ESTRO 2025 - Abstract Book

S3289

Physics - Intra-fraction motion management and real-time adaptive radiotherapy

ESTRO 2025

2936

Proffered Paper First clinical online real-time motion-including target and bladder dose reconstruction during prostate SBRT delivery Karolina A Klucznik 1 , Thomas Ravkilde 2 , Simon Skouboe 1 , Paul Keal 3 , Laura Happersett 4 , Hai Pham 4 , Brian Leong 4 , Pengpeng Zhang 4 , Grace Tang 4 , Per R Poulsen 1,2 1 Danish Centre for Particle Therapy, Aarhus University Hospital, Aarhus, Denmark. 2 Department of Oncology, Aarhus University Hospital, Aarhus, Denmark. 3 Image X Institute, Sydney Medical School, Sydney, Australia. 4 Department of Medical Physics, Memorial Sloan Kettering Cancer Center, New York, USA Purpose/Objective: Organ motion during prostate radiotherapy leads to degraded target dose and variable normal tissue dose. Real time prostate motion monitoring to guide patient position corrections during treatment is currently possible on conventional linear accelerators using research software. However, treatment outcomes are more directly linked to the actual delivered dose, which is currently unavailable in real-time. This study aimed to perform the first real-time, online motion-including dose reconstruction during prostate stereotactic body radiotherapy (SBRT). Material/Methods: Twenty patients received prostate SBRT of 40Gy in five fractions. The treatment plans were optimized to meet the planning objectives: D 95% between 100%-93.75% for the prostate PTV, D 95% >95% for the prostate CTV, and V 36Gy <10% for the bladder. All patients were treated with volumetric modulated arc therapy (VMAT). MV-kV image pairs were acquired at every 20deg of gantry rotation during treatment for real-time prostate localization using an in-house method [1]. The patients were repositioned when the prostate position error exceeded 1.5mm (MV-kV guidance). In-house developed dose reconstruction software [2] was integrated into this workflow and performed real-time motion including prostate and bladder dose reconstruction during the treatments. Motion-induced dose distortions were quantified at each fraction as differences between delivered and planned prostate CTV D 95% and bladder V 36Gy . Post-treatment, treatments without intrafractional patient repositioning were simulated to assess the effect of MV-kV guidance. Real-time dose reconstruction was validated against treatment planning system (TPS) calculations by emulating the motion with multiple isocenter shifts [3]. Results: Figure 1 presents the real-time reconstructed doses for a fraction with large motion. Under MV-kV guidance, the planning criterium of prostate CTV D 95% >95% was fulfilled at all 92 fractions except one (Figure 2a, blue boxplots) and the criterium of bladder V 36Gy <10% was fulfilled at all fractions (Figure 2b) for this patient cohort. Simulations without MV-kV guidance revealed that the planning criteria would have been violated at 9/92 fractions (CTV D 95% ) and 4/92 fractions (bladder V 36Gy ) (Figure 2, red boxplots). The mean (standard deviation, range) motion-induced dose difference was -0.6% (1.1%, [-5.6,+1.6]%) (CTV D 95% ) and +0.1% (0.5%, [-1.0,+1.3]%) (bladder V 36Gy ) with MV-kV guidance and increased to -2.0% (6.6%, [-42.5,+1.8]%) (prostate) and +4.1% (2.0%, [-5.8,+10.1]%) (bladder) without MV-kV guidance. The root-mean-square error between real-time and TPS doses was 0.8% for the prostate CTV D 95% and 0.3% for the bladder V 36Gy .