X-ray computed tomography

• Limitations of using conventional X-ray technique

– Super-imposition of the three-dimensional information
onto a single plane makes
– Photographic film has limited dynamic range – hence
sufficient contrast is necessary to be distinguished by the
eye
• Details of bony structures can be clearly seen but it is difficult to
discern the shape and composition of soft tissue organs
accurately: growths and abnormalities within tissue only show a
very small contrast difference on the film
• Difficult to study soft tissues in the brain as it is shielded by the
dense bone skull
• Computed tomography developed by G.N. Hounsfield,
introduced on a commercial scale in 1972.
• Awarded the Nobel Prize for Medicine in 1979,
jointly with Allan Cormack
⚫Tomos - slice
graphein - to write

⚫imaging by sections or
sectioning, through the
use of any kind of
penetrating wave
 Computed tomography (CT)
• X-ray CT was invented by Godfrey Hounsfield
in 1972
• Allan Cormack independently worked earlier
on the mathematics of techniques
• Both were jointly awarded the Nobel prize in
Medicine in 1979
 X-ray CT
• 2D images of thin slices through the body
• Reasonable contrast between soft tissues compared to
radiography
– As X-rays transmitted through each organ are no longer
superimposed on one another at the detector as in planar
radiography
• 2D image of an object is reconstructed from a series of
1D projections of the object acquired at different
angles
• To obtain an image from a thin slice of tissue, the X-ray
beam is collimated to give a thin beam
• Stack of 2D slices => reconstruct 3D volume
⚫2D internal structure of an object is reconstructed from a series of 1D “projections” of
the object acquired at different angles
1D-CT

• I = Io exp(-  i xi)
• CT machine computes ln(I/Io)
giving  i xi
• Assuming Xi = 1, a line
projection is given by i
 Contrast scale
• Quality of the reconstructed image is a matter of the
differentiation between  (X-ray attenuation coefficient) at
different points and of the size of each pixel
• The differential attenuation coefficient is correlated with
the specific gravity
• The image reconstructed as a result of computerized
tomography can be considered as the mapping of densities
with respect to that of water
• The linear attenuation coefficient of tissue is represented
by the scanner computer as integers that usually range in
values from -1000 to +1000. These integers have been
given the name ‘Hounsfield’ units, and are abbreviated as H.
They are also denoted by CT numbers.
• H=((-water) / water)1000
• CT: 0 is assigned to water, and -1000 to air. The value +1000
represents highly dense materials.
 H=((-water) / water)1000

CT number scale
 CT resolution
• Spatial resolution around 2 mm
• Noise level: 0.5% change in the X-ray
absorption coefficient
 Scanning system
• First generation – parallel beam geometry:
– A collimated X-ray beam is passes through the body and its
attenuation is detected by a sensor that moves on a gantry
along with the X-ray tube
– The tube and detector move in a straight line, sampling the
data 180 times
– At the end of the travel, a 1 tilt is made and a new linear
scan begins.
– This assembly travels 180  around the patient’s position
– The procedure results in 32,400 independent
measurements of attenuation
– Fairly slow procedure, scan time of about 5 minutes
– It is essential for the patient to keep still during the entire
scan period and for this reason, the early scanners were
limited in their use to only brain studies
First generation – parallel beam geometry
 CT-Scan

Source and detector Source and detector are rotated 180/N

move M linear step degree, where N is no. of positions in a 180-
to get 1D projection degree rotation
• Total of MxN data points acquired for inversion
• Projection data is back-projected to reconstruct the spatial distribution of
X-ray attenuation
• Each pixel in the projection gives a CT number (Hounsfield number)
• Resolution can be improved by using finer step size.
• However, the limiting value is based on the X-ray focal spot size
Improvements in CT scanner
2nd Generation:
Using an array of detectors and a fan
beam of X-rays
Permits greater index angle
Faster scan, 18-20 s/scan

3rd Generation:
• Wider fan beam, More detectors, Faster scan
• Two orthogonal collimators
• One for controlling angular spread
• another for slice thickness control
• Slice thickness:1-5 mm, 360rotation, 512-768
detectors, 1 s/rotation
Limitation: it has a fixed geometry. With a fan
beam set for the largest patient, the arrangement
is inefficient for smaller objects.
Improvements in CT scanner
4th Generation:
• Complete ring of detectors
• Cone beam
• X-ray tube rotates 360
• Not much reduction in scan time
• Spiral/helical CT
• Second generation:
– Using an array of
detectors and a fan beam
of X-rays
– Effectively takes several
profiles with each
traverse and thus permits
greater index angles
– e.g. by using 10 fan
beam, it is possible to
take 10 profiles at 1
intervals, with each
traverse and then index
through 10 before taking
the next set of profiles
– System operating in the
18-20 s range
Construction of CT

Modern CT scanner (2021), photon-

counting CT (Siemens NAEOTOM Alpha)

https://en.wikipedia.org/wiki/CT_scan
 CT scan modes
• Spiral/Helical scanning: This is a scanning technique in
which the X-ray tube rotates continuously around the
patient while the patient is continuously translated
through the fan beam.
• The operation is often referred to as helix, spiral,
volume or three-dimensional CT scanning.

https://www.howradiologyworks.com/scan-modes-on-modern-ct/
 Detectors
• The detectors used in CT systems must have a
high overall efficiency in order to minimize the
patient radiation dose, have large dynamic
range, be very stable with time and insensitive
to temperature variations within the gantry
• Three main types of detectors:
– Scintillation crystal and photomultiplier tube
– Xenon
– Scintillarc (Scintillation crystal and photodiode)
Detectors
⚫ Xenon-filled ionization chambers

Electrode plates typically 10 cm in length, with

a 1mm gap between adjacent plates
 Ionization chamber working
• Metal electrodes separate chambers filled with gas
• X-rays incident in the chamber ionize gas and produce electron-ion pairs
• These are attracted to the electrodes
by an applied electric potential difference
between them
• Flow or electron-ion pairs causing current flow
• Current flow  incident X-ray photons
• Signal from each detector electrode is amplified
and digitized.
• Electrode plates also act as ant-scatter grids
Image Processing

x = dimension of each pixel

 Data processing
1) Pre-Processing: Projections are filtered to avoid artifacts in the image

2) Back-Project or Iterative algorithms are used

(i) Algebraic iterative reconstruction (ART) algorithms are used to reconstruct

(x,y)
(ii) Radon transform

3) 2D Image stacking
 Iterative
reconstruction
Projections
Actual Data

Step 1 Step 2

Iteration 1

Step 3 Step 4
⚫ After reconstruction, the image displayed as a map of the
tissue CT number

H=((-water) / water)1000

where
⚫ H is the CT number
⚫ μ is the linear attenuation coefficient of the tissue.
⚫ CT numbers varying in value from +1000 to−1000
Clinical Applications

⚫ Cerebral scans
⚫ Pulmonary disease
⚫ Abdominal Imaging
 Health dangers from X-ray
• It is desirable to obtain a diagnostic CT scan while administering a
minimal dose to the patient. However there is a trade-off between
resolution, noise and dose.
• The effects of cumulative X-ray dosage of ionizing radiation may
results in mutations –genetic changes resulting from damage to
chromosomes
• Physical illness: vomiting, headache, dizziness, loss of hair , and
burns
• Over dosage: destruction of vital physiological systems such as
nervous, cardiovascular, respiratory, renal and digestive systems
and tissues

Typical head and body doses: for average patients scanned on the CT systems
have been about 2.5 rads for heads and 1.0 – 2.0 rads for bodies. In multiple-
slice operations, the overall dose profile is the sum of the individual skin dose
profiles.