Opengate meeting 2017 – Marseille, France

Slides:



Advertisements
Similar presentations
Computer Simulation for Emission Tomography: Geant4 and GATE Xiao Han Aug
Advertisements

The GATE-LAB system Sorina Camarasu-Pop, Pierre Gueth, Tristan Glatard, Rafael Silva, David Sarrut VIP Workshop December 2012.
Hydrological information systems Svein Taksdal Head of section, Section for Hydroinformatics Hydrology department Norwegian Water Resources and Energy.
Rocchio’s Algorithm 1. Motivation Naïve Bayes is unusual as a learner: – Only one pass through data – Order doesn’t matter 2.
Wavelets Fast Multiresolution Image Querying Jacobs et.al. SIGGRAPH95.
TOMOGRAPHIC IMAGE RECONSTRUCTION FOR PARTIALLY-KNOWN SYSTEMS AND IMAGE SEQUENCES M.S. Thesis Defense :Jovan Brankov.
Simulation System for Emission Tomography (SimSET): using the new block detector feature. Robert L. Harrison, William C. J. Hunter, Steven B. Gillispie,
ELEC 7250 Term Project Presentation Khushboo Sheth Department of Electrical and Computer Engineering Auburn University, Auburn, AL.
Kirov A S, MSKCC Overview of Geant4 Use and Issues in Imaging: Emission Tomography (PET and SPECT) Assen S. Kirov Department of Medical Physics Memorial.
Matrix Row-Column Sampling for the Many-Light Problem Miloš Hašan (Cornell University) Fabio Pellacini (Dartmouth College) Kavita Bala (Cornell University)
Author: Texas Instruments ®, Sitara™ ARM ® Processors Building Blocks for PRU Development Module 2 PRU Firmware Development This session covers how to.
Tissue inhomogeneities in Monte Carlo treatment planning for proton therapy L. Beaulieu 1, M. Bazalova 2,3, C. Furstoss 4, F. Verhaegen 2,5 (1) Centre.
PHYSICS IN NUCLEAR MEDICINE: QUANTITAITVE SPECT AND CLINICAL APPLICATIONS Kathy Willowson Department of Nuclear Medicine, Royal North Shore Hospital University.
Utilising software to enhance your research Eamonn Hynes 5 th November, 2012.
The LiC Detector Toy M. Valentan, M. Regler, R. Frühwirth Austrian Academy of Sciences Institute of High Energy Physics, Vienna InputSimulation ReconstructionOutput.
Design and simulation of micro-SPECT: A small animal imaging system Freek Beekman and Brendan Vastenhouw Section tomographic reconstruction and instrumentation.
0 A Fast Time Incorporating Monte-Carlo Simulation of Wire Chamber Based Small Animal PET Scanners for Detector Scatter Correction M. Dawood 1, Don Vernekohl.
1 08/06/2007 ThIS : a Geant4-based Therapeutic.
Ian C. Smith 1 A portal-based system for quality assurance of radiotherapy treatment plans using Grid-enabled High Performance Computing clusters CR Baker.
1 PARTNER/ENLIGHT meeting – Valencia – June Treatment Planning: Current Challenges & Research in Progress in Lyon Nicolas Freud & David Sarrut.
Enabling Grids for E-sciencE Grid enabling ThIS CAVIAR Hugues BENOIT-CATTIN LRMN.
S. Guatelli, CPS Innovations, Knoxville, 13 th -21 st January Brachytherapy exercise.
S. Guatelli, M.G Pia, INFN Genova G. Cosmo, S. Guatelli, M.G Pia Salamanca, July 2002
Today’s lecture 2-Dimensional indexing Color Format Thread Synchronization within for- loops Shared Memory Tiling Review example programs Using Printf.
ORDER OF CONTENT AND INSTRUCTIONS A program in its simplest form usually contains three kinds of activity:  INPUT : The program asks the user for some.
1 Everyday Statistics in Monte Carlo Shielding Calculations  One Key Statistics: ERROR, and why it can’t tell the whole story  Biased Sampling vs. Random.
O AK R IDGE N ATIONAL LABORATORY U.S. DEPARTMENT OF ENERGY Image Reconstruction of Restraint-Free Small Animals with Parallel and Multipinhole Collimation:
TPS & Simulations within PARTNER D. Bertrand, D. Prieels Valencia, SPAIN 19 JUNE 2009.
