ESTRO 2025 - Abstract Book

S2633

Physics - Detectors, dose measurement and phantoms

ESTRO 2025

Material/Methods: A custom in-house phantom was 3D printed to enable accurate positioning of a spinal titanium implant, comprising one bridge, two screws in different directions and one bar perpendicular to axial plane. The phantom had a rectangular shape and was filled with water (Figure1a). CT images were acquired in a Revolution CT (General Electric) with department's standard head protocol and slice thickness of 0.625mm. A secondary reconstruction was performed using the smallest possible FOV (SFOV) that allowed full visualization of the implant (Figure1b). Metallic implant was contoured in Raystation TPS (Raysearch). Window and level settings were adjusted to match the bar's actual diameter. These settings were used to contour the implant automatically via thresholding in the SFOV CT. After review, the contoured region was overrriden to titanium with density 4.54 g/cm 3 (Figure1c) and mapped to standard calculation CT. A single-energy 150MeV proton beam with 10x20 cm 2 size was calculated using the clinical TPS Monte Carlo dose engine, with a 1mm dosegrid and 0.2% uncertainty (Figure1c). Beam direction was perpendicular to one screw axis and parallel to the second screw axis. The beam was delivered with gantry at 90º in a Proteus One (IBA) system. The dose plane was measured 4 mm distal to the implant (Figure1d), using the portable EPID device Sphinx Compact (IBA Dosimetry), calibrated for absorbed dose in proton beams.

Results: The contouring procedure allowed the geometric definition of the metallic implant contour with 0.2 mm resolution, corresponding to the SFOV pixel size. This methodology can also be applied to clinical cases, allowing for the accurate and efficient contouring of implants with complex geometries. Figure2 compares dose plane and profiles across the implant. TPS calculations replicated the peaks measured for the different components of the implants. The maximum dose difference was observed in the screw parallel to beam axis (0-4.5%). Smaller differences were noted in the bridge (0-2.1%), the perpendicular screw (1%) and the bar