Methods for High Resolution PET

Transcription

1 Methods for High Resolution PET Neal Clinthorne Radiology / Nuclear Medicine University of Michigan Ann Arbor

2 PET Basics Ring of Photon Detectors Inject positron emitting radiotracer into patient Tracer localizes according to its metabolic properties Radionuclide decays, emitting β +. β + annihilates with e from tissue, forming back-to-back 511 kev photon pair Photon pairs detected via time coincidence (<5ns) indicate line along which positron annihilated By collecting many (>10 6 ) events, activity distribution can be reconstructed 2

3 PET Brain Images PET tracers use elements of life (N, C, O, F) and can be designed to follow specific metabolic pathways Temporal change in distribution is used to estimate parameters in kinetic models 3

4 FDG PET for Cancer / Oncology Brain Heart Metastases Shown with Red Arrows Normal Uptake in Other Organs Shown in Blue Bladder 4 Many tumors have higher than normal uptake. Image the whole body to find metastases.

5 PET / CT for Anatomic Correlation PET CT Fused PET + CT 5 *Data courtesy of David Townsend, U. Tenn.

6 Small Animal [ 18 F]FLT No tumor control 35 Gy PET 35 Gy total (no tumor control) Mean tumor volume increase (+60%) Decrease in [ 18 F]FLT uptake Increase in [ 18 F]FAZA uptake [ 18 F]FAZA Pre-RT 1 week 2 weeks 50 Gy total (tumor control) Mean tumor volume decrease (-20%) Decrease in [ 18 F]FLT uptake Decrease in [ 18 F]FAZA uptake [ 18 F]FLT Tumor control Gy 417 [ 18 F]FAZA Inhomogeneity of intratumoral uptake Pre-RT 1 week 2 weeks 6

7 PET Cameras Patient port ~60 cm diameter. 15 cm axial coverage (patient bed moved for larger span). 4 5 mm fwhm intrinsic spatial resolution. ~2% solid angle coverage. ~ $2 million dollars or more with CT 7 Images courtesy of GE Medical Systems and Siemens / CTI PET Systems

8 Early PET Detectors The Cyclotron Corporation PCT 4600 ¾ PMT BGO + Parallel Operation Expensive Low spatial resolution 8

9 Counts Profile through Row 2 Y-Ratio Early BGO Block Detectors Uniformly illuminate block. For each event, compute X-Ratio and Y-Ratio, then plot 2-D position. Individual crystals show up as dark regions. Profile shows overlap (i.e. identification not perfect). 9 X-Ratio Many individual crystals multiplexed among a few (4) photomultipliers LSO or LYSO scintillators generally used now

10 Small Animal PET Cameras Position Sensitive Photomultiplier Tube Fiber Optic Bundle LSO Scintillator Crystals (2x2x10 mm) 17 cm Detector Ring Diameter *Image courtesy of Simon Cherry, UC Davis 10 Miniature Version of Standard PET Camera Many crystals coupled to PSPMT

11 Distortions

12 Photon Counting Noise 1M Events 10 M Events Higher efficiency is better! 12

13 Volume ( 3D ) Acquisition Increases Efficiency Inter-Plane Septa No Septa 13 2-D (w/ Septa) + Septa Reduce Scatter + Simple image reconstruction Smaller Solid Angle for Trues 3-D (w/o Septa) No Scatter Suppression More difficult reconstruction + Larger Solid Angle for Trues

14 Annihilation Photon Attenuation P 1 = e!µ"d1 P 2 = e!µ"d2 P = e!µ"(d1+d2) d1 d2 Event detection probability is product of individual photon detection probabilities. Attenuation depends on entire path length through object 14

15 Distortion Caused by Attenuation Measured Sinogram Emission Distribution Linear Attenuation Distribution True Emission Sinogram and Image Reconstruction Distorted Image 15

16 Random Coincidences Simultaneous decays can cause erroneous coincident events ( randoms ) For 3-D PET, randoms can be as high as 50% of image. Random Rate is Rate 1 x Rate 2 x 2 Δt Randoms reduced by narrow coincidence window Δt. Time of flight across tomograph ring requires Δt > 4 ns. 16 Random Rate (Activity Density) 2

17 Compton Scatter Compton scatter in patient produces erroneous coincidence events ~15% of events are scattered in 2-D PET (i.e. if tungsten septa used) ~50% of events are scattered in 3-D Whole Body PET ~30% of events are scattered in 3-D Brain PET Correspondingly small in small animal PET 17

18 Depth-of-Interaction Uncertainty Penetration of 511 kev photons into crystal ring blurs measured position Tangential Projection Blurring worsens as attenuation length increases Can be eliminated by measuring depth of interaction 18 Radial Projection

