ESTRO 2025 - Abstract Book

S3051

Physics - Image acquisition and processing

ESTRO 2025

Conclusion: This study demonstrates PCCT’s accuracy in physical quantity estimation, particularly in RSP uncertainties and tissue separation. Although comparable results with DECT were demonstrated, literature shows dose reduction possibilities and the PCCT potential is not yet entirely exploited as the methodology is optimized for DECT.

Keywords: Photon Counting CT, VMIs, Quantitative imaging

3440

Digital Poster Bone bulk density based on classical, Direct Density and synthetic CT for prostate MR guided SBRT: Is there a relevant dosimetric impact? Akos Gulyban 1 , Zelda Paquier 1 , Kevin Tihon 1 , Sara Poeta 1 , Nicolas Jullian 2 , Robbe Van den Begin 2 , Madeline Michel 2 , Christelle Bouchart 2 , Manuela Burghelea 1 , Nick Reynaert 1 1 Medical Physics, Institute Jules Bordet, Brussels, Belgium. 2 Radiation Oncology, Institute Jules Bordet, Brussels, Belgium Purpose/Objective: During MR-guided online adaptive radiotherapy (MRgRT) the lack of electron density information is usually compensated by assigning a pre-determined average value to the entire organ (bulk density) based on the available CT scans. In our investigation, we evaluated the bulk density variation of the bone for prostate MRgRT, based on three different CT scans: the classical (C-CT), the direct density (DD-CT) CT and the synthetic CT (S-CT). Additionally dosimetric effect of bone density variation was assessed. Material/Methods: Twenty prostate patients treated with prostate MRgRT with 40 Gy in 5 fractions using 1.5 T MR-Linac daily online adaptive workflow were included in this study. Treatment planning simulations were performed on Somatom goPro CT-scanner and on Magnetom Aera 1.5T MR scanner (Siemens Healthineers, Germany). C-CT and DD-CT scans were reconstructed with two different CT kernel (Br40 and Sd40) using carekV protocol based on the topogram, while the synthetic CT were generated using the syngo.via platform[1] based on the T1 Vibe DIXON sequence trained to mimic CT with 120 kVp acquisition. Bone structure were semi-automatically segmented and transferred to the reference T2 MR scans for treatment planning using MonacoUnity (version 5.51, Elekta AB, Sweden). Bone density were assigned based on the dedicated HU-Relative Electron Density tables. For CT acquisitions the kVp and CT dose index (CTDI) were also extracted. Clinical treatment plans were generated using the DD-CT based bone bulk-density, followed by a recalculation using average bone ED±standard deviation (SD) values. For each relevant target and organ at risk, the DVHs were compared and dose deviation from the reference plan for a given volume were calculated. Results: Large variation of kVp selection and CTDI values: 80/90/120/130 kVp were used for 7/6/5/2 patients respectively, with an average of 14.1 mGy CTDI (range: 2.9-34.9). The bone bulk density showed also moderate differences: 1.21±0.05 (Average±SD), 1.21±0.04 and 1.17±0.03 for C-CT, DD-CT and S-CT respectively, resulting for S-CT slightly lower average values, while making it more consistent (figure 1). Based on the density variation all clinical plans were recalculated with 1.15 and 1.25 bone bulk density. Dose difference distribution of a given structure showed less than 2% deviation from the reference values (figure 2).