SPECT: Single Photon Emission Computed Tomography

1
SPECTSingle Photon Emission Computed Tomography

• Cristina Rentas
• November 24, 2009
• BE 581

2
SPECT Design and Principles

• Transverse images created
• Depict x or gamma rays emitting nuclides in
  patients
• Images acquired from arc of 180 or 360
• Design
• Revolving scintillation camera
• Reconstruction by digital computer
• Filtered Backprojection (like CT system)
• Iterative Reconstruction Methods

3
Picture of SPECT System
4
Image Acquisition

• Camera head revolves around patient acquiring
  images from evenly spaced angles
• Continuous vs. Step and Shoot Acquisition
• Attenuation greatly reduces of photons from
  activity in half of the patient opposite the
  camera
• Noncardiac studies 360 and cardiac 180
• 180 will give superior contrast, the other 180
  will have poor resolution due to greater distance
  and attenuation

5
Image Acquisition (con.)

• Pixel Format either 64² or 128²
• Brain SPECT- smaller radius used then in body
  SPECT, therefore higher spatial resolution

6
Transverse Image Reconstruction 1

• Filtered Backprojection
• Projection images are mathematically filtered
• Then, simple backprojection is performed of the
  row, corresponding to the transverse image.
• Ex 5th row of image 5th transverse image
• Based on the assumption that projection images
  are perfect projections of 3D object

7
Transverse Image Reconstruction 2

• Iterative Reconstruction
• Initial activity distribution in patient is
  assumed
• Projection images are calculated from the assumed
  distribution using known imaging characteristics
  of the scintillation camera
• Calculated images compared with actual, and based
  on this activity distribution is adjusted
• Repeated until calculated images approximate
  actual images
• Computationally less efficient but computers have
  made it feasible

8
Attenuation Correction in SPECT

• More severe in body SPECT than in brain
• Attenuation is not uniform throughout
• Cameras w/ radioactive sources can measure
• Acquire transmission data from projection
• Data reconstructed to provide maps of tissue
  attenuation
• Maps used to provide attenuation-corrected SPECT
  images
• Chang Method (most common)
• Assumes constant attenuation coefficient

9
SPECT Collimator

• Most common High-resolution parallel-hole
  collimator
• Fan-beam collimator
• Parallel in y direction, converging in x
  direction
• Brain SPECT
• Artifacts are created when used for body SPECT

10
Multihead SPECT Cameras

• 2 or 3 scintillation camera heads
• Permits use of higher resolution collimators
• However, imposes technical difficulties
• Electrical and Mechanical stability
• Configurations
• Double head in 180º
• Triple head, fixed angle
• Double head, variable angle

11
Performance

• Spatial Resolution
• Measured by acquiring a SPECT study of a line
  source placed parallel to the axis of rotation
  (ex capillary tube filled w/ solution of
  technetium)
• Deteriorates as the radius of the camera orbit
  increases
• Brain SPECT gt Body SPECT
• Noncircular gt Circular

12
X and Y Magnification Factors

• Also called X and Y gains
• Relate distances in the object being imaged to
  of pixels between corresponding points in
  resultant image
• Values for X and Y should be equal
• Important!
• Uniformity
• Axis of Rotation Corrections

13
Uniformity

• Lack of uniformity can cause significant
  artifacts
• Ring Artifacts images not acquired by all heads
  over 360º
• Primary causes
• Spatial nonlinearities stretch
• Variations in light collection

14
Axis of Rotation

• Imaginary reference line about which the head of
  the SPECT camera rotates
• Misalignment
• Mechanical or Electrical
• Loss of spatial resolution
• Corrected by shifting in x direction prior to
  filtered backprojection

15
Camera Head Tilt

• Camera exactly parallel to angle of rotation
• Left counts collected in pixel backprojected
  onto same slice
• Right counts from activity outside of a
  transverse slice (lighter) to be backprojected
  into transverse slice (darker)

16
Summary PET versus SPECT

• Principle of Projection Data Collection
• PET Annihilation Coincidence Detection
• SPECT Collimation
• Transverse Image Reconstruction
• Both Filtered Backprojection / Iterative
  Methods
• Radionuclides
• PET Positron emitters only
• SPECT Any emitting x-rays, gamma rays, or
  annihilation photons

17
Summary PET versus SPECT

• Cost
• PET 1 Million to 2 million, SPECT 500,000
• Attenuation
• PET More severe, correction possible
• SPECT Less severe, correction sources
  available, utility not yet established
• Spatial Resolution
• PET Relatively constant across transaxial
  image, best at center
• SPECT Depends on collimator and camera orbit

18
References

• The Essential Physics of Medical Imaging by
  Jerrold T. Bushberg, et al.

19
Picture References

• http//www.medical.siemens.com/siemens/en_US/rg_ma
  rcom_FBAs/images/press_room_images/2006/045.06_MD_
  Anderson.jpg
• http//www.clementsclinic.com/public/userfiles/ima
  ges/SPECT/TraumaticBrainInjury.jpg
• http//neutra.web.psi.ch/images/collimator_e.gif