Development of an End-to-end Phantom and Procedure for Computed Tomography and Cone-beam Computed Tomography-guided Online Adaptive Radiotherapy

Purpose: The purpose of this study was to develop an end-to-end (E2E) testing procedure to verify in-room computed tomography (CT)-on-rails and cone-beam CT (CBCT)-guided online adaptive radiation therapy (OART) workflows. Materials and Methods: A water tank-type lung phantom was developed to mimic the lung tissue and tumors, to evaluate deformable image registration (DIR) and dosimetry. Tumor shrinkage was mimicked by shape, position, and material changes. DIR was performed from the ellipsoidal tumor contoured on the initial CT image to multiple tumor cases on the CT and CBCT images. DIR accuracy was evaluated using the dice similarity coefficient (DSC) for manually delineated contours and those propagated by the adaptation anatomy algorithm. Dosimetric accuracy was evaluated by comparing the calculated and measured doses. An ionization chamber was used for point dose measurement, and a GafChromic film was used for gamma analysis using a 3%/2 mm criterion. Results: The averaged DSC values for the CT-guided and CBCT-guided OART workflows were 0.988 and 0.982, respectively. We confirmed high DSCs across different tumor conditions. The relative point dose differences between the measured and calculated doses for the CT-guided and CBCT-guided OART workflows were 0.02% and 0.86%, respectively. In the sagittal planes, the gamma passing rates were 97.8% and 95.6% for the CT-guided and CBCT-guided OART workflows, respectively, indicating good agreement between the measurements and calculations. Conclusion: We developed a novel lung phantom for E2E testing and established workflows for both CT-guided and CBCT-guided OART.