Discuss why quality assurance testing is required for CT equipment.
List some of the governing agencies that guide CT quality control programs.
Define quality control and quality assurance as applied to CT.
List the 3 basic tenets of an acceptable quality control program.
Discuss the two basic types of phantoms used in CT quality assurance testing.
Discuss Hounsfield units (HU) and give the HU for various substances such as water, air, fat, and bone.
List the 8 required QA test performed at equipment acceptance, daily, or annually.
State why testing the CT number for water and image noise is done daily.
Discuss how performance testing for high and low contrast resolution is done.
Discuss the importance of modulation transfer function for spatial resolution, and how this test is performed.
Define the terms spatial resolution and spatial frequency.
Discuss how the slice thickness test is performed.
Discuss why it is important for the laser light localizations to be accurate, and how they are tested for in a QA program.
Discuss the various units associated with calculating patient dose
in CT: computed tomography dose index (CTDI) and dose length product
(DLP) for examples.
Define and discuss terms found on the patient’s CT dose report.
Discuss relative risk from various CT exams by comparison to other x-ray procedures such as the chest x-ray or fluoroscopy.
History of CT and Evolution of Spiral Scanners
The term tomography stems from the Greek word "tomos" meaning "section".1
Scientists and mathematicians have described, "body section
radiography" in many different ways since the 1920's. It wasn't until
the 1960's after much research, that the world's first CT scanner
emerged.1,2 The inventor was Godfrey Newbold Hounsfield, born in England 1919.1
He and Alan Cormack, a medical physicist, together developed and placed
the first brain scanner into operation in 1971 for a company called EMI
Ltd. 1,2 In 1979, they were awarded the Nobel Prize in medicine and physiology.1 Initially data acquisition in CT scanning was very slow. The first experimental brain scan in 1967 took 9 days.1 By 1971 they had reduced the scan time to 20 minutes.1 In 1989, the helical (spiral) concept was considered one of the most significant developments in CT.1,2 This development meant continuous rotation of the x-ray tube without reversal between images.2 The new continuous motion was given the name "slip-ring" technology, and it reduced brain scan times to as low as 0.8 seconds.2 As technology continues to develop with multi-slice systems, times are getting even shorter (0.4 sec.).2
Data from spiral or helical scanners is often referred to as "volumes" of tissue rather than individual cross-sectional slices.2
These images are "overlapped" and do not have a gap in between them.
This allows complete coverage with no "missed" areas of tissue. This
modern equipment has allowed more rapid image production for trauma and
pediatric patients; however, they can sometimes deliver a higher dose of
ionizing radiation.2 Since Hounsfield's first operational
scanner in 1971 to today's modern "multi-slice" spiral scanning devices,
a lot of progress has been made to improve image quality and most
importantly, to decrease patient ionizing radiation dosage. With this
evolution of technology has come the need for more comprehensive quality
assurance programs, new phantoms specific to spiral CT, and higher
standard safety guidelines. Not only are specific tests conducted to
maintain equipment operation at an acceptable level, but these programs
are also designed to recognize and create a corrective action for
quality assurance issues.3
The development of faster scan times has also created some drawbacks
such as: the need for x-ray tubes with higher heat ratings; more
powerful generators to sustain added heat volumes; and, increased image
noise consistent with the rapid reconstruction of images.2 This "image noise" can cause an artifact known as the venetian blind artifact.4 This occurs with multi-slice scanners and appears as bright and dark bands superimposed on three dimensional images.4
Another drawback from helical scanners is the notable difference in
low contrast resolution. This problem has created the need for
additional test tools and more suitable phantoms for spiral and
multi-slice scanners.5 Even after a close review; the benefits of spiral scanners definitely still outweigh the drawbacks.
Governing Quality Assurance Agencies for the CT Helical (Spiral) Scanner
There are many agencies that guide our medical physicists or QA
program coordinators in creating the best quality control program for
their facilities both at the national and local level. The AAPM
(Association of Physicists in Medicine), The ICRU (International
Commission on Radiation Units and Measurements), and the ACR (American
College of Radiology), just to name a few, are involved in setting x-ray
equipment standards. The use of CT for medical purposes in the U.S. is
primarily controlled at the state or local government level.6
However, the guidelines set forth in the equipment manufacturer's
manuals for quality assurance are the primary resource for compliance
concerns in diagnostic x-ray systems. The Food and Drug Administration
(FDA), as well as the applicable state health department, should first
approve all equipment prior to being put into operation for patient use.6
Each equipment manufacturer should include a list of standard tests
to be performed and the proper phantoms for their equipment.
