CT made easy

1
CT made easy

• ?

2
Introduction

• The computed tomograpic (CT) scanner is
  revolutionary.
• It does not use an ordinary image reseptor, but
  instead a well-collimated x-ray beam directed on
  the patient, and the attenuated image response is
  transmitted to a computer.
• The computer reconstructs the image and displays
  it on a monitor.
• The reconstructions are accomplished with
  algorithms adapted for computer processing.

3
History of CT

• The first demonstration of the technique was done
  in 1970, but the components to construct a CT
  scanner was avalible 20 years before this.
• In 1982, Godfrey Houndsfield shared the Nobel
  prize in physics with Alan Cormack. Cormack had
  earlier developed the mathematics used to
  reconstruct CT images.
• No other x-ray equipment development are as
  important as the CT. Some say that MRI and UL are
  as equally important, but they are, however, not
  x-ray equipments.

4
Principles of use

• CT is a tranaxional/transverese image. That means
  that you get axial pictures of the body.
• It is extremly complicated to understand the
  precise method on how the CT makes these images.
  You need to have good knowlegde of physics,
  engineering and computer science.
• The basic principles can be demonstrated if you
  consider the simplest CT systems.

5
Principles of use

• The x-ray source and and detector are connected
  so that they move at the same time.
• When the machine makes one sweep over the patient
  the internal structures of the body attenuate the
  x-ray beam according to their mass density and
  effective atomic number.
• The machine takes several sweeps of the body and
  collects this in a computer, this computer then
  reconstructs the images of the anatomic
  structures in that slice.

6
Principles of use

• The detector signal during each sweep is
  registered in increments with values as high as
  1000. The value og each increment is related to
  the x-ray attenuation coefficient of the total
  path trough the tissue.
• Through the use of simultaneous equations, a
  matrix of values is obtained that represents a
  cross section of anatomy.

7
The different generations of CT

• There are four/five generations of CT scanners.
  The fifth is still under development.
• The first generation translate-rotate
  configuration, pencil- shaped beam, single
  detector, 5-min scan time
• Second generation translate- rotate
  configuration, fan-shaped beam, detector array,
  30 s- scan time

8
The different generations of CT

• Third generation rotate- rotate configuration,
  fan-shaped beam, detector array, 1-s scan time,
  disadvantage ring artefacts
• Fourth generation rotate-stationary
  configuration, fan-shaped beam, detector array,
  1-s scan time

9
Third generation scanner

• In these scanners the x-ray tube and detector
  array are rotated concentrically about the
  patient.
• They can produce an image in one sec.
• It uses a curvilinear array containing many
  detectors and a fan beam.
• The curvilinear detector array results in a
  constant source-to-detector path length, which
  was an advantage for good image reconstruction.
• This also allows for better x-ray beam
  collimation to reduce the effect of scatter
  radiation.

10
Third generation scanner

• One disadvantage is the ring artefacts. They
  occur for several reasons.
• Each detector views a ring of anatomy , so if any
  single detector malfunctions, the aquired signal
  will result in a ring on the reconstructed image.
  
• Software-corrected image reconstruction
  algorithms minimize such artefacts.

11
Fourth generation scanners

• Radiation detection is accomplished through a
  fixed circular array, which contains as many as
  8000 individual elements.
• The fixed detector array does not result in a
  constant beam path from the source to all the
  detectors, but it allowes each detector to be
  calibrated and its signal normalized during a
  scan.
• They are generally without ring artefacts
• Disadvantage patient dose and cost of buying

12
Fifth generation scanners

• Development of CT is always going on. The
  producers wish to make a CT scanner with improved
  image quality at a lesser patient dose.
• Rotate-nutate scanners Toshiba has produced a
  novel extension of the fourth generation. To
  maintain the x-ray source at the same distance
  from the patient as the detectors, the detector
  array nutates, as the x-ray source rotates.

13
Fifth generation scanners

• Electron-beam CT (EBCT) is a fundamentally
  different way to produce CT images. Imatron came
  up with the idea for scanning the heart.
• Currently , EBCT is used to scan all tissues, but
  especially when ultrafast imaging is helpful.
• EBCT images are produced in 50 ms.

14
System components

• The gantry
• The computer
• The operating console

15
The gantry

• Includes the x-ray tube, the detector array, the
  high-voltage generator, the patient support couch
  and the mechanical support for each.
• X-ray tube it has special requirements. The
  power capacity must be high. The anode heating
  capacity must be atleast several million heat
  units (MHU).

16
The gantry

• High speed rotors are used in most tubes for the
  best heat dissipation.
• Focal-spot size is important. CT scanners
  designed for imaging using high spatial
  resolution incorporate x-ray tubes with small
  focus-spot.
• Detector assembly Early scanners had one
  detector. Modern scanners have up to 8000,
  devided in to groups scintillation detectors and
  gas-filled detectors.

