ESTRO 2025 - Abstract Book

S3584

Physics - Quality assurance and auditing

ESTRO 2025

Conclusion: Although J-CTCA and EPTN use different tissue-equivalent materials and methods to create CTN-to-SPR conversion curves, the mean differences in the CTN-to-SPR conversion curves evaluated for each tissue type remained within 1%.

Keywords: CT-SPR conversion table, proton therapy

References: 1. Nakao M, et al. CT number calibration audit in photon radiation therapy. Med Phys . 2024;51(3):1571-1582. 2. Peters N, et al. Experimental assessment of inter-centre variation in stopping-power and range prediction in particle therapy. Radiother Oncol . 2021;163:7-13. 3. Nakao M, et al. Stoichiometric CT number calibration using three-parameter fit model for ion therapy. Phys Medica . 2022;99:22-30. 4. Kanematsu N, et al. Modeling of body tissues for Monte Carlo simulation of radiotherapy treatments planned with conventional x-ray CT systems. Phys Med Biol . 2016;61(13):5037-5050. 5. Peters N, et al. Consensus guide on CT-based prediction of stopping-power ratio using a Hounsfield look-up table for proton therapy. Radiother Oncol . 2023;184.

650

Digital Poster Patient specific quality assurance of sCT in a MR-only workflow: a comparison among generators Stephane Dufreneix 1 , Mathilde Levardon 2 , Maxence Rayer 1 , Thomas Le Dorze 1 , Camille Guillerminet 1 , Damien Autret 1 1 Medical Physics, Institut de Cancerologie de l'Ouest, Angers, France. 2 Medical Physics, Centre Hospitalier Universitaire, Angers, France Purpose/Objective: Magnetic Resonance Imaging only workflow has gained in interest with the introduction of artificial intelligence for synthetic CT (sCT) generation. These sCTs have been extensively evaluated against CT but there is still a need to perform patient-specific quality assurance (PSQA) in clinical routine where CT is missing by definition. It was previously suggested to evaluate a sCT against another sCT [1,2]. The aim of this study was to evaluate 3 generators commercially available and based on AI or bulk density assignment for the PSQA of another AI-based generator in clinical routine. Material/Methods: 42 brain patients were retrospectively enrolled in the study and underwent a MR and a CT exam. 4 sCT were generated (Siemens Artificial Intelligence, Spectronic, Therapanacea and Siemens Bulk Density). One sCT (Siemens_AI) was considered as the clinical sCT and compared against CT for reference and against the 3 other sCT for PSQA. Dose metrics like the mean error, dose-volume histogram metrics and 1%/1mm gamma analysis were evaluated. Results: The comparison against CT allowed to set tolerance levels to 1 % for ME (Figure 1) and DVH metrics and 90% for gamma pass rate (Figure 2). All sCT used for PSQA gave false negative for at least one metric studied and none of the sCT was able to clearly identify a true sCT failure for a patient with a metal artifact (patient 14). The comparison against the bulk density sCT gave the worst results and this simplistic generator should not be considered for PSQA.