Upload
gladys-shelton
View
219
Download
0
Tags:
Embed Size (px)
Citation preview
New Detectors for PET Imaging of Small, Awake AnimalsNew Detectors for PET Imaging of Small, Awake Animals
• RatCAP• Non-invasive wrist monitor• Beta microprobe
Craig WoodyInstrumentation Seminar
March 12, 2003
2
Positron Emission TomographyPositron Emission Tomography
g
e+
g
e+
e-
Radioisotope
PET ScannerPET detects coincident 511 keV gamma rays
Isotopes emit positrons with energies of a few hundred keV
range ~ few mm
3
Positron Emitting RadiotracersPositron Emitting Radiotracers
Organic molecules are labeled with positron emitting isotopes
and used as tracers
Tracers can be used to study neurotransmitter activity in the
brain
4
Neurotransmitter Activity in the Brain Neurotransmitter Activity in the Brain
DA
DA
DA
DA DA
MAO A
DA
DA
signal
DA
DA
DA
DADADA DA
DA
DA DA
DA
DA
DA
Drugs like cocaine can block the re-uptake sites for neurotransmitters like dopamine which upsets the normal equilibrium and can cause effects of addiction
11C-Cocaine
5
The Problem The Problem
• Cannot study animal behavior while under anesthesia • Anesthesia can greatly depress brain functions and affect the neurochemistry that one is trying to study
One wants to use PET to study the neurophysiological activity and behavior in laboratory animals in order to understand and treat these effects in humans.However, animals currently need to be anesthetized during PET imaging.
6
Effects of AnesthesiaEffects of Anesthesia
The effect of anesthesia on the uptake of β-CFT on dopamine transporters in the monkey brain.
H.Hideo et.al., Synapse 42 (2001) 273-280
Reduction in glucose metabolism with isoflurane in humans Similar effects are seen in the rat
D.B. Stout et al, UCLA 1998
7
RatCAP: Rat Concious Animal PETRatCAP: Rat Concious Animal PET
Mockup of the portable ring on the head of a rat
A septa-less, full-ring tomograph with a diameter of 4 cm and an axial extent of 2 cm, suspended by a tether, which would allow nearly free movement of the awake animal
8
The Collaboration The Collaboration
Part of a larger project for Imaging The Awake Animalalso involving motion tracking in both PET and MRI
9
The PeopleThe People
PhysicsCraig Woody
Sean StollBill Lenz
Mike Lenz
InstrumentationVeljko RadekaPaul O’Connor
Jean-Francois PratteBo Yu
SUNY Stony BrookSepideh ShokouhiAzael Villanueva
Aarti Kriplani
MedicalPaul Vaska
Nora Volkow
ChemistryDavid Schlyer
Richard FerrieriMike SchuellerJoanna Fowler
10
Design RequirementsDesign Requirements
• The tomograph ring must be light enough to be supported by the rat and allow reasonable freedom of movement
• Light weight detectors (~ 125 g total weight)• Light weight electronics with low power dissipation New custom ASIC • High data rates and large singles background• Small field of view and large parallax effects• Limited sampling due to space and weight requirements• Must be rugged enough withstand activity of the rat
11
Support StructureSupport Structure
Similar support structures are used in microdialysis experiments
“Ratturn Bowl”Prototype support tether
12
Tomograph RingTomograph Ring
2x2 mm2 LSO crystalsread out with APD arrays
Ring containing 12 block detectors Up to two layers of 5 mm deep crystals with APDs and integrated readout electronics
13
SensitivitySensitivity
Comparison of sensitivity with other small animal PET scanners
Relative Scanner Sensitivities