17
The gantry

• Scintillation detectors
• Containes scintillation crystal-photodiode
  assemblies. They convert light into electronic
  signals. They are highly efficient at detecting
  x-rays, almost 90 of the x-rays are absorbed
  and contribute to the output signal.
• But the space between each detector is big, so
  the overall detection efficiency may only be 50
  . They give dose to patient but do not
  contribute to the image.

18
The gantry

• Gas-filled detectors
• Contructed of a large metallic chamber with
  baffles spaced with 1 mm intervals.
• The baffles are like grid stripes and devide the
  large chambers into small ones.
• Each small chamber is one detector.
• It is sealed and filled under preassure with an
  inert gas with high atomic number
  (xenon/xenon-krypton mixture)
• The overall total detetction efficiency is 45 ,
  almost the same as scintillation detectors.

19
The gantry

• Collimators Required for the same reason as
  conventional x-ray. Correct collimation reduces
  patient dose and improves image quality due to
  less scattered radiation.
• In CT there are normally two collimators.
• One is the prepatient collimator on the x-ray
  tube housing/adjacent to it.
• It limits the area of the patient that intercepts
  the useful beam and thereby the slice thickness
  and the patient dose.

20
The gantry

• Improper adjustment of this collimator is the
  cause of most of the un-necessary dose to
  patient.
• The predetector collimator located under the
  patient, over the detector array.
• Reduces scatter radiation? improves image quality
• When coupled correctly with the prepatient
  detector, it defines the slice thickness.
• Has nothing to do with patient dose.

21
The gantry

• High-voltage generator
• All CT scanners operate on three-phase or
  high-frequenzy power.
• Most manufactors built them into the gantry or by
  mounting on the rotating wheel of the gantry. It
  reduces the amount of space needed, and winding
  and unwinding a power cable is unnecessary.

22
The gantry

• Patient positioning and support couch
• It has to be made of a material with a low atomic
  number (carbon fiber) so that is does not
  interfere with x-ray beam transmission and
  patient imaging.
• It should move smoothly for accurate patient
  positioning, and is especially important for
  spiral CT

23
Computer

• It is unique for the CT and a must! A ultra-high
  speed digital computer is needed for making CT
  images.
• Depending on the format the computer has to do up
  to 250 000 equations at the same time!
• In the computer there is a microprocessor and a
  primary memory. These determine the
  reconstruction time the time from end of
  scanning to image appearance.
• Array processors are becoming more common. They
  are faster than the microprocessor and can
  reconstruct an image in less than 1 s.

24
Operating console

• Many CT scanners have 2 or 3 consoles.
• One for the CT radiologic technologist to operate
  the scanner.
• One for an other technologist to postprocess
  images.
• One for the physician to view the image,
  manipulate contrast, size and general visual
  appearance.
• A typical operating console contains controls and
  monitors for the various technique factors.

25
Image characteristics

• With CT, the x-rays form a stored electronic
  image that is displayed as a matrix of
  intensities.
• The CT scan format consists of many cells with
  its own number which is shown as a brightness
  level.
• A matrix of 512 x 512 262 144 cells of
  information.

26
Image characteristics

• Each cell is a pixel (picture element)
• The numerical information in each pixel is a CT
  number/ Houndsfield Unit (HU)
• It is a two dimensional representation of a
  corresponding tissue volume.
• The diameter of image reconstruction is called
  the field of view (FOV)

27
Image characteristics

• When the FOV is increased for a fixed matrix (
  for example from 12 to 20 cm) the size of each
  pixel is increased proportionately.
• When the matrix size is increased for a fixed FOV
  (for example 512 x 512 to 1024 x1024)the pixel
  size grows smaller.
• Pixel size FOV/matrix size
• The tissue volume is known as a voxel (volume
  element)
• Voxel size Pixel size x slice thickness

28
CT numbers

• Each pixel is displayed on the video monitor as a
  level of brightness and on the photographic image
  as a level of optical density.
• The levels correspond to a range of CT numbers
  from -1000 up to 1000 for each pixel.
• -1000 is air, 1000 is dense bone and 0 is water.
  

29
CT numbers

• The CT number is related to the x-ray attenuation
  coefficient of the tissue contained in the voxel.
  
• Remember the degree of x-ray attenuation is
  determined by the avarage energy of the x-ray
  beam and the effective atomic number of the
  absorber and is expressed by the attenuation
  coefficient.

30
CT numbers

• By the scale of HU there is a range of 2000
  different gray scales with imformation, but most
  of it goes lost.
• The screen only shows 32 grayscales.

