VASCPROG 560

Vascular Imaging Techniques

Module 3 Computed Tomography (CT)

CIHR Strategic Training Program

In Vascular Research

CIHR Strategic Training Program in Vascular Research

Navigation through this Module

This module was generated using Microsoft PowerPoint and then converted to

Adobe Acrobat. You will need Adobe Acrobat Reader to view the content. Different

web browsers may display WebCT content differently. Please contact Jackie

Williams at the email address below if you experience difficulties viewing any

module.

Instead of a course textbook, all the modules contain links to excellent information

that can be found on the internet. It is important that you visit these links to get more

background on the topics. These also may be printed out to read in more detail later,

or to be saved for future reference.

If you have any difficulty in accessing any of the links within these modules please

send an email to jwilliams@robarts.ca. Sometimes the sources of the links change

and adjustments will be made to correct this.

When you have finished the module, please go to the Module 3 Quiz under the Quizzes icon on the Course Home Page.

1

Credits

This module was written by the Course

Instructor, Dr. Terry Peters, Imaging

Research Laboratories, Robarts Research

Institute, one of the early pioneers of CT,

who wrote the first PhD on computerized

tomography in the 1970’s.

Jackie Williams also wrote some of the

content and organized the module in

WebCT.

Some of the content also was based on

information from the following source:

W. Huda & R. Slone. Review of Radiologic

Physics, 2nd edition. Lippincott Williams &

Wilkins. 2003.

Many of the graphics (with the identifying

mark Ka 2000) appear with the permission

of Dr. Willi A. Kalender, Professor and

Chairman of the Institute of Medical

Physics (IMP) at the University Erlangen-

Nuremberg, and his publisher, from his

book, “Computed Tomography:

Fundamentals, System Technology, Image

Quality, Applications”, 2000, Publicis MCD

Verlag, Munich.

2

Introduction

CT was the first clinical imaging modality that provided computed digital images instead of

analog images. It allowed clinicians to view single discrete slices of the body instead of

superposition images of large body areas.

The basic principles of medical CT are straightforward, and are a progression from x-ray

imaging. X-rays project through a cross-section of the body, are attenuated by the body’s

tissues, and then are detected after leaving the body. CT scanning also records the intensity

of the x-rays behind the object, with the x-rays being emitted in many different directions. The

attenuated signals are recorded and converted to projections of the linear attenuation

coefficient distribution of the body. These multiple projection data are reconstructed by the

computer using certain mathematical algorithms, which will be described in more detail later

in this module. CT relies heavily on geometry to calculate the image specifications. As will be

described in this module, most of the different configurations of CT scanners are referred to

as geometries.

It is expected that students will have studied the preceding module on x-ray imaging and so

will be conversant with the basic terms, which are not repeated in this module.

The following website from the University of Colorado gives a very basic, interactive section

on how CT works, and if you know nothing about CT it is an excellent place to start.

CAT Scans - An Overview

3

What is Computed Tomography (CT)?

Like conventional x-ray images, the CT or

CAT (Computerized Axial Tomography)

scanner uses an x-ray beam to form images

of the anatomy of the body. Unlike

conventional x-ray imaging, which produces

radiographs or “x-ray shadows” of a volume,

CT provides computed digital images of

anatomical cross-sections.

Most soft tissues have x-ray attenuation

coefficients that are similar, and when the

image is presented as a conventional

radiograph, the differences in attenuation (or

absorption) are lost since the shadows from

various organs overlay each other.

However, when viewed in cross-section (by

CT), the small differences in attenuation

between tissues (fat and muscle for

example) can be easily discriminated.

Typical radiograph of the

head (top), and CT-scan of

the brain (bottom) 4

The Concept of Slices

CT uses computers to transform flat two-dimensional x-ray images into three dimensions,

one slice at a time...