0
5
10
15
20
25
-4 -2 0 2 4
Axi
al
Acc
ep
tan
ce
An
gle
(d
eg)
ConcordeUCLA
BNL nanoPET
rat brain region
BNL RatCAP4 cm Ø x 2 cm
Concord mPET R415 cm Ø x 8 cm
UCLA mPET 17 cm Ø x 2 cm
• Intrinsic detector sensitivities are essentially the same
• Coincidence sensitivities are proportional to axial acceptance
centimeters
P.Vaska
14
Parallax EffectsParallax Effects
Due to the small diameter of the ring, the parallax error is a major factor in determining the spatial resolution
2 cm
15
Spatial Resolution and Field of ViewSpatial Resolution and Field of View
Calculated Image Spatial Resolution of BNL nanoPET
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0 2 4 6 8 10 12 14 16 18 20
Radial Position (mm)
Rad
ial F
WH
M S
pat
ial
Res
olu
tio
n (
mm
) 10mm crystals
5mm crystals
~edge of rat brain
UCLA microPET
Resolution at theedge of the rat brain
Single layer - 10 mm ~ 2.5 mmDouble layer - 5 mm ~ 1.9 mm
First prototype will be a single layer of 5 mm crystalsWill sacrifice sensitivity for improved spatial resolution
Rad
ial S
patia
l Res
olut
ion
FWH
M (m
m)
P.Vaska
16
Inter-Crystal Scatter and Cross TalkInter-Crystal Scatter and Cross Talk
P.Vaska et.al., 2002 MIC, to be published in IEEE TNS
Conclusion: The slight improvement in resolution and increased sensitivity does not justify the
added complexity in the ASIC
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
Energy information in each crystal could be used to recover Compton events and correct for cross-talk between pixels
Collected data with ADC information and compared energy, position, time resolution and sensitivity :• DE (FWHM) : 21% 19% • time & position resolution ~ same• ~25% increase in coincidence sensitivity
17
Light Output and Energy ResolutionLight Output and Energy Resolution
Studying different types of crystal arrays to optimize light output and
energy resolution
CTI white powder reflector
Proteus unbonded 3M reflector
Proteus glued 3M reflector
18
Contributions to Energy ResolutionContributions to Energy Resolution
22222
he
noise
he
he
col
col
sc
sc
NN
N
N
N
N
N
E
E
Nsc = Number of scintillation photons produced
Ncol = Number of photons collected
Ne-h= Number of electron-hole pairs produced in APD
dnoise = Dark current noise S.Stoll
ADC channels
200 300 400 500
Cou
nts
0
1000
2000
3000
4000
5000
6000
137Cs 662 keV g
s/m ~ 7.5%
19
Monte Carlo Study of Light CollectionMonte Carlo Study of Light Collection
OPTICAD Program
Maximum light collection near ends of crystal
Very dependent on geometry and reflector properties S.Shokouhi
80% Lambertian Reflector
Distance from APD (mm)
0 2 4 6 8 10 12E
ffic
ien
cy0.0
0.2
0.4
0.6
0.8
1.0
20
Comparison of Light Output and Energy Resolution of Different CrystalsComparison of Light Output and Energy Resolution of Different Crystals
manufacturer length # meas. pe/MeV resolutionCTI 8mm slotted block 3 2512 0.23
Proteus 5mm glued 2 1402 0.1810mm glued 2 2272 0.17
Proteus 5mm not glued 1 2265 0.1210mm not glued 1 2660 0.13
Average Light Output and Resolution of 4x8 LSO arrays
Conclusion: Crystals wrapped with 3M reflector and air gap give the highest light output and best energy resolution
S.Stoll
3M Reflector
3M Reflector
FWHM
21
Avalanche PhotodiodesAvalanche Photodiodes
4x8 array 1.6 x 1.6 mm2
active pixel areaCT~ 10 pF
Hamamatsu S8550 Improvements in Gain and
Dark Current
Jan ‘02
Jan ‘01
Typical G ~ 50Npe ~ 1200
