GATE Monte Carlo simulation platform Lydia MAIGNE Laboratoire de Physique Corpusculaire, Clermont-Fd CNRS/IN2P3 maigne@clermont.in2p3.fr

Medical physics simulations PET camera Ocular brachytherapy treatment Radiotherapy OpenGATE collaboration http://www.opengatecollaboration.org

The birth of GATE July 2001 Concept proposed by C. Morel at the « Monte Carlo in nuclear medicine workshop » organised by I.Buvat in Paris Autumn 2001 First specification documents Initiation of developments December 2001 First collaboration meeting Participants : University of Lausanne, University of Geneva, LPC Clermont-Ferrand IN2P3 May 2002 OpenGATE public meeting May 2004: First Gate release May 2009: Gate version 4.0.0 June 2009: GATE version 5.0.0 for nuclear medicine and radiotherapy June 2010: GATE version 6.0.0 with improved RT features and CT imaging March 2010: GATE version 6.1.0 (new hadronic processes)

The OpenGATE collaboration Current status 20 groups Fields: SPECT, PET, µPET, radiotherapy, hadrontherapy All members contribute According to their interest and know-how OpenGATE meetings: 2/3 per year Coordination Irène BUVAT, spokesman Sébastien JAN, technical coordinator Jan S et al, GATE V6: a major enhancement of the GATE simulation platform enabling modelling of CT and radiotherapy. Phys. Med. Biol. 56 (2011) 881-901 CRUMP INSTITUTE FOR MOLECULA IMAGING

How it works Based on Geant4: version 4.9.3 & 4.9.4 C++ object oriented langage Reliable cross sections Core layer: C++ classes close to the G4 kernel GATE development: C++ classes derived from the core layer classes modelisation of detectors, sources, patient movement (detector, patient) time-dependent processes (radioactive decay, movement management, biological kinetics) Ease of use Command scripts to define all the parameters of the simulation (geometry construction, specification of physical processes, sources) User interface Application layer Core layer Geant4

GATE architecture GATE 6.1 Geometry/ General/ Physics/ Digit_hits/ GateClock,GateHounsfield, GateIDFilter, GateParticleFilter, GatePrimaryGeneratorAction, GateRunManager.... CLHEP 2.0.4.2 Geometry/ GateBox, GateCylinder... GateGeometryVoxel, GateImageNestedParametrisation, GateImageRegionalized, GateCPETSystem.cc, GateCylindricalPETSystem.cc, GatePETScannerSystem.cc, GateScannerSystem.cc, ROOT 5.24 Physics/ GateBremsstrahlung, GateElectronionisation, GateHadron..., GateCompton..., GatePhotoelectric..., GateRayleigh, GatePositron.... GEANT4.9.3 & GEANT4 9.4 Digit_hits/ GateActor, GateCrystalHit, GateCrystalSD, GateDoseActor, GateKillActor, GateToAscii, GateToRoot, GateToInterfile, GateToSinogram... GEANT4 Data G4EMLOW6.2 G4NDL-3.13 PhotonEvaporation-2.0 G4RadioactiveDecay-3.2 G4ABLA-3.0 ELASTIC-1.1

Physics processes EM processes Standard and low energy EM processes available Photoelectric, compton scattering, Rayleigh scattering, pair production Ionization (for e-, e+, hadrons, ions..) Bremsstrahlung, positron and e- anihilation, single and multiple scattering Hadronic processes Elastic scattering Inelastic process for proton Inelastic process for ion Pions Neutrons Radiative capture Inelastic scattering Fission Particle decay Radioactive decay

GATE for medical imaging IMNC-IN2P3 (CNRS UMR 8165) Simulation of the MOBY phantom in a PET/CT ImXgam prototype Simulation of a FDG PET of a whole body scan as aquired on a GEMINI GXL PET scanner XCAT simulation Clinical scan (PET, SPECT, CT)

Systems for scans Scanner SPECTHead PET-ECAT scan CylindricalPET CPET CTscan CTscanner scanner SPECThead ecat cylindricalPET CPET module level1 crystal block rsector crystal cluster level2 pixel crystal module pixel level3 crystal level4 layer0 level5 layer1

PET-MR camera: Key Jo Hong, Yong Choi

Fields of research in nuclear medicine Caracterization of detector efficiency Caracterization of medical scans and small animal scans Molecular imaging Quantification Dosimetry (low energy) Computing time efficiency

GATE for radiation therapy All linac geometry possible Advanced features with dedicated scorer (actors) doseDistributionActor, KillActor, PhaseSpaceActor Variance reduction technique Voxelized phantom/patient Output = Images in ROOT or Analyze format of: The dose distribution The statitistical uncertainty The number of hits in each voxels…

Variance reduction technique The weight of secondary particles is recalculated in function of the number of secondaries generated. User can also define filters to increase the efficiency of these techniques. Splitting to split 100 times the electron bremsstrahlung photon Russian roulette Russian roulette is played on secondary particles. The survival probability is 1/N and the weight of each secondary is N. Selective splitting and russian roulette user can add selections criteria on the incident (primary) or secondary particles. The selection is done with filters

Regionalized Volume (RV) CT images are segmented into multiple homogeneous regions (with no density gradient greater than a user specified threshold) A first image of homogeneous regions is created : RV image A voxelized image describing the patient is created and associated to the RV image The particles are tracked in the RV image, the dose scored in the voxelized image Less boundaries, less steps, gain of CPU time D. Sarrut et al. Med Phys 35, April 2008 Representation with PV GATE example : 1 million 6MV photons Field 5x5 cm² one CPU 2.33 GHz NPV RV 40 materials 13.5 min 5.1 min 7 materials 2.5 min Representation with RV CT image

Fields of research in radiation therapy Validation Dosimetry (10 keV to 20 MeV) Standard and Low energy G4 packages Linac modelling + treatment plans New techniques: IMRT, IGRT, cyberknife, pencil beam Computing time efficiency (grid): GateLab Radiobiology Bacteria survivals (E. Coli), human cell survivals DNA repare

GATE for hadron therapy In France 2 projects: ARCHADE: European centre for R&D Located in Caen Partnership with IBA company Cyclotron for acceleration of protons and carbon ions Research in physics and radiobiology and then patients ETOILE: treatment of patients (2015) Located in Lyon Collaborations with IBA company on protontherapy applications

Fields of research in hadron therapy Online beam imaging to control the dose during the treatment Novel in beam PET, in vivo tumour response imaging Innovative gantry designs Detectors & electronics: Silicium PM, MCPPMT Dosimetry Validation of physic processes Validation of treatment plans Radiobiology Cell survival, DNA repare GSI Darmstat European NoVel Imaging Systems for ION therapy

GEANT4/GATE and Grid tutorial – Seoul July 2010 ~50 participants NCC, Samsung Medical Center… Strong interest by the medical physics community of Seoul for radiation therapy and medical imaging applications Involvment in new collaborations Protontherapy applications (treatment plans, validation): future collaboration with NCC Radiobiology GEANT4-DNA & GATE: multi-scales approach

Next step: Multi-scale simulations GATE-G4DNA Nano scale: LPC-CENBG (G4 DNA) Modelling the DNA, DNA strand breaks, DNA repair, validation Trung Quang PHAM (2011-2014) Micro scale: LPC-LIMOS Cell survival after radiation modelling Benjamin Louvet (2010-2013) Macro scale: GATE collaboration Dose to human tissue, organs GATE collaboration Yann Perrot (2009-2012) 3 PhD students in our group working on the 3 levels of simulation