Nuclear Imaging: Emission Tomography II

1
Nuclear ImagingEmission Tomography II

• SPECT peformance
• Positron Emission Tomography (PET)

2
SPECT performance

• Spatial resolution
• X- and Y-magnification factors and multienergy
  spatial registration
• Alignment of projection images to
  axis-of-rotation
• Uniformity
• Camera head tilt

3
Spatial resolution

• Can be measured by acquiring a SPECT study of a
  line source (capillary tube filled with a
  solution of Tc-99m, placed parallel to axis of
  rotation)
• National Electrical Manufacturers Association
  (NEMA) specifies a cylincrical plastic
  water-filled phantom, 22 cm in diameter,
  containing 3 line sources
• FWHM of the line sources are determined from the
  reconstructed transverse images (ramp filter)

4
(No Transcript)
5
Spatial resolution (cont.)

• NEMA spatial resolution measurements are
  primarily determined by the collimator used
• Tangential resolution for peripheral sources (7
  to 8 mm FWHM for LEHR or LEUHR collimators)
  superior to central resolution (9.5 to 12 mm)
• Tangential resolution for peripheral sources
  better than radial resolution (9.4 to 12 mm) for
  peripheral sources

6
Spatial resolution (cont.)

• NEMA measurements not necessarily representative
  of clinical performance
• Studies can be acquired using longer imaging
  times and closer orbits than would be possible in
  a patient
• Patient studies may require use of lower
  resolution (higher efficiency) collimators to
  obtain adequate image statistics
• Filters used for clinical studies have lower
  spatial frequency cutoffs than the ramp filters
  used in NEMA measurements

7
Comparison with conventional planar scintillation
camera imaging

• In theory, SPECT should produce spatial
  resolution similar to that of planar
  scintillation camera imaging
• In clinical imaging, its resolution is usually
  slightly worse
• Camera head is closer to patient in conventional
  planar imaging than in SPECT
• Short time per view of SPECT may mandate use of
  lower resolution collimator to obtain adequate
  number of counts

8
Comparison (cont.)

• In planar imaging, radioactivity in tissues in
  front of and behind an organ of interest causes a
  reduction in contrast
• Main advantage of SPECT is markedly improved
  contrast and reduced structural noise produced by
  eliminating the activity in overlapping
  structures
• SPECT also offers promise of partial correction
  for effects of attenuation and scattering of
  photons in the patient

9
Magnification factors

• Magnification factors determined from a digital
  image of two point sources placed against the
  cameras collimator
• If X- and Y-magnification factors are unequal,
  the projection images will be distorted in shape,
  as will coronal, sagittal, and oblique images
• Transverse images, however, will not be distorted

10
Multienergy spatial registration

• A measure of the cameras ability to maintain the
  same image magnification, regardless of the
  energies of the photons forming the image
• Important when imaging radionuclides such as
  Ga-67 and In-111 which emit useful photons of
  more than one energy
• Uniformity and axis-of-rotation corrections that
  are determined with one radionuclide will only
  be valid for others if multienergy spatial
  registration is correct

11
COR calibration

• The axis of rotation (AOR) is an imaginary
  reference line about which the head or heads of a
  SPECT camera rotate
• If a radioactive line source were placed on the
  AOR, each projection image would depict a
  vertical straight line near the center of the
  image
• This projection of the AOR into the image is
  called the center of rotation (COR)

12
COR calibration (cont.)

• Ideally, the COR is aligned with the center, in
  the x-direction, of each projection image
• Misalignment may be mechanical
• Camera head may not be exactly centered in the
  gantry
• Misalignment may also be electronic
• May be the same amount in all projection images
  from a single camera head, or may vary with angle
  of the projection image

13
(No Transcript)
14
COR calibration (cont.)

• Misalignment may be corrected by shifting each
  image in the x-direction by the proper number of
  pixels prior to filtered backprojection
• If COR misalignment varies with camera head
  angle, it can only be corrected if computer
  permits angle-by-angle corrections
• Separate assessments of COR correction must be
  made for different collimators (and possibly
  different camera zoom factors and image formats)

15
Uniformity

• Nonuniformities that are not apparent in
  low-count daily uniformity studies can cause
  significant artifacts in SPECT
• Artifact appears in transverse images as a ring
  centered about the AOR
• High-count uniformity images used to determine
  pixel correction factors
• At least 30 million counts for 64 x 64 images
• At least 120 million counts for 128 x 128 images
• Collected every 1 or 2 weeks separate images for
  each camera head

