Purpose/Objective(s):
CBCT-guided adaptive radiotherapy enables daily plan adaptation based on the anatomy of the day. Earlier software generations relied on deformable image registration to generate synthetic CT (sCT) datasets for dose calculation. Prior phantom studies demonstrated clinically relevant dose deviations in the presence of substantial density changes, particularly in lung treatments [1]. This work evaluates whether a next-generation CBCT-guided adaptive workflow incorporating advanced CBCT imaging and direct CBCT-based dose calculation improves dosimetric agreement compared to the former sCT-based approach.
Materials/Methods:
Measurements were performed on a CBCT-guided adaptive system using advanced CBCT imaging (H-CBCT). A dynamic thorax phantom with interchangeable spherical inserts (1, 2, and 3 cm diameter) simulated anatomical and density variations. Treatment plans were generated on planning CT (pCT) datasets. Adaptive sessions were conducted using H-CBCT. Dose distributions from the CBCT-guided adaptive system were compared with independent recalculations in a reference treatment planning system. In addition, results were compared with sCT-based dose calculations and corresponding results from analogous phantom measurements [1]. Evaluated parameters included PTV Dmean, D2%, D98%, and monitor units (MU).
Results:
Dosimetric differences between the earlier sCT-based workflow and the next-generation H-CBCT-based workflow are summarized in Table 1. Using the earlier software generation, substantial deviations were observed for scenarios involving pronounced density changes. For adaptation from a plan with a 3 cm sphere in the pCT to a 1 cm sphere, the PTV Dmean difference reached −15.9 %. In contrast, the next-generation software with direct H-CBCT dose calculation reduced the maximum deviation to 0.7 %. Across all insert combinations, dose agreement improved markedly. Furthermore, MU variability observed with the earlier workflow was substantially reduced with the H-CBCT-based approach.
Conclusion:
The CBCT-guided adaptive workflow incorporating advanced CBCT imaging and next-generation software improves dosimetric accuracy for lung treatments. The pronounced reduction of density-related dose errors highlights the robustness of direct CBCT-based dose calculation and mitigates limitations previously associated with sCT generation.
References: [1] J Appl Clin Med Phys. 2024, 25(7):e14311
Table 1: Comparison of PTV Dmean deviations in % between sCT- (top) and direct H-CBCT-based dose calculation (bottom) across all sphere insert combinations (pCT vs. on-couch session).