~ 60K signal electrons
Expected noisein final ASIC
~ 500-600 e’s
22
Gain and Quantum Efficiency VariationGain and Quantum Efficiency Variation
Measured with N2 laser + optical fiber
1.01 1.08 1.14 1.18
1.00 1.00 1.14 1.12
1.00 1.02 1.07 1.04
0.97 0.96 1.02 0.99
0.92 0.94 1.01 0.99
0.94 0.96 1.00 0.98
0.95 1.00 0.94 0.95
0.92 0.93 0.90 0.93
APD 024. ratio of laser signals at "<gain>=50"/gain=1
gain
50 55 60 65 70 75 80
coun
ts
0
2
4
6
8
10
12
14
mean = 66.5sigma = 2.4sig/mean = 3.6%
adc channels
90 100 110 120 130 140 150 160 170
coun
ts
0
2
4
6
8 mean = 134.8sigma = 9.2sig/mean = 6.8%
adc channels
5000 6000 7000 8000 9000 10000 11000 12000
coun
ts
0
2
4
6
8
10
12
14mean = 8964sigma = 578sig/mean = 6.4%
at applied voltage = + 50V (gain = 1)
at applied voltage = + 370V (avg. gain = 50)
3/20/02
Quantum efficiency variation
Gain + QE variation
Channel to channel
gain variation
Channel to channel differences dominated by
quantum efficiency variation
s/m ~ 3.6 %
s/m ~ 6.8 %
s/m ~ 6.4 %
S.Stoll
23
Timing ResolutionTiming Resolution
Timing resolution of detector.standard set-up (APD's 023,024 w/LC3420 disc., Na22 source)
time (ns)
-15 -10 -5 0 5 10 15
0
20
40
60
80
100
120
all channels in coin. trigger ............fwhm: 8.4nscentral 4 channels on each side in coinc. trigger.....fwhm: 4.4ns
DT~ 5 ns fwhm
Presently dominated by noise in electronics setup
S.Stoll
Expect time resolution to improve considerably with new ASIC
Need good time resolution (~ few ns) for coincidence timing to reject singles background
B.Yu
1 ns
t ~ 70 ns
24
Test Setup with Crystal and APD ArraysTest Setup with Crystal and APD Arrays
Two arrays of 4x8 APDs and crystals using hybrid preamps
and shapers and CAMAC DAQ system
Lab 2-93 in Physics Challenge is to put all of this on a chip !
25
Rotatable Stage for Taking Tomographic DataRotatable Stage for Taking Tomographic Data
Source phantoms mounted on rotatable stage to simulate full
tomographic ring
26
Tomographic ImagesTomographic Images
Measured with a ~2 mm diameter 68Ge point source
and two gamma coincidence 52 samples (6.9 degrees)
Spatial resolution is 2.1 mm FWHM
2 mm
S.Shokouhi
27
Comparison with MicroPETComparison with MicroPET
MicroPET resolution measured with the same ~ 2 mm point
source
Spatial Resolution
MicroPET ~ 2.7 mm
RatCAP ~ 2.1 mm
P.Vaska & S.Shokouhi
28
Modeled Reconstructed ImagesModeled Reconstructed Images
Reconstruction Simulations using SimSet
Fully sampled image of four circular point
sources
Image reconstructed using incomplete data set with interpolation
Image reconstructed
using incomplete data set
S.Shokouhi
29
Electronics & Readout SystemElectronics & Readout System
Tomograph ring with detectors and front end readout electronics are mounted to the head of
the rat Tether carries
discriminator pulses, encoded addresses, high and low voltage power to a Data Collection Module which adds time stamp information
30
Tomograph Ring with Readout Electronics Tomograph Ring with Readout Electronics
LSO blocks, APDs and front end readout electronics will be
mounted on a PCB / Multi Chip Module with interconnecting flexstrip
cable
A.Kandasamy
31
So you want me to put my head in here ?….So you want me to put my head in here ?….
A.Kandasamy
32
Electronics + Readout SystemElectronics + Readout System
enc losure
B LO C K 0
B LO C K 1
B LO C K 2
B LO C K n
T D C
A D D RE NC O D E
controls
pow er
LV D S
to co inc . p roc .