5

Standard X-Ray Views

Standard radiography produces

images by irradiating the body

with x rays, and exposing a

photographic plate to the rays

that penetrate the body. Because

various tissues absorb or scatter

radiation differently (bones more

than muscle, which in turn

attenuates x rays more than

water), the resulting pattern on

the x-ray film is a shadow of the

internal structures. Thus high-

density objects can obscure

lower density organs. As depicted

on the next page, this is

particularly important in the brain

where the overlying skull

completely overwhelms any

detail that might be in the brain

tissue.

(PA = Posterior-Anterior)

6

The CT Contrast Advantage

a) A tumour shown in a typical CT scan image of the head. Note that is is well visualized

against the surrounding brain tissue. In terms of Hounsfield Units, (described later in

this module) the contrast is approximately 50%.

b) A skull x-ray image of the same region. Because of the overlying bone the contrast is

much lower (0.23% as shown in the diagram (c) on the following page), and the

tumour is not detectable by the human eye.

a) b)

Ka 2000 Ka 2000

7

The CT Contrast Advantage

c)

8

Instrumentation

The main component of the CT

scanner is the gantry, which

contains the x-ray source, the

detector array and lots of

electronics to control the collection

of the data from the detector.

Generally the gantry rotates

continuously around the patient.

The patient lies on a table that

resembles a bed or stretcher,

which slides into a gantry. The

table slides the patient in or out of

the scanner, depending on which

part of the body is being scanned,

and is under the control of the

computer, which can be

programmed to move the table to

the desired slice.

The computer takes all the

information from the detectors and

reconstructs an image of the

required slice(s).

Gantry

Sliding Table

(Courtesy of Elscint)

Computer

X-ray

tube/detector

assembly

inside the

gantry cover

X-ray Source

Beam

Motorized

Table Detectors

9

The CT-Gantry

The x-ray tube sweeps around the

body emitting a fan-shaped beam of

x rays through the body, so that they

fall upon an x-ray detector on the

opposite side. This detector can have

in excess of 1000 elements.

Measurements are made at each of

several hundred angular positions as

the x-ray detector assembly rotates

about the patient. Each set of

measurements consists of an

independent “view” of the slice,

These views are then filtered and

back-projected, as will be described

later in the module, to reconstruct the

image.

Interior view of CT gantry

X-ray tube

Detector

Fan-beams

10

Pixels and Voxels

The individual elements within a

slice are called “pixels” (or picture

elements). Even though a pixel is a

2D quantity, it represents the

average attenuation coefficient in a

volume of tissue x.y.s. (See

figure on right). If however the

pixels are deemed to be part of a

3D volume, they are called voxels

(volume elements).

A typical CT image consists of an

array of 512 512 pixels, and a

volumetric scan might produce a

block of 512 512 512 voxels.

Usually the voxel is square in the

x-y plane, but depending on the

slice thickness selected, the

dimension in the axial (z) direction

could be several times the in-plane

(x-y) dimension. x

11

Pixel Size and Slice Thickness

a) Transverse image of

the brain, large

pixels, thick slice.

b) Sagittal image

reconstructed from a

series of such

images.

c) Transverse image of

the brain, small

pixels, thin slice.

d) Sagittal image

reconstructed from a

series of such

images.

Note the clearer

representation of

images in c) and d).

12

Volumetric Data

Originally, CT volumes were acquired by

scanning patients slice-by slice, where each

slice was approximately 5 to15 mm thick.

After the data for each slice was acquired,

the patient was physically moved through

the gantry to be in the correct position for

the next slice. This process was continued

slice-by slice until the entire volume was

collected. When the slices are stacked, the

result is a 3D volumetric image, where each

individual element is referred to as a voxel.

Because of the geometry of the scanner

gantry, the images in CT are almost always

acquired as 2-D transverse (or Axial – the

A in CAT) x-y plane slices. This means

that the z-axis is perpendicular to the scan

and parallel to the body’s longitudinal axis.

However, if a sufficient number of slices are

stacked up, they can be represented by the

computer as sagittal or coronal images

also.

x

y

z

Sagittal images are generated by the y-z

plane.

Coronal images are generated by the x-z

plane.

