ESTRO 2025 - Abstract Book

S3314

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

ESTRO 2025

Purpose/Objective: Online adaptive radiation therapy (oART) enables the treatment of metastases by using planning and delineation performed on a recent diagnostic CT (dCT). For emergency treatments, we studied an CT-free oART workflow without immediate access to prior diagnostic imaging. Material/Methods: In ten patients, two workflows were simulated digitally (Ethos v2.0, Varian a Siemens Healthineers Company). One workflow with dCT images generated during oART in the FAST-METS trial [1] versus a newly developed CT-free approach. The ten patients were randomly selected with targets in the cervical spine (n=1), thoracic spine (n=3), lumbar spine (n=3), and pelvis (n=3) and gave informed consent for the use of their data. The CBCT-guided oART workflow utilizes a fixed optimization template that cannot be modified during the on-couch session. A generic planning template was created for all anatomical sites. A digital cylindrical phantom with a density of HU=0 and dimensions of 40x80 cm (diameter x length) was created. We evaluated two workflows: (1) treatment planning using the dCT and on-couch optimization on synthetic CT (TP_dCT); (2) treatment planning with a phantom replacing the dCT (Figure 1A) followed by on-couch optimization on a Hypersight CBCT, allowing accurate dose optimization (Figure 1B; TP_CTfree). During treatment planning, the original clinical target volume (CTV) was used, which was rigidly transferred onto the phantom in the TP_CTfree workflow. Standard equidistant 9-field or lateral 7-field IMRT beam setup was used. During the on-couch procedure, deformable registration and AI auto-contouring were used for organ-at-risk (OAR) delineations, while the CTV was rigidly translated to the CBCT images in both workflows without manual adjustments. The clinical objectives and monitor units between the workflows were compared (Wilcoxon signed rank) (Table 1).

Results: The oART workflow was successfully executed for both TP_dCT and TP_CTfree. For all patients, adaptive treatment in the TP_CTfree workflow met all clinical objectives (Table 1), while maintaining clinically acceptable OAR doses. The clinical goals and monitor units for TP_CTfree were not significantly different from the TP_dCT oART. AI derived delineation of OAR (kidney, spinal canal, and esophagus) required manual modification in 5 cases.