- Email: [email protected]
279 COMPARISON OF PROMPT-GAMMA AND POSITRON IMAGING FOR HADRON-THERAPY MONITORING
S145
ICTR-PHE 2012
Results:
Comparison between our results, previous data found in the literature and
Cross-section measurements have been done on the energy range from 18 MeV to 68 MeV.. Figure 1 shows our results for the natTi(p,X)47Sc as stars compared with previous data found in literature. We found an overall good agreement. A comparison with the talys code (Koning A.J., Hilaire S. and Duijvestijn M.C., “TALYS1.0”, Proceedings of the International Conference on Nuclear Data for Science and Technology - ND2007, 2007 Nice France. (http://www.talys.eu/)) has been made. It has been found that the default parameters of the code do not allow a good prediction of the data. By changing some of the parameter values, it is possible to get better calculation results. We have spent some time to found the better set of parameters for this reaction (see dashed line on Figure1) as well as for the 40 different cross sections measured in this experiment in the target foils as well as in the monitor foils. Conclusion: An experimental device has been design and built to measure production cross section at ARRONAX using the stacked foils technique. A first experiment using titanium foils and monitor foils has allowed extracting cross section for 40 different reactions. A comparison has been made with the talys code and a set of parameters determined which allows better calculation results than the default parameters. Measurements are ongoing for the 68Zn(p,x)67Cu and a measurement campaign using deuterons as projectile is foreseen for 2012. ACKNOWLEDGMENTS: The cyclotron ARRONAX is supported by the Regional Council of Pays de la Loire, local authorities, the French government and the European Union. 278 RECONSTRUCTING COMPTON CAMERA IMAGES FOR ION THERAPY MONITORING - CHALLENGES AND APPROACHES FACING THEM S. Schoene1, T. Kormoll, H. Rohling1, W. Enghardt2, F. Fiedler1 1 Helmholtz-Zentrum Dresden-Rossendorf) 2 OncoRay Dresden, Dresden, Germany Radiation therapy by means of protons or heavier ions can improve the performance of radiotherapy for cancer treatment by delivering dose more locally to
tumor tissue. Due to the inherent precision of this irradiation modality a dose deposition monitoring is desirable. It has been shown that positron emission tomography (PET) can be used for that by exploiting the decay of positron emitters which arise from nuclear fragmentations of projectile and target nuclei. This monitoring was implemented and successfully applied to the treatment by carbon ions for more than 440 patients at the pilot facility at GSI Helmholtz Centre for Heavy Ion Research, Darmstadt, Germany. However, it comes with inherent drawbacks, e.g. spatial blurring of the reconstructed dose distribution due to unpredictable metabolically driven transport of the positron emitters within the patient. Alternatively, the imaging of prompt gamma-rays -photons emitted at time and location of the beam-tissue interaction -- is proposed to estimate dose deposition distributions more precisely and probably in real-time. Determined by their origin, these gamma-ray emission distributions have certain properties, i.e. smoothly extended in the order of a decimeter and a continuous energy spectrum up to 10 MeV. A Compton camera is a single photon imaging device measuring position and energy deposition according to incoherent scattering (Compton scattering) of a photon, and its trajectory afterwards. Since the scattering angle is related to the photon energy before and after scattering a conical surface can be spanned in patient space which covers all possible source locations of the gamma-ray. This surface of response is the Compton camera equivalent to the well known line of response in PET. Spanning this surface (which means performing a backprojection and constructing the system matrix) is the major part of the image reconstruction process of Compton camera imaging (CCI) in terms of design and implementation efforts, and computation time. As shown previously, its precision also strongly impacts the quality of the image reconstruction results. Furthermore, in CCI a realistic system matrix (SM) has to take into account the physical and technological characteristics of the camera like limited energy and spatial resolution as well as Compton camera specific features as: (i) the event and image space are of high dimensionality, i.e. due to the energy of the gammarays and multi attribute measurements, (ii) the backprojections are surfaces instead of lines (Anger SPECT, PET) and therefore the SM occupancy is high, and (iii) incompletely absorbed photons may lead to incorrect conclusions on the initial gamma energies and scattering angles and subsequently to misarranged backprojections. These challenges of CCI including their impacts will be discussed. Approaches to overcome them will be proposed and evaluated. This is done by means of measurements using radioactive sources as well as via in-beam tests at the proton beam facility AGOR in at KVI Groningen, Netherlands. Complementary results by means of simulations will be shown. 279 COMPARISON OF PROMPT-GAMMA AND POSITRON IMAGING FOR HADRON-THERAPY MONITORING I. Torres-Espallardo1, J. Gillam1, P. Solevi2, G. Llosa Lacer2, V. Stankova1, J. Barrio1, C. Solaz1, C. Lacasta1, M. Rafecas1 1 IFIC, Univ. Valencia/CSIC, Spain
S146 2
Universidad de Valencia, Spain
The ENVISION project studies and compares different modalities for on-line monitoring of particle tumor therapy. The present work will in particular focus on two techniques: prompt-gamma and positron imaging. During therapy using protons both positrons and single, high energy photons, are produced along the beam path due to nuclear reactions, fragmentation and nuclear desexcitations. Their production through the beam path is correlated with the position of the Bragg peak. Positrons travel a short distance and then undergo annihilation producing two photons, that can be used to provide images related with the positron distribution within the patient if detected in coincidence. Single photons, whose energies are from a few MeVs up to 20 MeV, can scape the patient and can also be used to image the therapeutic beam path along the patient by means of single photon emission tomography or Compton cameras. However, as partialring geometries are required (to allow place for the beam), and statistics (the positron yield is low) are limited, Positron Emission Tomography using Time of Flight (TOF-PET) is the preferred candidate. TOF will increase the signal-to-noise ratio in the image and also it reduces the artifacts due to the limited angular coverage. In the case of the single photons, large mechanical collimators reduce the sensitivity so that Compton collimation is identified as a possible solution. In order to acquire and reconstruct optimal images, for TOF-PET time information is critical, whilst in Compton imaging it is the photon cross-section that is thought to be of greatest importance. At 511 keV Resistive Plate Chambers (RPC) detection provides exquisite timing resolution, whilst for single photons Lanthanum Bromide (LaBr3) provides the best crosssections for detection of high-energy photons. In this study a comparison is drawn between TOF PET using RPC-based detection and a Compton telescope composed of LaBr3. For this work we will use a partial ring RPC-based PET of 80 cm diameter consisting of 6 head sectors at each side. Each sector has a layered structure and the layers are grouped in modules. The number of modules within each sector (from 20 to 60) will be optimized to get the highest sensitivity and keeping a good spatial resolution. Each module will contain 5 glass (ρ = 2.4 g/cm3) layers of 400 μm and 4 gas layers of 300 μm, having a thickness of 3.2 mm. The dimensions of the module are 125x300 mm2 and the sensitive area is 120x300 mm2 . The used gas is C2F4H2 and has a density of 0.004275 g/cm3 . The system was simulated with no spatial resolution, but with time resolution. Several time resolution values were considered 40, 200 and 400 ps. The reconstruction code is based on MLEM and probabilities for the system matrix elements were calculated using ray-tracing base on the Siddon algorithm. We have modified it in order to include time of flight (TOF) information. The simulations were performed using Gate 6.1. In a previous study, we have not simulated a proton or ion beam, but an activity distribution whose geometry is similar to the one produced by a hadron beam. Therefore the source employed to evaluate the performance in terms of time resolution of the RPCsystem consists of a series of line sources (background), 30 cm long, covering the FOV that contains several peaks of 1, 2, 3 and 4 mm long, emulating the increase