Under the Radiation Control for Health and Safety Act, all CT equipment is subject to an equipment performance standard.6
This consists of: minimum radiation safety requirements and
manufacturer requirements regarding safety standards and performance.6 Manufacturers must ensure that their equipment meets the standards set forth by this act.6
Any additional tests are chosen at the facility's discretion and in
accordance with national and their state regulatory agencies.6
There are typically three categories for regulation and guidance:
laws- which are passed through legislation and provide an authority on
the subject matter; regulations - which provide specific requirements
that are authorized or required by law; and, guidance -which are agency
decisions or policies that describe how to comply with a regulation.7
These laws, regulations, and guidelines ultimately allow the nation�s
medical professionals to maintain high quality images with the lowest
possible patient dose.
Regulations and laws governing CT scanners insure both quality
assurance (QA) and quality control (QC) programs are implemented.
Quality assurance is the component that measures CT scanner performance
to assure operation is at an acceptable level. Quality control takes
action to correct inadequacies before they are problematic. These
measures assure high imaging performance standards, and reduce the risk
of patient harm due to change in equipment performance characteristics.
There are three basic tenets of an acceptable quality control
program. QA must be performed on a regular basis, there must be prompt
interpretation of test results, and the third tenet is accurate
bookkeeping. Some test are required daily, others monthly; annually or
at equipment acceptance. The CT technologist usually performs daily
tests, which means for these tests they must recognize when results are
out of range. Test results must be recorded in a logbook, data log, or
computerized record for as long as the scanner is in use. Daily, weekly
and monthly results can be compared to acceptance data. This can be
very useful especially if there appears to be a malfunction of the
equipment. Often the CT technologist is too busy to perform daily
tests; however, you should always find time to perform daily tests since
a properly performing CT scanner eliminates the equipment as the cause
of an improper interpretation of CT images.
Phantoms and Required Test Tools for Spiral Scanners
The ICRU defines tissue substitutes and phantoms as the following:
tissue substitute - any material that simulates body tissue interaction
with ionizing radiation8; and, phantom - any structure that
contains tissue substitutes, at least one, which can simulate radiation
interactions in the human body.8 There are two categories of phantoms: calibration phantoms and imaging phantoms.8 Calibration phantoms are for testing detectors and correcting quantitative information obtained from digital images.8 Imaging phantoms assess image quality and are usually further classified as head, body, standard, or reference phantoms.8
This photograph of a CT phantom used
to perform quality assurance testing. Phantoms are made of Plexiglas,
which has a density of 120 Hounsfield units. The black holder is also
used to perform fine alignments of the phantom in the gantry and to the
localizing laser lights. There are several types of phantoms, for
example, one type may be used for head dose calculations, and another
for body dose calculations. Different manufacturers will recommend
testing their equipment with specific phantoms. All phantoms must meet
performance standards set by the Food and Drug Administration (FDA).
Quality assurance for the spiral CT scanner consists of these basic
required elements of testing: contrast scale and mean (standard
deviation), CT number for water, high-contrast resolution, low contrast
resolution, laser light alignment and accuracy, image noise, uniformity
and artifacts; slice thickness and localization, and patient dose.9,10
There is a wide array of tests that may be performed as well as test
tools that can be used. The facility's quality assurance manager or
medical physicist generally decides this. The selection of these tests
should be based upon the type of equipment and the frequency in which
the equipment is to be utilized. This will usually limit some of the
more complex tests to an annual survey.
Here is a brief description of the basic QA tests done for a CT
helical scanner. Again, these may vary from facility to facility and
from state to state. There are daily, monthly, semi-annual, and annual
test categories, as well as acceptance testing.2,9,10 Tests
are routinely performed by the medical physicist or staff QA manager
upon receiving new equipment. Special tests may also be required if
equipment is relocated, or the x-ray tube is replaced.2
A good QA program will provide regular testing, prompt interpretation of test data, and faithful record keeping.3
A current logbook or computer file should be readily available for
viewing by any regulatory agency in the event of an unscheduled
inspection. In addition to the following tests, daily visual checks
should be done on table/gantry movement, cables, cords, operating
console, controls, and print system if installed.2 These visual tests should be documented as well, for future comparison.
Acceptance testing includes the following baseline type tests:11
- Light alignment
- Slice thickness
- Noise
- Contrast resolution
- Heat Unit linearity
- Uniformity
- Spatial resolution
- Patient dose
Quality assurance and performance phantoms assess system
performance and allows for the establishment of an ongoing quality
assurance program. Phantoms are designed so that maximum performance
information is gathered with minimum effort. These tests are simple to
perform and provide unquestionable accuracy for quality control.
