Validation of an Offline Adaptive Radiotherapy Workflow on Halcyon 4.0 Using Raystation v11A
Abstract
Purpose
HyperSight CBCT (hCBCT) on Halcyon 4.0 provides image quality comparable to planning CT (pCT), supporting development of an offline adaptive radiotherapy (OART) workflow. This study evaluated the clinical feasibility of implementing an hCBCT-based OART workflow using RayStation v11A.
Methods
A multi-energy CT phantom was scanned using HyperSight CBCT Acuros protocols for head, thorax, and pelvis on a Halcyon 4.0 linac. HU-to-density relationships were derived per protocol and configured in RayStation v11A for hCBCT-based treatment planning. To validate the workflow, pCT and hCBCT images of an anthropomorphic head phantom were rigidly registered for contour propagation. A brain VMAT plan optimized on pCT was recalculated on hCBCT and evaluated using point, volumetric, and gamma metrics to assess dosimetric agreement. Retrospectively, the same workflow was applied to four clinical Halcyon brain VMAT patients. For each patient, six hCBCTs acquired at five-fraction intervals were replanned, and target and OAR doses were averaged and compared with the original clinical plans.
Results
HU-to-density calibrations showed small variations (maximum < 51 HU) across protocols. In phantom validation, the hCBCT plan had higher target coverage than the pCT plan, with increases of +0.7% for GTV D100% and +0.5% for PTV D95%. Optic chiasm D0.03cc increased by +0.5%, whereas brainstem D0.03cc decreased by -0.6%. Gamma values were 90.0% (1%/1mm), 99.5% (2%/1mm), and 100% (2%/2mm), indicating strong agreement between pCT and hCBCT plans. On average, clinical hCBCT plans exhibited higher target coverage (0.6%±0.4%, p=0.04) and higher PTV D0.03cc (+1.1%±0.2%, p = 0.006). OARs differences were minimal and not statistically significant.
Conclusion
A Halcyon-based OART workflow using RayStation v11A was demonstrated to be clinically feasible, with phantom validation confirming that dosimetry between pCT and hCBCT plans was closely matched. Application of this workflow in clinical patients further supported safe clinical implementation.