ESTRO 2025 - Abstract Book

S3144

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2025

Keywords: Dose accumulation QA, proton therapy, thorax

2503

Digital Poster 3D markerless kV tracking of the spine using different kV-based techniques Martijn Simons 1 , Daan Hoffmans 1 , Anna Dinkla 1 , Max Dahele 1 , Wilko Verbakel 1,2 1 Radiation Oncology, Amsterdam UMC location Vrije Universiteit Amsterdam, Amsterdam, Netherlands. 2 Radiotherapy Solutions, Varian Medical Systems, Palo Alto, USA Purpose/Objective: 3D spine position during SBRT can be monitored by markerless tracking of the spine using 2D template matching followed by triangulation of intrafraction kV images. Clinically, we use 2D kV fluoroscopy at 7 fps. An alternative imaging modality could be Triggered Imaging (TI) at 2 fps. Since different kV settings influences tracking rate, accuracy, and dose we performed a phantom study and analyzed these factors for both fluoroscopy and TI-based methods. Material/Methods: A single-arc VMAT plan was delivered (70 seconds). All tracking was done on a spinal target in a static 3D-printed thorax phantom. With RapidTrack Realtime (RTR) non-clinical software (Varian Medical Systems) the target position was tracked by matching an angle-dependent template generated from the planning CT-scan with each kV-image, resulting in a 2D match position and a match score (MS) [1]. An MS<0.15 was used for rejection of matches. The 3D position was obtained by triangulating each match with matches at previous gantry angles. We tested different acquisition modes on a TrueBeam LINAC: Fluoroscopy at 7fps and 2fps (to assess influence of lower imaging frequency) and Triggered Imaging at 2 fps. We applied different kV and mAs modes in order to reduce the dose (some with titanium foil filter). For each kV mode we analyzed the position in 3 dimensions, match score, and rejected matches. kV imaging dose was measured in a 32 cm diameter CTDI phantom with the detector positioned in the periphery at isocenter [2]. Results: 9 different kV/mAs settings were analyzed (table 1). Fluoroscopy with 2fps resulted in 3.5 times lower dose compared to 7 fps fluoroscopy, while maintaining accuracy at the expense of slightly higher percentage of rejected images. TI with the 2 fps default settings (100kV, 10mAs), resulted in 7 times more dose than 2fps fluoroscopy. At lower TI levels (<100kV), the number of rejected matches increases without reaching the lower dose level of fluoroscopy.