13

Acquiring the CT data

In the simplest CT configuration, a

Fan-beam of radiation is rotated around

the body, so that only a single slice is

illuminated with x rays.

Views of this slice are acquired over a

range of at least 180 degrees by an

array of electronic x-ray detectors.

The x-ray intensity registered by each

detector is the sum of the attenuation of

the tissues the beam has passed

through, and is known as the ray sum.

The collection of ray sums for all the

detectors at one tube position is called

a projection. Projection data sets are

acquired at different angles around the

body.

A typical projection will have up to 1,000

data points, and a CT image generally

consists of about 1,000 projections.

The projection data sets for all

detectors, and from all directions, is

known as a “sinogram”.

14

Sinogram

Scanned Object

View Data(Sinogram)

The sinogram is the

representation of raw data

obtained when projection-

reconstruction imaging is used,

and is the representation of the

signal measured at a given

angle in the imaging plane at

varying distances along the

detector array. The signal

intensity measured is a map of

intensities for each angle .

The figure at the right shows a

cross-section of an orange (a),

along with the sinogram of the

cross-section (b). Note the

sinusoidal appearance in the

sinogram of the pip in the

orange.

a)

b)

Beam direction

15

Reconstruction

Sinogram

Reconstructed Image

Computer

Reconstruction

Algorithm

The task of the CT

reconstruction program is to

take the data in the sinogram,

and mathematically convert

them into a picture of the cross-

section.

16

Classical Tomography

X-ray Tube

Start position End position

Film Start position Film End position

Object being imaged

Fulcrum plane A A

B B

Motion of Film

Motion of X-ray Tube

Before the advent of CT, cross-

sectional imaging was still possible

using classical tomographic

techniques. Transverse tomography

attempted to simulate the focusing

of x-rays by moving the x-ray source

and the detector (film) relative the

patient during x-ray exposure, in

such a way that only a single plane

would remain in “focus”.

In the figure on the right, as the x-

ray tube and the film moved, all of

the details in plane A-A (the fulcrum

plane) remained stationary with

respect to the film, while detail in

other planes (B-B for example)

would move relative to the film and

would therefore be blurred. Thus,

although the details in the plane of

interest would be imaged sharply,

the detail was obscured by the

blurred representations of the details

in the other planes. 17

x-ray source

Object being imaged

Imaged

(fulcrum)

plane

Direction of rotation

Film plane

Transverse Tomography

The same principles were applied to the

transverse tomograph that could create images of

planes perpendicular, rather than parallel, to the

patient’s longitudinal axis. In this case, while there

was still motion of the x-ray film, the object

(patient) rather than the x-ray tube moved (see

figure at right).

This geometry also resulted in a sharp image of

the fulcrum plane, but which was again degraded

by blurred renditions of overlying structures.

The figure below shows a transverse tomogram

through the thorax.

18

Modified Transverse Tomograph

Let’s examine what happens when we

modify the geometry of the transverse

tomograph somewhat so that the beam

only passes through the slice of

interest, and there for avoids the

possibility of shadows of structures

from other planes interfering.

To appreciate what happens in this

case, consider a single spot within the

desired cross-section. As shown on the

right, the shadow of this spot is

“smeared” across the film, resulting in a

thin line.

After three such views are “back-

projected”, the image of the point

consists of three lines intersecting at a

point (b). After a large number of views

have been back-projected in this

manner, the original point is

reconstructed as a diffuse “blob” (c)

Film plane

x-ray source

Object being imaged

Imaged plane

Direction of rotation

(Single disc) Aperture to form

laminar beam

“Image” of

object

Back-projection of

many views onto

reconstruction

plane

Back-projection of

three views onto

reconstruction

plane

a)

b) c)

19

Measuring an Object in CT This simple diagram shows the basic procedure as was used in the first EMI CT scanners

introduced in the early 1970’s. The geometry was actually quite similar to that introduced in the

previous pages. A pencil beam is emitted from the x-ray source and the intensity that has been

attenuated by the body is recorded by the detector. The x-ray tube/detector combination is

translated across the object, and then rotated around the object by 1o and a second projection is

recorded. This is repeated at 1o intervals for 180 projections. This generates an intensity profile

for parallel rays by taking the logarithms of the primary intensity and the attenuated intensity

behind the body.