(Please see your equipment manufacturer specifications manual or refer
to the appropriate requirements and guidelines for your state to learn
what QA tests will best fit your program). This module will explore a
phantom type used to perform quality assurance testing on General
Electric HiSpeed CT/I scanners. While the specific methods of testing
described in this module may differ slightly from one typer of CT
scanner to another, these test are required on all scanners and their
results are uniformly regulated.
This CT phantom (left photograph) is
used to make various images used to gather quality assurance
measurements. Image "A" is used for high contrast resolution, contrast
scale, slice thickness, and laser accuracy. Image "B" is used to
measure low contrast detectability. Image "C" is used to measure noise
and uniformity of the scanner. We will look at each of these and other
images and their respective tests in detail in this module.
Basic Quality Assurance Tests
Test 1 - CT number for water (average & standard deviation)1,3
Computed tomography involves complex processing of digital data
using mathematical principles called reconstruction algorithms. Scan
data is based on penetration and attenuation measurements of photons as
they traverse matter. This raw data must then be converted to digital
data and displayed as a CT image. Each pixel in the image is assigned a
number collectively referred to as a CT number. CT numbers are
referred to as Hounsfield Unit (HU) named after the inventor of computed
tomography. The Hounsfield scale (H scale) is a calculation from the
linear attenuation coefficients of tissues that make up an image slice.
CT numbers are relative to the attenuation of water, which is assigned a
value of 0 HU. Using water as the reference the maximum brightness of a
pixel is -1000 HU and will appear white on the CT image. The opposite
end of the H scale is maximum darkness, which is +1000 HU and the image
will be black. Between these extremes are various shades of gray that
make a diagnostic CT image. Therefore, when we check the CT number for
water we are effectively checking the reconstruction algorithm that
computes CT numbers across the image.
This chart shows the various CT
numbers (Hounsfield units) calculated for various tissues and substances
based on the density of water. Notice that at the extremes is bone
(+1000 HU) and air (-1000 HU). Water has a CT number of zero, which is
used to test for the function of the algorithm that calculates CT
numbers.
To perform the CT number test a water phantom test tool is used.
The phantom is a water-filled cylinder with a 20cm diameter for head
technique calibration, or 32 cm diameter for body simulated calibration.
Scan with usual technique for the body part represented by the phantom
type. A 20 cm diameter phantom and head technique is most frequently
used. Select a ROI (region of interest) of about 2-3 cm or containing
about 200-300 pixels and measure the average CT number. Air measures
-1000 HU; water should be close to 0 (zero). Water should not exceed
+/-3 HU at the center of the image, and no more than +/- 5 HU from
center to periphery. If the CT number in the center of the image is not
within 3 HU the scanner fails this test. Recalibrate and retest if not
within limits. Keep in mind that although the CT number for water is
measured daily there are two media used in calibration, these are water
and air. At least once a month a ROI outside the phantom image in a
region representing air is taken. This reading should measure -1000 HU,
+/- 5 HU. This reading is to check contrast since water is zero and
air measures –1000 HU. When the algorithm that calculates CT numbers is
accurate, water is within +/- 3 HU of zero, and air is within +/- 5 HU
of –1000 HU.
The CT number for water- (average & standard deviation) test
is done to ensure equipment manufacturer specifications for CT number,
field uniformity, and noise. The test for CT number of water is done
daily. Possible causes for the CT number of water to be out of range is
miscalibration of the algorithm generating CT numbers. This is a type
of problem that needs immediate attention of the biomedical engineer or
radiation safety officer. When the CT number for water and air fail the
recommended range it must be immediately corrected to insure accuracy
of the displayed CT image. If the CT number fluxuates significantly, but
remains within the acceptable range, this too should be brought to the
attention of the radiation safety officer.
CT Number for Water
- Water filled plastic cylinder (20 cm diameter)
- Take scan, reconstruct image, place ROI 200-300 pixels in center of field and take measurement.
- Expected results: CT number of water
equal to zero, but range of +/- 3 at center of image is acceptable, and
+/- 5 HU at peripheral locations.
- Cause of failure is usually miscalibration of the algorithm that generates CT numbers.
- CT number test is Performed Daily.