LV , HV
tem p senseHV ad jus tthresho ld ad j.anlg . m onito r
9b
24b
330 Mb it/s
P d iss ~ 1W (2 .6 m W /chan) P d iss 3 - 10 W
tim ing c lock
te the r10^7 events /s
on-de tec to r e lec tronics T D C & p rocess ing box No analog information Single ZCD per channel Serial transmission from
on-block electronics to a Data Collection Module (DCM) at the top of the tether
DCM adds time stamp to each event and sends
address and time information to a remote
coincidence processor Individual links from each
block to DCM
DCM
~1.5 watts totalpower on ring
P.O’Connor
33
Custom Front End Readout ChipCustom Front End Readout Chip
A 0 A 1 A 2 A 3 A 4
A 0 A 1 A 2 A 3 A 4
C L O C K
C F D
T IMINGE D G E
C HA NA D D R E S S
E D G E +A D D R E S S
T se rial
T c lk
• Discriminator pulse is encoded to give 5 bit address • Leading edge of encoded serial pulse train gives time information
Custom ASIC in 0.18 μm CMOS gives ~ 4 mW per channel
(~125 mW per chip)
J-F. Pratte
P.O’Connor
34
ASIC ProductionASIC Production
First test chip delivered Feb ‘03 and ready for testing
2nd version submitted Nov ‘02 3rd version to be submitted March
‘03 Final version to be submitted fall
‘03
Preamp circuit is contained within the yellow circle
Final size will be ~ 4.3 x 1.6 mm
J.-F. Pratte
35
Alternative Approach: Light-sharing LSO/APD Alternative Approach: Light-sharing LSO/APD
LSO slab ~ 6 cm Ø, 10mm thick obtained from CTI
Large Area APDsRequires low-noise
Suppliers• Adv Phot, P-E, RMD• 3 RMD 8x8mm now in hand
Currently: setting up to measure energy & transverse spatial resolution
APDsAPDsLSO slab
The depth of interaction is determined directly without
pixelating the crystals
P.Vaska
36
Non-Invasive Wrist MonitorNon-Invasive Wrist Monitor
• A wide range of quantitative PET studies using tracer kinetic modeling demand accurately measured radiotracer concentration in arterial blood as a function of time after injection, known as the Arterial Input Function. • The common method of measuring the input function is the invasive withdrawal of blood from a wrist artery. However, because of its health risks for both patients and hospital personnel, it is not compatible with clinical studies.
• A small ring tomograph similar to the RatCAP can be used to image the artery and measure the input function
37
Wrist Anatomy Wrist Anatomy
For human studies, the input function will be taken from the radial artery in the wrist.
Activity in the surrounding veins produce a significant background which can be rejected using the good spatial resolution of the wrist monitor.
Phantom based on MRI images of the wrist with anatomically correct placement of the vessels
A.Villanueva
38
Planar ImagesPlanar Images
Measured with a 1 mm diameter 68Ge line source
and two gamma coincidence
1 x 1 mm2 pixels
S.Shokouhi, 2002 MIC, submitted to IEEE TNS
39
Simulated Input Function Measurement Simulated Input Function Measurement
0
10
20
30
40
50
60
70
80
0 100 200 300 400 500 600 700 800
Scatter Evaluation
Time (seconds)
Measure activity as an aqueous solution of a 11C - labeled
radioisotope passes between the two block detectors
0
1
2
3
4
5
6
7
8
0 0.5 1 1.5 2 2.5
Measured Input Function
Time (minutes)
Expected Input Function resolution for Wrist
Monitor
A.Villanueva
w/scatter w/o scatter
40
Beta Scintillation MicroprobeBeta Scintillation Microprobe
• Radioisotopes used in PET emit positrons with energies of a few hundred keV which have a range of several mm in blood or tissue. This range is comparable to the spatial resolution obtained in most PET cameras. • These positrons can be detected directly using plastic or crystal scintillators. With crystal scintillators which have high density, the positrons will range out, depositing all of their energy in the crystal and producing a large signal.
• Small scintillation probes can be used to directly measure the radiotracer concentration in the blood or tissue.
C.L.Woody et.al., IEEE Trans. Nucl. Sci NS-49 (2002) 2208-2212.
41
Comparison of LSO vs Plastic ScintillatorComparison of LSO vs Plastic ScintillatorR
ange
(m
m)
0.001
0.01
0.1
1
10
100
1000
Polystyrene
LSO
Energy (MeV)
0.001 0.01 0.1 1 10 100 1000 10000
dE/d
x (M
eV/m
m)
0.1
1
10
100photoelectrons
0 50 100 150 200 250 300 350 400
co
un
ts
1
10
100
1000
10000
LSO - BetasLSO - GammasPlastic - BetasPlastic - Gammas
Response of LSO and plastic scintillation probes to betas (32P) and gamma
rays (137Cs).
Range and energy loss of positrons in LSO and plastic scintillator
42
Sensitivity and Position Resolution Sensitivity and Position Resolution
Distance (mm)
0 1 2 3 4 5 6 7 8 9 10 11
Co
un
ts/m
in
1
10
100
1000
10000 30 mV, b=1.33 mm-1
600 mV, b=2.28 mm-1
Material LSO LSO Plastic
Size 0.3 x 0.5 mm 0.5 x 1.5 mm 1.0x1.0 mmVolume 0.035 mm3 0.295 mm3 0.795 mm3
Sensitivity 10.1 20.7 51.2(Hz/mCi/cc)
Region of sensitivity around the probe can be selected by
adjusting the readout threshold to improve spatial
resolution
43
Probe ConstructionProbe Construction
LSO microprobe consisting of 0.5 mm diameter x 1.5 mm long LSO crystal wrapped with
several layers of white reflecting teflon and covered with polyester shrink tubing.