20

Attenuation coefficients in the slice

CT measures not only the intensity of the x rays within the body, but also the x-ray intensity

behind the body. So it measures the intensity I attenuated by the body, and the primary

intensity Io to calculate the attenuation value along each ray from the source to the detector.

Following are two cases that use the basic formulae to calculate the attenuation value.

Attenuation is defined as the natural logarithm of the ratio of primary intensity to attenuated

intensity.

Where:

I = intensity

Io = primary intensity

d = absorber thickness

= linear attenuation coefficient

P = projection value

21

Attenuation coefficients in the slice

In a realistic situation, different tissues within organs attenuate x rays to differing degrees. In

addition, the x-ray beam is polychromatic (contains a spectrum of beam energies – like white

light, rather than a single pure colour (see section on artifacts later in the module), and since

the attenuation also depends on the energy, the attenuation coefficient of a pixel is a rather

complicated function of the beam energy as well. Correction algorithms are incorporated into

the reconstruction process to compensate for errors introduced by “beam hardening” (see

artifacts section), where the spectrum of the beam is modified as low energy photons are

preferentially absorbed over high energy photons.

22

Filtered Back-projection

Direct back-projection of the views generated by the

scanning operation in the previous page results in

blurred image. The objective of the reconstruction

algorithm in CT is equivalent to turning this blurred

image of the cross-section into a sharp image of the

same slice. This is equivalent to a mathematical “de-

blurring” problem. Blurring occurs when the fine detail

in an image is smeared out. De-blurring re-sharpens

this detail using a mathematical filtering operation.

We can back project all the views to get a blurred

image, and then filter the 2-D image to obtain a

reconstruction, or we can filter each of the one-

dimensional views prior to back- projection. Generally

in CT scanners, the latter approach is employed.

If we filter a single point in our view, the filtered view

appears as we see on the right hand side of a). If we

perform the same operation of the view of the disc in

our cross-section, the resulting filtered view is as

seen b). In the filtered view, positive numbers have

been represented in red while negative values are

shown in blue.

+

- -

a)

b)

Filter

23

Back-projecting Filtered Projections

Whereas the unfiltered projections

yielded three intersecting stripes

after three views had been back-

projected (a), and a blurred blob

after many views had been

processed, (a,b); after filtering it can

be seen in c) that the positive and

negative components of the filtered

views cancel on certain places, and

even after only three views have

been back-projected, the original

spot is becoming isolated. After

many views have been treated in

this manner, the original spot is

reconstructed faithfully (d).

a)

c) d)

b)

24

Reconstructing the Image

Original Slice

Single unfiltered view Reconstruction after

back-projecting multiple

unfiltered views

Single filtered view

Reconstruction after

back-projecting multiple

filtered views

By collecting views from around

the object and filtering each one

prior to back-projecting them, a

reconstruction is formed in which

each filtered view interferes (i.e.

re-enforces or cancels) with all the

others in exactly the right way to

reconstruct a well defined image of

the original slice. Note however

that without the filtering step, a

very blurred image results that is

reminiscent of the axial tomogram

presented earlier.

25

How is the Image Reconstructed from the Attenuation

Profile?

To summarize, the method described here is known as the “convolution-back-

projection” technique. The information from each vies is first filtered with an edge-

detection filter (convolution), and this data is then projected backwards into a new

image (backprojection). The need for these two steps lies in the mathematics of

reconstruction and Fourier techniques.

In order to process each vertical slice individually, we must first extract the data about

each slice from the initial images. This data is stored in "sinograms." A single sinogram

contain the information from all angles about a particular slice, with the information from

each angle in its own row. In the picture above, the first row contains the data from 0

degrees, and the last row contains the data from 180 degrees.

