S2257
Physics - Intra-fraction motion management and real-time adaptive radiotherapy
ESTRO 2026
Poster Discussion 3745
Tumor localization accuracy evaluation on fast ring-gantry CBCT using patient-specific breathing curves and a dynamic anthropomorphic thorax phantom Lars H.B.A. Daenen 1 , Didier Lustermans 1 , Tim H.A. Stassen 1 , Juliane Szkitsak 2 , Roua Abdulrahim 1,3 , Jo Goossens 4 , Gina Rishmawi 1 , Richard Canters 1 , Ilaria Rinaldi 1 , Frank Verhaegen 1 , Gabriel Paiva Fonseca 1 1 Department of Radiation Oncology (MAASTRO), GROW Research Institute for Oncology and Reproduction, Maastricht University Medical Centre+, Maastricht, Netherlands. 2 Department of Radiation Oncology, Universitätsklinikum Erlangen, Friedrich- Alexander-Universität Erlangen-Nürnberg, Erlangen, Germany. 3 Research group NuTeC, Centre for Environmental Sciences, Hasselt University, Diepenbeek, Belgium. 4 Iridium Kankernetwerk, University of Antwerp, Antwerp, Belgium Purpose/Objective: HyperSight CBCT (HS-CBCT) offers enhanced image quality and fast 6-second acquisition on ring-gantry systems1, enabling improved CT number accuracies suitable for dose calculation1,2 and opening avenues for online adaptive treatment. However, in regions affected by motion, such as the thorax, short acquisitions may hinder precise tumor localization and ITV definition. Especially in patients with slow (period >6 s) or irregular breathing, capturing the full extent of motion may not be guaranteed. To bridge the gap between simplified phantoms and complex patient geometry, an in-house dynamic thorax phantom, with synchronized imaging, and patient-derived breathing patterns were used to compare localization and volumetric accuracy of 6-second (fast) and 60-second
Results: For sinusoidal motion, both fast- and slow HS-CBCT acquisitions accurately captured target motion compared to 4DCT (Figure 2a). In slow breathing scenarios, fast HS-CBCT showed phase-dependent discrepancies, with center-of-mass and volume difference up to 4.60 mm (SI) and − 31% , during inhalation or exhalation-only phases (Figure 2b). Capturing both maximum inhalation and exhalation showed similar results to 4DCT. For irregular amplitude breathing, center-of-mass deviations up to 3.51 mm and volumetric discrepancies ranging from - 24% up to 21% were observed, relative to 4DCT, depending on the captured breathing pattern (Figure 2c). For both patient-specific breathing curves, the slow HS-CBCT showed superior localization and volumetric accuracy, compared to the fast HS-CBCT acquisitions.
(slow) HS-CBCT scans. Material/Methods:
A dynamic thorax phantom, with compressible lungs and spherical tumors3, was used to represent patient anatomy. Regular (sinusoidal) and two patient-derived (slow and irregular) breathing patterns were used to drive lung motion (Figure 1). The phantom was scanned on a SOMATOM Definition Drive (Siemens Healthineers) CT scanner, to obtain an average 4DCT scan. Fast and slow HS-CBCT scans were acquired on a Halcyon linac (v4.0, Varian). For the patient-derived patterns, fast acquisitions were synchronized to phases with maximum expected discrepancies, while the sinusoidal pattern was scanned randomly. The lower left-lung tumor was manually contoured in all scans, and the fast and slow HS-CBCT acquisitions were compared to 4DCT in terms of target location and volume, to assess motion capture accuracy.
Made with FlippingBook - Share PDF online