COMPUTED

TOMOGRAPHY
 COMPILED BY DR. SARAH MARYAM
ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 INTRODUCTION

• Tomography literally means slice view of the

patient.

• Computed tomography has had many names.

• Two of the more popular names are CAT

(computerised axial tomography) and CT
(computed tomography). CT is currently the
preferred name.
• Computed tomography generates images in
transaxial sections,i.e perpendicular to the axis of
rotation of the x-ray tube about the body and
generally perpendicular to the craniocaudal.
• The basic principle behind
CT is that the internal
structure of the object can
be reconstructed from
multiple projections of the
object.
• The ray- projections are formed by scanning the
thin cross-section of the body with a narrow x-ray
beam.

• The transmitted radiation is then measured with

sensitive radiation detectors.
• The detector does not form an image. It merely
adds up the energy of all the transmitted photons.

• The numerical data from multiple ray sums are

then computer processed to reconstruct an
image.
COMPUTED TOMOGRAPHY IMAGE: is most
commonly calculated on a 512x512 pixel matrix,
although 256x256 and 1024x1024 matrices may
also be used.

Each square in a matrix is a pixel and it represents a

tiny elongated block of tissue called a voxel…
• Each pixel is in fact a ‘voxel’. . it’s a volume
element having three dimensions having depth
that is equal to the thickness of the section.

• The value stored in the pixel (CT number/ CTn)

represent the average linear attenuation
coefficient of the tissues within the voxel.
• Water has a CT number equal to zero.

• Air, because of it’s very low density (attenuation

coefficient of zero), has a CT number equal to -
1000.

• Air and water are are used for the calibration of

the CT number scale of the scanner.
• The CT numbers of the body tissues themselves
are variable because of heterogeneity of the
tissues themselves and the variation in
attenuation coefficient of each tissue relative to
water, being dependant on the kV and filtration of
the x-ray beam.
APPROXIMATE RANGE OF CT NUMBERS FOR
VARIOUS TISSUES

Bone 500-1500

Muscle 40-60

Brain (grey matter) 35-45

Brain (white matter) 20-30

Fat -60 to -150

Lung -300 to -800

CT number of cortical bone can be as high as 1500,
although higher values may be displayed.

• The range of CT number values is most

commonly from -1024 to +3071 (4096 levels of
grey).
PARTIAL VOLUME EFFECT:

• The Ct number stored in each voxel represents

the average attenuation coefficient in that voxel.

• A high contrast object that is smaller than the

object may therefore be seen even though its
dimensions in the transaxial plane may be less
than the displayed pixel size.
• More significantly in CT scanning, a thin high
contrast structure, such as a vessel filled with
contrast or the skull that crosses the transaxial
plane at an oblique angle will be visualized in
several adjacent pixels giving an appearance of
much larger structure.
• The extent to which partial volume affects the
image depends on the thickness of the transaxial
slice, thinner slices providing better resolution.

• Partial volume effect may also reduce the visibility

of low contrast detail.
• If the detail only partially fills the voxel, then the
CT number will lie between the CT number of the
detail and the surrounding structure.

• There may then be insufficient difference in CT

number to visualise the low contrast detail.
IMAGE DISPLAY:

A CT image is usually displayed on a television

monitor for immediate viewing and recorded on a
film for interpretation and permanent storage.

In CT image, pixels have CT number in approximate

range (-1000 to +3000), i.e.4000 levels of grey.
However, human eye is limited to a range that is
not much greater than 50.

• CT scan displayed with CTn -1000 represented as

black and +3000 as white.
• All the shades of grey are displayed between,
would appear very flat with little differentiation.

Such a CT number might be -200 for air (lung),

+20 to +40 for retroperitoneal soft tissues, or +200
for bone.
• WINDOWING is a common technique in digital
imaging to bring about the hidden detail in the
image.

• Generally, the viewing window is set

automatically depending on the imaging protocol.

• However, the viewer has the opportunity to alter

this manually.
 The computer may then be instructed to assign
one shade of grey to each of the 128 CT numbers
below and each of the 128 CT numbers above the
baseline CT number.

For example, a general view of the abdomen will

centre at CT number 20 and display CT numbers -
108 to +148 as 256 grey levels.
The centre CT number is called the ‘window level’
and the range of CT numbers above and below the
Window level is called the ‘Window width’.

It is possible to set the window level at any desired

CT number.
 Window width is also variable to any width desired
by the operator.

Generally, one does not pick a window width of

exactly 256 and a window width will vary widely
depending on the type of examination or pathology.

Window obviously limits viewing to a narrow

portion of the total information available.
 In practice, multiple window levels and multiple
window widths may be examined in an effort to
extract maximum diagnostic information from each
examination.
• In lung window,

window range has been set to -900 to -400

to display the detailed structure in the lungs.

• All pixels with CTn of 100 and greater are

displayed as white and therefore little detail can
be seen in mediastinum or the chest wall.
• In mediastinal window,

the window level has been set from -240 to 300,

providing no information in the lungs but good

structural

detail outside.
 Typical window width
and level values

Although this varies somewhat from institution to

institution and vendor to vendor, window width and
centres are generally fairly similar. The values below are
written as width and level in Hounsfield units (HU).
• head and neck

◦ brain W:80 L:40

◦ subdural W:130-300 L:50-100
◦ stroke W:8 L:32 or W:40 L:40
◦ temporal bones W:2800 L:600
◦ soft tissues: W:350–400 L:20–60
• chest
◦ lungs W:1500 L:-600
◦ mediastinum W:350 L:50

• abdomen
◦ soft tissues W:400 L:50
◦ liver W:150 L:30

• spine
◦ soft tissues W:250 L:50
◦ bone W:1800 L:400
EQUIPMENT FOR COMPUTED TOMOGRAPHY

• The present generation of multislice scanners are

based on 3rd generation scanner geometry.
• THE PRINCIPAL COMPONENTS OF THE CT
SCANNER ARE:

• X-ray tube

• collimator

• detector array,

all are mounted on a rotating gantry.

The rotation axis is refer to as

“z-axis”.

The x-ray beam is collimated as a

wide fan-beam

sufficient to cover the patient cross-section at it’s

widest.
 It has a narrow width parallel to the z-axis that, for
a single slice scanner generally defines the imaged
slice thickness.

Behind the patient is an arc of detectors, with the

radius of the arc being equal to the focal distance
so, that each detector is the same distance from the
source.
• The total number of detectors is in the range 500-
1000.

• The high voltage generator is also mounted on

the gantry alongside the x-ray tube.

• On a helical scanner, the gantry can rotate around

the patient in 1s or faster, the minimum time on
the newest scanners being about 0.3s for a full
360˚ rotation.
• Rotating gantry weighs about 500kg.

• The patient lies on a couch that has a tabletop

that can be moved longitudinally through the
gantry aperture.

• Normally, plane of gantry rotation is perpendicular

to the long axis of the tabletop.
• However, it is possible to tilt the gantry usually up
to 30˚ about the vertical.

• This allows transaxial scans to be produced at an

angle in the body.

• The main application is in scanning of the head,

for which the scanning planes may be made
parallel to the base of the skull.
 X-RAY TUBE

• The x-ray tube is mounted with its anode-cathode

axis parallel to the axis of rotation of the scanner.

• This is so to minimise the influence of anode heel

effect because the size of the beam parallel to the
axis of rotation (z-axis) is more than a few cm
even with a multislice scanner.
• Tube of CT scanning have to be capable of
prolonged exposure time at high mA.

• Typically tube has 2 focal spot sizes,

the smallest being about 0.6mm.

• Many are designed to run continuously for periods
of 90 sec or greater at 120kV and 200mA.

• To, achieve this performance, they have heat

capacities of 4MJ or more.

• CT scanners incorporate heat exchangers to cool

the oil and the air within the gantry enclosure is
maintained at a low temperature.
Typically scanners are operated at fixed 120kV

but

a range of kV is usually available with three or

four fixed settings between

about 80 and 140kV.

• Ideally, monoenergetic photon beam would be
used for CT but this cannot be achieved with an x-
ray source.

• Earlier scanners used additional copper filters to

remove low energy photons.
• But because of improvements in reconstruction
algorithms this is no longer necessary and
generally scanners have a total tube filtration of
6mm Al.

• The collimator is mounted on the x-ray tube.

 The beam is collimated to a fixed width

that is generally about 50cm

at the axis of rotation,

sufficient to cover the full cross-section of the

patient.
• On the z-axis , the size is variable and in the case
of single slice scanner, defines the imaged slice
thickness.

• Typically there are fixed width settings between 1

and 10mm.
• For single slice scanners, there may also be a
patient collimator.

• This is mounted in front of the detectors.

• It’s purpose is to reduce scatter radiation from

reaching the detectors when slice thickness is less
than width of the detector.
• Some manufacturers have also used post patient
collimation to get accurate thickness selection for
the smallest slice widths.

• For multislice scanners, post-patient collimation is

neither necessary nor possible because the full
width of detectors in each row is used to form the
image.
• Shielding from scatter radiation would also shield
from direct radiation.

• The cross-section of patient is generally elliptical

so, that at the edges of the patient the ray path
from focus to detector passes through a relatively
low tissue thickness.
 This has the effect that noise levels

are poorly matched over the area of the transaxial

section,

being highest at the centre where doses are low

and also that the dose at the periphery

and on the skin is unnecessary high.

• To minimise this effect, some manufacturers
introduce additional shaped filters that are thin at
the centre and progressively thicker towards the
edges of the fan beam.

• These serve to equalise the transmitted

intensities emerging from the patient.
• They also even out the beam-hardening effect
across the projected fan beam.

• Because of their shape , such filters are

sometimes referred to as bow-tie filters.

• Generally different sizes of filter are used for head

and body scanning.
 “WEEK 2”

COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 DETECTORS
REQUIREMENTS FOR SCANNER
DETECTORS ARE AS FOLLOWS:
1. To be small in order to allow
good spatial resolution. For
single slice scanners with 600 to
900 individual detectors in the
detector bank, the width of each
detector is no more than about
1.5mm.
2.To have a high detection efficiency.

