Abstract Book

S1109

ESTRO 37

respiratory cycle and internal fiducial markers to measure the movement of the tumor and set up the patient. The tumor position is measured intermittently during the treatment via stereoscopic x-ray images to compensate the baseline drift. Therefore, the accumulative changes in the couch position correspond to the baseline drift in the tumor motion. Results The average change in position of the treatment couch during the treatment time was 0.2 ± 0.8 mm (mean ± standard deviation), -0.7 ± 2.1 mm, and 2.3 ± 5.3 mm in the left-right (LR), antero-posterior (AP) and cranio- caudal (CC), directions respectively. Overall the baseline shift/drift occurs toward the posterior and cranial directions. The incidence of a baseline drift exceeding 3 mm was 0% and 22%, for the AP and CC directions respectively, within 10 minutes of the start of treatment, and 43% and 86% within 25 minutes. On the other hand, the incidence of a baseline drift exceeding 5 mm was 0% and 10% for the AP and CC directions respectively, within 10 minutes of the start of treatment, and 14% and 71% within 25 minutes. The intra-fractional uncertainties due to baseline drift of liver tumors are: Direction M (mm) Σ (mm) σ (mm) LR 0.1 0.2 0.4 AP -0.3 0.6 0.8 CC 1.2 2.2 1.5 M: the overall mean or group systematic error, Σ: the standard deviation (SD) of the systematic error, σ: the SD of the random error. In the absence of intra-fractional IGRT, the baseline drift uncertainties imply the use of increased margins of up to 4 mm and 3 mm in the cranial and caudal directions respectively, in SBRT for liver tumors, when the rest of uncertainties are minimized. Conclusion Real-time monitoring and frequent adjustments of the couch position are suggested to be essential to compensate for possible underdosage in CC direction due to baseline drift in SBRT for liver tumors. EP-2030 Examination of the automatic fiducial marker detection on the Vero system M. Ziegler 1 , T. Brandt 1 , S. Lettmaier 1 , R. Fietkau 1 , C. Bert 1 1 University Hospital Erlangen, Radiation Oncology, Erlangen, Germany Purpose or Objective The Vero SBRT system (Brainlab, Munich, Germany) can treat intra-fractionally moving tumors via dynamic tumor tracking (DTT). During a DTT treatment, the tumor position is determined by extracting the 3D-position of implanted radio-opaque fiducial markers (FM) from two orthogonal kV X-ray images. At the University Hospital Erlangen two Visicoil markers (IBA, Schwarzenbruck, Germany) with a diameter of 0.75 mm and a length of 20 mm are implanted in close vicinity to the tumor one week before pre-treatment imaging and used as FM during DTT. In this work the detection rate of the FM at the Vero system is investigated and possible influencing Nine male and two female patients between the age of 51 and 90 years were treated with DTT at the University Hospital in Erlangen with a total of 132 fractions. On median 173 X-ray image pairs were taken during every fraction of the treatment. In 64.6 % of these, the FMs are detected by the Vero system. For each kV image the average intensity I of the tissue surrounding the FM in a 3 cm radius was calculated (within the acquired images a factors are determined. Material and Methods

high intensity corresponds to low tissue density), as well as the relative angle θ of the FM towards the imager. The resulting values were then evaluated regarding the success of the FM (2D-)detection of every imager individually as well as against the combined 3D- detection. Based on the obtained information, the density of the successful FM detection S(X) and of the failed detections F(X) depending X ∈ {I,θ} on were calculated. The ratio S(X)/F(X) was used to check for dependencies of the FM detection rate depending on either the surrounding intensity or the relative angle of the marker towards the imager. By performing this evaluation not only for the combined detection, but for every imager individually as well, a difference of the two imagers could be made visible. Results The ratio S(X)/F(X) shows, that the detection probability does not depend on the relative angle of the FM towards the imager. However an increase of this ratio for increasing intensities, and thus higher marker contrast, could be observed. This shows that a FM implantation in line of sight with tissues of low density can improve the FM detection. Additionally our results show that the detection is more reliable within images taken by imager 1. A following analysis with the TOR 18FG (Leeds Test Objects, Boroughbridge, United Kingdom) phantom revealed that during DTT the contrast of imager 2 is about 3 times lower compared to imager 1. This effect is not visible in kV images acquired without using DTT. Conclusion It could be shown that the FM detection of the Vero system is not dependent on the relative angle of the marker towards the imager, but shows a dependency on the intensity surrounding the FM within the acquired kV images. A higher FM detection rate within imager 1 leads to further examinations of the imaging process during DTT. EP-2031 Kidney motion in children and young people quantified using 4DCT deformable registration N. Lavan 1 , D. Mc Quaid 2 , G. Smyth 2 , S. Vaidya 3 , F. Saran 4 , U. Oelfke 5 , H. Mandeville 3 1 Institute of Cancer Research, Radiotherapy and imaging, London, United Kingdom 2 The Royal Marsden NHS Foundation Trust, Joint Department of Physics, London, United Kingdom 3 The Royal Marsden NHS Foundation Trust, Children's and Young Person's Unit, London, United Kingdom 4 The Royal Marsden NHS Foundation Trust, Department of Neuro-Oncology, London, United Kingdom 5 Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, Joint Department of Physics, London, United Kingdom Purpose or Objective Kidneys are the principle dose-limiting organ for abdominal RT in children and young people. Centre of mass (COM) measurement is a frequently reported descriptor of respiratory-related organ motion but is insensitive to changes in organ shape. Deformable image registration of 4DCT is an alternative method of measuring intra-fraction changes in structure geometry reflecting both deformation and translation in x, y and z. This work is the first to apply this methodology to 4DCT in children and young people. Material and Methods 15 patients having upper abdominal RT and simulation 4DCT were analysed; median age 4 (range: 1 - 17 years). 13 right and 10 left kidneys in total were included. Following acquisition, 4DCTs were retrospectively sorted into 10 phase bins; 0 – 90%. Kidneys were contoured on all image sets by the same clinician to minimise intraobserver delineation variation. Kidney structures were divided into three segments; upper, middle and lower, with respect to the renal hilum Figure 1 .