Abstract Book

S1140

ESTRO 37

stoichiometric fit was performed for each set of tissue- substitutes. Based on that, the CT number and proton stopping power (relative to water) were calculated for different sets of biological tissues and tissue-substitutes. Results The figure shows different SECT calibration curves. The disconnected symbols correspond to experimental SECT calibration curves based on tissue-substitutes, i.e. HU are measured. The connected symbols correspond to stoichiometric SECT calibration curves based on biological substitutes, i.e. the HU of biological tissues are calculated using the stoichiometric coefficients obtained using different calibration phantoms (Gammex, CIRS, Catphan).

were extracted using a B-splines based deformable image registration (DIR), with the full exhale phase as a reference. The quality of DIR is assessed for selected landmarks whose locations were defined manually. A rigid registration was applied to match the coordinates of CT and MR images of the exhale phase. The extracted vector fields from 4DMRI were resampled to obtain the same resolution as CT, and were then applied to the full exhale CT image to achieve the simulated 4DCT-MRI images, as depicted in Figure 1. The motion vector fields of the 4DMRI were interpolated in time to match the phases of the 4DCT. The extent of motion extracted from 4DCT and 4DMRI was analysed by comparing the motion vectors in the region of the tumour (as shown in Figure 2a).

Results Figure 2a) shows a comparison of the synthetic 4DCT-MRI for two phases, with respect to full exhale and the corresponding phase of the original 4DCT. The differences between the real and the simulated CT are rather limited, in comparison to the initial differences between the phases and the full exhale CT. The extent of motion of the selected region (mean +/- standard deviation) for the 4DCT (red circles) and the resampled 4DMRI (blue crosses) is shown in Figure 2b), where the differences of the vector fields is given by the black dashed line. Although the reconstruction principle for the time resolved images is fundamentally different for 4DCT and 4DMRI, very similar motion vector fields could be obtained when phantom motion was well controlled. An error of 2-4mm was found for the landmarks, which is in the order of the voxel size of the MRI.

It was found that, contrary to common belief, the stoichiometric fit depends on the elemental composition of the tissue-substitutes used in the calibration, leading to CT calibration curves which differ up to 15% in predicting the relative stopping power (RSP) of bone tissues. For a given RSP, only Gammex tissue-substitutes were found to reproduce the CT numbers of biological tissues within the whole energy range relevant to computed tomography. Finally, it was found that, for Gammex tissue-substitutes, the CT calibration curve resulting from the stoichiometric method agrees with that obtained by simple interpolation of experimental data. Conclusion The stoichiometric method for SECT calibration seems to depend on the tissue-substitutes used in the calibration, which may be regarded as an additional source of systematic uncertainty in proton range for bone tissues. Furthermore, Gammex tissue-substitutes appear to be a good representative of biological tissues within the energy range relevant to computed tomography—making in this case the stoichiometric method unnecessary. EP-2078 Experimental validation of a synthetic 4DCT- MRI approach using an anthropomorphic breathing phantom M. Krieger 1 , K. Klucznik 2 , C. Emma 1 , M. Peroni 1 , O. Bieri 3 , Z. Celicanin 3 , D.C. Weber 1 , A.J. Lomax 1 , Y. Zhang 1 1 Paul Scherrer Institute, Centre for Proton Therapy, Villigen PSI, Switzerland 2 ETH Zürich, Department of Physics, Zürich, Switzerland 3 University Hospital of Basel, Department of Radiology, Basel, Switzerland Purpose or Objective To validate the synthetic 4DCT-MRI approach for lung proton treatment planning using an anthropomorphic phantom under well controlled conditions. Material and Methods Temporally resolved CT and MR images of an anthropomorphic moving phantom were acquired. The pressure applied to the phantom followed a sin 4 curve to simulate human respiration, which resulted in mean motion amplitude of 7mm for the embedded tumour. The motion vector fields of the resulting 4DCT and 4DMRI

Conclusion The idea of creating a synthetic 4DCT-MRI by warping a static CT with the extracted motion fields from 4DMR was shown to be feasible for the lung anatomy. The good validation results under the controlled conditions show that it is a reliable approach to be used for patients and volunteers, making way for the testing of lung treatment plans under different breathing scenarios without delivering additional imaging dose to the patient. EP-2079 Feasibility of MR-only rectum radiotherapy using a commercial prostate sCT generation solution M. Maspero 1,2 , M.D. Tyyger 1,3 , P.R. Seevinck 2 , R.H.N. Tijssen 1 , M.P.W. Intven 1 , C.A.T. Van den Berg 1,2