- Test an area outside the image representing air for CT number of air, monthly. The acceptable range is -1000 HU, +/-5 HU.
|
The water phantom test images shown
here demonstrate a clear uniform field free of artifacts. The image on
the right shows a ROI placed in the center of the image to measure the
CT number for water. In this example the ROI measured 0.07 HU with a
standard deviation of 3.33 HU. This measurement is within acceptable
performance guidelines of the manufacturer. This is a simple test that
confirms the proper functioning of algorithms that calculate CT numbers
and provides a quick check of the field for artifacts.
Test 2 - Noise and Field Uniformity1
When we think of image noise in traditional radiographic imaging
using screen-film imaging we are referring to the overall graininess of
the image. CT is a type of digital imaging processing in which image
noise can be caused by a variety of factors. Noise in CT is mainly
related to the following: (1) number of detected photons; (2) matrix
size (pixel size); (3) slice thickness; (4) algorithm; (5) electronic
noise (detector electronics); (6) scattered radiation; and (7) object
size.1 Noise limits low contrast resolution and may hide
anatomy similar to surrounding tissue. Most pathology imaged in CT is
seen in soft tissues such as the lungs, kidney, liver, and brain. To
test for image noise a simple cylinder or container of about 20 cm. in
diameter is used. The phantom used to calculate the CT number for water
and air is used for the noise uniformity test. The phantom is scanned
at different slice thicknesses, and gradual increases in mAs. Measure an
ROI of about 200-300 pixels and find the standard deviation of the CT
number at the center of each image. The image field is sampled along
the periphery as well as in the center of the image. There should be
uniformity in the CT numbers throughout the image. The noise level can
be stated as a percentage of image contrast in CT numbers. The stand
deviation for noise should be +/- 3. Since CT numbers range from
+/-1000 HU, noise is less than 0.3%. The maximum standard deviation
between the center ROI and any peripheral ROI is less than +/- 5 HU.
The result should show the noise in the image being directly
proportional to the standard deviation of the CT number of water. The
number should decrease as the slice width and mAs are increased.
Increased noise can be a result of poor beam/detector alignment, reduced
detector sensitivity, or reduced output from the tube.1,2 Be sure that
the noise levels produced by the equipment do not increase with age. A
higher noise level will result in a lower dose to the patient and lower
noise level results in a higher dose. There must be a balance between
the two to maintain quality images and low doses. Out of range CT
numbers for standard deviation could be caused by decreased dose or
increased electronic noise. Standard deviation describes the difference
between the lowest ROI value and the maximum ROI value. Testing is
performed daily with the CT number for water and also upon acceptance of
new equipment.
These two CT images taken of a water
phantom show an ROI in the center and periphery of the field to measure
field uniformity. The same locations for measuring should be used each
day. The image on the right shows a grid that has been inserted on the
image to help with consistent daily placement of the ROI's. The image
on the left shows the ROI's without the grid. The center ROI measured
0.25 HU and the four ROI's in the periphery measured within the
acceptable < +/- 5 HU of center measurement. Comparing peripheral
field HU's is necessary since attenuation of x-rays by different body
tissues when displayed in different parts of the image field must be
accurate.
Causes of Image noise
- Mechanical, mathematical, or electronic differences in
detected x-ray energies, electronic outputs or reconstruction
algorithms.
Reducing Image noise
- Performing tube warm-up, proper calibrations, and daily Fastcals reduces noise
|
Region of interest selected from the scanners monitor is also
capable of measuring the standard deviation within the ROI. Keep in
mind that the standard deviation is dependent on several factors
including kVp, mA, scan time, slice thickness, phantom size, and ROI
position. This is why the technical factors used to measure HU for
phantom tests must be standardized so that they are constant each day.
Likewise, when measuring field uniformity it is important not to place
peripheral ROI's too close to the edges of the image since at this
location standard deviation is lower than towards the center. Ideally,
the standard deviation should be small; however, more importantly it
should be compared to the latest service record standard. It should not
vary by much from day to day and an increasing pattern should be
brought to the attention of the service department. An increase in the
standard deviation indicates the image is becoming "noisier." Possible
causes include decrease in tube output, increased detector,
amplifier(s), analog-to-digital converter, or other issues.
This CT image shows two ROI’s taken
during the CT number for water phantom test to measure image noise. The
standard deviation within each ROI was measured by the computer’s
quality assurance software. The CT number for water is shown in the
three measurements along with their standard deviations. For ROI 3, 4,
and 5 the standard deviations are 02.71, 02.88, and 02.85 respectively.
A low SD indicates low image noise, which is desirable for diagnostic
image interpretation. In this CT image, the mean CT number is given
along with the standard deviations for selected peripheral ROI’s. A
center and peripheral ROI is demonstrated.
