Download Lecture 3 - IQ scatter and contrast agents 2013

FRCR: Physics Lectures
Diagnostic Radiology
Lecture 3
Image quality, scatter and contrast agents
Dr Tim Wood
Clinical Scientist
Overview
• What is image quality?
• Contrast
– What is contrast?
– Contrast agents
– Scatter vs contrast
• Spatial Resolution
– What is limiting spatial resolution?
– What determines spatial resolution?
• Noise
– What is noise?
– Quantum noise vs patient dose
The story so far…
Interaction Processes
• Elastic scattering
• Photoelectric effect
• Compton effect
The Mass Attenuation
Interaction Coefficient
X-ray tube design - basic principles
• Electrons generated by thermionic emission
from a heated filament (cathode)
• Accelerating voltage (kVp) displaces space
charge towards a metal target (anode)
• X-rays are produced when fast-moving electrons
are suddenly stopped by impact on the metal
target
• The kinetic energy is converted into X-rays
(~1%) and heat (~99%)
The diagnostic X-ray tube
Factors Affecting Patient Dose
• Tube Current (mA)/Exposure Factor (mAs)
– Double the mA/mAs, double the intensity
– Beam quality not affected
• Tube Voltage (kVp)
– Intensity α kVp2
– Penetrating power increases with kVp
– Higher kVp reduces skin dose
• Filtration (mm Al)
• Focus-to-skin distance
Filtration
Focus-to-skin Distance:
The Inverse Square Law
• For a point source,
and in the absence
of attenuation,
intensity decreases
as the inverse of the
square of the
distance
• This is a statement of
the conservation of
energy
2
2
2
1
D1 r

D2 r
Image Quality
What is ‘image quality’
• Image quality describes the overall appearance
of the image and its fitness for purpose
– Remember: There is always a play-off between
image quality and patient dose
– We only need images that are of diagnostic quality (fit
for purpose) – not pretty pieces of art!
• The main factors to consider are:
– Contrast
– Spatial resolution
– Noise
Contrast
• Most medical images presented as shades of
grey from black to white (greyscale)
• Contrast resolution is the ability to distinguish
between regions of the image
• The amount of contrast between tissues is
intrinsically linked to their properties and the
imaging modality being used
• Also influenced by the technique and
specification of the equipment being used, and
how it is displayed
– Digital imaging techniques allow image processing to
be applied to improve the visibility of details
Image contrast
• The final contrast in the image will depend on a
number of factors, such as;
– Subject contrast – an inherent property of the patient
being imaged that will depend on the attenuation
coefficients of the tissues (or contrast media), the
thickness of structures, the nature of any overlapping
tissues and the incident X-ray spectrum (kVp,
filtration, etc – discussed previously)
– Detector properties – film and digital detectors each
have different implications for the contrast in the final
image (these will be discussed in the next lectures)
– Scattered radiation – scatter can degrade image
contrast if it reaches the detector as it conveys no
information about where it came from. Scatter
rejection techniques may be used to remove this.
Subject contrast
Subject contrast
• Visibility of object structure – requires physical
contrast between the object to be imaged and
the surrounding tissues
• The contrast may be based on a difference in
densities or atomic number (chemical
composition), coupled with the thickness of the
particular material through which the radiation
has to pass
– Proportional to object density x object thickness
• Also remember, higher kVp = less photoelectric
absorption = poorer contrast
Subject contrast
Material
Effective Atomic
Number Z
Density g cm-3
Contrast Index
Density x
thickness = 4
Soft Tissue
- Fat
- Water
- Muscle
7
6
7.4
7.5
1
0.9
1
1
4cm
4.4cm
4cm
4cm
Air
7.6
0.00129
3100cm
Bone
14
1.85
2.1cm
Calcium
20
1.55
Tooth
2.4
Iodine
53
4.9
Barium
56
3.5
Barium Contrast Media
56
4.25
Lead
82
11.3
1.7cm
1.1cm
0.35cm
Subject contrast
Subject contrast
• Contrast between bone and muscle is large at
low kV, but decreases with increasing kV
• Contrast between low atomic number soft
tissues is low, even at low kV
• The contrast between air and tissue (similar Z) is
large due to the difference in density
Balancing Contrast & Patient Dose
• Penetration, contrast and patient dose depend
on the x-ray beam spectrum. The ‘best’
spectrum provides adequate penetration and
contrast, whilst keeping the patient dose as low
as possible
• The spectrum produced is dependent upon the
target material, inherent and added filtration, and
kilo-voltage
Balancing Contrast & Patient Dose
Low Energy (kV) = Low Penetration = High
Contrast = High Skin Dose
High Energy (kV) = High Penetration = Low
Contrast = Low Skin Dose
Optimal Situation = kV that gives adequate
contrast for an acceptable patient skin dose
– will vary for the clinical application
Contrast agents
• Soft tissue contrast is a problem in radiography
• Two possible solutions;
– Use lowest possible kVp e.g. mammography
– Use a high Z contrast agent in the tissue of interest
• The high Z of contrast agents maximises the
photoelectric absorption of X-rays
• Ideally, absorption edges lie just to the left of the
major part of the X-ray spectrum
• Two common contrast agents are Iodine (Z = 53,
EK = 33 keV) and Barium (Z = 56, EK = 37 keV)
• Air (similar Z, but low density) can also be used,
but much less common now
Contrast agents
Contrast agents in clinical practice
http://health.siemens.com/AX/Artis_zee_cardiology_image_quality/clinicalresults.swf
Contrast agents in clinical practice
Clinical Examples
Clinical Examples
Clinical Examples
Contrast and scatter
• Primary radiation carries the information to be
imaged
• Scatter obscures this as it carries no information
about where it came from
• The amount of scatter (S) may be several times
the amount of primary (P) in the same position
• The ratio S/P depends on thickness of patient
– For typical PA chest = 4:1 (20% of photons carry
useful information)
– For typical lateral pelvis = 9:1 (10% of photons…)
Contrast and scatter
• Scatter is generally quite uniform across the
image
• Reduces contrast that would otherwise be
produced by the primary by a factor of (1+S/P) –
up to 10 times!