3. To have a fast response with negligible afterglow

so as to keep up with fast scanning times and rapid
changes in radiation intensity.

4. To have wide dynamic range. The x-ray intensity

may vary over range of 5000 to 1 between the
situation in which the beam passes by the side of
the patient with no attenuation to that in which it
passes through the lateral projection of a heavy
patient.
 5. To have a stable noise free response.

• There are two types of detectors used in CT

scanners. These are:

1. scintillation crystals

2. Xenon gas ionisation chambers

 Prior to the development of multislice scanners,
the detector of choice for most manufacturers was
the ionisation chamber.

• To have a reasonably high detection efficiency

linked with a small size, chambers had to be
relatively deep,i.e. to the direction parallel to
the x-ray beam.
• They used xenon gas because of their high
atomic number (Z=54) and K-shell energy
(35keV) and gas was kept at high pressure.

• These detectors have a detection efficiency

about 60%.
• Ionisation chamber detectors are not suitable for
multislice scanners and greater detection
efficiency can be achieved by solid state devices.

• Scanners with ionisation detectors are no longer

manufactured.
SOLID STATE DETECTORS

They incorporate a scintillant with an embedded

silicon photodiode to detect the light output.

• The scintillant may be cadmium tungstate,

bismuth germinate or rare earth ceramic.
 Detection efficiency may be very high,

up to 98%,

although, it is effectively less than this

because the detectors have to be separated

to prevent light crossover / cross-talk.

 DETECTOR CROSS-TALK
Occurs when a photon strikes a detector, is partially
absorbed and then enters the adjacent detector and is
detected again.
Cross-talk produces two weak signals coming from two
different detectors.
Cross-talk is bad because it decreases resolution.
Cross-talk is minimised by using a crystal that is highly
efficient in absorbing x-rays.
• If this geometrical factor is included, the efficiency
is closer to 80%.

• Solid state detectors can be made very small

down to about 1x0.5mm2

• They have negligible afterglow.

• They have a stable response.

 SCANNER GENERATIONS

• The earliest clinical computed tomography

scanner had an x-ray source and a single detector.

• Data acquisition involved moving both the tube

and and detector across the scanning plane to
acquire a series of transmission measurements.
• Detector and tube were than rotated 1˚ and
process repeated.

• In all data were collected through a 180˚ rotation.

• This scanner may be referred to as rotate-

translate type.
• To reduce scan times next generations of scanners
had, instead of a single detector, a bank of 30
detectors that could measure data simultaneously.

• But that were still insufficient to cover ..full cross-

section of the patient.
• Therefore rotate-translate procedure was still
needed but it became possible to reduce scan
times from just under 5min from the earliest
scanner to less than 20s.

• These are the data acquisition times for a single

slice.
• These two early designs came to be known as
FIRST and SECOND GENERATIONS.

• The next step in scanner technology was the

development of a scanner with large number of
smaller detectors arranged in an arc to cover the
complete cross section of the patient.
• This eliminated the requirement for the linear
translation of the tube and detectors and allowed
for the continuous data collection through a full
360˚ rotation.

• This is described as rotate-rotate or third

generation scanners.
• One of the technological problems with early
scanners was detector stability and this was made
worse by movement of gantry.

• This and other problems led to the development

of rotate-stationary scanner or the fourth
generation, in which the detectors were arranged
in stationary ring outside the path of the rotating
tube.

• This overcome some of the problems of detector

stability and and made reconstruction simpler.
• An additional advantage was that the outer part
of the fan beam could always pass outside the
patient and during each rotation every detector
would be able to measure the un-attenuated
radiation.

• This measurement could be used to adjust the

calibration of each detector throughout the
scanning cycle.
• However, the downside to the design is that the
total number of detectors is increased by a factor
of about 6 and this becomes prohibitive
particularly with multislice scanners.

• In addition, higher doses are required because of

the increased distance between the patient and
the detectors due to the fact that the tube has to
rotate within the detector ring.
A so-called 5th generation scanner more correctly

referred to as electron beam scanner was

introduced in

early 1980’s.
• It employed an electron source that produced an
electron beam that could be focused on to and
swept round a high-voltage target ring that
covered a 210˚ arc below the patient.

• X-rays would be produced and following

collimation detected above the patient on an
offset 216˚ ring of detectors.
• Because there were no mechanical parts, the
electron beam could be swept across the full arc
in no more than 50ms.

• This rapid imaging time permits imaging of the

heart, for which the system was designed.

• The generation terminology for describing

scanners is now largely redundant.
• With the advent of multislice scanners, the rotate-
rotate geometry of the third generation scanner
has become an industry standard.

• These scanners with up to 64 sub millimetre

detector rows and scan times 0.4s or less, linked
with gated data acquisition techniques and
improved reconstruction algorithms, provide
devices that produce genuine three-dimensional
images even of the beating heart.
 OTHER SCAN
CONFIGURATIONS

• Rotate rotate and rotate fixed CT scanners cannot

achieve scan times much shorter than 1sec
because of mechanical constraints.

• Rotating a heavy x-ray tube takes time.

Interest in faster scan times evolves from a desire
to

image moving structures, such as the wall of the

heart or

contrast material in blood vessels and heart

chambers.
• One proposed solution to the problem of
producing an image in a very short time is to use
multiple x-ray tubes.

• An example of this concept is the Mayo Clinic’s

dynamic spatial reconstructor, which uses 28 x-
ray tubes positioned around a semi-circular
gantry.
• The x-ray tubes are aligned with 28 light
amplifiers and Tv cameras that are placed behind
a single curved fluorescent screen that surrounds
the patient.

• The gantry of x-ray tubes and imaging systems

rotates about the patient at the rate of 15
rev/min.

• This system can acquire the data of an image in

about 16ms.
• Disadvantage includes high cost and the fact that
the mechanical motion is not eliminated.

• Another approach to very fast CT scans eliminate

all motion of the x-ray tube or detectors.
• Magnetic focusing and deflection of an electronic
beam replaces x-ray tube motion.

• In effect this system positions the patient within

the curvature of a giant x-ray tube.

• This device is commonly referred to as CVCT

(cardiovascular computed tomography scanner)
 IMAGE MATRIX SIZE FOR
DIFFERENT IMAGING
MODALITIES

Different matrix sizes are used for the different

imaging modalities, this is to produce a pixel size
that is compatible with the blurring and detail
characteristics of each modality.
Also, with many modalities, the matrix size can
be adjusted by the operator to optimize image
quality and the imaging procedure.
 EFFECT OF FIELD OF
VIEW ON DIGITAL IMAGE
DETAIL

When the field of view (FOV) is reduced, but not

changing the matrix size, the pixels become
smaller and the visibility of detail is improved.
A practical issue is that larger images (such as a
chest radiograph) require a larger matrix (more
pixels) than a smaller image in order to have
good detail.
 IMAGE COMPRESSION
• Image compression is
the process of
reducing the
numerical size of
digital images.

• There are many

different
mathematical
methods used for
image compression.
The level of compression is the factor by which
the numerical size is reduced.

It depends on the compression method and the

selected level of compression.
 Lossless
compression is when there is
no loss of image quality, and
is commonly used in many medical applications.
 Lossee
compression results in some loss
of image quality and must be used with care for
diagnostic images.
 “WEEK 3”

–COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 CARDIOVASCULAR COMPUTED
TOMOGRAPHY (courtesy; wikipedia)

CT scanning of the heart (CT coronary

angiogram) is a procedure used to assess the
extent of occlusion in the coronary arteries,
usually in order to diagnose
coronary artery disease.
 The patient is injected with an intravenous
contrast (iodine) and then their heart is scanned
using a high speed CT scanner, allowing
radiologists to assess the blood flow to their
heart muscle.
• useful in the diagnosis of suspected coronary
heart disease, for follow-up of a
coronary artery bypass, for the evaluation of
valvular heart disease and for the evaluation of
cardiac masses.
• The positive predictive value of cardiac CTA is
approximately 82% and the
negative predictive value is around 93%.

• In addition to the diagnostic abilities, cardiac

CTA beholds important prognostic information.
• Stenosis severity and extent of coronary artery
disease are important prognostic indicators.

• However, one of the unique features of cardiac

CTA is the fact that it enables the visualisation
of the vessel wall, in a non-invasive manner.
• Therefore, the technique is able to identify
characteristics of coronary artery disease that
are associated to the development of acute
coronary syndrome.
• Because the heart is effectively imaged more
than once (described above), cardiac CT
angiography can result in a relatively high
radiation exposure (around 12 millisievert).

• Although newer acquisition protocols, have

recently been developed which drastically
reduce this exposure to around 1 mSv.
• With the advent of subsecond rotation
combined with multi-slice CT (up to 320 slices),
high resolution and high speed can be obtained
at the same time, allowing excellent imaging of
the coronary arteries (cardiac CT angiography).
• Images with an even higher temporal resolution
can be formed using retrospective ECG gating.

• In this technique, each portion of the heart is

imaged more than once while an ECG trace is
recorded.
• The ECG is then used to correlate the CT data
with their corresponding phases of cardiac
contraction.

• Once this correlation is complete, all data that

were recorded while the heart was in motion
(systole) can be ignored.
• Images can be made from the remaining data
that happened to be acquired while the heart
was at rest (diastole).