Nuclear Medicine: Tomographic Imaging – SPECT, SPECT-CT and PET-CT Katrina Cockburn Nuclear Medicine Physicist.
EGR 115 Introduction to Computing for Engineers Loops and Vectorization – Part 2 Wednesday 15 Oct 2014 EGR 115 Introduction to Computing for Engineers.
SoLID simulation with GEMC Zhiwen Zhao 2015/03/26.
Introduction In positron emission tomography (PET), each line of response (LOR) has a different sensitivity due to the scanner's geometry and detector.
Nuclear Chemistry: The Heart of Matter. 2 Radioisotopes Radioactive decay Radioactive decay – Many isotopes are unstable – Many isotopes are unstable.
Comparison of MC and TP Dosimetry Simulations for Cancer Treatment Raymond Mumme May 12, 2014 Supervisor: Thiago Lima (PhD student) Head Supervisor: Dr.
Improving Matlab Performance CS1114
1 ParadisEO-MOEO for a Bi-objective Flow-Shop Scheduling Problem May 2007 E.-G. Talbi and the ParadisEO team
1 Giuseppe G. Daquino 26 th January 2005 SoFTware Development for Experiments Group Physics Department, CERN Background radiation studies using Geant4.
INFSO-RI Enabling Grids for E-sciencE Grid based telemedicine application for GATE Monte Carlo dosimetric studies using HOPE (Hospital Platform.
Flair development for the MC TPS Wioletta Kozłowska CERN / Medical University of Vienna.
Dae-Hyun Kim Dept. of Biomedical Engineering The Catholic University of Korea Department of Biomedical Engineering Research Institute.
GATE application example using SuperComputer 국립암센터 박 세 준 연구원.
1 National Technical University of Athens School of Electrical and Computer Engineering National Technical University of Athens School of Electrical and.
Segmentation of 3D microPET Images of the Rat Brain by Hybrid GMM and KDE Tai-Been Chen Department of Medical Imaging and Radiological Science,
Cross-frequency coupling
Enabling Grids for E-sciencE LRMN ThIS on the Grid Sorina CAMARASU.
A radioactive isotope is an atom that has a nucleus that is not stable (will change to form a nucleus of a different element). The process by which the.
Source catalog generation Aim: Build the LAT source catalog (1, 3, 5 years) Jean Ballet, CEA SaclaySLAC, 23 May 2005 Four main functions: Find unknown.
3D Reconstruction Based on 3D/2D Registration Longwei Fang 29/1/2016.
2013 Annual Scientific Meeting of the SSRMP
Solar gamma-ray and neutron registration capabilities of the GRIS instrument onboard the International Space Station Yu. A. Trofimov, Yu. D. Kotov, V.
Simulation Tools for Test Beam
CCS Engineering Tools The tools are used help development and debugging of VLT SW control applications This presentation will provide a general view of.
The Use of AMET and Automated Scripts for Model Evaluation
The Lund R3B prototype: In-beam proton tests and simulations
MongoDB Er. Shiva K. Shrestha ME Computer, NCIT
Evaluation of mA Switching Method with Penalized Weighted Least-Square Noise Reduction for Low-dose CT Yunjeong Lee, Hyekyun Chung, and Seungryong Cho.
Releases and developments
Team leader  Luc SIMON, INSERM 1037, Centre de Recherche en Cancérologie de Toulouse, France Medical Physics Department,
OpenGate Technical Meeting May 11th 2017 Clermont-Ferrand
A Brachytherapy Treatment Planning Software Based on Monte Carlo Simulations and Artificial Neural Network Algorithm Amir Moghadam.
Advanced applications of the GLM: Cross-frequency coupling
N. Stoffle University of Houston
Tianfang Li Quantitative Reconstruction for Brain SPECT with Fan-Beam Collimators Nov. 24th, 2003 SPECT system: * Non-uniform attenuation Detector.
Anti-Neutrino Simulations
SoLID simulation with GEMC
The Hadrontherapy Geant4 advanced example
TransCAD Working with Matrices 2019/4/29.
Nuclear Physics 243 NMT Dr. Abdo Mansour
The Geant4 Hadrontherapy Advanced Example
Presentation transcript:

GATE activities @ Creatis, Lyon Opengate meeting 2017 – Marseille, France GATE activities @ Creatis, Lyon D. Sarrut Thursday, 30th November 2017

Projects @ CREATIS, Lyon Gate RT Ion (Medaustron) OpenDose (Toulouse) Still working on VRT Fast simulation combining determinist and stochastic methods Fast low energy photon dose [Baldacci et al. zMedPhys 2015] Faster low energy photon dose [Smekens et al. PMB 2014] Fast X-ray FFD simulation [Arbor et al, PMB 2015] Fast Prompt-Gamma emission [Huisman et al. PMB 2016] Fast SPECT FFD+ARF simulation [Cajgfinger et al. PMB, 2018] neutronTLE Arbor Currently working on 3D SPECT simulation Next project: Compton Camera module For proton dose monitoring For nuclear medicine with high E radioisotope Hermann Maxime Nothing new Still future

Back to basis the DoseActor

DoseActor: back to basis Uncertainty Conventional “batch method”: 10-50 simulations, mean, standard deviation “History by history” estimation [Walters2002] k =pixel /gate/actor/addActor DoseActor da /gate/actor/da/save output/dose_map.mhd /gate/actor/da/attachTo phantom /gate/actor/da/setVoxelSize 2 2 2 mm /gate/actor/da/enableEdep true /gate/actor/da/enableDose true /gate/actor/da/enableUncertaintyDose true i = event

DoseActor: “uncertainty trick” Compute the squared dose, by event i Brute force approach: Take the square after each events, for all voxels Slow ! Approach in Gate [Walters2002]: Store id of the last event in every voxel Only take the square when the event change Need a final loop over all voxels before final output writing Fast. Small memory increase.

What if need output for a region ? CT (Hounsfield Unit) Image of labels

What if need output for a region ? ROI: group of voxels e.g. image of labels For the edep/dose: easy Also sum over k Consider mass per voxel For the uncertainty ? The squared edep in ROI by event i is not available This is not the SD of dk in the region Need a new option !

DoseActor by regions With Maxime Chauvin Macros Label image must have the same size/spacing than phantom Output is a txt file /gate/actor/addActor DoseActor da /gate/actor/da/save output/dose_map.mhd /gate/actor/da/attachTo phantom /gate/actor/da/setVoxelSize 2 2 2 mm /gate/actor/da/enableEdep true /gate/actor/da/enableDose true /gate/actor/da/enableUncertaintyDose true /gate/actor/da/inputDoseByRegions data/label.mhd /gate/actor/da/outputDoseByRegions output/dose_regions.txt

DoseActor by regions Example of output (TODO: add a column for mass) (TODO: change column name) #id vol(mm3) edep(MeV) std_edep sq_edep dose(Gy) std_dose sq_dose n_hit n_event_roi 0 158092650.3 137406.6687 0.4382735288 3641907760 1.155833108e-07 0.4382735288 2.576935695e-15 803 7 1 5565.894125 0 1 0 0 1 0 1 1 2 7060.298575 120754.0866 1 1.458154942e+10 2.660430382e-06 1 7.077889818e-12 27 2 … 237320.5761 4225237.117 0.3122243818 1.756459091e+12 2.612181852e-06 0.3122243818 6.713402446e-13 625 22 52 223184.7299 3093411.653 0.444229193 1.896061827e+12 1.896385818e-06 0.444229193 7.125747973e-13 330 16 Volume in mm3 edep & dose Nb of hits / last event id

How does it work ? Manage a map of label - region statistic std::map<label, regionStat> label: integer regionStat: data structure with: edep, dose, squared values, last_event_id, nb_hits At each step: Get label at current voxel (same index than phantom) Get region statistics thanks to the map Update statistics, use similar “trick” Also need final loop before writing output

What if group of regions ? Want edep+uncertainty in user-defined regions composed of several labels ? /gate/actor/addActor DoseActor da /gate/actor/da/save output/dose_map.mhd /gate/actor/da/attachTo phantom /gate/actor/da/setVoxelSize 2 2 2 mm /gate/actor/da/enableEdep true /gate/actor/da/enableDose true /gate/actor/da/enableUncertaintyDose true /gate/actor/da/inputDoseByRegions data/label.mhd /gate/actor/da/outputDoseByRegions output/dose_regions.txt /gate/actor/da/addRegion 1000: 2 22 37 56

Aggregated region labels Compute a map of vector To each label associate a vector of regions std::map<label, std::vector<regionStat>> Vector contains: The region with unique label All other “aggregated regions” that contain the label Update all regions Need to build this map during initialization (closure) Flexible Timing: DoseActor cost: ~14% additional time DoseActor with region: ~1% additional time

Conclusion on DoseActor GateDoseActor is a very used/old tool introduced in ~2006 … still alive !! Improvements: edep, dose, uncertainty, nb of hits dose_to_water vs dose_to_medium [Grevillot2012,Simon/Fuchs2017] Adaptation to large CT voxels [Deschler, Arbor2015] Adaptation to regions [Chauvin, Sarrut2017] Need finalize code, documentation, example Will be included in next Gate release Code in github Docs in the wiki Example in GateContrib DoseActor

That’s it ! Thanks for your attention Acknowledgements: Laurent Guigues Simon Rit Thibault Frisson Jean-Michel Létang Maxime Chauvin Herman Fuchs Adrien Halty Loic Grevillot Thomas Baudier Luc Simon Jean-Noel Badel Thomas Deschler Thomas Cajgfinger Nicolas Arbor