Intensity
10
% contrast reduced
8
S+P
S
Scatter = 4 x Primary (PA Chest)
S+P contrast = (10-9)/10 x 100 = 10%
Primary contrast = (2-1)/2 x 100 = 50%
2
P
Position
Scatter reduction
• Scatter and patient dose may be reduced by;
– Reducing the field area with collimation (reduce
volume of tissue generating scatter)
– Compress the tissue to minimise overlying structures
(reduce volume of tissue generating scatter)
• Scatter may be reduced at the expense of
increased patient dose by;
– Reducing kVp – less forward scatter is produced and
is much less penetrating (less scatter produced)
– Use an anti-scatter grid to remove scatter from the Xray beam
– Use an air-gap to reduce the intensity of scatter
reaching the detector
Anti-scatter grids
• Series of thin strips of high Z metal (e.g. Pb)
sandwiched between low attenuation spacer
material (e.g. C fibre)
– Typically 0.05-0.07 mm wide with 30-80 strips per cm
(the line density)
• X-rays that hit the grid at oblique angles (i.e. not
along the path of the primary beam) are absorbed
• High proportion (but not all) of the primary beam
passes straight through to the film
• The grid ratio is the depth of the channels divided
by its width – typically 8:1
– Larger grid ratios = smaller angle of acceptance and
better rejection of scatter
Anti-scatter grids
depth
width
separation
Anti-scatter grids
• Large fields tend to require higher grid ratios
(12:1 or 16:1)
• Grids generally not used for thin parts of the
body or children as less scatter is generated
(and increase in dose not justified)
• Contrast improvement factor is defined as
contrast with a grid divided by contrast without
– Typically between 2 and 4
Anti-scatter grids
• Parallel lead strips = unfocussed grid
– Away from the centre of the beam, primary X-rays will
strike the grid obliquely and be attenuated giving grid
cut-off
– Use longer FFD or lower grid ratio to avoid this
• Focussed grid uses strips that are angled to
match the divergence of the X-ray beam
– All strips point to the tube focus
– Ensure grid is the right way round!
– Only to be used at a specified distance (or range of
distances)
– Tube must be centred on grid with no tilt, otherwise
get cut-off
Focussed anti-scatter grids
Anti-scatter grids
• Grid lines will result from the shadow of the lead
strips for a stationary grid
• If sufficiently high line density, may not be visible
without magnification
– However, stationary grids used in digital imaging may
produce interference patterns
• Alternatively, may use a moving grid to blur out
the grid lines during exposure
– Movement must start before exposure
– Must be fast for blurring to occur on shortest exposure
times
Anti scatter grids
• The use of grids necessitates increase exposure
factors (mAs) due to the attenuation of primary
beam
• The grid factor is the ratio of exposure
necessary with grid to that without
– Typically in the range 3-5
– Hence, corresponding increase in patient dose!
• The grid, cassette holder and AEC system is
often referred to as the ‘Bucky’
• Alternatively, can use slot scanning techniques,
but this is at the expense of exposure time
– e.g. CT and mammography
Air-gaps
• If the detector is moved 20-30 cm away from the
patient, a significant proportion of the scatter will
miss the detector
• Due to inverse square law, small reduction in
primary intensity (hence, increase mAs to
compensate)
• Also results in a magnified image
Air-gaps
Magnification
• Magnification is the
result of the diverging Xray beams travelling in
straight lines
• Reduce by using longer
focus-to-film distance or
shorter object-to-film
distance
• Magnification exploited
in some modalities e.g.