Test 3 - High-contrast (spatial) resolution1,2,3,5,11,12
Spatial resolution is important in detecting the edges of
structures, margins of tumors, small foreign bodies, and small bony
structures. This test measures spatial resolution by measuring the
high-resolution pattern in a phantom image. This test measures how the
scanner distinguishes between two high contrast objects placed close
together, and how small an object that can be visualized. Resolution
phantoms come in a variety of test patterns. Generally, the test tool
will have a bar pattern or series of holes cut in the Plexiglas within
the phantom. Generally, each bar patter contains a set of 5 holes or
bars and spaces of constant equal dimension. Each block decreases in
size from one pattern to another. Measurements are taken of the depths
of different drilled holes usually into an acrylic or hard resin-like
substance. The holes may be filled with air giving 100% contrast or
water giving 12 to 20% contrast. All holes or bars should be seen on the
scan image; however, we are only interested in the smallest row in
which all five bars and spaces can be clearly seen. The resolution
block contains the following bar sizes 1.6mm, 1.3mm, 1.0mm, 0.6mm, and
0.5mm. The smallest the row clearly seen indicates better performance
of the scanner and image quality. Expected result is that complete set
of bars or holes and accompanying spaces will be in the range of 0.75 to
1.0 mm. All modern scanners should have a resolution of 0.5% contrast
for 5mm. The minimum size of the holes visualized should not increase
over the life of the equipment.
High-contrast resolution tests should be performed upon
acceptance of equipment and monthly. Baseline is established at
acceptance or referenced to manufacturer's specifications. Test failure
is related to an enlarging x-ray tube focal spot, poor registration,
detector failures, mechanical misalignments, mechanical wear and so
forth. In any case the biomedical engineer should be notified if
resolution degrades from baseline measurement.
The photograph on the left is of a
phantom designed for multiple QA testing. The yellow arrow points to
the bar pattern in the phantom that is used to measure scanner
high-resolution. The resulting image is seen on the right. Notice the
bar pattern is filled with air giving 100% contrast. The smallest row
of bars is recorded and compared to baseline value in this test.
Expected result is that a complete set of bars or holes in some rows in
the range of 0.75 to 1.0 mm or 0.5% contrast for 5 mm for modern CT
scanners.
This CT image is used to determine
high contrast spatial resolution. Spatial resolution is a measure of
how well two high contrast objects placed close together are
distinguished. In this phantom image the two high resolution objects
are the Plexiglas (120 HU) and water (HU 0) seen in the spaces between
the bars. Each pattern consists of five bars and spaces called line
pairs. The size of each pattern is 1.6mm being the largest and 0.5mm
being the smallest. The smallest line pairs discernable are seen at
0.8mm.
To validate this test the 1.6mm bar
pattern is measured using a box ROI (yellow arrow). The box ROI should
be sized until it fits into the pattern. Measure the standard deviation
of the pixels in this ROI to get a quantitative assessment of changes
in system resolution. The standard deviation measurement should be 40
+/- 4. This should be compared to the baseline measurement at equipment
acceptance for accuracy. Then, take a measurement of the other
patterns, or alternately of the smallest pattern with discernable line
pairs and record.
High-Contrast (Spatial) Resolution
- Requires high-resolution test pattern in a phantom.
- Determining the smallest row of holes or bars that can be clearly seen gives measurement.
- Expect results to demonstrate complete set of bars or holes in some rows in the range of 0.75 to 1.0 mm.
- Test failure could be caused by enlarged
x-ray tube focal spot, excessive mechanical wear in gantry motion,
misalignment of electromechanical components, or detector failures.
- Any change in test from baseline should be reported to the service person immediately
|
As we have discussed, spatial resolution is a measure of detail
resolution. Sometimes it is practical to evaluate the spread of
information within the CT system. To do this we look at what is known
as the modulation transfer function (MTF). The MTF is the most common
method of describing spatial resolution in CT, digital radiography, and
film-screen radiography systems. The MTF analysis allows us to compare
system performance on a day-to-day basis, or to compare a system's
performance against another CT system. Modulation transfer function is
expressed in line pairs per centimeter (lp/cm). When counting a line
pair, one line and its adjacent space are called a line pair. To
measure MTF directly a line pair phantom is imaged and the number of
line pairs is counted. If 5 line pairs are counted, the spatial
resolution is reported as 5 lp/cm. If 10, 15, or 20 line pairs are
seen, the spatial resolution is reported as 10 lp/cm, 15 lp/cm or 20
lp/cm respectively. The number of line pairs seen in a given length is
known as the spatial frequency.