19 Resolution Across FOV Point Source Images in 60 cm Ring Diameter Camera 1 cm Near Tomograph Center 14 cm from Tomograph Center Resolution significantly worse at edge of FOV 19

20 Positron Range Before Annihilation Positron Range Distribution Positron Annihilation Point Distribution F-18 C-11 N-13 O-15 Max energy (MeV) Mean energy (MeV) FWHM (mm) FWTM (mm) [by Levin and Hoffman] 20

21 Combined Resolution Effects r tot = rdet + racol + r! + rmot + r 2 rec Detector resolution (3D) Acolinearity (0.5 FWHM 2.2mm / 1000mm) Positron range (<1mm for F-18, ~5mm for Rb-82) Subject motion during scan Additional blurring in reconstruction 21

22 Goals: Submillimeter PET for Mice and Rats 1-2 mm PET for humans in specific regions of interest 22

23 Detour What makes one imaging system better than another? (Especially in light of resolution recovery )

24 Example Coded Apertures Which aperture is best? 3 different apertures same overall counting efficiency 1 Pinhole, 10 pixel FWHM resolution 10 Pinholes, 10 pixel FWHM 100 Pinholes, 1 pixel FWHM 24

25 25 Expected Measurements

26 26 Expected Measurements

27 27 Expected Measurements

28 Reconstructions Original 28 OK, which is best??

29 It Depends! Images were reconstructed at different resolutions Need to decide on desired image resolution to accomplish task as well as tolerable noise Need to compare noise at desired resolution (or viceversa) 29

30 Noise vs. Resolution Results n o i t a i v e D 10 pinholes perform better if PSF widith > their natural resolution is desired n o 1000 i t a 500 i v e 12 D Single pinhole performs better if PSF width > its natural resolution is desired d r a d n a t S g n i t i m i L Operating Resolution (mm FWHM) d r a d n a t S g n i t i m i L Operating Resolution (mm FWHM) 30 Each system has a different noise-resolution tradeoff!

31 Desirable PET Detector Characteristics Circumferential resolution 1mm or less Sufficient depth-of-interaction (DOI) resolution Timing resolution < 5ns FWHM (even less is better) Low deadtime (system singles countrates >10 7 ) Good energy resolution (<15% FWHM) High detection efficiency Ability to resolve multiple hits due to Compton scatter 31

32 Interactions at 511 kev Compton scatter is most prevalent interaction of 511 kev photons in PET Detectors! Material Coherent Compton Photoelectric Attenuation/cm BGO 6% 53% 41% 0.97 LSO 6% 61% 33% 0.88 NaI 5% 77% 18% 0.34 Si <1% >99% <<1% 0.20 The most desirable coincidence interactions are photoelectricphotoelectric Photo-Photo coincidence efficiency less than 17%, 10%, and 4% for BGO, LSO, and NaI, respectively 32

33 Centroid of Energy Distribution BGO LSO NaI Crystal array 16 mm PSPMT 511 kev photons mm 2mm x 2mm x 20mm thick crystals EGS4 Monte Carlo Simulations

34 Does It Affect Performance? - EGS4 Monte Carlo simulations - BGO PET ( 17.6 cm I.D. 16 cm length segmented with 3 mm x 3mm x 20 mm crystals) - Point sources at 0, 3, 6, 9, 12, 15, and 18 mm from center of FOV - Filtered back projection reconstruction True first interaction position Centroid of scattered E distribution 34 DOI uncertainty included in both images

35 High Resolution Small Animal PET Artist s Conception 1st detector (high resolution) 2nd detector (non position sensitive) First Concept

36 Compton PET BGO detector Si-Si Si detector Si-BGO BGO-BGO 36 Very high resolution achievable in small field-of-view

37 High Resolution Imaging Probes Position tracker PET Probe detector FOV Si - BGO BGO - BGO 511 kev photon Scattered photon 37 LOR (position uncertainty) Do not need complete inner detector partial high resolution detector sufficient in many cases Can potentially be used in conjunction with existing PET instruments Probes for head & neck cancer and prostate imaging currently under development

38 Effects of Detector Resolution Already large uncertainty along path of annihilation photons (undone by tomographic reconstruction) Resolution determined primarily by uncertainty transverse to the photon paths R D % ( 1$ #) ( sin "! + cos "! ) + # ( sin "! + cos "! ) D1 1 C1 2 D2 2 C 2 Detectors α (1-α) θ 1 θ 2 R D Annihilation D 1 Photon Path D 2 38