mammography
Spatial Resolution
• Spatial resolution describes
the ability to see fine detail
within an image
• Fine detail is clearer when the
contrast is high
– e.g. microcalcifications
• May be expressed as the
smallest visible detail, but
most common descriptor is
the highest frequency of
lines that can be resolved in
a high-contrast bar pattern
Spatial Resolution
• Bars of Pb or W, with width of
bar equal to space between
them
• A bar and space make up a
line pair (lp), and the spatial
frequency of the pattern is
given as lp mm-1 (or lp cm-1
for CT)
• Observer determines the
highest spatial frequency that
can be resolved = limiting
spatial resolution
Limiting factors
• Geometric Unsharpness – due to finite focal
spot size (discussed previously)
• Movement Unsharpness – imaging moving
structures may result in blur e.g. heart, lungs, etc
– Minimise by immobilisation e.g. compression in
mammography
– Breath hold techniques
– Use shortest possible exposure times (may be at the
expense of high mA)
• Absorption Unsharpness – gradual changes in
absorption near tapered/rounded structures
– Minimise by careful patient positioning
• Detector resolution – will be discussed in
following lectures
Geometric unsharpness –
The ideal point source
Ideal point source
of X-rays
FFD
Object
OFD
Film/detector
Geometric unsharpness –
A ‘real’ focal spot
Focal spot of
finite size, f
FFD
Object
OFD
Film/detector
Penumbra
Noise
• Noise is a random, usually unwanted, variation
in brightness or colour information in a visible
image
• Noise is one of the most important limiting
factors to contrast and spatial resolution
• The most significant source is quantum noise
(or mottle) due to the low levels of radiation used
to form an image
• Other sources include film grain and electronic
noise in the image receptor
Noise
• Image noise is most apparent in image regions
with low signal level, such as shadow regions or
underexposed images
• Noise gives a grainy, mottled, textured or snowy
appearance to an image
• Noise can mask fine detail in a radiograph
• Noise reduces visibility of parts of a radiographic
image
• Noise is particularly an issue with image details
that are already of low contrast
Types of noise
• Several types of noise contribute to the overall
image quality, these are:
– Quantum Noise – from statistical nature of the
interactions of the X-ray beam
– Secondary Quantum Noise – associated with
secondary carriers at each conversion stage in image
formation e.g. CR, Image Intensifiers, etc
– Structural Noise – films/screens/intensifier tube
screens/digital receptors. Only tends to be important
at high doses where all other sources are negligible
– Electronic Noise – circuit noise, thermal noise,
external electrical device signals. Only obvious at low
doses
Quantum noise (mottle)
• X-ray beam is made up of individual quanta or
packets of radiation energy – Photons
• Quantum Noise describes the random
statistical nature of the interactions of the
radiation falling on a detector surface
• The number of photons absorbed varies from
pixel to pixel
– This variation is described by the standard deviation
• In all digital imaging modalities, quantum noise
should be the main contributor to the overall
noise at clinical dose levels
– ‘Quantum Limited’
Quantum noise (mottle)
• The higher the number of photons falling on a unit
area, the lower the deviation is relative to the mean
number, and the lower the Quantum Noise level.
• Signal-to-noise ratio (SNR) is proportional to the
square root of dose
Quantum noise (mottle)
• Therefore, to increase the concentration of
photons impinging on a unit area of a detector
(to reduce the noise level) the radiation
exposure needs to be increased
• Level of exposure required will depend on the
clinical task and receptor sensitivity
• The consequences of increasing the exposure is
to increase the dose to the patient
• As SNR is proportional to the square root of
dose, four times the dose will only double the
SNR
80kV 40mAs
80kV 0.5mAs
Striking the right balance
Reduced Exposure = Low Patient Dose =
Increased Q-Noise = Reduced Image Quality
Increased Exposure = High Patient Dose =
Reduced Q-Noise = Improved Image Quality
Optimal Situation = Exposure high enough to
give a diagnostic image with acceptably low
quantum noise, whilst maintaining an
acceptable patient dose.
So what does this all really mean?
Signal on detector
Something with lots of
contrast is easier to see
than…
Position on detector
Signal on detector
… something with little
contrast
Position on detector
So what does this all really mean?
Signal on detector
But there is some loss
of sharpness due to
various process, so the
sharp edge…
Position on detector
Signal on detector
… actually becomes
blurred, which may
not matter for large
structures…
Position on detector
So what does this all really mean?
Signal on detector
… but if the structure is
small…
Position on detector
Signal on detector
… it may start to
disappear into the
background…
Position on detector
So what does this all really mean?
Signal on detector
Up to now we haven’t
considered the third
component, noise…
Position on detector
Signal on detector
… which may not be a
problem if there isn’t
too much…
Position on detector
So what does this all really mean?
Signal on detector
… but if there is lots of
noise in the system, the
detail may be
obscured…
Position on detector
Signal on detector
… whilst the low level
of noise may be
problematic if contrast
is low to start with…
Position on detector
So what does this all really mean?
Signal on detector
… or resultion is poor
Position on detector