ESTRO 2025 - Abstract Book

S3094

Physics - Inter-fraction motion management and offline adaptive radiotherapy

ESTRO 2025

1089

Digital Poster Experimental evaluation of phase-specific target volume repositioning using time-resolved CT/CBCT Max Kelm 1 , Hermann Krümpelmann 1 , Laura Buettgen 1 , René Werner 2 , Tobias Gauer 1 1 Department of Radiotherapy and Radiation Oncology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany. 2 Institute for Applied Medical Informatics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany Purpose/Objective: Respiratory-induced target motion poses challenges for image-guided radiotherapy, complicating the alignment between planned and actual target positions. Time-resolved CBCT and CT imaging capture target positions across respiratory phases, providing probabilistic estimates of target location and motion range. This study evaluates the impact of respiratory motion variability on the accuracy of rigid image registration methods for determining pre treatment target volume repositioning (via couch shifts). It focuses on identifying optimal respiratory phase data to enhance rigid registration accuracy, thereby refining target repositioning in lung and liver SBRT, especially in cases of irregular breathing patterns. Material/Methods: Target volume repositioning was simulated using a CIRS thorax phantom with a target insert mounted on a 4D motion platform. Patient-derived breathing curves were translated into 1D longitudinal phantom motion. The moving phantom was imaged using time-resolved CBCT and CT, followed by phase-specific rigid registration. Determined shifts were compared to optimal shifts calculated from the corresponding motion trajectories to evaluate residual displacements across various registration methods. The analyzed phase images included conventional average, mid-ventilation, and maximum exhalation reconstructions. To assess the impact of respiratory variability, both regular and highly irregular breathing patterns were examined. Irregular breathing patterns were predefined using a classification model, which associated them with an elevated risk of local tumor recurrence. Results: As shown in Figure 1, across 42 measurements resulting in 126 registrations, residual positioning displacements averaged between 1 and 3 mm, rarely exceeding 3 mm. Green markers indicate instances of minimal residual displacement among the evaluated reconstruction methods, while red markers denote the least accurate implementations. Square markers represent regular motion measurements, while triangles indicate irregular ones. Registration using conventional averaged phase images demonstrated the highest accuracy for patients with irregular breathing (1.35 ± 0.90 mm), while maximum exhalation phase images resulted in the lowest displacements for regular breathing patterns (0.69 ± 0.74 mm). The maximum exhalation phase was highly reproducible during regular breathing, whereas irregular patterns exhibited greater variability, primarily in inspiration but also in expiration depths.