S1678
Physics - Detectors, dose measurement and phantoms
ESTRO 2026
1 Department of Physics and Astronomy “Galileo Galilei”, University of Padova, Padova, Italy. 2 Medical Physics Department, Veneto Institute of Oncology IOV - IRCCS, Padova, Italy Purpose/Objective: Radiochromic films are widely used for QA and dosimetry in radiotherapy; however, their clinical implementation remains highly heterogeneous across centers. This variability is mainly due to differences in calibration strategies and software tools, which limit inter-institutional consistency [1,2]. This work presents a reproducible single-channel dosimetry pipeline grounded in reference guidelines. Implemented in an open-source software, the workflow ensures transparent and reproducible dose evaluation, including rigorous uncertainty estimation via full error propagation, a feature typically absent in commercial software. The pipeline was tested on Gafchromic EBT- 4 films, focusing on an extensive uncertainty and residual analyses of various calibration functions to identify the best model. The results were benchmarked against the manufacturer’s proprietary software. Material/Methods: The calibration pipeline followed procedures reported in the literature and guidelines for radiochromic film dosimetry [1,3]. Four of the most common calibration functions found in literature were evaluated to determine the best model. For each function, residuals, defined as the relative difference between calculated and measured doses on validation films, were computed and associated uncertainties derived via full error propagation. For benchmarking, the same films were processed with the manufacturer’s proprietary software to compare residuals, although explicit dose uncertainty could not be extracted. All steps were implemented in AGAR (Automated Gafchromic film Analysis for Radiotherapy), a Python- based GUI software that will be released as open source after final validation to promote reproducible and transparent film dosimetry. Results: The 2.5th-order polynomial function performed best, with dose errors on validation points ranging from 1.6% to 4.4%, a mean residual of 0.25%, and a mean error on residuals of 2.9%, the lowest among the tested models. For comparison, the other polynomial functions showed dose errors between 2.6% and 4.6%, while the rational model exhibited significantly higher uncertainties, from 6.5% to 7.3% across the investigated dose range. AGAR produced residuals comparable to those obtained with the commercial software (0.25% vs –0.12%, standard deviation 1.39% vs 1.82%), although a complete comparison would require explicit uncertainty estimation, which the proprietary tool does not provide.
Conclusion: The single-channel calibration approach implemented in AGAR offers a transparent and reproducible solution that simplifies radiochromic film dosimetry and supports harmonization across institutions. Results are comparable to commercial software, but AGAR provides key advantages: a simpler, fully open, verifiable workflow aligned with reference guidelines, avoiding black-box limitations, and enabling rigorous, explicit dose uncertainty estimation through complete error propagation. References: [1] Devic, S.; Tomic, N.; Lewis, D. 2016. Reference radiochromic film dosimetry: Review of technical aspects. Physica Medica 32: 541–556.[2] Beveridge, S.; Alves, A.; Hussein, M.; Clark, C.H.; Jornet, N.; Viegas, C.C.B.; et al. 2024. An international film dosimetry intercomparison to establish a multi ‑ center audit framework. Medical Physics 51: 9071–9087.[3] Niroomand ‑ Rad, A.; Chiu ‑ Tsao, S.; Grams, M.P.; Lewis, D.F.; Soares, C.G.; Van Battum, L.J.; et al. 2020. Report of AAPM Task Group 235 Radiochromic Film Dosimetry: An Update to TG ‑ 55. Medical Physics 47: 5986–6025. Keywords: Radiochromic film, EBT-4, open-source
Made with FlippingBook - Share PDF online