16
(No Transcript)
17
Camera head tilt

• Camera head or heads must be exactly parallel to
  the AOR
• If not, loss of spatial resolution and contrast
  results from out-of-slice activity being
  backprojected into each transverse image slice
• Loss will be less toward the center of the image
  and greatest toward the edge of the image
• Can assess using a point source in cameras FOV,
  centered in the axial (y) direction, but near the
  edge in the transverse (x) direction
• If there is head tilt, position of point source
  will vary in y-direction from image to image

18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
Positron emission tomography

• PET generates images depicting the distributions
  of positron-emitting nuclides in patients
• In the typical scanner, several rings of
  detectors surround the patient
• PET scanners use annihilation coincidence
  detection (ACD) instead of collimation to obtain
  projections of the activity distribution in the
  subject

22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
Annihilation coincidence detection

• Positrons emitted in matter lose most of their
  kinetic energy by causing ionization and
  excitation
• When a positron has lost most of its kinetic
  energy, it interacts with an electron by
  annihilation
• The entire mass of the electron-positron pair is
  converted into two 511-keV photons, which are
  emitted in nearly opposite directions

27
(No Transcript)
28
ACD (cont.)

• If both annihilation photons interact with
  detectors, the annihilation occurred close to the
  line connecting the two interactions
• Circuitry within the scanner identifies
  interactions occurring at nearly the same time, a
  process called annihilation coincidence detection
• Circuitry of the scanner then determines the line
  in space connecting the locations of the two
  detector interactions

29
ACD (cont.)

• ACD establishes the trajectories of the detected
  photons, a function performed by collimation in
  SPECT systems
• Much less wasteful of photons than collimation
• Avoids degradation of spatial resolution with
  distance from the detector that occurs when
  collimation is used to form projection images

30
True, random, and scatter coincidences

• A true coincidence is the simultaneous
  interaction of emissions resulting from a single
  nuclear transformation
• A random coincidence, which mimics a true
  coincidence, occurs when emissions from different
  nuclear transformations interact simultaneously
  with the detectors
• A scatter coincidence occurs when one or both of
  the photons from a single annihilation are
  scattered, but both are detected

31
(No Transcript)
32
Design of a PET scanner

• Scintillation crystals coupled to PMTs are used
  as detectors in PET
• Signals from PMTs are processed in pulse mode to
  create signals identifying the position,
  deposited energy, and time of each interaction
• Energy signal is used for energy discrimination
  to reduce mispositioned events due to scatter and
  the time signal is used for coincidence detection

33
Design (cont.)

• Early PET scanners coupled each scintillation
  crystal to a single PMT
• Size of individual crystal largely determined
  spatial resolution of the system
• Modern designs couple larger crystals to more
  than one PMT
• Relative magnitudes of the signals from the PMTs
  coupled to a single crystal used to determine the
  position of the interaction in the crystal

34
(No Transcript)
35
Scintillation materials

• Material must emit light very promptly to permit
  true coincident interactions to be distinguished
  from random coincidences and to minimize
  dead-time count losses at high interaction rates
• Must have high linear attenuation coefficient for
  511-keV photons in order to maximize counting
  efficiency

36
Materials (cont.)

• Most PET systems use crystals of bismuth
  germanate (Bi4Ge3O12, abbreviated BGO)
• Light output 12 to 14 that of NaI(Tl), but
  greater density and average atomic number give it
  higher efficiency at detecting 511-keV photons
• Other inorganic scintillators being investigated
  lutetium oxyorthosilicate and gadolinium
  oxyorthosilicate faster light emission than BGO
  produces better performance at high interaction
  rates

37
(No Transcript)
38
Energy signals

• Energy signals sent to energy discrimination
  circuits which can reject events in which the
  deposited energy differs significantly from 511
  keV to reduce effect of photon scatter in patient
• Energy window may be adjusted to include part of
  the Compton continuum, increasing sensitivity but
  also increasing the number of scattered photons
  detected

39
Time signals

• Time signals of interactions not rejected by the
  energy discrimination circuits are used for
  coincidence detection
• When a coincidence is detected, the circuitry or
  computer of the scanner determines a line in
  space connecting the two interactions
• PET system collects data for all projections
  simultaneously
• Projection data used to produce transverse images
  of the radionuclide distribution as in x-ray CT
  or SPECT

40
(No Transcript)