1. SPECT imaging with semiconductor detectors
Andy Boston

2. Outline of presentation
What is SPECT?
What detector technology can we consider?
The ProSPECTus project & links to fundamental research
The future prospects

3. What is SPECT?
Functional imaging modality

4. What SPECT Radionuclides?
t1/2=65.94h
t1/2=6.01h
141 keV
910-5%
2.1105y
>99%
stable

5. Tomographic Imaging The sinogram is what we aim to measure
- Measure of intensity as a function of projection, θ and position, r
- Often seen plotted as a 2d grey scale image θ x r y
f(x ,y)
p(r , θ)
Measured result – “Sinogram”
(256 projections, 363 positions per projection)
Note : We measure from 0 to 180°
Underlying source distribution
“Shepp-Logan Phantom”

6. SPECT : Problems/Opportunities
Technical
Collimator Limits Spatial Resolution & Efficiency
Collimator is heavy and bulky
Energy of radioisotope limited to low energy
NaI:Tl Dominant for >40 Years...
MRI  Existing PMTs will not easily operate
Would like to be able to image a larger
fraction of events.
Common radionuclides: 99mTc, 123I, 131I

7. What are the detector requirements?
Ideally would want:
Good energy resolution (Good light yield/charge collection) < few%
High efficiency (High Z)
Position resolution
Timing resolution
Detector materials:
Semiconductors (Si, Ge, CdZnTe)
Scintillators (LaBr3, CsI(Tl), NaI(Tl), BaFl, BGO)

8. Next generation Single Photon Emission Computed Tomography
ProSPECTus
Next generation Single Photon Emission Computed Tomography
Nuclear Physics Group, Dept of Physics, University of Liverpool, Nuclear Physics & Technology Groups, STFC Daresbury Laboratory, MARIARC & Royal Liverpool University NHS Trust

9. ProSPECTus: What is new?
ProSPECTus is a Compton Imager
Radical change  No mechanical collimator
Utilising semiconductor sensors
Segmented technology and PSA and digital electronics (AGATA)
Image resolution 7-10mm  2-3mm
Efficiency factor ~100 larger
Simultaneous SPECT/MRI

10. What’s new? Conventional SPECT Compton camera Source E0
Gamma rays detected by a gamma camera
Inefficient detection method
Incompatible with MRI
Gamma rays detected by a Compton camera
Positions and energies of interactions used to locate the source
Factors that limit the performance of a Compton Imager:
Energy resolution, Detector position resolution, Doppler Broadening

11. System Configuration GEANT4 simulations L. Harkness 1cm 2cm 5cm 141keV
Si(Li) Ge
Total Coincident ~3.49%
SPECT ~ 0.025% (typical value)
Factor of ~140
Event Type %
Single / Single
2.23
Single / Multiple
0.33
Multiple / Single
0.61
Multiple / Multiple
0.04
Not absorbed
0.28

12. HPGe Germanium Excellent energy resolution Medium Z (32)
Lithographic electrode segmentation
Requires cooling to LN2
HPGe growth still presents challenges
Technology drivers: large scale physics projects (AGATA/GRETA/GERDA/MAJORANA)

13. AGATA (Advanced GAmma Tracking Array) 4 -array for Nuclear Physics Experiments at European accelerators providing radioactive and high-intensity stable beams
Main features of AGATA
Efficiency: 43% (M =1) 28% (M =30)
today’s arrays ~10% (gain ~4) 5% (gain ~1000)
Peak/Total: 58% (M=1) 49% (M=30)
today ~55% %
Angular Resolution: ~1º 
FWHM (1 MeV, v/c=50%) ~ 6 keV !!!
today ~40 keV
Rates: 3 MHz (M=1) 300 kHz (M =30)
today MHz kHz
Summary
List few items
Agata very powerful instrument for nuclear spectroscopy
Exciting science programme subject of this meeting
Aim to propose and start realisation of the full AGATA from 2008
180 large volume 36-fold segmented Ge crystals in 60 triple-clusters
Digital electronics and sophisticated Pulse Shape Analysis algorithms allow
Operation of Ge detectors in position sensitive mode  -ray tracking

14. Ingredients of g-Tracking4 1
Identified
interaction points
Reconstruction of tracks
e.g. by evaluation of permutations
of interaction points
Highly segmented
HPGe detectors
(x,y,z,E,t)i g
Pulse Shape Analysis to decompose recorded waves 2 3
What has to be done to perform tracking
What are the ingredients
Detectors highly segmented, encapsulated, cryostats
Electronics Digital
PSA to extract Energy Time and Position
Algorithm development particularly for position
Area where effort is required Thorsten Kroell. See him.
Tracking algorithms, reconstruction of the events for full array
Obtain full energy
Many technical development in all areas.
Digital electronics
to record and process segment signals
reconstructed g-rays