• In this way, individual frames in a cardiac CT

investigation have a better temporal resolution
than the shortest tube rotation time.
• Dual Source CT scanners, introduced in 2005,
allow higher temporal resolution by acquiring a
full CT slice in only half a rotation, thus
reducing motion blurring at high heart rates
and potentially allowing for shorter breath-hold
time.
• This is particularly useful for ill patients having
difficulty holding their breath or unable to take
heart-rate lowering medication.
• The speed advantages of 64-slice MSCT have
rapidly established it as the minimum standard
for newly installed CT scanners intended for
cardiac scanning.
• Manufacturers have developed 320-slice and
true 'volumetric' scanners, primarily for their
improved cardiac scanning performance.
• Introduction of a CT scanner with a 160 mm
detector in 2014 allows for imaging of the
whole heart in a single beat without motion of
the coronary arteries, regardless of patient
heart rate.
• The latest MSCT scanners acquire images only
at 70-80% of the R-R interval (late diastole).

• This prospective gating can reduce effective

dose from 10-15 mSv to as little as 1.2 mSv in
follow-up patients acquiring at 75% of the R-R
interval.
• Effective dose using MSCT coronary imaging
can average less than the dose in conventional
coronary angiography.
• This prospective gating can reduce effective
dose from 10-15 mSv to as little as 1.2 mSv in
follow-up patients acquiring at 75% of the R-R
interval.

• Effective dose using MSCT coronary imaging

can average less than the dose in conventional
coronary angiography.
 CONTRAINDICATIONS
Pregnancy is considered an absolute
contraindication.
Since an iodine-containing contrast agent is
used, contrast agent allergy, hyperthyroidism or
renal function impairment are relative
contraindications.
Cardiac arrhythmias, coronary artery stents and
tachycardia may result in a reduced image
quality.
 IMAGE
RECONSTRUCTION

• The tube and detectors rotate smoothly around

the patient.

• In the modern scanner, x-rays are produced

continuously and the detectors are sampled (i.e.
measurements are taken) approximately 1000
times during the course of 360˚ rotation.
• A matrix of tissue voxels traversed by the fan
beam of x-rays are then incident on the array of
detectors.

• The signal measured by each detector depends on

attenuation co-efficient of the overlying voxels
that are traversed by the beam of x-rays as it
passes through the patient.
• In calculation process, account is taken of the
fraction of the volume of the voxel through which
the pencil beam passes.

• Each individual voxel is traversed by one or more

x-ray pencil beams for every measurement taken
during the full 360˚ rotation of the gantry.
• The attenuation of each voxel therefore
contributes to the measured transmission for a
large number of the ray sums.

• The logarithm of the ratio of the intensity of the

unattenuated pencil beam to that measured by
each detector after transmission through the
patient is equal to the sum of the linear
attenuation coefficients multiplied by the average
path length of the pencil beam in each voxel
through which the pencil beam has passed before
reaching the detector.
• The number of measurements taken in scanning
single section of the patients depend up on the
number of detectors and on the number of
measurements that are taken in the full rotation.

• With 800 detectors and with measurements taken

at 0.5˚ intervals, this amounts to some 576000
individual measurements.
• That is more than the total number of voxels for
which the attenuation co-efficient is to be
calculated, approximately 260 000 for a 512x512
matrix

• In theory CT numbers for each voxel could be

calculated from a series of simultaneous
equations.
• In practice this would not work for many reason,
including the influence of individual measurement
uncertainty on the final result.

• For the purpose of reconstruction, complex

mathematical techniques referred to as
‘algorithms’ are required.
(radiopaedia)
• HRCT images have very high levels of noise (due
to thin sections and high-resolution algorithm),
which may make them non-diagnostic for the soft-
tissues of the mediastinum.
• Intravenous contrast agents are not used for HRCT
as the lung inherently has very high contrast (soft
tissue against air), and the technique itself is
unsuitable for assessment of the soft tissues and
blood vessels, which are the major targets of
contrast agents.

• Inter-slice gap varies from 10-20mm,depending

on the examination.
 INSPIRATORY SCAN: Routine HRCT is obtained in
full inspiration.

It optimises contrast and reduces transient

atelectasis, a finding that may mimic or obscure
significant abnormalities.
EXPIRATORY SCAN:

is useful in obstructive airway disease or airway

abnormalities.
• Because the bases of the lungs lie posteriorly in
the chest, a mild degree of collapse under the
lungs' own weight can occur when the patient lies
on their back.
• As the very base of the lungs may be the first
region affected in several lung diseases, most
notably asbestosis or Usual interstitial pneumonia,
the patient may be asked to lie prone to improve
sensitivity to early changes of these conditions.
• The lung bases are often inconsistent in
appearance in patients due to the potential for
atelectasis causing positional ground glass or
consolidative opacities.

• When the patient is positioned prone, or on their

belly, the lung bases can expand further and help
distinguish atelectasis from early fibrosis.
• High resolution CT is a scanning protocol in which
thin sections (usually 0.625 to 1.25 mm) are
acquired and reconstructed using a sharp
algorithm (e.g. bone algorithm).

• It has been used for:

• lung imaging

• temporal bone imaging

• Two techniques have been used:

• spaced axial (non-helical)

• volumetric HRCT imaging.

• Spaced Axial

• Thin sections are acquired with an interval of 1-2

cm between the two sets of images. It has been
considered sufficient to detect abnormalities in
diffuse lung diseases. This technique is most
useful when single detector CT is being used. The
patient dose is less as compared to volumetric
imaging.
• Volumetric HRCT

• Thin sections are acquired continuously using

multidetector row CT scanners in a single breath
hold. Volume imaging with thinner slices allows
detection of a greater degree of pathology and
also allow reconstruction in any plane. Patient
dose is higher compared to axial
• Fundamental technical protocols

• slice thickness: 0.625-1.25 mm

• scan time: 0.5-1 second

• kV: 120

• mAs: 100-200

• collimation: 1.5-3 mm

• matrix size: 768 x 768 or the largest available

• FOV: 35 cm

• reconstruction algorithm: high spatial

frequency

• window: lung window

• patient position: supine (routinely) or prone (if
suspected ILD)

• level of inspiration: full inspiration (routinely

recommended) expiratory HRCT scans at three or
more levels in patients with obstructive lung
diseases.
 CARDIAC CT
CARDIAC CT FOR CALCIUM
SCORING
• Cardiac computed tomography (CT) for Calcium Scoring
uses special x-ray equipment to produce pictures of the
coronary arteries to determine if they are blocked or
narrowed by the buildup of plaque – an indicator for
atherosclerosis or coronary artery disease (CAD).
 The information obtained can help evaluate the risk for
heart attack.
A cardiac CT scan for coronary calcium is a non-invasive
way of obtaining information about the presence, location
and extent of calcified plaque in the coronary arteries—the
vessels that supply oxygen-containing blood to the heart
muscle.
• Calcified plaque results when there is a build-
up of fat and other substances under the inner
layer of the artery.

• This material can calcify which signals the

presence of atherosclerosis, a disease of the
vessel wall, also called coronary artery disease
(CAD).
• Theses people have an increased risk for heart
attacks. In addition, over time, progression of
plaque build up (CAD) can narrow the arteries
or even close off blood flow to the heart.
The result may be chest pain, sometimes called
"angina," or a heart attack.

Because calcium is a marker of CAD, the amount

of calcium detected on a cardiac CT scan is a
helpful prognostic tool.
The findings on cardiac CT are expressed as a
‘Calcium score’.

Another name for this test is

coronary artery calcium scoring.
The major risk factors for CAD are:

• high blood cholesterol levels

• family history of heart attacks
• diabetes
• high blood pressure
• cigarette smoking
• overweight or obese
• physical inactivity
 PROCEDURE

• The technologist begins by positioning the

patient on the CT examination table, usually
lying flat on back.

• Straps and pillows may be used to help

maintain the correct position.
• Electrodes will be attached to patient’s chest
and to an ECG machine that records the
electrical activity of the heart.

• This makes it possible to record CT scans when

the heart is not actively contracting.
• Next, the table will move quickly through the
scanner to determine the correct starting
position for the scans.

• Then, the table will move slowly through the

machine as the actual CT scan is performed.

• Patients are asked to hold their breath for a

period of 10 to 20 seconds while images are
recorded.
The extent of CAD is graded according to calcium
score:
Calcium Score
Presence of CAD
0
No evidence of CAD
1-10
Minimal evidence of CAD
11-100
Mild evidence of CAD
101-400
Moderate evidence of CAD
Over 400
Extensive evidence of CAD
 LIMITATIONS FOR CT CARDIAC SCORING

A person who is very large may not fit into the

opening of a conventional CT scanner or may be
over the weight limit—usually 450 pounds—for
the moving table.
CAD, especially in people below 50 years of age
can

be present without calcium (non-calcified plaque)

and

may not be detected by this exam.

 “WEEK 4”

–COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 BACK PROJECTION
• Most algorithms are based on the method of
back projection.

• Consider a cylindrical uniform body

incorporating a hole down the centre filled with
air.

• A beam passing through this body from one

direction will have a transmitted profile in it’s
central region.
• This single measurement cannot determine
position of the hole other than identifying that
it is in the line of the pencil beam passing
through the centre of the body.

• Pixel values along this line are decreased by

the amount calculated from the measurement
and these values are projected back as a band
of light grey.
• A second projection provides a second band of
grey.

• Progressive addition of projections produces the

reconstruction that clearly shows the hole
positioned at the centre of the body but with a
surrounding star like pattern.
 MATHEMETICAL
ILLUSTRATION OF BACK
PROJECTION

The object is
represented by a 5x5
matrix with cell values.

Considering just the two

projections that are
represented by by the
sums of the values in
each row and in each
column (b) and ©
The two back projected images are

then summed to give

the final representation, shown in (d)

In both examples, the final representation is not
an accurate reproduction of the original.

However, an approximate representation of the

relative values in each cell is demonstrated.
 FILTERED BACK
PROJECTION

• The images produced by back projection are a

heavily blurred version of the original object.

• To overcome this effect, a mathematical

process called ‘filtering’ is used.
• For each pencil beam measurement, the values
that are projected back into the overlying
voxels use the data not just from the pencil
beam itself but also from the neighbouring
pencil beams.
• Data from the neighbouring pencil beams can
be either subtracted from or added to the
projection to an extent that depends on the
distance between the beams, the closest
neighbouring beams having the greatest effect.
 FBP
Filtered back
projection is the most
common method
used for image
reconstruction.

