ESTRO 38 Abstract book

S532 ESTRO 38

planning-CT fiducial marker registration with 6 degrees of freedom. The performance of the SCSSS for positioning accuracy for these specific phantoms was also benchmarked against a commercial surface scanning system (AlignRT, VisionRT Ltd., UK) on a C-arm system.

Conclusion Inter-DIBH and intra-DIBH uncertainties varies greatly between patients. For some patients, the size of the uncertainties will undermine the advantages of DIBH. DIBH-CTs acquired before treatment do not predict the inter-DIBH uncertainty observed during treatment. PO-0975 Feasibility of single-camera intra-bore surface scanning in an O-ring linac L. Delombaerde 1 , S. Petillion 1 , C. Weltens 1,2 , T. Depuydt 1,2 1 University Hospital Gasthuisberg, Department of Radiation Oncology, Leuven, Belgium ; 2 KU Leuven - University of Leuven, Department of Oncology, Leuven, Belgium Purpose or Objective Patient surface scanning inside closed-bore linacs or imaging systems is difficult using ceiling-mounted projector/camera assemblies. A compact single camera surface scanning system (SCSSS) for motion monitoring inside an O-ring linac is proposed and characterized. Material and Methods A Microsoft Kinect depth camera is mounted on a movable rack placed at the back-end of the bore of a Halcyon (Varian Medical Systems, USA) (fig. 1). A transformation to the linac coordinate system retrieved from a cube detection calibration, is applied to the surface point cloud data acquired by the SCSSS. The real-time surface data is registered using an Iterative Closest Point algorithm to a region-of-interest (a) of a 3D CAD-model or (b) of a reference surface acquisition from the first fraction. The positioning accuracy is investigated using anthropomorphic 3D printed phantoms with clinically relevant surface complexity: a head, hand and breast (fig. 2). Each phantom was placed 24 times to within 0 - 11 mm and 0 - 7.6° from the planned position to emulate clinically observed positioning errors. At every position a CBCT is acquired and a surface registration is performed. The surface registration result is compared to a CBCT-to-

Results The registration errors of the SCSSS, compared to CBCT of the 3 phantoms combined were: Lat: 0.30 ± 0.76 mm, Vert: -0.25 ± 0.17 mm, Long: 0.31 ± 0.48 mm and Yaw: - 0.31 ± 0.83°, Pitch: 0.10 ± 0.26°, Roll: 0.51 ± 0.59° for the CAD-model reference, and Lat: -0.65 ± 0.74 mm, Vert: 0.32 ± 0.15 mm, Long: 0.21 ± 0.50 mm and Yaw: -0.49 ± 0.46°, Pitch: 0.13 ± 0.25°, Roll: -0.66 ± 0.65° for the surface acquisition reference. The AlignRT system had combined registration errors of: Lat: 0.74 ± 0.64 mm, Vert: -0.57 ± 1.40 mm, Long: 0.61 ± 0.82 mm and Yaw: - 0.33 ± 0.31°, Pitch: -0.15 ± 0.34°, Roll: 0.38 ± 0.29°. No correlation between displacement magnitude from the planned position and SCSSS registration errors were found, p > 0.14. Conclusion The prototype single camera intra-bore surface system is capable of accurately detecting displacements from the planned position of the phantoms. The compact system, proposed in this work, will allow surface guided radiotherapy and monitoring of intra-fraction motion inside the bore of O-ring gantries.