Unfortunately, to reconstruct we have to do more than just convolution and

backprojection. There are several real-life issues that we must deal with before we can

actually reconstruct the image. These include the non-uniformity of the x-ray beam, that

the original images contain x-ray intensity, not "apparent thickness" (how much stuff the

beam is passing through), that the detector may have bad elements, and that the beam

may not be centered on the sample’s rotation axis.

26

Filtering Techniques

Many CT scanners give the operator some control over the exact nature of the filter being

applied prior to back-projection. If for example, it was determined that soft-tissue contrast was

important, it might be appropriate to employ a “smoothing” filter (b), or if on the other hand it

was desired to enhance fine structure in bone, it might be more appropriate to use an “edge-

enhancing” kernel (c).

a)

b)

c)

27

Summary of the Reconstruction Process

28

The CT Number or Hounsfield Unit

1000-

w

wt

tHU

CT depends on the same interaction of x-

ray energy with tissue as conventional

radiography (see Module 2), that is, it

relies on the attenuation of x rays by

tissue. Since CT reconstructs cross-

sections of the body, there is the possibility

of calculating the attenuation coefficients

directly for each voxel in the image.

The symbol is used to represent the

attenuation coefficient, but rather that

expressing these values directly, they are

represented as values relative to the

attenuation coefficient of water multiplied

by 1000. This number is called the

Hounsfield Unit (HU), after the inventor of

CT.

So, if t is the attenuation of the tissue of

interest, and w is the attenuation

coefficient of water, then the CT number

(in HU) of this tissue is given by the

formula on the right.

28

CT Numbers of Various Tissues

29

Window Level and Width

Reconstructed images can have

pixels or voxels having different

CT numbers (Hounsfield

numbers) with a range of over

4000. For example, air has a

value of

–1000 HU, air is 0, and dense

bone has a HU value of around

3000. If each number is

represented by a gray level on

the screen, there would be over

4000 levels to discriminate. Since

the human eye can resolve less

than 64 levels, often only a small

subset of the entire range is

displayed on the screen.

All CT scanners have interactive

controls that allow the user to set

the window width (the range of

numbers displayed in an image)

and the window level (the CT

number at the centre of the

range).

Voxels with a CT number less than the lower limit are

displayed as black, CT numbers greater that the upper

limit are displayed as white, and the CT number at the

centre of the range as represented as gray. 30

Classification According to CT Generation

As computed tomography technology developed, researchers created improved methods of

collecting the x-ray projections through an object. CT Generation refers to the order in which

new CT scanners were introduced. They are classified based on how the different components

are laid out and the type of mechanical motion they use. However, it is not necessarily true that

a higher generation number means a higher performance system.

First Generation

In 1972, the first clinical CT scanner was used in clinical practice; the EMI scanner, and is now

known as a first generation scanner. Its design used a single x-ray source emitting a pencil

beam with two detectors that translated (moved) across the patient. The source/detector pair

were then rotated one degree and a subsequent set of measurements are obtained. This

process was repeated 180 times for each projection angle, taking about 5 minutes to rotate

180o. Because of the translation and rotation process, this generation of scanners is referred to

as a translate/rotate scanner.

Second Generation

Second generation scanners are also referred to as a translate/rotate scanner, but they use

multiple detectors and a fan-shaped beam. The other major difference was that they larger

rotational increments, making them more efficient and faster than the original 1st generation

scanner.

31

Classification According to CT Generation

Third Generation

Larger detectors and improved data acquisition technology, made it possible to measure the

entire patient cross-section simultaneously using a fan-beam projection. The geometric

relationship between the tube and the detectors stays the same as it rotates 360o around the

patient. The imaging process is significantly faster than 1st or 2nd generation systems. This

generation of scanners are often referred to as having rotate-rotate scanner geometry.

