Purpose: Dose accumulation is a key component in adaptive radiotherapy (ART). The purpose of this study is to develop a multi-layer quality assurance system to verify dose accumulation performed in an ART program.
Methods: The development involves the following steps: converting HU and dose in daily and planning CTs to mass and energy, generating a pseudo-inverse map φ-1 for each displacement vector field φ generated by the ART program, and using the map φ-1 and a mass-and-energy congruent mapping (EMCM) method to reconstruct dose on the planning CT. To generate φ-1, the source voxels of φ were divided into three categories: empty, unique and clustered voxels, and the center-of-mass and tetrahedron-based continuous mapping methods were used to generate the inverse displacements for the clustered and empty source voxels, respectively. The HU-mass density tables used in clinic were adopted for the dose-to-energy conversions. An inverse consistency error (ICE) metric, energy conservation and Gamma index were taken as criteria to evaluate the displacements, energy and dose reconstructed by the ART program. Five registration cases were evaluated: three performed from daily MVCTs to planning CTs using PreciseART and two performed between the end-inhale and end-exhale of 4DCTs using MIM software.
Results: The results showed that the empty, unique, and clustered voxels are 45.6%, 10.2% and 44.2% on average for the three daily image registrations, and 20.7%, 65.9% and 13.4% for the 4DCT registrations. Their average ICEs are 1.05±0.92 mm and 1.26±1.22 mm, respectively. After dose reconstruction, the average energy in PTV changed by 0.4±0.13 % for EMCM and 3.22±0.58% for PreciseART. Compared with the results of EMCM, the dose and energy reconstructed by PreciseART show the 5%/5 mm gamma passing rates of 96.9±2.1% and 90.6±3.3%, respectively.
Conclusion: A multi-layer quality assurance system has been developed to verify deformable dose accumulations performed in ART.
Funding Support, Disclosures, and Conflict of Interest: This project was supported in part by NIH/NIBIB R01-EB028324.