There are different

filters and back
projection algorithms
that may be applied
that depend on the
imaging task, e.g. to
detect low contrast
detail or to enhance
resolution in bone.
 It should be noted that although standard
reconstruction algorithms are based on data collected
from a full 360˚ rotation.

It is possible to reconstruct data from data collected

over a reduced arc down to 180˚.

Such reconstructions may be used when imaging

rapidly changing scenes, such as for CT fluoroscopy
and for cardiac imaging.
https://www.youtube.com/watch?v=Bh
OMbjXzjP8
 ITERATIVE
RECONSTRUCTION

• An iterative reconstruction starts with an

assumption (for example all the points in the
matrix have the same value).

• It compares this assumption with the measured

values.
• Then makes corrections to bring the two into
agreement.

• Then repeats the process over and over until

the assumed and measured values are the
same or within acceptable limits.
There are three variations of iterative
reconstructions, depending on wether;
the correction sequence involves:

1. the whole matrix,

2. one ray or

3. a single point.
SIMULTANOEOUS RECONSTRUCTION:
All
projections for the

entire matrix are calculated at the beginning of

the iteration and

corrections are made simultaneously for each

iteration.
RAY BY RAY CORRECTION:
One ray sum is
calculated and

corrected and these corrections are incorporated

into future ray

sums, with the process being repeated for every

ray in each

iteration.
 POINT BY POINT CORRECTION:
The calculations and
corrections

are made for all rays passing through one point.

These corrections are used in ensuing calculations, again

with

the process being repeated for every point.

The figure in the previous slide shows a ray by
ray iterative reconstruction for a four element
square.
1. horizontal,
2. vertical and
3. diagonal ray sums,
are shown in adjacent blocks.
 In the first step,
the two horizontal ray sums (16 and 6 in the
hatched blocks)
are divided equally among the
two elements in the ray.
• If the ray sums had represented 10 elements,
the sum would have been divided equally
among all 10 elements.

• Next, the new numbers in the vertical row are

added to produce new ray sums (11 and 11)
and compared with the original ray sums.
• The difference
between the
original and
the new ray
sums

• (10-11= -1
and 12-11=
+1) is divided
by the number
of
elements in
the ray ( -1 ÷ 2
= -0.5 and +1 ÷
2 = +0.5).
• These differences are algebraically added to
each element.
• ( 8 - 0.5 = 7.5, 3 - 0.5 = 2.5, 8 + 0.5 = 8.5 and
3 + 0.5 = 3.5 )
• The process is repeated for diagonal ray sums
to complete the first iteration.
• In this example, the first iteration produces a
perfect reconstruction.

• With more complex data, iteration may have to

be repeated 6-12 times to reach an
acceptable level of agreement between
the calculated and measured values.
 COMPARISON OF
MATHEMATICAL
METHODS
No general agreement exists as to which
mathematical method is superior and indeed a
system’s superiority under one set of
circumstances may disappear under new
circumstances.

In terms of speed, analytical reconstructions are

faster than iterative methods.
A projection can be processed immediately
after recording, so the entire reconstruction can
be displayed seconds after the completion of
the last projection.
Iterative reconstruction may take several
minutes to process after scan completion.

Both methods are equally accurate if the

projection data are complete.
With incomplete data (i.e. , insufficient to
determine the image fully), iterative methods
are superior.

Analytic methods must do time-consuming

interpolations to fill in missing data.

Whereas iterative methods simply average

adjacent points.
 Thus, with incomplete data, analytical methods
are slowed

while iterative methods are actually made faster

than they

would be with complete data.

 SCANNED PROJECTION
RADIOGRAPHS

• A standard feature on CT scanners is the

facility to produce scanned projection
radiographs.

• These are known by a number of names:

scanogram, scout view and topogram.
• The scanned projection radiograph is a
transmission image taken at a fixed projection
angle.

• The collimator is generally set to the narrowest

slice width and the image is acquired as the
patient table is moved through the gantry.
• The resultant image has a relatively poor spatial
resolution when compared with standard
radiograph.

• However, one advantage of the technique is that

there is minimal scatter.

This achieved through continuous rotation of the

gantry without table movement.
• The appearance of the image is somewhat
different from the conventional radiograph
because there is no beam divergence along the z-
axis.

• Scanned projection radiographs may in principle

be taken at any projection angle, although in
practice only the AP, PA or lateral views are
normally required.

• The scans are used for planning the CT sequence,

i.e. selecting scan start and end points and
displaying scan slice positions.
COMPUTED TOMOGRAPHY FLUOROSCOPY:

scanners commonly have the facility for so-called

CT fluoroscopy; that is the display of CT image in
real time.

• Using fast reconstruction techniques, generally

from 180º data sets from frame rates of 5
frames/s or greater may be achieved.
• Most commonly such techniques are used for
biopsy needle placement.

• It should be noted, however, that although the

effective dose may be less than for a standard
diagnostic scan, there is the potential for
relatively high patient skin doses because
scanning is confined to a narrow region of the
body.

• Operators have to adopt techniques that will

minimise the risk of putting their fingers in the
beam.
 HELICAL AND MULTISLICE
SCANNING

• The simplest form of CT scanner can be used to

produce sequential transaxial slices with the
patient being moved along the z-axis of the
scanner between each succeeding slice.
• This is generally termed as axial scanning (step
and shoot term may also be used)

• Scanning like this is now a rarity other than a

few specific clinical protocols.
 To scan the lungs the CT scanner moves the
patient and scans from up to down.

For this the scanning area is divided into slices.

By scanning the slices one at the time, it takes a
long time to finish the scan, the patient might
move causing artefacts and the x-ray tube might
heat up too much.

By scanning with multiple slices and multiple

detectors, multiple images are made at once, so
the scan takes less time.
HELICAL SCANNING:

A major advance in scanner design was the

introduction of slip ring technology.

The x-ray tube has to be provided with a high

voltage supply and the detectors have to pass
their signals to the computer.
• If the equipment on the gantry was wired
directly to the external equipment, then the
gantry would be able to rotate for not vey
much more than 360º without the cabling
becoming entangled or over-stretched.
• In older equipment following a single rotation
of the gantry it would have to be returned to
the start position, resulting in a delay of
several seconds between each slice.

• Slip ring technology was introduced in the early

1980s.
• In it’s simplest form, slip ring comprises a
metal ring mounted on the gantry.

• This is connected to the signal output from the

detectors.
• Adjacent to the gantry there is a connector that
is able to retain good contact with the ring as
the gantry rotates.

• This connector takes the signal and pass it to

the computer system for image reconstruction.
• To, facilitate this, the detector signals have to
be sent sequentially at very high frequency in
order to accommodate total number of
individual detectors on the gantry and
sampling frequency.

• Additional rings are used to connect the x-ray

generator to the external voltage supply.
• Connecting the system by slip ring rather than
through hard wiring permits continuous,
unidirectional rotation of the x-ray tubes and
detectors around the patient.

• A basic advantage of continuous rotation is to

permit faster rotation times.
• If the gantry can
rotate through little
more than 360º, then
maximum rotation
speed is severely
limited, requiring 5s or
more per rotation.
• The more significant advantage is to permit
continuous acquisition of data while the table
moves the patient through the scan plane.
• Data for complete volume to be imaged can be
collected in a single exposure.
• This technique introduced in the early 1990s is
known as; volume/helical/spiral CT.
• It has largely replaced conventional axial
scanning.
 “WEEK 5”

COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
pitch=tabletop movement per rotation/ slice
thickness
 A helical scan can be visualized as a ribbon
wrapped around the body, e.g. scanner was set
up with a slice width 10mm and it had a rotation
time of 1s and if the table movement were at a
speed of 10mm/s then, with the analogy of the
ribbon wrapped around the patient, the edges of
the 10mm wide ribbon would be touching and
there would be full coverage of the body.
So, in this case pitch=1
• For a pitch >1, the ribbon edges would be
separated; if <1, the ribbon edges would
overlap.

• To reconstruct an image in helical scanning, the

measured data has to be interpolated from the
data collected.
• The relative weight given to each being
dependant on the relative distances from the
plane of reconstruction.
• The positions at which section scan be
reconstructed can be anywhere within the
scanned volume other than for a short distance
at the start and at the end of the scan for
which the data are not available for
interpolation.
• This means e.g. that if a 300mm long volume is
scanned with a 10mm slice width and a pitch of
2, there will have been only 15 rotations (plus
an extra one at each end for the purpose of
interpolation).
• From the measured
data, 30 contiguous
sections, each 10mm
thick, can be
reconstructed.
• Although the position of the reconstructed
sections can be selected retrospectively as
required, the slice width cannot be less than
the detector width.
• It’s not possible to extract finer details from a
10mm thick voxel.
• It is possible, however, to reconstruct on
greater slice widths that are multiples of the
detector width, effectively by adding adjacent
slices.This technique is used in multislice
scanners.
• Helical scanning has a number of advantages.
The principal advantage is;

• scan speed: which is beneficial for many

reasons, e.g. a scan of chest at pitch of 1.5
and a slice thickness of 10mm would take
between 15 and 20s at rotation speed of
1rotation/s.
• For example, a scan of chest at pitch of 1.5 and
a slice thickness of 10mm would take between
15 and 20s at rotation speed of 1rotation/s.

• For many patients, this is possible in a single

breadth hold.
• This helps minimise the risk of slice
misregistration.

• In sequential scanning, patients are asked to

hold their breath for each separate slice.
• Because the depth of breathing is liable to vary
over the time of the examination, anatomical
detail, particularly close to diaphragm, may
shift in a direction parallel to the scanner axis
between successive slices or to be missed
entirely.
• With a single breath hold, such slice mis-
registration cannot occur.