Fourth Generation

Fourth generation scanners use a rotating tube with a fixed array of detectors that might have

as many as 4,800 detectors. They are said to have a rotate-fixed system.They were developed

to overcome one of the disadvantages of third generation scanners, which is a tendency to

create ring artifacts, but the performance of third and fourth generation scanners is about the

same. Unfortunately, the increased number of detectors in the fourth generation scanners also

increased the cost. Both third and fourth generation scanners can acquire a single section in

0.5 to 2 seconds.

Multislice CT

The advent of multislice CT made fourth generation scanners obsolete. After this, the modern

scanners (fifth generation) that use electron beam technology were developed. They are

described on the following page.

32

Modern CT Scanners

Modern Scanners are divided into:

Axial Computed Tomography Scanners

Helical, or Spiral Computed Tomography

Multi-slice Computed Tomography (MSCT)

Computed Tomography Fluoroscopy

Cone-beam CT

The features of these scanners are described in the following pages.

33

Axial Computed Tomography

All CT acquisitions begin with a topographic or scout image obtained by advancing the

patient table through the gantry with the x-ray tube in a fixed position. The use of cables

that supply the necessary high voltage to the x-ray tube limit the gantry’s rotation to one

revolution.

There are several parameters that the operator selects to obtain the optimum image.

These include:

– The scan time – usually from 0.5 to 2 seconds

– The tube current – 50 to 400 mA

– The voltage – from 120 to 140 kVp

– The section (slice) thickness – from 1 to 10 mm

– The field of view (FOV) and

– The image reconstruction filter.

Most modern CT scanners use slip ring technology in which contact rings inside the

gantry supply the high voltage to the tube. Slip ring scanners speed the process of data

acquisition and several sections can be scanned in a single breath-hold. The floating

contact of the slip ring allows the passage of signals between the continuous rotating x-ray

tube and stationary components of the scanner. The development of slip ring technology

in CT made spiral CT possible.

34

Spiral CT and Multi-slice CT

Spiral CT scanners image entire anatomic regions (like the abdomen or lungs) in 20

to 30 seconds. The scanner rotates continuously as the patient couch glides. And at

that speed, most patients can hold their breath for the entire imaging session. That

eliminates the possibility that image quality will suffer due to the motion associated

with breathing.

The continuous nature of the images-there are no gaps between slices obtained

through spiral scanning-means that the data can be reconstructed to provide three-

dimensional images, displaying the entire volume of organs and vessels. This

increases the likelihood that very small lesions will be detected.

Spiral CT has become the primary imaging technique for the chest, lungs, abdomen,

and bones because of its ability to combine fast data acquisition and high resolution.

"Multi-Slice" Spiral CT Scanners

The newest "multi-slice" spiral CT scanners can acquire up to four slices in a single

rotation and collect as much as eight times more data than previous state-of-the-art

spiral CT scanners. This new technology will provide for more non-invasive imaging of

a wider range of conditions in less time and with greater patient comfort.

The section thickness is determined by the detector width rather than the collimation

thickness, as in conventional CT.

35

Computed Tomography Fluoroscopy

CT fluoroscopy is sometimes referred to as real-time CT scanning. It displays

constantly updated images produced almost in real-time by continuous rotation of the

CT tube.

Images are typically updated at the rate of 6 per second, which gives excellent

temporal resolution.

As motion in the patient can be followed in the constantly updated reconstruction, it is

very useful for following the passage of needles for biopsy or drainage procedures.

It is generally performed at the same voltage (kV), but at lower currents (mA) than

conventional CT to minimize radiation doses. The usual protocol would be:

– 120 kV, and between 20 to 50 mA for CT Fluoroscopy

– 120 kV, and between 200 to 300 mA for conventional CT

36

Cone-Beam CT (Volume CT)

Cone-Beam CT (CBCT)

Also called volume CT or 3-D scanning, this is a recently developed technique that can greatly

speed up collection of CT data from an object. The term Volume CT has been in use for many

years to refer to the 3-D visualization of an object from a contiguous stack of separately

scanned CT images, or "slices" made up of a conventional CT system. True Volume CT

Scanning is a variation of that method which uses an area detector (such as a camera coupled

to an image intensifier, or scintillating screen) to collect 2-D radiographic projections of the

object while it is rotated 360°.

