ESTRO 2021 Abstract Book

S260

ESTRO 2021

Conclusion A common methodology for measuring the accuracy of MLTT approaches has been developed and used to benchmark academic and commercial approaches retrospectively and prospectively. The study outcome paves the way for clinical implementation guidelines for MLTT hardware and software, commissioning and quality assurance recommendations, and identify potential for future developments. The MATCH is live, with datasets and analysis software being available online at https://www.aapm.org/GrandChallenge/MATCH/ to support future research. We thank the AAPM, Scandidos, Washington University and Varian for their support. OC-0358 Development and validation of 4DCT(MRI) numerical lung phantoms for 4D radiotherapy investigations A. Duetschler 1,2 , M. Krieger 1,2 , D.C. Weber 1,3,4 , A.J. Lomax 1,2 , Y. Zhang 1 1 Paul Scherrer Institute, Center for Proton Therapy, Villigen PSI, Switzerland; 2 ETH Zurich, Department of Physics, Zurich, Switzerland; 3 University Hospital of Zurich, Department of Radiation Oncology, Zurich, Switzerland; 4 University of Bern, Department of Radiation Oncology, Bern, Switzerland Purpose or Objective Respiratory motion is a challenge for radiotherapy. Based on CTs from lung cancer patients and motion extracted from volunteer 4DMRIs, we have developed realistic 4D lung phantoms considering a wide range of motion irregularity with accessible ground truth motion fields. The phantoms can be used to investigate and improve treatment strategies for moving tumor treatments. Materials and Methods Subject-specific motion is extracted from 4DMRIs using deformable image registration (DIR). DIR is also used to register binary lung masks from any reference CT and MRI to establish inter-subject correspondence of the lung mesh based on the resulting deformation vector field. Consequently, multiple cycle synthetic 4DCTs, referred to as 4DCT(MRI)s, are obtained by deforming the corresponding CT lung mesh according to the 4DMRI motion fields and then warping the reference CT image according to the mesh motion. For validation of the warping procedure, the same procedure is followed but using motion extracted from a 4DCT to generate a synthetic 4DCT(CT). The end-expiration phase of the 4DCT was chosen as the reference CT. If the motion extraction and warping procedure is correct, the 4DCT(CT) should be similar to the original 4DCT. We have calculated digitally reconstructed radiographs (DRR) simulating fluoroscopy for the end-inspiration phase of two 4DCTs and the corresponding 4DCT(CT)s. Additionally, 4D proton dose calculations have been employed to compare 4D dose distributions calculated on the two 4DCTs with those on the synthetic 4DCT(CT)s. Results Different phases of the 4DCTs and 4DCT(CT)s are compared in Figure 1. An overlay of the DRRs and the relative change of the DRR intensities from the 4DCT(CT) relative to the 4DCT is also shown. The appearance of the 4DCT and 4DCT(CT) and the DRRs thereof are similar. Some differences are visible due to the static ribcage of the 4DCT(CT) and density differences inside the lung due to the image warping. The relative change inside the body is below 5% for at least 81.2% of the pixels for both cases. The comparison of 4D dose distributions for the original 4DCTs and the synthetic 4DCT(CT)s showed good agreement (Figure 2), with a 3%/3mm gamma pass rate of 90.8% and 97.9%, respectively, for the studied cases, thus validating the data processing methods used. Consequently, using 13 lung cancer CTs covering large variations of lung and tumor shapes, locations and sizes, combined with motions extracted from 4DMRIs of 5 volunteers, 65 4DCT(MRI)s have been created, incorporating a total of 871 different breathing cycles.