Because CT scanners are not created equally, how well a given
scanner displays an object is also a function of the size of that object
and the spatial resolution of the scanner. An object's size in a given
length is also known as its' spatial frequency. What this means is
that how frequently an object fits into a given space is it spatial
frequency. Generally speaking, the smaller an object is the higher
spatial frequency and the more difficult it is to be displayed
accurately. Likewise, a large object will have low spatial frequency
and will be more accurately displayed. If an object is displayed
accurately as it is, then the MFT is given a value of 1.0. The
modulation transfer function scale is from zero to 1. A MTF value of 0
would mean the image is blank and contains no information about the
object scanned. Scanned objects will have values between 0 and 1;
however, the closer to 1 an object is the better the MTF of the scanner.
This picture demonstrates how object
size is related to their spatial frequency. Small objects have low
spatial frequency since more of them fit into a prescribed length.
Large objects have high spatial frequency since they fit fewer times
into a given length.
In practical terms, the information needed during quality
assurance testing or when comparing CT scanners for purchase appear in
graph form called MTF graphs. Modular transfer function is plotted
along the y-axis and object spatial frequency along the x-axis. When
comparing the function of a scanner over time, or when comparing the
performance of different scanners, we look at what is called limiting
resolution. Limiting resolution is the spatial frequency at MTF of 0.1
for any scanner. MTF of 0.1 is referenced because it is the lowest MTF
that will result in a visible CT image. A scanner with a higher spatial
frequency will be able to image small objects.
This graph of MTF and spatial
frequency for three scanners is shown. At limiting resolution (0.1 MTF)
scanner "A" have a spatial frequency of 11.5 lp/cm. Scanner "B" at a
MTF of 0.1 gives a spatial frequency of 17.0, and scanner "C" the
spatial frequency is 20.0. Interpretation of the graph implies that
scanner "C" is better able to display small objects than scanners "B or
C".
Test 4 - Low Contrast Resolution/Detectability1,2,3,5,11,12
Purpose of this test is to determine the ability of the scanner
to discriminate low contrast objects. Keep in mind that most relevant
detail in the human body is made of soft tissue, and is low contrast.
Arguably, this is perhaps the most important quality control performance
test. This test measures the scanners ability to detect objects that
vary only slightly from its background. This is especially important
when trying to detect low-density tumors that lie in soft tissues. The
most common areas where this is important are the brain, kidney, and
liver. Often intravenous or oral contrast media is used to increase
density difference that can be detected by x-ray. In other words,
contrast agents increase the sensitivity of low contrast resolution.
The visibility of low contrast objects is constrained mainly by
amplitude and frequency characteristics of the image noise11.
Subject contrast is a product of both high and low contrast
within and surrounding a structure. Subject contrast in a CT image is in
simple terms the difference in average CT numbers between two adjacent
regions of the image. So we must be assured that low-contrast resolution
is accurate and there is accuracy in the display of high-contrast image
resolution. Low contrast sensitivity is where CT excels over
conventional radiography. CT can resolve small differences in tissue
densities because there must be at least a 10% density difference to be
detected with conventional x-ray imaging. This is why we use a 15%
increase or decrease in the kVp to change the scale of contrast in
conventional radiography. CT on the other hand is able to resolve
density differences as little as 0.1%. Notwithstanding, the size of a
low contrast object, its inherent density (calcium vs. fat), image
noise, and viewing window setting will in part determine its
detectability.
To perform this test various types of low-contrast inserts are
available for the CT phantom: therefore, scanning for this test is
manufacturer specific. Our low contrast detectability test phantom
image is defined by the smallest hole size visible for a given contrast
level and dose. The phantom contains a doped polystyrene membrane
suspended in water. The membrane is pierced with holes ranging from
10.0mm, 7.5mm, 5.0mm, 3.0mm and 1.0mm. The basic of this test is that
the number of object visualized on the phantom image is determined, and
the mean value of each visualized hole and surrounding material is
recorded. The smallest holes that should be visualized is 5 mm in
diameter or smaller for 5% contrast objects.
This low contrast detectability
phantom image displays various sized holes used to determine low
contrast (left). The various sizes are labeled on the right image;
however, it is difficult to see the smallest holes. This test measures
the scanners ability to detect an objects density when it is close to
background density.