• The other advantages of speed are patient

throughput and the reduced use of contrast
medium.
• Increasing pitch above 1 is clearly
advantageous both in reducing exposure
timing and in reducing patient dose.

• However, as pitch is increased, resolution is

lost because of the need of greater
interpolation.
 It is a general rule that increasing pitch

beyond 2 gives unacceptable image quality

and for most clinical applications pitch

is not increased beyond 1.5.

• One of the consequences of helical scanning
and the longer continuous exposure times
associated with the technique (modern
scanners generally give exposure continuously
for 90s or longer) is that a greater load is put
on the X-ray tube.
 MULTISLICE SCANNERS

• In it’s original form, CT scanning was a two

dimensional imaging technique generating
sectional views of the body on a transaxial plane.

• It could only be thought of as a 3-D technique in

so far as multiple parallel sections could be
imaged.
 True 3-D imaging technique

requires isotropy;

i.e. the voxel size must be equal in

all three dimensions.

• Under those circumstances, data generated in
a three-D matrix can be reconstructed on any
plane and are not restricted to transaxial slices.
 In addition, with the development of ever more
powerful image-processing software it is possible
to display surface-rendered images, e.g. for
reconstruction surgery, maximum intensity
projection for angiography and virtual endoscopy
images among many other potential
applications.
• The voxel size in the trans-axial plane is
determined by the matrix size and the field of
view (FOV) but it’s typically in the region of
1mm.

• Single slice helical scanners generally have the

capability of collimating down to a slice width
of 1mm.
 Therefore a single slice scanner

theoretically has capability of a true 3—D

imaging but this is restricted by scanning time.

• The key to true 3-D images lies with the
multislice scanner.
• These scanners use solid-state detectors with
multiple rows of detectors.
• Total scan time is limited to about 90s because
of X-ray tube limitations.
• Therefore with 1s rotation speeds the volume
to be imaged is restricted to 90mm if maximum
resolution were to be achieved.
The scanner in fig. has 12 curved detector rows
with each row typically having about 800
detectors, as in conventional scanner.
• Unlike the conventional scanner, however,
each row is narrow because the length of the
detector along the Z-axis determines the
reconstructed slice width and not the
collimated width of fan beam.
• The detector rows are constructed to have the
minimum possible gap between them,
consistent with a separator to prevent light
crossover between detectors and to reduce the
effects of scatter produced in the detectors.
• In fig. the configuration is typical, with the
central eight rows having half the length of the
outer rows that are equally distributed on
either side.

• The length of the individual detectors (parallel

to the z-axis) in the central rows is typically
between 0.5mm and 1mm.
• The detectors rows may be used separately or
in combination.

• If, the central detector length (w.) were 1mm,

then the total detector length would be 16mm.
• This configuration could give the following
options;

(a) 8x1mm slices, (b)8x2mm slices, (c)4x4mm

slices, or (d)2x8mm slices.
The eight slice scanner (described here) is
capable of producing scans at four slice widths:

1, 2, 4, or 8mm

In each case slice width is determined by the

detector size and not by collimation.
• (in theory) collimator length is set to 8mm in
option (a) above or to 16mm for the other
three options.

• (in practice) collimation will be somewhat

greater in order to ensure that the outer two
rows are uniformly covered by the x-ray beam.

• Scanners with 16, 32, 40, 64 and even more

rows of detectors are now available.
MULTISLICE PITCH

• In single slice helical CT, pitch is defined as the

ratio of table movement during one full rotation
to slice (or collimator) width.

• In multislice CT it may be defined as:

Pitch = tabletop movement per rotation/ collimator

length
• That is essentially the same definition as the
single-slice helical pitch, except that slice width
is replaced by the collimator length.

• Strictly this should be written as nominal

collimator length that is equal to the total
detector length used for the scan.
 Values greater >1

implies that there is a gap in data collection,

whereas

pitch <1 implies an overlap.

• Some manufacturer’s have adopted an alternative
definition in which the pitch is defined as:

ratio of the tabletop movement to slice width.

The former definition may be referred to as

a beam pitch and the

later as

slice or row pitch.

MULTISLICE RECONSTRUCTION:

In principle, multislice and single-slice helical

reconstruction techniques are the same.

• They involved interpolation of measured data

between the adjacent data sets.
• However, in multislice scanning to reconstruct
in single plane within the scanned volume the
data have to be interpolated for different pairs
of detectors as the rotation progresses around
the full 360º cycle.

• This is further complicated by the “cone beam

effect”.
 CONE BEAM EFFECT

• Standard
reconstruction
algorithms assume
that the x-ray beam
is non divergent in
the z-direction.
• This a reasonable assumption with a single-
slice scanner, as the measurement made by
the detector is due to voxels lying in flat plane
centred on the plane of rotation of the gantry.
• In opposing projections, the beam length at the
rotation axis that defines the slice width exactly
matches the length of voxel on the axis.
• For a voxel displaced from the axis, the beam
length is marginally smaller or larger than the
slice width.

• This comprises the accuracy of calculation of

CT number in that voxel, but even from a
10mm slice, the maximum generally
encountered on a single slice scanner, this
effect is marginal.
• In a multislice scanner (8 slice), the object on
rotation axis contributes to the ray sum of
detector 2 and would be imaged on the slice
plane defined by that detector.
• However, the off-axis object that lies within the
same slice plane contributes to the ray sums of
detector 1 for the downward projection and
detectors 2 and 3 for the upwards projection.

• In effect the slice reconstructed from detector

2 relates to voxels within the shaded regions,
that would trace out a double in shaded region.
• The cone beam effect
becomes more
significant with
increased number of
slices and with
increasing total detector
length.
• Cone beam algorithms
had to be introduced to
minimise the
significance of this
effect.
CONE BEAM ARTEFACT:

if there is inadequate correction for this, an

artefact is seen as a blurring of the boundaries
between high-contrast details (e.g, at the
boundary of bone and soft tissue).
• As the tube and detectors rotate around the
patient (in a plane perpendicular to the
diagram), the data collected by each detector
correspond to a volume contained between two
cones, instead of the ideal flat plane.
• This leads to artefacts similar to those caused
by partial volume around off-axis objects.
• The artefacts are more pronounced for the
outer detector rows than for the inner ones,
where the data collected correspond more
closely to a plane.
• Cone beam effects get worse for increasing
numbers of detector rows.
• Thus, 16-section scanners should potentially be
more badly affected by artefacts than 4-section
scanners.
 THREE-D IMAGE PROCESSING AND
DISPLAY

There are a number of image processing

and display techniques that are

commonly available for three dimensional

imaging with multislice computed tomography

scanners.
MULTIPLE REFORMATTING: permits display on
any plane.

• This is commonly used to display images on

the coronal and sagittal planes but oblique
plane display is also possible.
CORONAL AXIAL SAGITTAL
 Figure 9. Sixty-eight year old female patient with left
flank pain. CT stone protocol through the left kidney in (a)
axial, (b) coronal, and (c) sagittal planes revealed a small
calcified left kidney, consistent with putty kidney.
 “WEEK 6”

COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 3-D IMAGING

Three dimensional reconstruction of the CT scan

data is an option now available on many
contemporary units.
Wether the technology will prove to be of
significant clinical usefulness is still being
evaluated.
There are two methods of CT reconstruction:

1. Surface reconstruction shows only the

surface of the object.

2. Volumetric reconstruction shows the surface

of the object ‘in relation to its surroundings’.
 3-D reconstruction is simply the result of
computer manipulation of the CT scan data
obtained in a routine manner.

That is, there is no additional information

obtained for the specific purpose of obtaining a
3-D image.
 It is true that a high resolution set of CT scans is
obtained for good 3-D reconstruction; a typical
scan might consist of 4mm thick slices with no
gap between slices.
 SURFACE
RECONSTRUCTION

Display of 3-D surface of an object can be

accomplished if the object is displayed with
shading.

By making surface appear opaque (e.g., one

cannot see the petrous pyramids through the
maxilla when looking at the front of a facial
image.
Essentially the computer is told to scan the
data bank of CT numbers and find all the CT
numbers corresponding to the CT number of
the object to be displayed.
 For example, we maybe interested in a bone
and the CT numbers searched for would be of
high value.

Similarly the computer may be told to search

for soft tissue CT numbers.
 The computer has the task of locating the
required CT numbers, then must keep tract of
the pixel and slice location at which those CT
numbers occur.
Once, the pixels have been selected and
located, 3-D reconstruction of the surface may
begin.

There is sufficient information to present the

object from any orientation and the selection of
which orientation must be made by the
operator.
In the skull, e.g,. one may choose anterior,
posterior or either lateral view.

In fact, some programs allow rotation of the

object, which is simply the selection of different
views in rapid sequence.
Once a view is selected, shading of pixels is
used to enhance the perception of depth in an
image.

Pixels that are perceived as being closer to the

viewer maybe shaded light.
Those that are perceived as farther away are
shaded darker.

Pixels that are perceived to be behind an

opaque structure will be disregarded.
The computer must at times interpolate to form
a continuous image from matrix data points.

Such things as artefacts, gaps between slices

and volume averaging will degrade the image.
 VOLUMETRIC RECONSTRUCTION

Volumetric reconstruction requires two additional

things;

1. the first is to give computer a window of CT

numbers to look for.

2. Second, the computer must display the pixels

as translucent.
The window represents the material
surrounding the area of interest in the
volumetric reconstruction. e.g,. one may wish
to display the muscles surrounding the bony
pelvis.

Therefore, at least two windows of CT numbers

would be set (one for bone and one for
muscles).
Displaying the pixels as translucent, means
that objects behind the displayed pixels will
also be displayed, e.g,. transparency will allow
us to see bone surfaces through overlying
muscles.
Surface reconstruction allows to see objects as
if they were anatomic specimens (i.e., their
surfaces are opaque).

