1 Medical Imaging Group StudyJanuary 19th, 2016 Staff: Prof David Parker, Dr Garry Tungate and Mr Jack Bishop Acknowledgements Professor Peter Jones designed this GS and has kindly allowed us to use his presentations and other material.
2 Overview Organisation Duration: 10 weeks, starting 19th January 2015.Timetabled hours: Tues 2-6 pm, Thurs 2-6 pm, Fri 4-6 pm. Location: Tues/Thurs Nuclear Lab (PTG-R11/12), Fri LAW-BOARD(220) An extra 100 hours (approx.) required. Items of assessment and deadlines Management Plan (Group Mark) – 5% (Wed 27th Jan) Worksheet (Individual Mark) – 15% (Mon 8th Feb) Project Work (Individual Mark) – 20% (Personal statement: Fri 18th Mar) Project Seminar (Sub-group Mark) – 10% (Thurs 19th Mar) Viva (Individual Mark) – 10% (Tues 22th Mar / TBC) Lab Books (Thurs 24th Mar) Project Report (Individual Mark) – 40% (Thurs 24th Mar)
3 Group Structure All activities are student organisedStaff available on request Nominate a group leader / Project Manager Decide on sub-groups (4-5 students) / Work Package leaders See handbook guidance on Project Management Project Manager is expected to be involved in one (or more) sub-groups Work as a team; help each other; you are not in competition Take your time to decide on sub-groups / Work Package leaders Spend the first week or so scoping out the problem It is important to identify all the tasks and prioritise
4 Medical Imaging Group StudyAn “industry” Group Study Originally introduced as part of an IOP / HE STEM initiative Idea was to develop projects linked with an external industrial partner Experience of working in a professional research environment Our “industry” Partner Medical Physics Department at UHB NHS Foundation Trust Expertise: nuclear medicine (imaging and therapy) This project focuses on potential advances in diagnostic imaging Aims: better images; faster; lower dose to patient Project is of genuine interest in the field of medical imaging Staff: Prof David Parker, Dr Garry Tungate and Mr Jack Bishop Prof Stuart Green (UHB), Dr Michael Wilson (UHB)
5 Medical Imaging How it works Proposed study areasMedical Physics Department will present the problem to you Project work will be carried out here in the Y3 Nuclear Laboratory They act as both the customer and as consultants You will present your results to them in your final seminar They will contribute to your feedback, but not your marks Proposed study areas Imaging with Compton Cameras Potential to achieve high resolution and high sensitivity with reduced patient dose Pixelated sensor technology Evaluate the performance of gamma-ray pixel sensors BGO PET camera detector 4 channel PMT with Plastic scintillator.
6 Medical Imaging Gamma (Anger) Camera Applications Planar ImagingOrgan emitting g-rays Applications Planar Imaging Planar Dynamic Imaging SPECT Imaging Gated SPECT Imaging SPECT = Single Photon Emission Computed Tomography Collimator NaI crystal PMT array Position logic Computer Problem: the collimator determines spatial resolution but limits sensitivity
7 SPECT How it works 360o 180o 0o sinogram
8 Compton Camera ConceptExample in 2-d Compton scattering – energy of scattered photon depends on angle of scatter Compton formula q2 q1 g1 g2 DE E Scattering plane (Compton scattering) Absorption plane (Photoelectric effect)
9 Compton Cameras Advantages Disadvantages Design issuesWide field of view; no need for collimation; Hence lower patient dose No trade off between spatial resolution and sensitivity In principle, 3-dimensional imaging without tomography Disadvantages 3-dimensional case is complicated Requires pixelated detectors – many readout channels Design issues Must work with commonly used radioisotopes? What determines the image resolution? What is the optimal choice of detector materials? What is optimal geometry? Will it work for extended objects? 3 1 2 4 99mTc – 140 keV 18F – 511 keV
10 Scintillator MaterialsSodium Iodide - NaI(Tl) Other materials LaBr3, LaCl3, BaF2, CeBr3, CsI, CaF2, BGO, CdWO4, Lu18Y2SiO5(Ce), Plastic… Considerations Compton cross section is proportional to Z Photoelectric effect cross section is proportional to Z4 and Eg-3 Density [g/cm3] Effective atomic number 50 Hygroscopic yes Wavelength of emission max. [nm] 415 Refractive emission max 1.85 Primary decay time [ns] Light yield [photons/keV] https://www.saint-gobain.co.jp/sites/default/files/download/pdf/Crystal _SGC_Scintillation_Materials_and_Assemblies_Saint-Gobain.pdf
11 Compton Scattering Klein-Nishina Formula 137Cs Eg = 662 keVArbitrary units 137Cs Eg = 662 keV Angle (degrees)
