1. What are PET basics?

2. The basic principle of PET
1. Positron-emitting tracer is injected into the body
2. Emitted positrons (+) travel 1 – 3 mm
3. Positrons collide with electrons (-) causing an “annihilation”
4. Annihilation emits energy in the form of two 511keV energy gamma rays at ~180 degrees
5. Gamma rays are detected by opposing detectors
6. Energy discrimination (an “energy window”) is used to ensure that each gamma is ~511 keV
7. Timing discrimination (a “coincidence time window”) is used to ensure that each gamma ray comes from the same annihilation, hence ensuring accurate localization of the tracer 5 1 6 4 2 3 7

3. Coincidence

4. Trues One annihilation Detection within coincidence window
E energy window
One annihilation
Detection within coincidence window
Energy within energy window
trues = const * activity

5. Randoms
Two annihilations
Detection within coincidence window
Energy within energy window
Randoms = const * activity * activity

6. Correction of randoms
Randoms are related to the single rate of each detector
Randoms are related to the length of the coincidence window
Randoms can be calculated when the singles for each detector are measured, and the coincidence window for each detector pair is known
Randoms can be measured and corrected in real time for each LOR, using a delayed coincidence window with exactly the same length as the “direct” coincidence window

7. Reduction of randoms Relevant parameters: Coincidence window 12 ns
random coincidences
12 ns
6 ns
4.5 ns (pico 3D)

8. Scatter
One annihilation
Detection within coincidence window
Energy loss due to scatter
But energy still within energy window
Scatter fraction is object dependent!

9. PET event energy spectra
PET events are distributed across a range of energy, not only in the 511 keV range. An energy window is employed to reject scatter.
ENERGY
WINDOW
425 – 650
ENERGY WINDOW
511 keV PHOTONS
LSO
SCATTER
BGO
Counts
350 – 650
100
200
300
400
500
600
700
Energy (keV)

10. Correction of scatter Scatter is related to mu map
Emission Transmission Scatter Corrected
Scatter is related to mu map
Scatter is patient dependent
Scatter needs to be measured for each patient
Scatter can be estimated by phantoms (but a cylindrical phantom may be a good approximation for the brain; everywhere else it is a very poor estimation)
Scatter can be precisely modeled for each patient using the mu map: Watson method

11. Correction of attenuation
Patient absorbs some of the 511 keV photons
Attenuation is patient dependent
mu map has to be measured for each patient
mu map can be measured with external sources
137Cs for estimated mu map
68Ge for precise definition of mu map
X-ray for high statistics and precise mu map

12. Noise Equivalent Countrate (NEC)
Main sources of statistical error in a PET system are randoms and scatter
Comparison to a system that is resistant to randoms and scatter
NEC describes the effective number of counts measured by the PET scanner as a function of the activity in the FOV

13. NEC – clinical performance
INJECTED DOSE RANGE
185 – 740 MBq 5 – 20 mCi
1 hour uptake
Biograph HI-REZ PICO 90 80
Biograph 70 60 50
Noise Equivalent Count Rate [per sec] 2D 40 30 20 10 2
0.1
0.2
0.3
0.4
0.5 4 6 8 10 12 14 16 18 20
Specific Activity kBq/cc [uCi/cc]
*Ring difference and energy window unspecified; for Biograph HI-REZ all measurements are clinical
Source: Carney, et Al., “Regionally dependent count rate performance analysis
of patient data acquired with a PET/CT scanner,” abstract 364, SNM 2003.

14. Sensitivity Septa employed No septa Low efficiency High efficiency
A measure of the number of coincidence events a scanner is able to detect, assuming no dead time. Four to five times improvement with 3D acquisition techniques.
2D acquisition mode
3D acquisition mode
Septa employed
Low efficiency
Higher dose required
Lengthy scan times
Fewer counts per dose (low count rate)
Low scatter
No septa
High efficiency
Lower dose required
Short scan times
Higher counts per dose (high count rate)
High scatter

15. PET•CT Protocol
The typical protocol begins with a CT topogram to identify the scan range.
This is followed by a spiral CT exam of the body part of interest.

16. PET•CT Protocol
The patient is then automatically positioned for the start of the PET exam.
The PET exam is a series of bed positions during which the radioactive emissions are collected.

17. attenuation correction
PET•CT scan protocol
Spiral CT: seconds CT
PET
Survey CT
Recon
Fused PET•CT
FUSION
scatter correction
attenuation correction
WB PET: min
PET CT
PET
Recon

18. Block detector components
169 crystal elements per detector block
4 photomultiplier tubes (PMTs)/detector block
Detector module
PMT
Channeled
scintillation light
Detector block

19. Attenuation artifacts
Conventional CT: 50 cm FOV
Note: arms not fully imaged, hardening at edges of field of view
Emission only PET
Note: arms fully imaged
Attenuation correction PET
Note: artifacts in liver and possible lesion distortion
Reduced image quality
Reduced accuracy
Increased artifacts
Potential diagnostic impact

20. ACPlus™ Attenuation Correction
Conventional attenuation scan
~120 sec scan time
106 counts
FULL FOV
Conventional CT attenuation scan
~10 sec scan time
1012 counts
TRUNCATED FOV
Siemens ACPlus
~10 sec scan time
1012 counts
FULL FOV (NOT TRUNCATED)
Extended 70 cm transverse FOV
Super fast attenuation scanning
Exceptionally high statistics
Unmatched attenuation image quality
Highest accuracy attenuation correction

21. Standard PET: filtered backprojection
DETECTOR ELECTRONICS
GANTRY CROSS SECTION
COINCIDENCE TIMING WINDOW (4.5 nsec)

22. Standard PET: filtered backprojection
DETECTOR ELECTRONICS
GANTRY CROSS SECTION
COINCIDENCE TIMING WINDOW (4.5 nsec)

23. Time of flight T, TIME DIFFERENCE OF DETECTION DETECTOR ELECTRONICS
CONVENTIONAL
TOF
COINCIDENCE TIMING WINDOW (4.5 nsec)
Source: Conti, et al., IEEE 2004

24. Complex schematic of a PET•CT