Volumetric reconstruction represents objects

more like they would appear during
fluoroscopy.
In reality, one can show volumetric
reconstruction with only one window of CT
numbers by requiring the computer to assign
transparency to an object perceived to be
closer to the viewer.
For example, one may choose to view the bony
pelvis isolated from the surrounding soft tissues
but still displayed with volumetric
reconstruction.

Such a display would allow visualisation of the

acetabulum through the femoral head.
VIRTUAL COLONOSCOPY

• Virtual colonoscopy using CT is a 3D method of

analysis.

• It is very much like a colonoscopy.

• With this procedure, the patient undergoes the
same bowel preparation procedure but instead
of undergoing traditional colonoscopy, they
undergo CT scanning.
 CARDIAC ANGIOGRAPHY

• Cardiac angiography with volume CT scanners

is also now a reality, and, in fact, because of
the 3D rendering capabilities, one is able to
make a better differential diagnosis and have a
much safer procedure relative to putting a
catheter internal to the body.
• "CTA for evaluation of the renal arteries,
evaluation of the aorta, the peripheral vascular
system from the iliac down to the lower
extremities, particularly down near the foot, is
immediate, painless, and minimally invasive.
• It is useful for diagnostic work to determine
which patients actually need a dedicated
angiogram with a catheter and a therapeutic
procedure.
• CTA is incredible, especially with the new
scanners that can get thinner and thinner
sections, providing exquisite detail.

It's replacing diagnostic catheterisation and in

the future, it can replace diagnostic peripheral
angiography.
• Potential pitfalls

• It requires careful evaluation of the images to

make sure that one isn't being fooled by the
reconstruction.
• 3D reconstruction is one more step away from
what is really there.

Every step away has the potential for

misinformation or lost information.
• In addition, certain 3D procedures require a
large bolus of contrast media, which can be a
concern in patients who have underlying renal
disease or in those with contrast sensitivity.
• There is also concern regarding increasing
radiation dose with more frequent use of MDCT
imaging.

• Generally, the end benefit outweighs the small

theoretical risk from radiation, but it should be
used judiciously.
THREE-DIMENSIONAL SURFACE
RENDERING

This surface-
rendered 3-
dimensional
reformatted
CT image
from a lateral
projection
shows the
displaced
radial head
fracture
(arrow).
• It’s a technique in which the surface of a
structure can be displayed.
• It’s achieved by selecting a suitable range of CT
numbers that are associated with that
structure.
• Shading can be added to the reconstructed
surface (shaded surface display) by considering
the effect of light on the structure coming from
a virtual light source.
Three-dimensional surface rendering of the colon
 “WEEK 7”

COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 MAXIMUM AND MINIMUM INTENSITY
PROJECTIONS
MIPs and MinIPs

• These are formed by

projecting the
volume of interest on
to a viewing plane
and displaying the
maximum (or
minimum) CT
number along each
projected ray that
forms the image.
maximum intensity
projection (MIP)
• In effect, it has the appearance of the standard
projection image with the background
subtracted.
• MIP images may be used for CT angiography
and
• MinIP for imaging the tracheo-bronchial region.
MPR (left), external volume rendering (middle) and internal volume r
 True tracheal bronchus (Type III) in a 49-year-old man with
chronic cough and dyspnea.
Coronal volume rendering (A) and minimum intensity projection
(B) images showing an entire right upper lobe bronchus (open
arrow) originat- ing from the lower trachea. The lower trachea
distal to the origin of tracheal bronchus is markedly narrowed.
The diameter of the right main bronchus is less than or equal to
that of the left main bronchus.
Coronal VR and MinIp
 THREE-DIMENSIONAL VOLUME RENDERING
TECHNIQUE

• 3-D techniques are

formed by projecting
the volume of interest
on the viewing plane.
• However, instead of displaying the maximum
voxel values along each ray path, selected
ranges of CT numbers may be displayed with
varying levels of opacity or with different
colours to demonstrate e.g, not just the
contrast filled structures but also the adjacent
structures.
3-D VOLUME RENDERED
IMAGES
 VIRTUAL ENDOSCOPY

• It’s also a three-

dimensional surface
rendering technique
used to display the
internal walls of body
structures such as
the colon.
• These may be referred as “fly through”
projections.
• A Fly-Through function reproduces the complex
rotation, required for an extended camera shot.
Diagram of approach to virtual endoscopy
using volume modelling
Preliminary results suggest virtual angioscopy using

volumetric 3D rendering techniques as a potentially

useful technique for the noninvasive evaluation of

vascular pathology.
 IMAGE QUALITY

SPATIAL RESOLUTION: The maximum high

contrast resolution that may be achieved with on a
CT scan is about 20 lp/cm.

• Resolution in CT is more commonly expressed in

terms of line pairs per cm rather than per mm.

• High contrast and spatial resolution within the

scan plane is clearly limited by pixel size.
• Pixel size is dependant on matrix size and FOV
e.g, a 512x512 matrix used with a 40cm FOV has
a pixel size of just under 0.8mm, implying a
resolution of 6 lp/cm.

• However, because FOV is selectable, so, is pixel

size.
• Therefore, pixel size may determine resolution on
a particular reconstructed image but it is not an
intrinsic limitation.

•There are other factors that determine the

maximum resolution that may be obtained for a
particular scanner.
• ALGORITHM

• is the 2nd reconstruction parameter that

affects resolution.

• Certain algorithms are designed specifically for

bony structures and serve to enhance the
edges of high-contrast structures.
• These algorithms provide improved spatial
resolution at the cost of increased noise.
WIDTH OF THE PROJECTION PATH OF X-RAY BEAM:

as it passes through the patient . . is a factor

intrinsic to the scanner, affecting resolution.

This is affected by:

• focal spot size,

• geometry of the scanner (i.e. source to isocentre
and source to detector distance) and

• the physical size of the sensitive area of the

detector itself.
SAMPLING FREQUENCY:

or number of projections sampled on each rotation.

• Within limits, spatial resolution increases with

sampling frequency.
• The spatial resolution in transaxial plane is not
significantly different when comparing axial and
helical scanning.

• Spatial resolution in the z-direction, parallel to the

rotation axis depends on slice thickness.
• In single slice helical scanning using the simple
form of interpolation, data collected outside the
reconstructed section contribute to the transaxial
image, the resolution in z-direction is thus
reduced.
• The greater the pitch, greater is this effect.

• It is seen to a much lesser extent in multislice

scanners because of the algorithms that are
applied to overcome the cone-beam effect.
• Which means that although multislice scanners
have a greater tendency that the images
acquired by them are more susceptible to cone
beam artefact “but their algorithms have saved
them”.
 NOISE

• Noise is the fundamental limit to the quality of the

CT image.

• It both reduces CONTRAST resolution of small

objects and worsens the spatial resolution of low
CONTRAST objects.

• There are three sources of noise.

1. Quantum noise: caused by random variations in
the number of photons detected.

2. Electronic noise: produced in the measuring

system.

3. Structural noise: that is affected by the

reconstruction algorithm.

The least important, is, electronic noise.

• The quantum mottle or noise, is a variation in the
number of x-ray photons absorbed by the detector.

• At the detector, noise is still determined by the number

of photons absorbed (detected).
• QUANTUM noise may be reduced by increasing
the number of photons contributing to each
projection.

• There are three principal factors that affect

quantum noise:
1. mA

2. scan time

3. slice width

The first two of these are essentially linked; they

both serve to increase mAs per rotation.
Increasing slice width, while reducing noise, has an
adverse effect on image quality because of the
reduced spatial resolution and the increased partial
volume effect.
 SNR ratio is directly related to the square root of
the number of photons detected, therefore, it’s
increased by the square root of mAs and of slice
width.
 QUANTUM NOISE

• QUANTUM NOISE is influenced by kV.

• At increased kV, penetration is greater and

therefore a greater number of photons are
detected.
• Even if mA are reduced to restore dose to the
same level as a scan at lower kV, the number
of photons detected would be increased and
noise reduced.

• However, the influence of increased kV on

noise may be offset by a reduction in contrast.
• The only way to decrease noise is to increase
the number of photons absorbed by the
detector.

• The way to increase the number of photons

absorbed is to increase X-ray dose to the
patient.
• Precision cannot be improved by a more
accurate mathematical reconstruction.
• In fact, mottle becomes more visible as the
accuracy of the reconstruction improves.

• When high quality images are essential, scan

times may have to be lengthened to produce a
statistically more accurate reconstruction.
FOV AND MATRIX SIZE: also influence noise.

• For a reduced FOV or,

• for a lager matrix,

there is a smaller detector area defining each pixel.

Thus, there are effectively fewer photons per pixel

and so, noise is increased.
• The image is displayed on a television monitor
as thousands of tiny pixels.

• Each representing a tiny specific cross-sectional

area in the patient.
• A pixel of size such as, 3mm refers to the size
of the patient.

• Their actual size in a video image is determined

by the size of the television monitor.
• A given size of pixel will appear larger on a
large TV monitor than it would on a small TV
monitor BUT in both cases it would represent
same area of the patient.

• A voxel adds third dimension to the area

represented by the pixel.

• A 1mm square pixel might represent a 5mm or

10mm thick section voxel.
• Field size dictates the maximum size of anatomic part
that can be examined.

• The smaller field size may be used for head scans while
the larger size could accommodate a larger part and
might be used for body scans.
• For single slice scanners pitch does not affect
noise.

• Increased pitch reduces the dose but the data

are interpolated between the successive
rotations for a 360º scan and the same number
of detected photons is used in the
reconstruction.
• The downside to increasing pitch is not an
increase in noise; rather it is an effective
broadening of slice width that has an
impact on spatial resolution and partial
volume effect.
• The same is not true for multislice
scanners.

• Reconstruction algorithms for multislice

scanners also interpolate data between
adjacent rotations.

• However, because of the cone beam

effect, reconstruction is more complex
than single slice scanners.
• Interpolation involves not only the two detector
rows spanning the reconstruction slice plane at
the particular gantry angle; instead data is
used from all detector rows.

