconstructed from portal images. 1. Given the error-prone nature of the iViewGT panel extraction procedure, the repeatability of panel positioning was assessed by delivering a 10 cm × 10 cm field. The panel was closed and extracted before each acquisition, and positioning errors were evaluated in both directions. Repeatability was tested under three conditions: short-term (5 consecutive repetitions), medium-term (19 repetitions over one month), and long-term (18 repetitions over 7 months). 2. Sensitivity to delivery errors was evaluated by delivering a modified IMRT breast plan: removing a single control point; editing the position of one leaf at each control point; rotating collimator by ±1°, ±2°, and ±3°. 3. Performances were compared to those of ArcCheck on 98 VMAT plans. 4. 211 clinical plans were verified. Before each session, we tested EPID and LINAC status by delivering benchmark and 10x10 plans. For points 2-4, gamma passing rates (GPR) were computed with TH10%, 3%/2mm, global criteria. Results: 1. Over more than 200 recorded positioning errors, only one exceeded the manufacturer's 1mm tolerance. Sigma was evaluated, obtaining maximum sigma (over 2 directions): 0.14mm σ short-term and 0.22mm σ medium term and σ long-term . 2. To remove the smallest segment in an IMRT breast plan, significantly reduced the gamma passing rate (GPR), from 100% to 73% for medial and to 29% for lateral beams. Single leaf positioning errors and collimator rotations also significantly affected GPR (fig.1). 3. EPIbeam showed better agreement than ArcCheck, likely due to its higher resolution. The fraction of QA with GPR higher than 90% was 97.5% for EPIbeam versus 77% with Arccheck. 4. Results are reported in fig. 2. Benchmark plans’ low variability (GPR σ ≤1.1%), suggests high LINAC and iViewGT reproducibility, while variability in clinical plans GPR may be related to RTplan complexity.