31
Image reconstruction

• Filtered back projection all the projections
  during on CT examination is stored in the
  computers memory, and the reconstructions are
  made by these.
• With filter we do not mean a metal filter as in
  the tube of the x-ray, but it is a mathematical
  function. A difficult one!

32
Image reconstruction

• In CT there over 250 000 pixels to reconstruct
  from, that means that the machine has to solve
  250 000 equations to find the solutions for the
  images.

33
Image quality

• Spatial resolution
• Contrast resolution
• Noise
• Linearity
• Uniformity

34
Spatial resolution

• If you scan a regular geometric structure that
  has a sharp interface,the image at the interface
  will be blurred.
• The degree of blurring is a measure of spatial
  resolution of the system and is controlled by
  several factors.
• If you take a scan over an area that has a high
  contrast interface, for example the brain and the
  skull, the image will be blurred.
• The system will fix some of the blurring, and
  smoothen the picture.

35
Spatial resolution

• This, however, reduses the spatial resolution
  because of some features of the scanner.
• The larger the pixel size and the lower the
  subject contrast, the poorer the spatial
  resolution will be.
• The detector size and design of prepatient and
  postpatient collimation affect the level of
  scatter radiation and influence the spatial
  resolution by affecting the contrast of the
  system.
• Also the x-rays focal spot has influence on
  spatial resolution.

36
Contrast resolution

• Contrast resolution the ability to distinguish
  one soft tissue from another without regard for
  size or shape.
• Contrast resolution is superior in CT,
  principally because of x-ray beam collimation.

37
Contrast resolution

• Imagine a scan over abdomen, where you have
  spine, liver and fat. The atomic numbers are
  different, but in conventional x- ray it is
  difficult to seperate them. With CT and the CT
  numbers it makes it easy! With HU the CT can
  amplify these contrast differences, and make the
  contrast high. Then we can cleary see differences
  between tissue.

38
Noise

• Noise the precentage of standard deviation of a
  large number of pixels obtained with a water-bath
  scan.
• Noise depends on the following factors
• 1 kilovolt peak filtration
• 2 Pixel size
• 3 Slice thickness
• 4 Detector efficiency
• 5 Patient dose

39
Noise

• Example
• If you scan a homogeneous medium like water the
  pixel value should be zero. But because the
  system is not perfect some pixel values will be
  both higher and lower than zero. These variations
  in HU will show in the image as graininess, and
  is what we call noise. The larger the variations
  in pixel value, the more noise you get in the
  image.

40
Linearity

• The CT must be calibrated frequently so that the
  HU are correct.
• There is a test you can do with a phantom and a
  water bucket.
• The result from this test should show a linear
  line passing through the CT number of water (0)
• If the test shows deviation from linearity its a
  sign of malfunction of the CT.
• It may not show on the visual image, but could
  greatly affect quantitative analysis of tissue,
  the determination of tissue composition based on
  CT number.

41
Uniformity

• When you scan a uniform object (water) the pixel
  value should be zero (for water!).But since the
  machine is very complicated mechanically this
  does not happen. The value may drift from day to
  day/hour by hour.
• If it is scanned, and the pixel value is constant
  in all regions of the reconstructed image, this
  is called Spatial uniformity

42
Uniformity

• There is also a test for this, where you scan a
  bucket of water and plot the numbers along an
  axis of the image. If this axis is within 2
  standard deviations of the mean value, the system
  has acceptable spatial uniformity.
• Because of the x-ray beam hardening, there may be
  a decrease of CT numbers, so the middle of the
  image is darker than it should.
• This is called cupping artefact.

43
Summary

• The collimated x-ray beam is directed to the
  patient.
• The attenuated image-forming x-ray beam is
  measured by a detector array.
• The signal from the detector array is measured by
  a computer.
• The image is reconstructed in the computer.
• The image is displayed on a TV monitor.

44
Summary

• CT makes transverse images (axial images)
• The internal structures of the body attenuate the
  x-ray beam according to their mass density and
  atomic number.
• All data are processed in digital form.
• The resulting computer image is an electronic
  matrix of intensities.
• Matrix size is generally 512x512 individual cells
  or pixels.

45
Summary

• In each pixel is numerical information called a
  CT number or HU.
• The pixel is a two-dimensional representation of
  a corrensponding tissue volume.
• The voxel (volume element) is determined by
  multiplying the square of the pixel size by the
  thickness of the CT scan slice.
• HU? -1000air, 0 water, 1000 dense bone

46
Summary

• The CT scanner has exellent contrast resolution
  because of the reduction of scatter radiation by
  the x-ray beam collimators.
• The ability to scan low-contrast anatomic
  structures is limited by the noise of the system.
  
• System noise is determined by the numbers of
  x-rays used by the detector array to produce the
  image.

47
Questions...

• ?
• ?