Synthetic CT and Dosimetry Uncertainty in CBCT-based Online Adaptive Radiotherapy For Pleural Metastases With Significant Inter-fractional Tumor Volume Reduction

Abstract

Abstract Purpose Online adaptive radiotherapy (oART) dose calculation relies on synthetic computed tomography (sCT), which is significantly affected by anatomical changes. In this study, we present how significant inter-fractional tumor volume reduction can lead to uncertainties in sCT generation and dose distribution. Methods We exported the sCT and cone-beam CT (CBCT) images of each treatment session and analyzed a total of 20 adaptive plans. On each sCT, we overrode the CT number of the dissipated tumor volume with the CT number of the lungs. We recalculated each adaptive plan using the overridden sCT and compared the dose-volume histogram (DVH), Dmean, Dmax, and Dmin between the adaptive plans with the override and the adaptive plans without the override. We also applied linear regression analysis, percentage dose difference, and 2D gamma analysis to quantify and qualify dose variability. Results As the tumor volume shrank, the sCT at each radiotherapy session resembled the simulation CT and increasingly deviated from daily CBCT. Recalculated plans using the overridden sCT showed right-shifted DVH curves for the targets and right lung. Not only did the Dmean and Dmax of the targets and the right lung increase, but also the location of hotspots shifted toward the direction of tumor shrinkage and beam arrangement in the plans with the override. Conclusion As the pleural tumor volume reduced, the adaptive plans underestimated dose distribution and shifted the location of hotspots. Thus, for synthetic CT-based oART, if targets show unexpected drastic volume change, adaptive plans should be evaluated with caution, and re-simulation should be considered.