ICTR-PHE 2012 of intensity due to the Bragg peak (see figure 1(a)). The diameter of all the line sources was 5 mm. The total activity of the source was set to 1 μCi and the activity concentration of each source was set in order to have a ratio 2:1 peak to background. The geometry of the source and reconstructed images can be seen in figure 1(b). For this preliminary study, we have seen better signal-to-noise ratio and contrast-recovery coefficient, when including TOF. The initial design of the Compton camera is made up of six layers of LaBr3 crystal (dimensions 1x20x20 cm3), separated by 10 cm. The scintillator crystals are monolithic read out by Silicon Photomultipliers. The data have been reconstructed using MLEM code adapted for Compton imaging considering three interaction events. The simulations for the Compton Camera have been done with Geant4.9.3. The simulated source in this case is a proton beam of 160 MeV hitting a PMMA cylindrical target along its axis. The Compton Camera is situated 90o to the beam axis. The emitted photons and the reconstructed images of the simulated can be seen at the bottom of the figure 1. The dashed line corresponds to the theoretical position of the Bragg peak for 160 MeV in PMMA. Both simulated systems and reconstruction software provide images that allow us to identify structures like the ones produced by the hadron beam. The next step will be to compare reconstructed images of proton beams in terms of range in PMMA for various initial energies of the incident projectile. This studies provides a firts investigation of the capabilities of both imaging modalities dedicated to hadron beam monitoring.
Fig. 1. (a) Schema of the simulated source for RPC. (b) Reconstructed image using MLEM with TOF for the RPC-based PET simulations. (c) Intensity of emitted prompt-gamma in PMMA for a 160 MeV proton beam. (d) Reconstructed image of the proton beam using the simulations of Compton Camera. 280 SIMULTANEOUS INTEGRATED BOOST IN CERVIX CANCER: TOO MUCH UNCERTAINTY F. Herrera, E. Delikgoz-Soykut, J.L. Soares-Rodrigues, M. Coskun, L. Porta, A. Joosten, J.F. Germond, L. Heym, R. Moeckli, M. Ozsahin Centre Hospitalier Universitaire Vaudoise, Department of Radiation Oncology. Correspondence: [email protected]
ICTR-PHE 2012
Results:
Comparison between our results, previous data found in the literature and
Cross-section measurements have been done on the energy range from 18 MeV to 68 MeV.. Figure 1 shows our results for the natTi(p,X)47Sc as stars compared with previous data found in literature. We found an overall good agreement. A comparison with the talys code (Koning A.J., Hilaire S. and Duijvestijn M.C., “TALYS1.0”, Proceedings of the International Conference on Nuclear Data for Science and Technology - ND2007, 2007 Nice France. (http://www.talys.eu/)) has been made. It has been found that the default parameters of the code do not allow a good prediction of the data. By changing some of the parameter values, it is possible to get better calculation results. We have spent some time to found the better set of parameters for this reaction (see dashed line on Figure1) as well as for the 40 different cross sections measured in this experiment in the target foils as well as in the monitor foils. Conclusion: An experimental device has been design and built to measure production cross section at ARRONAX using the stacked foils technique. A first experiment using titanium foils and monitor foils has allowed extracting cross section for 40 different reactions. A comparison has been made with the talys code and a set of parameters determined which allows better calculation results than the default parameters. Measurements are ongoing for the 68Zn(p,x)67Cu and a measurement campaign using deuterons as projectile is foreseen for 2012. ACKNOWLEDGMENTS: The cyclotron ARRONAX is supported by the Regional Council of Pays de la Loire, local authorities, the French government and the European Union. 278 RECONSTRUCTING COMPTON CAMERA IMAGES FOR ION THERAPY MONITORING - CHALLENGES AND APPROACHES FACING THEM S. Schoene1, T. Kormoll, H. Rohling1, W. Enghardt2, F. Fiedler1 1 Helmholtz-Zentrum Dresden-Rossendorf) 2 OncoRay Dresden, Dresden, Germany Radiation therapy by means of protons or heavier ions can improve the performance of radiotherapy for cancer treatment by delivering dose more locally to