39 Spatial Resolution Detectors of 1mm, 2mm, and 3mm FWHM in coincidence with 6mm FWHM detector H W F 6 m m ( n o i t u l o s e R mm 2 mm 1 Probe resolution = 1mm FWHM Distance from probe face (cm) Spatial resolution improves close to detector with good resolution 39

40 Can It Have Enough Efficiency? Absolute efficiency (%) Si-BGO Si-Si Absolute efficiency (%) BGO-BGO Si-BGO Si-Si Silicon thickness (cm) BGO diameter (cm) Further evaluate a system having the following characteristics Silicon 4cm ID, 7.2cm OD (16 layers of 0.3 mm 0.3 mm 1 mm elements) BGO detector 17.6 cm diameter, 16 cm length, and 2 cm thickness segmented into 3 mm 3 mm 20 mm crystals 40

41 Simulated System and Results 0.3 x 0.3 x 1 mm 3 Silicon Pixels, 16 layers 2 cm Si-Si Sensitivity: 1.0 % *FWHM = 230 µm 1.6 cm 17.6 cm 4 cm 4 cm Si-BGO Sensitivity : 9.0% *FWHM = 790 µm 16 cm 3 x 3 x 20 mm 3 BGO Crystals Lower E Threshold: 350 kev Interaction Selection Method: BGO crystal with Maximum E BGO-BGO Sensitivity: 21.0 % *FWHM = 1.45 mm Image reconstruction: FBP * Does not include acolinearity and positron range

42 Overall Spatial Resolution Si-Si Si-BGO BGO-BGO Spatial Resolution (mm FWHM) Event Geometric Geometric Overall + Acolinearity F-18 C-11 N-13 O-15 Si-Si Si-BGO BGO-BGO

43 Simulated Multiple Disk Sources Object -2-1 y (cm) x (cm) Diameters of Disks : 1, 2, 4, 6, 8, and 10 mm Center of disks : 1 cm from center of FOV

44 2D Images of Simulated Multiple Disk Sources - Filtered back projection (FBP) - BGO crystal with Maximum Energy was used - Images were reconstructed with different system efficiencies Si-Si (100k) Si-BGO (800k) BGO-BGO (1.9M)

45 Combined Reconstruction Using ML - Maximum likelihood Expectation Maximization (ML-EM) - Iteration number = cm x 4 cm Si-Si (160k) + Si-BGO (1.4M) + BGO-BGO (3.1M) 45

46 Can a Small Amount of Hi Res Data Have an Effect on Performance? Low & Med Res Desired resolution = 1mm FWHM Gaussian Low Res data only ) e c n a i r a v ( t r q s mm FWHM Gaussian Low Res Low + Med Med + High All 10 All data / Med & Hi Res data norm(f - g) / norm(f) 46

47 47 Experiments

48 Silicon Pad Detectors for Compton Camera Double-metal Si pad detectors 1.4mm x 1.4mm pads, 16 x 32 array 0.5mm and 1mm thicknesses Full depletion ~180V for 1mm Readout via 4 x Ideas VATA GP-3 ASICs 48

49 VATA Readout ASICs GP-3: µs shaping in slow channel 200 ns peaking time in fast channel Serial, sparse, sparse + adjacent channel readout 49

50 Energy Resolution Am-241 (59.5 kev) Tc-99m (140.5 kev) FWHM = 1.49 kev (2.5 %) FWHM = 1.39 kev (0.99%) Pb Kα1 = kev, Kα2 = kev, Kβ1 = kev, and Compton edge = 49.8 kev

51 Experimental Setup Silicon detector Silicon detector Lead shielding 1mm Tungsten Slit Source turntable Laser

52 Experimental System with BGO BGO detectors flank silicon for energy resolution 52

53 Lines of Response for Point Source Lots of randoms coincidence window was ~2.5 µs LOR Silicon detector 1 F-18 Source Silicon detector 2 Random Coincidences 53

54 Point Source Comparison Compton PET MicroPET R4 Source F-18 in glass capillary tubes F ~1.2 mm 0.4 mm cm ML-EM Image reconstruction (no detector modeling) Si-Si coincidence events only Glass wall 0.2 mm cm MAP Reconstruction

55 Intrinisic Resolution Measurment F mm mm SS_steel wall Needle 25G (ID = mm, OD = 0.5mm, SS_steel wall = mm) Image Resolution = 700 µm FWHM cm

56 Resolution Uniformity cm Source pairs at 5, 10, 15, & 20mm off-axis Sinogram The sources in each pair are clearly separated at appropriate sinogram angles

57 New Pad Detectors 1040 (26 x 40) 1mm x 1mm pads, 1mm thick Co-57 Spectrum Should allow mm FWHM spatial resolution 57

58 Challenges Number of electronics channels (1.25M for simulated system) Packaging Triggering threshold Time resolution Event classification Comparing performance with more conventional PET 58

59 Packaging New stackable hybrid Stacked hybrids, side view 59 Double-sided 14-layer hybrid (FR4) allows 4mm active detector in 5.6mm (~70%) Won t work for 1mm x 1mm 1040 pad detectors!