Low-resolution contrast is determined as the difference in HU of
objects and background. High noise in the image will cause a decrease
in low-contrast resolution. To get an accurate measure of low contrast
we need to know the CT number for the polystyrene membrane. This is
accomplished by taking the CT number for water over an area that does
not include the membrane. A second measurement is taken over an area
that includes the membrane superimposed on water (called water plus the
membrane). The CT number taken for water is subtracted from the water
plus membrane to get the CT number for the membrane. To perform this
test, measure using a box ROI above and below the membrane in the water
section (labeled A and B in the phantom image below). Take a box ROI in
the polystyrene membrane above the holes (labeled B) and below the
holes (labeled C). Subtract "A" from "B" and subtract "D" from "C".
When the measurements are completed and recorded adjust the Window Width
to 20 and the Window Level to the CT number recorded for water. This
will allow for an accurate reading of the number of holes visible.
This CT phantom image shows boxed
ROI's placed over the polystyrene membrane and over water when
calculating low contrast detectability. The membrane shown by yellow
arrow contains spaced holes of different densities. There are various
types of low contrast testing phantom inserts that are equipment
specific. This on is used for the GE Lightspeed scanners.
Low contrast test tools are still being perfected as there
remain questions as to the most optimal means of testing. The low
contrast sensitivity test shows ideal results with the new polyurethane
resin material phantoms. The low-contrast sensitivity in the image plane
was easily measured with no dependence on temperature or beam quality.
Additional phantoms are also available for testing low resolution in
three dimensions. Contrast measurements should be within equipment
manufacturer specifications. Tests for contrast are done monthly and
upon acceptance.
Test 5 - Slice thickness (sensitivity profile) 1,2,3,13,14
The purpose for the slice thickness test is to determine if
collimators, which shape the x-ray beam, correctly open to the
appropriate size set at the console. Traditionally, 45 degree tilted
ramps (aluminum or plastic), a spiral, or step wedge are used. There is a
hole drilled in them that allows the beam to pass through it projecting
an image of the width and length of the hole needed to match the width
of the x-ray beam exactly. For example: slice widths of 7mm or greater
should match nominal slice width within 2mm or less. But for narrower
slice widths, there is a larger margin for error, possibly even doubled.
This is generally just a matter of adjusting the calibration. This
traditional test was once fairly time-consuming however; today's spiral
and multi-slice scanners, new phantoms have been developed that can be
used as an insert with a CT performance phantom. It has virtually no
set-up time and is much more accurate. The new phantoms also reduce
partial volume averaging and do not compromise z-axis (patient
coverage). They can even perform accurately when not parallel to the
imaging plane.
After the phantom and test tool have been imaged, the stainless
steel bearings within the phantom create graphs on the test monitor
providing information as to whether or not the equipment is within
limits. The pixel values obtained from each image are plotted and the
data processed and normalized. This test is done to determine if the
beam is actually creating an exact match with the specified slice
programmed. So let�s look at how this test is performed on the GE
Lightspeed scanner.
For this test the phantom insert contains a block pattern of air
filled holes designed to demonstrate slice thickness. Each visible
hole or line in the phantom is at 1mm thickness that is aligned
perpendicular to the scan plane. It is important when determining slice
thickness that the display image is viewed at the recommended window
level and width for counting visible lines. The width is always set at
250 HU, and for a slice thickness of 10mm the level is set at 10 HU.
For 5.0mm slice thickness viewing is at 250/0 HU; 3.0mm slice thickness
viewing is at 250/-50 HU; 1.5mm slice thickness is viewed at 250/-100
HU. Standard limits are +/- 1mm. This test is done semi-annually and
upon acceptance.
The quality assurance phantom on the
right is used to perform several tests including the slice thickness
test. On the left is the resulting image used to determine slice
thickness. The block pattern showing line thickness appears along the
edges (yellow oval) of the pattern.
Because several factors affect viewing
of the line pairs for width sensitivity profile, it is recommended on
some scanners that the window/level setting be standardized. This chart
shows the standard viewing windows for the GE Lightspeed scanners for
each slice thickness being tested for. A slice thickness of 10mm is
viewed at W/L 250/50, whereas for 3mm the recommended viewing is 250/-50
W/L. When this test is properly performed and the collimators working
correctly, the number of visible lines should equal the chosen slice
thickness.
This image taken on the CT phantom
shows 10 one-millimeter lines on the corners of the image. The slice
thickness is 10mm according to this image, which is what was set at the
console for the slice thickness test. The window and level setting for
this image was set at 250/50 according to the manufacturer�s
recommendation. The results of this test confirm that the collimators
that shape the x-ray beam are open to the appropriate size. One can
also vary the slice thickness to test for linearity of the system.