tumor tissue. Due to the inherent precision of this irradiation modality a dose deposition monitoring is desirable. It has been shown that positron emission tomography (PET) can be used for that by exploiting the decay of positron emitters which arise from nuclear fragmentations of projectile and target nuclei. This monitoring was implemented and successfully applied to the treatment by carbon ions for more than 440 patients at the pilot facility at GSI Helmholtz Centre for Heavy Ion Research, Darmstadt, Germany. However, it comes with inherent drawbacks, e.g. spatial blurring of the reconstructed dose distribution due to unpredictable metabolically driven transport of the positron emitters within the patient. Alternatively, the imaging of prompt gamma-rays -photons emitted at time and location of the beam-tissue interaction -- is proposed to estimate dose deposition distributions more precisely and probably in real-time. Determined by their origin, these gamma-ray emission distributions have certain properties, i.e. smoothly extended in the order of a decimeter and a continuous energy spectrum up to 10 MeV. A Compton camera is a single photon imaging device measuring position and energy deposition according to incoherent scattering (Compton scattering) of a photon, and its trajectory afterwards. Since the scattering angle is related to the photon energy before and after scattering a conical surface can be spanned in patient space which covers all possible source locations of the gamma-ray. This surface of response is the Compton camera equivalent to the well known line of response in PET. Spanning this surface (which means performing a backprojection and constructing the system matrix) is the major part of the image reconstruction process of Compton camera imaging (CCI) in terms of design and implementation efforts, and computation time. As shown previously, its precision also strongly impacts the quality of the image reconstruction results. Furthermore, in CCI a realistic system matrix (SM) has to take into account the physical and technological characteristics of the camera like limited energy and spatial resolution as well as Compton camera specific features as: (i) the event and image space are of high dimensionality, i.e. due to the energy of the gammarays and multi attribute measurements, (ii) the backprojections are surfaces instead of lines (Anger SPECT, PET) and therefore the SM occupancy is high, and (iii) incompletely absorbed photons may lead to incorrect conclusions on the initial gamma energies and scattering angles and subsequently to misarranged backprojections. These challenges of CCI including their impacts will be discussed. Approaches to overcome them will be proposed and evaluated. This is done by means of measurements using radioactive sources as well as via in-beam tests at the proton beam facility AGOR in at KVI Groningen, Netherlands. Complementary results by means of simulations will be shown. 279 COMPARISON OF PROMPT-GAMMA AND POSITRON IMAGING FOR HADRON-THERAPY MONITORING I. Torres-Espallardo1, J. Gillam1, P. Solevi2, G. Llosa Lacer2, V. Stankova1, J. Barrio1, C. Solaz1, C. Lacasta1, M. Rafecas1 1 IFIC, Univ. Valencia/CSIC, Spain
S146 2
Universidad de Valencia, Spain
The ENVISION project studies and compares different modalities for on-line monitoring of particle tumor therapy. The present work will in particular focus on two techniques: prompt-gamma and positron imaging. During therapy using protons both positrons and single, high energy photons, are produced along the beam path due to nuclear reactions, fragmentation and nuclear desexcitations. Their production through the beam path is correlated with the position of the Bragg peak. Positrons travel a short distance and then undergo annihilation producing two photons, that can be used to provide images related with the positron distribution within the patient if detected in coincidence. Single photons, whose energies are from a few MeVs up to 20 MeV, can scape the patient and can also be used to image the therapeutic beam path along the patient by means of single photon emission tomography or Compton cameras. However, as partialring geometries are required (to allow place for the beam), and statistics (the positron yield is low) are limited, Positron Emission Tomography using Time of Flight (TOF-PET) is the preferred candidate. TOF will increase the signal-to-noise ratio in the image and also it reduces the artifacts