• In the interpolation algorithm, values may be

either added or subtracted at each projection
angle.
• The net effect is to produce a scan slice
that has minimal broadening in the z-
direction but this is at the cost of
increasing noise with increasing pitch.
• Noise become more apparent as window width
is reduced.

• However, noise is not increasing and the SNR

ratio remain the same.

• Progressive reduction in window width will

reveal low contrast detail.
But if, contrast between adjacent structures
is comparable with the magnitude of the
noise, no amount of windowing will improve
the ability to resolve two structures.
 CONTRAST RESOLUTION

• It may be defined as the ability of an imaging

system to display an image of relatively large
(2 or 3mm) object that is only sightly different
in density from it’s surroundings.
• For the image to be visible, the object must
produce enough change in the number of
transmitted photons to overcome statistical
fluctuations in transmitted photons caused by
noise.
Low contrast visibility is determined by noise.

The more homogeneous the background. .

the better the visibility of low contrast images

and of course

homogeneity of the background is a function of

noise.
• The only way that resolution can be
improved for low contrast objects is with
better counting statistics, which produces
a more homogeneous background.
• With film-screen radiography, a density
difference of about 10% is required for contrast
resolution.

• Contemporary CT scanners can display objects

about 3mm in diameter with density
differences of 0.5 % or less.
• Even with improved low contrast resolution some
pathologic processes remain undetected.

• Use of iodinated contrast materials to enhance subject

contrast is an accepted practice in most head and body
CT scanning protocols.
 • Attempts to increase spatial resolution may cause
increased noise levels that decrease contrast resolution.

• Attempts to increase contrast resolution by increasing

scan times (larger dose) may decrease spatial resolution
by increasing the effects of patient motion.
 IMAGE ARTEFACTS

MOTION ARTEFACT

Cardiac motion produces

streak artefact (black and
white bands).

• The reconstruction process

is misled by a moving
structure occupying
different voxels during the
scan.

• Mechanical misalignment
and movement of the
patient have similar effect.
• Patient motion is the primary reason for developing
faster scan units.

• When the patient moves during scan acquisition,

reconstruction programme has no ability to make
appropriate corrections because motion is random and
unpredictable.

• The reconstructed image will display an object in motion

as a streak in the direction of motion.
• On reconstruction the scanner will average the density of
pixels covering the motion area.

• In some fashion, the intensity of the streak artefact will

depend on the density of the object in motion.

• Motion of objects that have densities much different from

their surroundings produces more intense artefacts.
• Thus, motion of metallic or gas containing structures
produce striking artefacts.
 HIGH ATTENUATION OBJECTS/
STREAK ARTEFACTS

• Metal implants,
dental amalgam, etc.
give rise to streak
artefacts that may
obscure the area of
interest.

• The streaks appear as

dark and light lines,
emanating from the
high attenuation
material.
• The effect is accentuated by motion.
• Small areas of bone or contrast medium can have a
similar effect.
• Modern scanners commonly include metal
correction algorithms that can be used to minimise
the appearance of this artefact.
 PHOTON STARVATION

• This is the variant of

the streak artefact.

• A typical example
might occur in the
pelvis scan of a
patient with bilateral
hip implants.
• On the lateral projection, the x-ray attenuation
through the two metal implants is very high and
may be outside the dynamic range of the
processing unit.
• As a result, the tissues between the hips are poorly
represented in the displayed image and horizontal
streaks appear across the image.
CT image shows streaking artifacts due to the beam
hardening effects of contrast medium.
LINK: https://pubs.rsna.org/doi/10.1148/rg.246045065#F4
 The most common method of avoiding photon
starvation artefacts is to employ tube current modulation
(or ‘mA modulation’).

In many modern scanners, the tube current (mA) can be

varied as the gantry rotates around the patient.

A higher mA is used for the more attenuating projections

and a lower mA for the less attenuating projections.

The mA to be used is calculated either in advance from

analysis of the scout view or
during the scan via a feedback system from the
detector.
 CT images of a patient with metal spine implants,
reconstructed without any correction (a) and with metal
artifact reduction (b)
https://doi.org/10.1148/rg.246045065
CT image of the body obtained with the patient’s arms down
but outside the scanning field shows streaking artifacts.
 “WEEK 8”

COMPILED BY DR. SARAH MARYAM

ASSISTANT PROFESSOR RADIOLOGY (UIRSMIT)
UNIVERSITY OF LAHORE
 BEAM HARDENING
ARTEFACT

• X- ray beam producing the CT image is not a

monochromatic beam.

• The beam contains x-rays with many energies, although

we speak of average energy of a CT x-ray beam as being
about 70 keV.
• As the beam passes
through the patient, the
low energy photons are
filtered (absorbed) out and
the beam becomes harder.

• The tissues towards the

centre of the patient are
crossed by hardened
beams.

• Whereas those nearer the

surface have not been
filtered to the same extent.
• This results in lower CT numbers in the centre of the
patient. This effect may also be described as
‘cupping’ / cup artefact.
• In some situations it may also result in dark streaks
in the image.

CORRECTION:

(1) beam-hardening algorithm and by using a

(2) bow-tie filter.

• Generally, beam hardening artefact is not a problem.

• Reconstruction programmes anticipate and correct for

variation in linear attenuation coefficients caused by
beam hardening but such corrections are not precise.

• The linear attenuation coefficients from the centre of the

anatomical region will be decreased compared to the
periphery and the reconstructed image will be less dark
centrally than peripherally.
STRATEGIES TO MINIMIZE BEAM HARDENING ARTEFACTS
INCLUDE:

1. patient positioning:
for example, tilting the gantry for head
scans to avoid imaging through regions of densest bone or
dental fillings

2. Iterative beam hardening corrections:

these may be
incorporated into the reconstruction on some scanners.
From an initial reconstruction, the image is segmented to
identify areas of bone.
 The beam hardening effects of the
bone are modelled and used to pre-correct the data.

This process may be repeated iteratively to produce a

final image essentially free from beam hardening artefacts.

However, the reconstructing time is consequently

increased.
RING ARTEFACTS
• Ring artefacts are the result of miscalibration of one
detector in rotate-rotate geometry scanner. (of course,
detector failure would also produce a ring artefact.)

• If, a detector is miscalibrated , it will record incorrect

data in every projection.

• This misinformation is reconstructed as a ring in the

image.

• The radius of the ring is determined by the position of

the faulty detector in the detector array.
 RING ARTEFACT

• The ray passing from the tube focus to a

particular detector in the row of detectors traces
out a circle as the gantry rotates about the
patient.
• CT numbers of the voxels that lie on that circle
are calculated from the signal derived from that
detector alone.
 An operational feature of CT scanners is that ,
when they’re switched on each day, automatic
calibration programs are run to check and adjust
the calibration of detectors to ensure that they
provide a balance response.
• Should that detector malfunction, CT numbers in
that ring will be incorrect and light or dark ring will
be seen in that image (depending on, whether that
detector is giving too high or too low a signal)
• This artefact can be seen on both single and MDCT
but in case of MDCT larger changes in sensitivity
are needed are needed to make artefact visible.
• Faulty detector in rotate-fixed units also record false
information.

• However, this information is not visible in the

reconstructed image because the faulty detector collects
data from many angles (rings may appear in rotate-fixed
geometry if the x-ray tube is not aligned correctly).

• Ring artefacts have virtually disappeared in

contemporary CT units.
• With contemporary CT units, artefacts due to equipment
defects are rare.
• Short scan times have diminished but not eliminated
motion artefacts.
• Artefacts at the end of high contrast objects may also be
seen (streak artefacts).
• Beam hardening artefacts are minimised by
reconstruction programmes.
• The one equipment defect that may be seen is a “ring
artefact".
 DOSE

• CT scanning is one of the highest dose

techniques used in medical imaging.

• Multislice scanners give somewhat higher

doses than single slice scanners.
DOSIMETRY PARAMETERS: CT dose index
(CTDI) is a measure of dose from a single
rotation of the gantry.

CTDI is measured using a pencil ionisation

chamber.
This has a diameter that is typically about 8mm
and length of 100mm, which is significantly
greater than the maximum slice width that is to
be measured.

The measured dose multiplied by the length of

the chamber represents the integration of the
dose profile and division by the width equals
CTDI.
• Figure shows typical dose profile along the z-axis for
an axial scan.
• It’s drawn for a beam collimated to a width of
10mm.

• It has a relatively flat peak and the peak value

is the maximum absorbed dose in the scan
plane at the position of measurement.

• CTDI is defined as:

CTDI= ∫D (z) • dz
T
 Geometric dose
efficiency.

(A) If MSCT detectors

configured to acquire
four 2.5-mm slices are
irradiated with 10-mm-
wide x-ray beam, as
specified for SSCT, outer
2 slices will receive lower
intensity and yield higher
image noise.

(B) To compensate, MSCT

beams are widened to
use only inner, non-
penumbra regions.
Penumbra regions that
were partially used in
SSCT are discarded in
MSCT, leading to reduced
 • In this, the dose at position z, D (z), is
integrated over the complete

dose profile and divided by the nominal slice

width, T.
• The peak dose in the profile shown in the figure
is virtually

independent of slice width, it depends only on:

1. scanner design;

2. filtration

3. kV and

4. mAs settings.
• So, if, collimation is accurate and the profile
width is equal to the nominal width, the
integrated dose is directly proportional to width
and CTDI is constant with width.

• CTDI is a useful parameter to assess the dose

within the scan plane.
• If, we consider the ideal square beam profile, it
describes approximately the same area as the
true profile.

• Therefore, CTDI is a close approximation of the

maximum dose at the position of the
measurement.
• There are internationally accepted standards
for the CTDI that involve measurement in
cylindrical perspex phantoms.