60 Packaging -- TAB Experiements Al traces 8 µm thick x 17 µm wide 40 µm dielectric inserts Packing fraction >90% 60

61 Triggering Threshold New hybrid routes analog and digital signals to separate cables Trigger threshold can be set as low as 3 kev 5.8 kev emission of Fe-55 clearly seen (source <10 cps) But threshold spread too broad for trimdac alignment on VATA GP-3 61

62 Timing Desired time resolution <10ns FWHM Marginal timing is evident Slower signal generation from events near backplane Large range of pulse-height coupled with leading-edge trigger is biggest issue Large time-walk is the result BGO-Silicon timing spectrum for 511 kev source 62

63 Simple Signal Generation Model 1.4 mm x n-bulk p+ implant holes electrons 1 mm Interaction Locations n+ backplane + Al electrode 0.5 Threshold Simplified Pad Detector x 10-9 Depth dependence of signal evident Non-linear effect with depth Higher bias helps Threshold

64 But It gets worse! Pad size is nearly same as detector thickness Weighting potential depends on x & y in addition to z Result is additional jitter due to unknown 3D interaction location 64

65 T-CAD Simulation / Measurements Energy deposited in Si detector vs. triggering time 65

66 Solutions Immediate VATA GP-7 with 50 ns peaking time in fast channel Low trigger threshold + secondary cut on energy Longer Term Use 2 pad-to-pad 0.5mm thick detectors Detector redesign readout from both sides Sum pad + 8 neighbors for more uniform triggering Complete redesign of readout ASICs 66

67 Event Classification -- Compton Kinematics Compton Kinematics BGO Si E Si Reduction of scatter, random, and misclassified events Si BGO Angular Uncertainty Factors E BGO Doppler Broadening Detector Element Size Energy Resolution Photon Acolinearity Si pad: 0.3 mm x.3 mm x 1 mm BGO crystal: 3 mm x 3 mm x 20 mm

68 Reducing Positron Range Embed PET FOV in strong magnetic field (Raylman, Hammer, etc.) Positrons spiral and range is reduced transverse to B-field vector Not very effective for F-18 positrons Potentially useful for emitters with higher endpoint energies (I-124, Tc-94m, etc.) increasingly being used in small animal imaging ) m m ( e c n a t s i D Tesla Tesla XZ-Plane Tesla XY-Plane Simulated PSF for I-124 at 0T and 9T

69 MRI-Safe Silicon PET Imager and inserted into bore of 7T MRI magnet at OSU System rebuilt using no ferromagnetic materials 69

70 70 Ga-68 Resolution Improvement at 7T

71 Next Steps for Compton PET? Complete BGO / silicon PET test bench Test 1mm detectors with VATA GP-7s Attempt comparison in terms of noise-resolution tradeoff with more conventional PET techniques 71

72 72 Prostate Cancer

73 C-11 Choline PET May be correlated with prostate cancer aggressiveness PET/CT Fusion Image PET / Histoloogy Fusion Promising, however, difficult to detect small tumors even with high uptake 73

74 Internal High Resolution Probe Internal probe External PET ring Geant 4 Monte Carlo simulation of internal prostate probe 74

75 Noise Advantage vs. Resolution a v d A e s i o N s e R - i H

76 Probe Construction and Performance 76 Courtesy of Stan Majewski and James Proffitt -- JLab

77 Acknowledgments UMich Les Rogers Scott Wilderman Li Han Sam Huh Sang-June Park Bob Koeppe Dave Raffel Morand Piert OSU Harris Kagan Klaus Honscheid Don Burdette Eric Cochran Shane Smith CERN Peter Weilhammer Enrico Chesi Alan Rudge IFIC/CSIC (Valencia) Carlos Lacasta Juan Fuster Gabriela Llosa IJS (Ljubljana) Marko Mikuz Gregor Kramberger Andrej Studen Dejan Zontar Gamma-Medica Ideas Einar Nygard Dirk Meier Bjørn Sundahl Sindre Mikkelsen etc. LEPSI (Strasbourg) Wojtek Dulinski LBNL Bill Moses JLAB Stan Majewski James Proffitt 77