due to the limited angular coverage. In the case of the single photons, large mechanical collimators reduce the sensitivity so that Compton collimation is identified as a possible solution. In order to acquire and reconstruct optimal images, for TOF-PET time information is critical, whilst in Compton imaging it is the photon cross-section that is thought to be of greatest importance. At 511 keV Resistive Plate Chambers (RPC) detection provides exquisite timing resolution, whilst for single photons Lanthanum Bromide (LaBr3) provides the best crosssections for detection of high-energy photons. In this study a comparison is drawn between TOF PET using RPC-based detection and a Compton telescope composed of LaBr3. For this work we will use a partial ring RPC-based PET of 80 cm diameter consisting of 6 head sectors at each side. Each sector has a layered structure and the layers are grouped in modules. The number of modules within each sector (from 20 to 60) will be optimized to get the highest sensitivity and keeping a good spatial resolution. Each module will contain 5 glass (ρ = 2.4 g/cm3) layers of 400 μm and 4 gas layers of 300 μm, having a thickness of 3.2 mm. The dimensions of the module are 125x300 mm2 and the sensitive area is 120x300 mm2 . The used gas is C2F4H2 and has a density of 0.004275 g/cm3 . The system was simulated with no spatial resolution, but with time resolution. Several time resolution values were considered 40, 200 and 400 ps. The reconstruction code is based on MLEM and probabilities for the system matrix elements were calculated using ray-tracing base on the Siddon algorithm. We have modified it in order to include time of flight (TOF) information. The simulations were performed using Gate 6.1. In a previous study, we have not simulated a proton or ion beam, but an activity distribution whose geometry is similar to the one produced by a hadron beam. Therefore the source employed to evaluate the performance in terms of time resolution of the RPCsystem consists of a series of line sources (background), 30 cm long, covering the FOV that contains several peaks of 1, 2, 3 and 4 mm long, emulating the increase
ICTR-PHE 2012 of intensity due to the Bragg peak (see figure 1(a)). The diameter of all the line sources was 5 mm. The total activity of the source was set to 1 μCi and the activity concentration of each source was set in order to have a ratio 2:1 peak to background. The geometry of the source and reconstructed images can be seen in figure 1(b). For this preliminary study, we have seen better signal-to-noise ratio and contrast-recovery coefficient, when including TOF. The initial design of the Compton camera is made up of six layers of LaBr3 crystal (dimensions 1x20x20 cm3), separated by 10 cm. The scintillator crystals are monolithic read out by Silicon Photomultipliers. The data have been reconstructed using MLEM code adapted for Compton imaging considering three interaction events. The simulations for the Compton Camera have been done with Geant4.9.3. The simulated source in this case is a proton beam of 160 MeV hitting a PMMA cylindrical target along its axis. The Compton Camera is situated 90o to the beam axis. The emitted photons and the reconstructed images of the simulated can be seen at the bottom of the figure 1. The dashed line corresponds to the theoretical position of the Bragg peak for 160 MeV in PMMA. Both simulated systems and reconstruction software provide images that allow us to identify structures like the ones produced by the hadron beam. The next step will be to compare reconstructed images of proton beams in terms of range in PMMA for various initial energies of the incident projectile. This studies provides a firts investigation of the capabilities of both imaging modalities dedicated to hadron beam monitoring.
Fig. 1. (a) Schema of the simulated source for RPC. (b) Reconstructed image using MLEM with TOF for the RPC-based PET simulations. (c) Intensity of emitted prompt-gamma in PMMA for a 160 MeV proton beam. (d) Reconstructed image of the proton beam using the simulations of Compton Camera. 280 SIMULTANEOUS INTEGRATED BOOST IN CERVIX CANCER: TOO MUCH UNCERTAINTY F. Herrera, E. Delikgoz-Soykut, J.L. Soares-Rodrigues, M. Coskun, L. Porta, A. Joosten, J.F. Germond, L. Heym, R. Moeckli, M. Ozsahin Centre Hospitalier Universitaire Vaudoise, Department of Radiation Oncology. Correspondence: [email protected]