Conventional CT scanners collect 1-D

projections from a fan beam of x-rays. In

Volume CT, the x-ray beam is opened from a

fan to a pyramid or cone. By collecting 2-D

projections, the Volume CT scanning method

makes more efficient use of the x-rays

emitted from the source. A special

reconstruction algorithm processes the 2-D

cone-beam projections into a volume that is

the equivalent of many simultaneous,

contiguous CT slices. A special feature of

this data set is that oblique slices through

any part of the object can be easily

generated for 2-D display. It is particularly

good for imaging the vasculature. 37

Contrast Agents

The contrast agents used in CT can

highlight specific areas so that the areas of

interest show more clearly. Contrast

agents are not always required and in

certain cases can actually make the

visibility worse. Also, they are expensive

and some patients may be allergic to

them. For such patients, an MR scan can

be used instead.

Depending on which body part is being

imaged, some of the more common

contrast agents used are iodine, barium,

barium sulfate and gastrografin. They can

be administered orally, rectally, by

intravenous injection, and, more rarely

xenon gas can be used by inhalation for

certain brain scans.

Intravenous Contrast

Intravenous contrast is used to highlight blood

vessels and to enhance the structure of

organs like the brain, spine, liver, and kidney.

The contrast agent (usually an iodine

compound) is clear, with a water-like

consistency. Typically the contrast is contained

in a special injector, which injects the contrast

through a small needle taped in place (usually

on the back of the hand) during a specific

period in the CT exam.

Once the contrast is injected into the

bloodstream, it circulates throughout the body.

The CT's x-ray beam is weakened as it passes

through the blood vessels and organs that

have "taken up" the contrast. These structures

are enhanced by this process and show up as

white areas on the CT images. When the test

is finished, the kidneys and liver quickly

eliminate the contrast from the body.

38

Artifacts

There are several different types of

artifacts that appear in CT images.

Motion artifacts

Motion artifacts are produced in a CT

image if an object moves, but is

assumed to be static during the

reconstruction process. Anatomical

structures move periodically due to

respiration or cardiac pulsation and it is

difficult for injured patients or children

to stay completely still during scanning.

The movements can cause significant

artifacts, but can be eliminated by spiral

CT single breath-hold scanning.

Beam hardening artifacts

The conventional x-ray sources in CT

scanners have polychromatic spectra,

so that the x-ray photons emitted do

not all have the same energy.

The x-ray attenuation of an object depends

on the photon energy, with lower energy

photons being attenuated more than high

energy photons. So, as the x-ray beam

projects farther through an object, the higher

the energy portion of the x-ray spectrum

increases, as the lower energy photons

become attenuated. This causes the

attenuation coefficient and if it occurs

manifests as a “cupping” in the gray-scale

image. This type of artifact can be

minimized by pre-filtering the x-rays, and

applying appropriate algorithms.

Ring artifacts

These are rare on modern CT systems, but

appeared on third-generation systems due

to miscalibrated detectors.

Partial volume artifacts (also called

Blurring artifacts)

This is caused when the linear attenuation

coefficient in a voxel is averaged, when in

fact, it is heterogeneous in composition. 39

Clinical Applications of CT

GENERAL

The CT scanner is a very important imaging

technique for clinical purposes. It produces

much clearer images than conventional x-

ray or ultrasound, and while it does not

produce images as well defined as an MRI

scan, this is not necessary to diagnose

most abnormalities. The costs associated

with CT are less than for MRI and there are

many more CT scanners in hospitals and

clinics than are MRI Scanners. The only

risk of a CT scan is radiation, the same risk

as all other x-ray tests. This risk increases

somewhat if intravenous contrast is used,

but in general the risks of CT are minimal.

CT is normally not used for breast

diagnosis (mammography is better),

obstetrics (ultrasound is safer), or for

imaging the various soft tissue structures in

joints. Other than those exceptions, CT is

most often the imaging modality of choice.