17. Next generation Single Photon Emission Computed Tomography
ProSPECTus
Next generation Single Photon Emission Computed Tomography
Nuclear Physics Group, Dept of Physics, University of Liverpool, Nuclear Physics & Technology Groups, STFC Daresbury Laboratory, MARIARC & Royal Liverpool University NHS Trust

18. The SmartPET DSGSD detectors
Detector Specification
Depletion at -1300V, Operation at -1800V
12 x12 Segmentation, 5mm strip pitch
1mm thick Aluminium entrance window
Warm FET configuration, 300mV/MeV pre-amps
Average energy resolution ~ 1.5keV 122keV

19. Am-241 AC x-y surface intensity distribution
AC AC12
DC12
DC1
The results are presented for 60 keV with 2 minutes of data per position.

20. Pulse Shape Analysis
PSA techniques developed through characterisation measurements
Calibration of variation in detector pulse shape response with position
Real Charge
Image Charge
Parameterisation of these pulse shapes provides increased position sensitivity

21. SmartPET detector depth response
“superpulse” pulse shapes for 137Cs (662 keV) events versus depth
DC signals
AC signals
DC signals
AC signals

22. Image Reconstruction
Sensors have excellent energy & position information.
Uniformity of sensor response
Optimise existing:
Analytical
Iterative
Stochastic
Requirement for GPU acceleration

23. Compton Imaging Use of the SmartPET detectors in
Compton Camera configuration
Typical measurements:
10μCi 152Eu
6 cm from SPET 1
Source rotated
Zero degrees in 15º steps up to 60º
Detector separation
3 – 11cm in 2cm steps
Gates set on energies
2 sources 152Eu and 22Na
at different x and y

24. Compton Imaging
Compton Cones of Response projected into image space

25. Compton Imaging
Compton Cones of Response projected into image space

26. Compton Imaging
Compton Cones of Response projected into image space

27. Compton Imaging
Compton Cones of Response projected into image space

28. Compton Imaging
Compton Cones of Response projected into image space

29. Compton Camera measurements (Ge/Ge)
E = 1408 keV, 30 keV gate
6 cm source to crystal
30 mm crystal to crystal
No PSA (5x5x20)
Iterative reconstruction
FWHM ~ 8mm

30. Multi-nuclide imaging
Compton Imaging
Multi-nuclide imaging
~7º Angular Resolution FWHM, central position
152Eu E = 1408 keV
22Na E = 1274 keV
152Eu
2cm source separation
No PSA (5x5x20)
Cone back projection

31. MRI compatibility & Status
Test existing gamma-ray detector in an MRI scanner
Does the detector cause distortions in the MRI image? No
Does the MRI system degrade the detector performance? In certain positions (which can be minimised)
Encouraging results!
ProSPECTus final construction stage
System in ~6 months

32. What are the next steps? Immediate priorities For the future:
We (almost) have an integrated Compton Gamma camera optimised for <500keV
Demonstrate sensitivity with phantoms
Commence trials including clinical evaluation
For the future:
Consider electron tracking Si scatterer
Possible use of large CZT analyser (requires large wafer material with 1cm thickness)

33. ProSPECTus : The Implication
Patient benefits:
Earlier and more effective diagnosis of tumours (higher probability of effective treatment).
Higher sensitivity offering the scope for shorter imaging time (more patients through one machine per day) or lower doses of radio pharmaceuticals.
Cardiac and brain imaging
Image larger patients
SPECT/MRI:
Functional/Anatomical
Image co-registration

34. Credit STFC Daresbury Laboratory, Daresbury, WA4 4AD, UK
Department of Physics, University of Liverpool, L69 7ZE, UK
MARIARC, University of Liverpool,
RLUH NHS Trust,
UK Industries
Funding agencies STFC, EPSRC, MRC
Many people have made significant contributions
Lots of UK PhD’s and Post Docs
Laura Harkness
University of Liverpool
2010 Shell and Institute of Physics
Very Early career Woman
Physicist of the Year

35. MRI imaging
Preliminary analysis of MRI images acquired with the detector at the entrance of the bore (b) show that the detector does not degrade the MRI performance. FLASH images and TSE MRI images were acquired. Image (a) is when there is no detector present.
FLASH - Fast Low Angle Shot TSE - Turbo Spin Echo

36. Assessing Image Quality
Geant4 energy and position
Experimental Factors
Generate realistic images
(i) 2 point sources
(ii) 5 point sources
(iii) 1 line source
Slice of a projection

37. 5 point sources: 1cm between each, alternating 141keV and 159keV
Point Source Images
5 point sources: 1cm between each, alternating 141keV and 159keV