ESTRO 38 Abstract book

S610 ESTRO 38

DICE value between RTO TB and RTT TB was 0.69±0.07 (range 0.53-0.81) in group1 and 0.37±0.18 (range 0-0.58) in group2 (p<0.05). Conclusion This study showed a decrease of the interoperator variability in the TB contouring with the use of surgical clips. The reduction of the volumes in the group with clips is closely related to the possibility of decrease side effects like fibrosis. The RTT following an appropriate training may become an important figure in the radiotherapy multidisciplinary team, able to support the RTO also in the contouring phase of the radiotherapy treatment. PO-1098 GTV definition agreement in brain metastasis radiosurgery using 1.5T MRI-sim: a multi-observer study J. Yuan 1 , G.G. Lo 2 , Y. Zhou 1 , O.L. Wong 1 , W.W.K. Fung 3 , K.F. Cheng 3 , K.Y. Cheung 1 , S.K. Yu 1 1 Hong Kong Sanatorium & Hospital, Medical Physics and Research Department, Happy Valley, Hong Kong SAR China ; 2 Hong Kong Sanatorium & Hospital, Department of Diagnostic and Interventional Radiology, Happy Valley, Hong Kong SAR China ; 3 Hong Kong Sanatorium & Hospital, Department of Radiotherapy, Happy Valley, Hong Kong SAR China Purpose or Objective We aim to assess the gross target volume (GTV) definition agreement in the treatment planning of frameless radiosurgery for brain metastasis by using 1.5T MRI- simulation data via a multi-observer study of MRI/CT registration and lesion delineation variability. Material and Methods 10 patients (brain metastases n=33) received CT- simulation (no contrast, 0.78x0.78x1mm 3 ) and 1.5T MR- simulation (3D-T1w-FSE with Gd contrast: 1x1x1 mm 3 ) scans in the identical thermoplastic mask fixed treatment position on the same day prior to their Cyberknife radiosurgery. Three observers were asked to blindly conduct rigid MRI/CT registration and then GTV contouring on the MRI fused CT images. Registration transformation matrix and the center position, volume, dice similarity coefficient (DSC) of the delineated lesions between the observers (obs) were compared using one-way ANOVA and intraclass correlation coefficient (ICC). Results The inter-observer registration translational variability (RMS of the registration shift SD between 3 observers) was 0.07, 0.07, 0.1 and 0.08mm in LR, AP, SI and 3D, respectively. The GTV center position variability (RMS of GTV coordinate SD of 3 observers) was 0.18, 0.21, 0.18 and 0.23mm in LR, AP, SI and 3D. All individual rigid registration shifts relative to the mean registration shifts of all observers were well within 0.3mm in all translations (Fig. 1 upper row). All GTV center positions relative to their mean delineated position by all observers were all within 0.5mm (Fig. 1 lower row). The volumes of the delineated GTVs between three observers (obs1: 1.23±1.51 cc, obs2: 1.35±1.67 cc, obs3: 1.30±1.59 cc) showed insignificant difference (one-way ANOVA: p =0.955;) and excellent (ICC>0.75) agreement (ICC=0.9986, 95% CI = [0.9975, 0.9993]). The DSC between the observers were 0.87±0.06 (obs1vs2), 0.85±0.07 (obs1vs3) and 0.85±0.12 (obs2vs3). The inter-observer agreement was excellent (ICC=0.9546, 95% CI = [0.9076, 0.9797]) for the tumors with volume >=0.2cc. However, large inter-observer variability in terms of DSC (ICC=0.2505, poor) was observed for the small targets with volume <0.2cc (Fig.2).

Conclusion Excellent inter-observer GTV definition agreement could be achieved in the frameless brain metastasis radiosurgery planning using 1.5T MR-sim data. PO-1099 A multi-center contouring study of spinal cord comparing myelo-CT and MRI fusion. H. Tanaka 1 , H. Shimizu 1 , T. Aoyama 1 , H. Tachibana 1 , Y. Koide 1 , D. Katou 1 , S. Adachi 1 , R. Miyauchi 1 , Y. Ooshima 2 , T. Koidaira 1 1 Aichi Cancer Center, Radiation Oncology, Nagoya, Japan ; 2 Aichi Medical University, Radiation Oncology, 1-1 Yazakokarimata- Nagakute-Aichi, Japan Purpose or Objective Spine stereotactic body radiotherapy (SBRT) is an emerging treatment for patients with spinal metastases and is rapidly being adopted in the clinic. Spinal cord is one of the most important organs at risk in Spine SBRT. To contour spinal cord, Myelo-CT or MRI fusion are used. To date, inter modality validation between Myelo-CT and MRI fusion in contouring spinal cord has not been evaluated. A multi-center contouring study was performed to evaluate inter modality variations between Myelo-CT and MRI fusion. Material and Methods Six matured radiation oncologists from 2 centers joined this study. Plain CT, myelo-CT and MRI scans were performed on 12 th thoracic to 5 th lumber vertebra of a patient who had bone metastases in 4 th lumber vertebra. The myelo-CT and MRI were registered to a plain CT by a radiation oncologist. Six radiation oncologists contoured lumber 1 st to 2 nd spinal cord without metastases in vertebra on registered myelo-CT and MRI independently for this study. To avoid inter Radiation Treatment Planning system variations, all registration and contouring were performed on RayStation. Inter modality validation between myelo- CT and MRI fusion, inter observer variations in myelo-CT, and inter observer variations in MRI were evaluated with Dice similarity coefficient. Results