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