Cranial CT has become a standard part of

any work-up for conditions such as trauma,

severe or persistent headache, and

changes in mental status. Radiologists

perform fine needle aspirations (biopsies)

of suspicious areas under CT guidance, as

well as under ultrasound guidance. An

abdominal CT scan can be an important

screening test for severe abdominal pain

when renal stones, appendicitis, or

diverticulitis are suspected.

VASCULAR APPLICATIONS

CT Angiography is now widely used to

visualize blood flow to study various

vascular diseases and abnormalities.

40

CT Angiography

CT (computed tomography) angiography (CTA)

uses CT to visualize blood flow in arterial vessels

throughout the body, from arteries serving the

brain to those bringing blood to the lungs, kidneys,

and the arms and legs. Compared to conventional

catheter-based angiography, which involves

injecting contrast material into an artery, CTA is

much less invasive and a more patient-friendly

procedure; contrast material is injected into a vein

rather than an artery. One of the most important

advantages of CT angiography over conventional

angiography however is that CT angiography is

inherently 3-dimensional, whereas conventional

angiography produces only 2D projection images.

This exam has been used to screen large

numbers of individuals for arterial disease. Most

patients have CT angiography without being

admitted to hospital.

Typical CT angiograms of the

head (top) and leg (bottom)

41

CT Angiography

CTA is commonly used to:

Examine the pulmonary arteries in the lungs to rule out pulmonary embolism, a serious

but treatable condition.

Visualize blood flow in the renal arteries, those supplying the kidneys, in patients with

hypertension and those suspected of having kidney disorders. Narrowing (stenosis) of a

renal artery is a cause of high blood pressure in some patients, and can be corrected

surgically. A special computerized method of viewing the images makes CT renal

angiography a very accurate examination. It is also done in prospective kidney donors.

Identify atherosclerotic disease, aneurysm, or dissection in the body's main artery, the

aorta and its major branches, the iliac arteries.

Detect atherosclerotic disease that has narrowed the arteries to the legs.

CTA also is used to detect narrowing or obstruction of arteries in the pelvis and in the

carotid arteries bringing blood from the heart to the brain. When a stent has been placed to

restore blood flow in a diseased artery, CT angiography will show whether it is serving its

purpose. Examining arteries in the brain may help reach a correct diagnosis in patients who

complain of headaches, dizziness, ringing in the ears, or fainting. Injured patients may

benefit from CTA if there is a possibility that one or more arteries have been damaged. In

patients with a tumour it may be helpful for the surgeon to know the details of arteries

feeding the growth. 42

Radiation Risks

Compared to most other diagnostic x-

ray procedures, CT scans result in

relatively high radiation exposure.

Normally, for patients with symptoms of

serious disease, the risks associated

with such exposure are greatly

outweighed by the benefits of diagnostic

and therapeutic CT.

The probability for absorbed x rays to

induce cancer is thought to be very

small for radiation doses of the

magnitude that are associated with CT

procedures. However, such estimates of

the cancer risk from x-ray exposure

have a broad range of statistical

uncertainty and there is some scientific

controversy regarding the effects.

The quantity most relevant for assessing

the risk of cancer detriment from a CT

procedure is the "effective dose".

Effective dose is evaluated in units of

millisieverts (abbreviated mSv).

Using the concept of effective dose allows

comparison of the risk estimates associated

with partial or whole-body radiation

exposures. This quantity also incorporates

the different radiation sensitivities of the

various organs in the body.

Radiation dose from CT procedures varies

from patient to patient. A particular radiation

dose will depend on the size of the body part

examined, the type of procedure, and the

type of CT equipment and its operation.

Typical values cited for radiation dose should

be considered as estimates that cannot be

precisely associated with any individual

patient, examination, or type of CT system.

Recently, a movement has sprung up in the

US for whole-body CT screening of

asymptomatic people, for early detection of

medical problems before they are noticeable.

The benefits of such screening are

questionable at this time, since healthy

people are undergoing what is likely to be

unnecessary radiation exposure. 43