ESTRO 2025 - Abstract Book

S3009

Physics - Image acquisition and processing

ESTRO 2025

Material/Methods: The trial aims to recruit patients with lung cancer (n=15) or head and neck cancer (n=10). The portable radiation therapy chair was positioned under a TrueBeam™ Linac (Varian, USA). Participants were setup on the chair with an arms-up, forward leaning position, and immobilised with vacuum-bags and a torso strap (Figure 1a). Upright CBCT (uCBCT) was reconstructed, from kV fluoroscopy (123kV, 257mAs, 15fps, 900 frames) acquired during a full rotation of the chair, using the reconstruction toolkit[5]. Orthogonal kV images (123kV, 4.8mAs) were acquired after participant setup, following a simulated treatment (intra-fraction motion), and after a short break from the chair and repeat setup (inter-fraction motion). Visibility of image guidance structures (carina, tumour) on uCBCT were recorded. Geometric accuracy of uCBCT was assessed by comparing manual contours of a rigid structure (the T6 vertebrae) on upright and supine CBCT images. To measure intra- and inter-fraction motion, orthogonal upright kV images were manually rigidly registered in the spinal region.

Results: Four participants receiving palliative radiation therapy for lung cancer were enrolled. The cohort was gender balanced, age range 69-89 years, ECOG performance status 0-3. All image guidance structures were visible on uCBCT (Figure 1b). Vertebrae contours had an average surface distance (mean [range]) of 0.6[0.4-1.0] mm, and 95th percentile Hausdorff distance of 1.8[1.0-3.0] mm. Participant translation (Figure 2a) had a mean absolute deviation (intra-/inter-fraction) of 2.0/1.2mm in left-right (LR), 2.8/6.7mm in anterior-posterior (AP), and 1.7/2.3mm in