Abstract Book

S210

ESTRO 37

Material and Methods Given the availability of the anthropomorphic, deformable and multimodal ADAM-pelvis phantom (Figure 1), the end-to-end test was simulated for prostate cancer patients. Two such patients, treated with SMART in 5 fractions have been selected as reference (TXSIM1 and TXSIM2). From these patients, the bladder and rectum filling was determined. At simulation the bladder filling was 320 and 220 cc for respectively TXSIM1 and TXSIM2, the rectum was air filled for P2. During treatments the average bladder filling was 150 cc (range 98 – 232 cc) for TXSIM1 and 195 cc (range 150-235 cc) for TXSIM2. Air filled rectum was observed during 1 respectively 5 fractions for TXSIM1 and TXSIM2. During this phantom study, these rectum and bladder filling volumes were recreated in the silicone bladder and rectum of the phantom. The clinical SMART workflow consists of generating a simulation CT- and MR (MR-SIM) scan, which are co-registered and contoured and were used to create an IMRT base plan of 10 Gy in 5 fractions. Prior to each fraction an MRI is acquired, which is deformable co- registered to MR-SIM and re-contoured. The transformation from the MR-SIM to the fraction MRI forms the basis for the creation of a synthetic electron density map by deformation of the CT-SIM. This synthetic CT is used for dose calculation and re-optimization using the same number and directions of beams. The phantom’s bladder and rectum are equipped with small pockets, which can hold 9 pieces of radiochromic film (GafChromic EBT3) with a size of 1x2 cm. After each fraction (of 2 Gy) the films were replaced, digitized and calibrated to dose. The integrated treatment planning system of the MRIdian was used for dose calculation. The calculated dose distributions in the planes of the film pieces were exported and registered to the film measurements. Per fraction, the total gamma pass rate over all films was calculated using 3%/3 mm criteria, and the mean dose difference between calculation and measurement over all fractions was calculated per film location.

OC-0410 Cone-beam CT based dose calculation in the thorax region L.P. Kaplan 1 , L. Hoffmann 2 , D.S. Møller 2 , U.V. Elstrøm 2 1 Aarhus University, Dept. of Physics and Astronomy, Aarhus C, Denmark 2 Aarhus University Hospital, Dept. of Medical Physics, Aarhus C, Denmark Purpose or Objective Anatomical changes during radiotherapy (RT) of lung cancer can lead to target under-dosage or over-dosage of organs at risk. Cone-beam CT scans (CBCT) are taken daily for patient setup and are used in adaptive RT to visually identify anatomical changes. The purpose of this study was to validate dose calculation based on CBCT images to make daily assessment of delivered dose possible. To overcome known limitations of CBCT dose calculation stemming primarily from scattered radiation, a stoichiometric calibration in the correct scatter conditions is made. Material and Methods Calibration: A published stoichiometric parametrization ("The calibration of CT Hounsfield units for radiotherapy treatment planning", Schneider et al. 1996) using three fitting parameters was used for calibration. The CIRS Electron Density Phantom model 062M containing eight tissue substitutes and a large air cavity was scanned using thorax protocols on a CT scanner (Philips Brilliance Big Bore) and five linacs (Clinac On-Board Imaging system, VMS). Phantom study: The anthropomorphic Alderson Radiation Therapy Phantom (ARTP) was scanned on the same scanners and a dose plan was simulated on these images using the Eclipse treatment planning system (VMS). The dose on CBCT images was calculated using a mean curve obtained by taking the mean of the individual curves from the five linac CBCTs. Six structures were delineated on CT images of the ARTP, rigidly transferred to the CBCTs and mean doses to these structures were compared. Patient study: Fifty lung cancer patients had control CT scans (cCT) taken at treatment fraction (F) 10 and 20. Twelve patients were selected for the analysis based on the criterion that the anatomy and patient positioning on cCT and CBCT acquired at F10 and F20 was unchanged. Target and organs at risk were transferred rigidly from cCT to CBCT. The planned dose was calculated on the cCTs and the CBCTs at F10 and F20 and compared. Results Phantom study: The median difference in mean dose to the structures delineated on the ARTP was -0.37% (max.: 3.12%, min.: -6.52%) and the 25th to 75th percentiles lie within ±1%. Patient study: Figure 1 shows 95% and 50% isodose levels for one dose plan on CT (left) and CBCT (right). Very similar anatomy and dose distributions are seen. Dose differences between CBCT and CT images for 24 patient image sets are shown in figure 2 (12 patients, two data points each). The median lies within ±2% and the 25th to 75th percentiles within ±3%. The largest deviation is - 7.9%.

Results The averaged dose differences per location over all re- optimized fractions are shown in Figure 2. The overall mean relative dose difference was 1.7% (SD = 3.2%). The average gamma pass rate for all fractions was 95.1% (SD = 3.1%) and 92.9% (SD = 3.4%) for TXSIM1 and TXSIM2, respectively.

Conclusion End-to-end testing of online daily adaptive treatment, including daily variation of patient’s geometry is shown to be feasible. Film measurements showed good agreement with dose calculation.