ESTRO 38 Abstract book

S1065 ESTRO 38

significantly smaller than at fraction 2 (6.07 vs. 7.25 cm3, p= <0.04, figure 2). Formal geometric verification confirmed that an isotropic 5 mm PTV margin reliably captured tumour, despite changes in volume and shape during treatment. The volume of tumour not captured by the PTV was trivial (range = 0.00-0.70 cm3, median = 0.00 cm3). 3 mm and 4 mm margins also consistently captured tumour and may be appropriate in clinical practice. Reasons for tumour falling outside the ITV included change in geometry, imperfect online match and CBCT artefact. Conclusion Semi-automated contours are effective for analysing Cone Beam CTs scans during SABR. Our cancer centre delivers SABR with excellent geometrical precision. 5 mm PTV contours effectively capture change in tumour dynamics during treatment. Smaller PTV margins may be appropriate in SABR to thoracic tumours. EP-1954 The role of 4D cone beam CT and abdominal compression in motion management for Liver SABR M. Nix 1 , G. Ward 1 , R. Goody 2 , J. Lilley 1 , N. Casanova 2 , R. Garratt 3 , K. Picken 3 , B. Al-Qaisieh 1 1 Leeds Cancer Centre, Radiotherapy Physics, Leeds, United Kingdom ; 2 Leeds Cancer Centre, Clinical Oncology, Leeds, United Kingdom ; 3 Leeds Cancer Centre, Radiotherapy, Leeds, United Kingdom Purpose or Objective Liver SABR relies on IGRT to ensure liver motion and organ at risk (OAR) position are reproducible from simulation through the treatment course. Motion management may include abdominal compression where possible. 4D cone beam CT (4DCBCT) on-set imaging allows verification of motion extent at each fraction and may lead to modifications prior to treatment delivery. Material and Methods A review of 19 patients treated with 5 fraction SABR for primary or metastatic liver cancer was performed. Total doses delivered were individualised based on OAR constraints and ranged from 40-50Gy. At simulation a multi-phase CT scan was obtained in voluntary exhale breath hold (ExBH), with bolus tracking to obtain arterial, portal venous and delayed phase scans. Oral contrast or water helped visualise duodenum. A 4DCT was acquired to determine extent of tumour motion. When required for target delineation, a treatment position contrast enhanced MRI was obtained. Where possible a CIVCO abdominal compression device was used for all imaging. The GTV was delineated on contrast enhanced ExBH CT or, where obtained, on MRI registered to ExBH CT. In most cases GTV was not visible on 4DCT datasets; ITV margins were determined based on assessment of surrogate liver motion. Maximum inhale and exhale liver contours were generated from 4DCT to aid motion assessment at 4DCBCT. As fiducials are not available in this centre, liver position was used as a surrogate for tumour location. Matching of liver position on maximum exhale phase of 4DCBCT to planning contour was performed. When a full liver match could not be obtained, accuracy of the match in the target region was favoured. A review of the motion and position of any dose-limiting OARs was then performed. Two pre-treatment 4DCBCTs were performed to ensure positioning and motion were stable. 4DCBCT was repeated after each fraction to ensure no change occurred during treatment. Cases were reviewed to assess impact of 4DCBCT. Results All 19 patients completed treatment. 6 (32%) patients were unsuitable for abdominal compression due to abdominal girth or intolerance of the device. 5 (38%) of the abdominal compression patients required modification to the compression device setting during treatment, to ensure liver motion observed on 4DCBCT was within limits

was rigidly registered to T2w (represented the position at 0min) based on mutual information and the intra-fraction motion shifts were calculated from transformation matrix. Results At the time point of 5min, the intra-fraction translation shifts (mm) were 0.03±0.29 (mean±SD), 0.08+0.11, - 0.53±0.72 and 0.71±0.62 in LR, AP, SI and 3D, and rotational shifts ( o ) were -0.06±0.19, 0.00±0.00, and - 0.01±0.20 in roll, pitch and yaw, respectively. At the 10min point, the corresponding translation became 0.02±0.35 (mean±SD), 0.10+0.15, -0.96±0.72 and 1.03±0.72 in LR, AP, SI and 3D, and rotation ( o ) -0.09±0.37, -0.00+0.01, and -0.26±0.33 in roll, pitch and yaw. The 3D intra-fraction shift at two time points was illustrated in Fig. 2. Most patients had excellent positional stability during the first 5min, but an apparent time trend of shift was observed at 10min. 7 out of 10 patients exhibited a much larger shift at 10min than at 5min. The superior soft tissue contrast of MRI facilitated more precise image registration and assessment of intra-fraction motion than X-ray skull tracking. So, our results might faithfully reveal the true intra-fraction shift magnitude of the intracranial target during radiosurgery and be helpful for planning margin setting. The limitations of this study included the off-line nature, small sample size, and limited duration and time points. Conclusion Our results suggested that thermoplastic mask immobilized frameless intracranial radiosurgery could achieve excellent positional stability within the initial 5min of treatment. The frameless immobilization might have high possibility of larger positional shifts that should be substantially compensated after 10min of treatment. EP-1953 Lung tumour dynamics during SABR: Analysis of 415 CBCTs using a semi-automated contouring technique E. Chandy 1 , L. Conway 2 , G. Distefano 2 , J. Earley 2 , K. Lamont 2 , M. Long 2 , I. Phillips 3 , H. Saxby 2 , C. South 2 , C. West 2 , V. Ezhil 2 1 Royal Marsden Hospital, Clinical Oncology, London, United Kingdom ; 2 Royal Surrey County Hospital, Radiotherapy, Guildford, United Kingdom ; 3 Western General Hospital, Clinical Oncology, Edinburgh, United Kingdom Purpose or Objective SABR (Stereotactic Ablative Body Radiotherapy) is an effective and increasingly utilised treatment in patients with early lung cancer. Because it relies on hypofractionation and a steep dose gradient, the ability to precisely target the tumour with tight margins is of paramount importance. Cone-Beam computed tomography allows online soft tissue matching during SABR to lung cancers. The data from these scans can be harvested to formally verify geometrical conformity and to extract radiobiological information about early 415 CBCTs of 44 lung cancer patients treated with SABR were contoured using a semi-automated technique. The volume of tumour at each CBCT was determined in order to verify geometric conformity with the planning scan and to elucidate the radiobiology of tumours during SABR. The contoured CBCT was registered to the planning CT scan with the online match used at radiation delivery. Formal geometric verification was performed and any volume of tumour not within the reference planning structures was recorded. Counterfactual planning tumour volumes (PTVs) with reduced margins were created to evaluate the potential consequences of smaller PTV margins in clinical practice. (Figure 1) Results Lung tumours increased before reducing in volume during SABR. The mean volume of tumour at fraction 5 was response to treatment. Material and Methods