12 Simulation A simulation in ROOT Photoelectric Compton Rayleigh 0.24for(Int_t i=0; i<100000; i++) { Double_t ephoton = egamma; SC = cs->Eval(ephoton,0,"S"); SP = pe->Eval(ephoton,0,"S"); SR = rs->Eval(ephoton,0,"S"); Double_t mu = SC+SP+SR; Double_t x = exp->Exp(1./mu); hDepth->Fill(x); Bool_t absorbed = kFALSE; Bool_t interacted = kFALSE; Int_t nc = 0; while( !absorbed && x < 5.) { interacted = kTRUE; Double_t probability = prob->Rndm()*(SP+SC+SR); if( probability <= SP ) { ephoton = 0; absorbed = kTRUE; } else if( probability <= SP+SC ) { ComptonScatter(&ephoton); hEPhoton->Fill(ephoton); nc += 1; } mu = SC+SP+SR; x += exp->Exp(1./mu); if (interacted) { DetectorResponse(egamma-ephoton); if (absorbed) hNComptonPE->Fill(nc); hNCompton->Fill(nc); Photoelectric TGraphErrors TRandom Compton TH1I 0.24 Rayleigh TRandom 0.03 0.01 TGraphErrors TRandom TH1I
13 Simulation Comparison with data 2” NaI(Tl) – Real data ( N x C ) + PEcounts ~250 x 1000 counts 2” NaI(Tl) – Real data ( N x C ) + PE At 662 keV, the probability of PE is 0.03/0.28 = 10.7% Observed ratio is ~ 50% Conclusion: size of detector matters
14 Possible Work BreakdownTask 1 – Hardware / Software Build a proof-of-principle Compton Camera using components in the nuclear lab Explore image reconstruction for one/two point source(s) in 2/3 dimensions Design and build the necessary analysis and reconstruction software Task 2 – Software Develop a simple Monte Carlo model of a Compton Camera Study optimal detector placement; choice of detectors; size of detectors etc Benchmark against measurements from the proof-of-principle device Task 3 – Hardware / Software Evaluate pixelated BGO Detector Evaluate Plastic scintillator Build Pixelated detector Programming environment C++ / ROOT
15 Multiple images Previous years looked at on axis point sourceLast year saw problems with multiple sources Not yet looked in 3D Need quick results Can you solve all outstanding problems? Choices need to be made.
16 Practical details Energy measurements HV * source Pre-amp AmplifierE PC/MCA (Multi-Channel Analyser) X-ray energy spectrum Proportional counter Tb X-rays – 50 keV, 44 keV Ka Kb
17 Practical details Timing (coincidence) measurementsCould be used as initial setup One detector used as a scatter detector; the other stops the scattered photon Difficult to get energy sum information or add more detectors HV HV * source Pre-amp Pre-amp Amplifier Energy Amplifier Discriminator Time Discriminator Time-to-Amplitude Converter (TAC) Delay Start Stop T PC/MCA (Multi-Channel Analyser)
18 Digital signal processingAnalogue versus digital Peak Sensing ADC Energy Charge Sensitive Preamp Shaping Amplifier Trigger Fast out DISC Logic Unit Position Threshold Timing TDC Scalar Counting CAEN DT5724, 4 channel, 14 bit, 100 MS/s Digitiser Digitiser Energy Charge Sensitive Preamp DPP Timing Readout Interface PC A/D Counting Shape High throughput
19 Safety Radioactive sources ImportantMost -ray sources are sealed; perfectly safe if handled intelligently Open , sources – risk of contamination You should not come into contact with these, but be aware they exist Ask a demonstrator if unsure about anything Keep your distance Do not sit close to sources for long periods Important All sources to be checked out/in by a demonstrator Sources must not leave the laboratory under any circumstances
20 Safety Pb bricks Use sensibly for background shieldingBeware of crushing injuries; do NOT overload bench; Carry one at a time; wash hands after handling High voltage Check detector/documentation – do NOT exceed maximum Do NOT connect/disconnect HV with HV switched on Turn on/off HV gradually – detector may be damaged otherwise Typical values: NaI V; check polarity Detectors Note that detectors are fragile – please handle with care
21 R12 is for our sole use. Can use all equipment and leave it set up.Resources We have use of R11 and R12 R11 shared with PTNR so we should not disturb any experimental equipment, but can use computers. R12 is for our sole use. Can use all equipment and leave it set up. Detectors inch NaI ~2 CeBr3 1 LaCl3 1 LaBr Ge ~2 BGO array of crystals + 4PMTs Plastic build your own PMT cluster of 4
22 Final Comments Challenges Any questions?Output of digitiser is list mode uses lots of memory Need to be willing to write some analysis code ROOT analysis framework is useful for histogramming Need to design experiments that will give you the information you want Lots of discussion / reading / thinking needed Some references on Canvas to get you started … We will be learning about these issues with you Any questions?