44
Rat Brain Studies with MicroprobeRat Brain Studies with Microprobe
Elapsed time (minutes)
-20 0 20 40 60 80 100 120 140
Co
un
ts/s
ec (
de
cay
corr
ecte
d)
0.1
1
10
100
0.5 mm x 1.5 mm LSO Microprobe
Uptake of 11C-methylphenidate in the nucleus accumbens region of
a rat brain with LSO probenucleus accumbens
~ 2 mm
45
Input Function Measured with MicroprobeInput Function Measured with Microprobe
Seconds
0 60 120 180
Co
un
ts/s
ec
(d
ec
ay c
orr
ecte
d)
0
10
20
30
40
50
60
ProbeWell counter
LSO microprobe (0.3 mm dia. x 0.5 mm) inserted inside an 18 gauge syringe needle for blood flow study.
Activity of 11C-tyrosine measured in baboon blood flow during a PET scan using a syringe mounted LSO microprobe.
Input function
46
SummarySummary
New detectors are needed for PET studies of the neurological behavior in live, awake animals The RatCap will provide a new tool for carrying out these studies in laboratory rats. Similar small ring tomographs can be used to measure the arterial input function in larger animals and humans Small scintillation microprobes can be used to directly measure the positron activity in blood and tissue in live, awake animals with a spatial resolution comparable to or better than current PET tomographs
47
Backup slidesBackup slides
48
Energy ResolutionEnergy ResolutionEi Si Wi Ei*Wi Si*Wi Si/Ei Wi*(Si/Ei)
1 0.656 0.0844 0.1629 0.1069 0.0138 0.1287 0.02102 0.552 0.1612 0.1468 0.0810 0.0237 0.2919 0.04293 0.602 0.0576 0.1314 0.0791 0.0076 0.0957 0.01264 0.530 0.1037 0.1123 0.0595 0.0116 0.1956 0.02205 0.566 0.0963 0.0990 0.0560 0.0095 0.1702 0.01686 0.452 0.1279 0.0865 0.0391 0.0111 0.2831 0.02457 0.556 0.0764 0.0786 0.0437 0.0060 0.1373 0.01088 0.520 0.0529 0.0676 0.0352 0.0036 0.1018 0.00699 0.600 0.2026 0.0607 0.0364 0.0123 0.3377 0.0205
10 0.650 0.0995 0.0541 0.0352 0.0054 0.1531 0.0083SUM 5.684 1.063 1.000 0.572 0.104 1.895 0.1862AVG 0.568 0.1063 0.1000 0.0572 0.0104 0.1895 0.0186
0.5720.1050.1830.186
Npe + ENF of APDNpe/MeV 2400Npe 1200D(Npe)/Npe 0.029F(=2)*D(Npe)/Npe 0.041
NoiseNe(noise) RMS 3000Ne(signal), G=50 60000Noise/signal 0.050
Npe+Noise 0.065
Measured (662 keV) 0.075
Remainder (Intrinsic + Light Collection) 0.038
Weighted <Ei> =Weighted <Si> =
Weighted <Si>/<Ei> =Avg weighted <Si/Ei> =
49
Neurochemistry 101Neurochemistry 101
MonoAmineOxydase-A (inside neuron)
Both MOA- A,B consume dopamine
MonoAmineOxydase-B (outside neuron)Deprenyl given to increase MOA-B
DA
DA
DA
DA DA
MAO A
DA
DA
signal
DA
DA
DA
DADADA DA
DA
DA DA
DA
DA
DA
MAO B
StriataHigh Dopamine concentration
Raclopride concentrates at D2
receptor sites
Ceretonin (neurrotransmitter)Present in frontal corrtex (Cerebellum)
Vesicules containing dopamine
Dopamine transporters
(re-uptake sites)(Cocaine, MP)
Dopamine receptors
(raclopride)Only few %
receptor sites occupied
Methylphenadate (MP) = Ritalin11C-Cocaine
50
The Human BrainThe Human Brain