KEY POINTS
- OSCAR is an online simulation-free CBCT-based adaptive workflow implemented on Ethos with HyperSight CBCT for urgent and palliative radiotherapy. Predefined templates covered seven anatomical regions, including cervical, thoracic and lumbar spine, sacrum, brain, thorax and cervix, with commonly used schedules ranging from 1 × 8 Gy to 10 × 3 Gy and 5 × 4 Gy.
- The first 11 patients were successfully treated without a dedicated planning CT for their initial fraction. HyperSight CBCT served as the final treatment-day anatomy for contour review, automated optimization and direct dose calculation; a recent diagnostic CT or PET/CT was generally used only as a reference dataset.
- Estimated total time from workflow initiation to treatment was approximately 1 hour 50 minutes. This included about 20 minutes for target delineation, 15 minutes for template-based planning, 30 minutes for physician/physicist review and approximately 20 minutes on the treatment couch, including setup, adaptation and irradiation.
- The workflow deliberately incorporated multiple safety layers: predefined RT-intent templates, physician/physicist four-eye review, a dedicated CBCT acquisition preset, automated OAR generation with manual review, target correction on CBCT and independent secondary dose calculation before delivery.
- Treatment indications were heterogeneous and included spinal and sacral disease, whole-brain radiotherapy, thoracic obstruction or venous congestion, and acute cervical bleeding. One patient had only MRI available and was initialized using an institutional generic whole-body CT reference, while the actual treatment dose was still calculated on patient-specific HyperSight CBCT.
- OSCAR was primarily a first-fraction acceleration strategy rather than necessarily a full-course adaptive workflow: only 2/11 patients continued exclusively with adaptive treatment, while the remaining 9/11 received subsequent fractions using a VMAT plan created from the first-fraction HyperSight CBCT and approved contours.
- Automated nine-field IMRT and retrospectively generated VMAT plans both met institutional target and OAR acceptance criteria. PTV coverage, conformity and homogeneity were broadly comparable, while VMAT provided better sparing for several OARs; because of the small heterogeneous cohort, the comparison was descriptive and no inferential statistics were performed.
CLINICAL TAKEAWAY
For selected urgent or palliative patients, modern CBCT-based adaptive platforms can compress the path from treatment decision to first fraction into roughly two hours without dedicated CT simulation. The concept is immediately clinically interesting, but the evidence remains an 11-patient implementation series with estimated timing and no prospective comparison of safety, symptoms or outcomes.