ESTRO 2021 Abstract Book

S1499

ESTRO 2021

excursions, the advantage of using Synchrony respiratory tracking was reduced for some patients. Prediction model errors were negligible. Correlation model errors were small but not negligible. The related DVH variations were clinically not relevant, indicating a good plan robustness to tracking errors when simulated on the planning CT. Uncertainties due to non-rigid anatomical variation should be included in future analysis.

PO-1771 Accuracy for patient setup positioning with Catalyst™ HD for deformed cases B. Kadman 1 , A. Takemura 2 , T. Ito 3 , N. Okada 4 , H. Kojima 5 , S. Ueda 5

1 Division of Health Sciences, Graduate School of Medical Sciences, Kanazawa University , Department of Quantum Medical Technology, Kanazawa , Japan; 2 Faculty of Health Sciences, Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Department of Quantum Medical Technology, Kanazawa, Japan; 3 Division of Health Sciences, Graduate School of Medical Sciences, Kanazawa University, Department of Quantum Medical Technology, Kanazawa, Japan; 4 School of Health Sciences, College of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Department of Quantum Medical Technology, Kanazawa, Japan; 5 Kanazawa University Hospital, Department of Radiation Therapy, Kanazawa, Japan Purpose or Objective Catalyst ™ HD (C-RAD Positioning AB, Uppsala, Sweden) is a commercial equipment for Surface-Guided Radiotherapy which employs a non-rigid deformable model. The challenge of optical surface scanning (OSS) system is the performance of the deformable image registration (DIR) algorithm when the body of the patient is deformed. The purpose of this study was to evaluate the impact of breast deformation on the accuracy of the surface imaging system. Materials and Methods In-house physical phantoms were applied to investigate the relationship between the position setup error from the OSS system Catalyst™ HD and the breast deformation. The in-house physical phantoms were the box and slab phantom to simulate the extension and shrinkage deformation (-30 to 30 mm) in six directions, and the in-house deformable breast phantom relating to the clinical situation, including small, standard, and large breast sizes. After changing the slab and breast piece of the phantoms to deform the shape, we collected the relative correction data of Catalyst™ HD and compared it with the original shape of each phantom. The Hausdorff distance (HD), volume difference, and the maximum value of each element in displacement vector field (DVF) were employed as the deformation metric. Results The characteristic of position accuracy of the OSS system was simplified by the box phantom, the position setup error increased during extension and shrinkage, but the cranial and caudal directions decreased when the magnitude deformed >25 mm. For the extension, the maximum translational setup error was 7.36 ± 4.80 mm in the right lateral, and rotation 3.28º ± 2.17º in the left lateral. For the shrinkage, the maximum translational setup error was 12.01 ± 6.24 mm in the cranial, and rotation 3.26º ± 1.80º in the right lateral. All rotations were < 3.26º and in almost all cases remained stable (Figure 1). In accordance with the results of position accuracy with three sizes of breast phantom, the position setup error increased when deformation increased, which occurred more in translation than rotation.

Conclusion When the reference surface was deformed, the position setup error in translation was more than in the rotation direction. The magnitude of deformation had an impact on the position accuracy of the Catalyst™ HD system. As one concern relates to the large surface deformation in the longitudinal axis, further prospective deformation evaluation of the phantom should define the positional errors for all translations and rotations. Investigation of the region that deforms in the longitudinal may clarify the position accuracy on this axis.