• These have diameters of 16 and 32cm to

represent head and body scans, respectively.
• Measurements are made in the centre of the
phantom and in several locations on the
periphery, in order to derive a weighted value
(CTDIw) for the complete cross-section that is
an approximation to the average dose.

• For most standard scanning protocols, CTDIw is

in the range 10-40mGy.
• Dividing CTDIw by pitch (beam pitch in case of
multislice scanner) gives CTDIvol , which is an
approximation to the average absorbed dose
within the scanned volume.

• Maximum skin dose is approximately equal to

CTDIw for scans of the head and 20% higher
for the body scans.
 E/DLP ratio is much lower for head than for
trunk. It’s also lower for chest than for scans
involving pelvis.
 Skin doses in CT are much higher than for
radiographic examinations but not as high as
doses from prolonged fluoroscopy.

• The CTDI is useful for describing the dose

efficiency of a scan protocol and for
comparisons between scan models.
 Typical equipment used for CT dose index
measurements. Courtesy of Thomas Flohr and Berndt
Ohnesorge, Siemens, Erlangen, Germany.
• CTDI, is less useful for for comparison of doses
to individual patients because attention has
also to be given to the length of scan and to the
number of times a particular region might be
scanned e.g. pre and post-administration of
contrast medium.
• Dose-length product (DLP) is defined as:

• DLP = CTDIvol X L,

• ‘L’ is the total scan length

• The length can be estimated from the total

table top movement.
• Helical scanning and particularly multislice
helical scanning, there’s some over-
scanning, with an additional rotation at
the beginning and end of the scan.
• This degree of over-scanning, may be
particularly significant for short scan lengths.

• Length may more correctly be estimated from

the number of rotations multiplied by
collimated length and pitch.
• Effective dose ‘E’ can be derived from DLP
using various published data sources.

• Conversion co-efficient depend on the body

region and on the scanner design.
These differences are due to the distribution
of the organs and tissues having the highest
weighting factors (WT)

that are used for the calculation of the

effective dose.
• Tissues and organs in the head have
relatively low weighting factors,whereas
the highest WT for the gonads, resulting in
the highest E/DLP factor being associated
with scans of the pelvis.
FACTORS AFFECTING PATIENT DOSE
(OPERATOR-SELECTABLE
PARAMETERS)
• All the selection parameters are
interdependent.

• In setting up a new protocol, the parameters

that the radiologist and radiographer may
consider are likely to include:

1. kV

2. mA

3. rotation time
4. Reconstructed slice width

5. helical or axial acquisition.

6. pitch

7. number of detector rows

8. reconstruction algorithm

9. FOV
The technologist monitoring the examination can
control most of these parameters and modulate
them to obtain the desired image quality.

In addition the dose depends on scanned length

and on patient build.
 kV
• It’s normally set to 120 kV, but for areas of
high attenuation a higher setting may be
required.

• Radiation output increases with kV (by about

40% for an increase of 20kV) but attenuation in
the patient is lowered.
 Therefore, higher kV increases dose but for the
same

level of noise in the image the mA maybe reduced

sufficiently to give an overall dose reduction to the

patient.
• The influence of mA and rotation time maybe
taken together as mAs per rotation.

• This is a key parameter to be selected when

establishing a scanning protocol because of the
direct relationship between mAs and dose and
the inverse dependance of noise on the square
root of mAs.
• Scan protocols should be based on the
maximum acceptable level of noise to give the
optimum balance of image quality and dose.

• Commonly scanners include an option for

mA modulation.
• In this technique, mA is adjusted with an
aim of achieving a near constant detector
signal so that changes in patient size
through the scan region do not
significantly influence the level of noise.
• The techniques used are manufacturer
dependant but in principle the aim is to use
higher mA on the thicker lateral projections
compared with anteroposterior.

• Adjust the mA in accordance with changes in

patient cross-section over the length of the
scan and for the patient to patient variations
in build.
• Modulation is based on measured transmission
through the patient, using either data from
prescan scanned projection radiographs or
from the previous rotation.

• Dose reduction in the range of 10-40% may be

achieved in comparison with constant mA
techniques.
 a b c

High-dose CT may not improve image quality

substantially.
Corresponding transverse CT images acquired at (a)
256 mAs, (b) 176 mAs, and (c) 88 mAs,
with remaining scanning parameters held constant,
in a 72-kg 62-year-old man.
Image quality was deemed acceptable at all three
tube currents.
• Any decrease in tube current should be
considered carefully, because such reduction
causes an increase in image noise, which may
affect the diagnostic outcome of the
examination.

• This is especially true in abdominal studies,

where low-contrast areas are severely affected
by an increase in image noise.
• The pelvis, however, with its greater inherent
contrast is usually not noticeably affected.

• According to a recent review of scanning

protocols diagnostic quality abdominal CT
scans can be obtained at lower tube currents
with a four detector row scanner.
 NUMBER OF DETECTOR
ROWS Strategies for CT Radiation
Dose Optimisation RSNA 2004
(Pub Med)

In principle, the number of detector rows used

together

in a multislice scanner has no influence on dose

but

some models of scanner have a better

geometrical

efficiency than others.

• Nor is there any direct influence of the
reconstruction algorithm or the FOV.

• However, both of these last two do influence

noise and image quality and influence the
choice of mA.
 SCANNER GEOMETRY

• The distance between the focal spot of the x-

ray tube and the isocenter of the scanner
depends on scanner geometry.

• Since single– or multi– detector row helical CT

scanners can have a long or a short geometric
configuration.
• According to the inverse square law, radiation
intensity varies with the inverse of the squared
distance between radiation source and patient.

• Thus, if all other scanning parameters are

identical, a short-geometry scanner can
produce more interaction of radiation with the
patient and lower image noise than a long-
geometry scanner can.
• The first step, usually performed by a
physicist, is to measure the weighted CTDI in a
phantom.

• This provides basic calibration data for the

scanner.

• By using the KV, Time, and mA values for a

specific patient and the calibration data the
weighted CTDI for the patient can be
calculated.
• The next step is to use the pitch value
(spiral/helical scanning) to calculate the
volume CTDI.

• This is often calculated and displayed by the

scanner as part of the patient information.
• Multiplying by the length of the scan provides a
value for the DLP, the relative total energy
imparted to the body.

• This is also calculated and displayed by many

scanners.
 The final step

is the calculation of the effective dose

that provides an estimate of the relative

biological effect on the patient.

• The next big and
important step is
relating radiation
dose to image
quality.
• That provides us
with the background
for optimising
imaging protocols
and producing
images with the
necessary clinical
quality without
unnecessary
radiation dose.
• The actual dose to a specific anatomical
location is generally determined by the values
selected by the operator for the protocol
factors shown here and the size of the patient.
• In principle, the dose can be reduced by setting
the Kv, mA, and time to lower values and
increasing the pitch.
• But. .changing these factors to reduce dose will
also have a significant effect on image quality.
• That is why an optimised protocol is one in
which the factors are selected to produce an
appropriate balance between image quality and
dose.
 It becomes somewhat complex because we
must

also develop a balance between two of the

image

quality characteristics, (detail) blurring and

noise.
•
Increasing the pitch has the effect of
"spreading" the ray and introducing blurring
along the length of the body.

• If an image with high detail is required "high

detail" scan data must be collected, generally
by scanning with a thin beam and low pitch
value.

• An image with high detail cannot be

reconstructed later if the "detail" is not in the
scan data.
• During the reconstruction phase there are two
additional sources of blurring that adds to the
blurring in the scan data.

• One is blurring that might be produced by ;

the mathematical filter in the reconstruction

calculation and the

other is the size of the tissue voxel.

• DLP, the product of CTDI and scan length,
should remain constant with changes in slice
width.

• For example, a 200mm long region maybe

scanned using 20 rotations and a 10mm slice
width or 100 rotations with a 2mm slice width.
The CTDI for the two geometries

is the same

as the total scan length.

• The first limitation to this general rule is
that for some single slice scanners the
narrowest beam, nominally 1mm, may be
defined by post patient collimation and
the beam size passing through the patient
might be significantly greater.

• For, such equipment use of the narrowest

slice width could produce a significantly
increased dose.
• However, this is a very special case that only
applies to a few single slice scanners.

• It does not apply to multislice scanners.

• The second caveat to the rule that slice
thickness does not increase dose is to
emphasise that this is only if the mA is not
changed.

• Reducing reconstructed slice thickness

increases noise.

• Therefore as a general rule the use of narrower

reconstructed slice is associated with an
increase in dose.
But this is not because of slice width per se;

rather it is due to the

associated increase in mA.

 CHOICE OF HELICAL AND AXIAL
SCANNING

• The choice of helical rather than axial scanning

does not influence dose other than the small
amount of over-scanning at the start and end
of the scan.

• The increase in dose is equivalent to an

additional rotation of the gantry.
Patient dose is inversely proportional to pitch

and therefore

the use of a pitch of 1.5

rather than 1 leads to a

33% dose reduction.

• It is generally accepted that doses with
multislice CT scanners are greater than with
single slice scanners.

• Some of these are intrinsic to the technology.

• One of these is the need for over-scanning at

the start and end of each scan.
• This is true for helical scanning in general but
this effect is larger for multislice because of the
greater length of detectors.

• Of greater significance is over-collimation.

• For a single slice scanner, the collimator length

defines the slice with.
• For an 8mm slice, the collimator is set precisely
to 8mm.

• If the beam width were any greater, a thicker

slice would be produced.
• FOR THE MULTISLICE SCANNER, the
reconstructed slice width is defined by the
length of detectors in the detector rows.

• The overall beam length is likely to be about

2mm greater than that required.

• This automatically increases dose to the

patient by more than 10%.
• Doses are also increased because of the
method of use.

• Whereas previously it might have been

standard practice to scan on 10mm slice
widths, it is now possible within shorter
scanning times to produce 5mm slices or less,
thus providing potentially improved diagnostic
information.
However, reconstruction of thinner slices

increases noise and to compensate

mAs and thus dose maybe increased.