Collimators that shape slice thickness should be accurate to +/- 1mm of
the setting at 10mm.
Test 6 - Localization device accuracy 1,2,3
There are two centering lights used in computed tomography
imaging, an internal and an external laser positioning markers. The
importance of the centering lights is to place the object at isocenter,
and to accurately represent slice location, especially during entry
needle placement for procedures such as a biopsy or fine needle
aspiration. Different scanners will have a different procedure and/or
test tool to evaluate localization light accuracy. The traditional
method uses a test device that has a "target." The beam is adjusted to
be on the center of the "X". Generally, the test tool consists of a
plastic phantom with 2 holes drilled at 45-degree angles crossing one
another but not touching. Centering the phantom markings to where the
two holes cross to the laser beam is how the scan is performed. The
resulting image should show the targeted image holes and the actual
image of the holes to be exactly aligned. If "L" (the distance from the
center of the CT slice to the target location) is 3mm., then
adjustments should be made. Because this is a clumsy method to identify
accuracy of the localization light newer more accurate methods are
used.
This photograph of a head phantom
demonstrates the "X" pattern seen with the localization lights. Notice
that there is an internal (green arrow) and an external (purple arrow)
light source that can be used to place the part in the isocenter of the
gantry. Coronal and sagittal (blue arrow) light field alignment must
also be tested. The photograph on the right shows the phantom setup
with the light localizer on to demonstrate testing procedure. Testing
the accuracy of the vertical and axial plane light localizers is
performed monthly and whenever new equipment is installed.
Localization light testing for the GE Lightspeed 16-slice
scanner, and most multidetector row CT scanners is simple and can be
tested for along with the slice thickness test. The insert used for the
slice thickness test has two deeper center holes on the reference that
are distinctly visible on the image. The position of the center holes
corresponds precisely to the line scribed on the circumference of the
phantom, which is aligned with the light field. When this alignment is
accurate to the axial light and vertical fields, the resulting image
should demonstrate a symmetrical hole around the center hole in the
slice thickness pattern. When using a line insert the longer line in
the pattern for slice thickness is the center alignment.
These CT phantom images demonstrate
light localization images from one type of phantom insert apparatus.
Notice the small holes in the image on the left that is aligned with the
laser light on the phantom. The holes are bored to the vertical center
of the phantom and aligned with the circumferential line of the
phantom. The white broken line on the right image connecting the
alignment holes is vertical alignment center. When a grid is placed on
the image the alignment with these holes indicate correct light
alignment. A grid placed on the phantom image allows us to evaluate
sagittal and coronal alignments of the light to the phantom. Both
internal and external lights are examined in this fashion for accuracy.
The CT image on the left with grid
lines shows the alignment of the center holes to be accurate with the
alignment of the laser lights in the axial and vertical planes. The
image on the left demonstrates both the slice thickness and the light
localization tests. The red arrows point to the center hole lines that
represent alignment of the test tool with the circumferential lines of
the phantom. Sagittal and coronal line accuracies must also be checked
using a grid to assure they are accurate. Laser light localization is
tested upon acceptance and monthly.
This CT phantom image on the left
shows good alignment of the axial laser lights to the center holes of
the phantom. Relative points on the image used to measure sagittal and
vertical alignment of the light field are marked with red and yellow
lines. These markings on the grid indicate relative points of
light-phantom alignment are within sagittal and vertical plane limits.
For QA purposes the measurement is taken from the visualized hole on one
side to a selected point on a grid; the same relative measurement is
performed on the opposite side too. Coronal alignment is verified by
visualization of the holes in the phantom image.
Test 7 - Table/bed indexing accuracy 1,2,3
This test is done to ensure the distance the bed is moving
between scans is accurate to what the equipment reads. This test is
performed at same time as localization test. One way to do the test is
with a piece of x-ray film taped to the table. The piece of film is
generally covered or in a holder to prevent exposure. A series of 10 to
12 scans are done 10mm apart. A scan can be done at any thickness to
check indexing. Using a ruler or tape measure, determine the distance
between bands. The distances should be equal between selected
increments. If there is more than a +/-2 mm. difference between any
slice, equipment needs to be serviced. This test is done monthly and
upon acceptance.
Patient dose 1,2,3,7,9,12,15,16
When we talk about image quality, especially contrast
resolution, noise, spatial resolution, image artifacts, it is important
to realize quality assurance assures high quality images and acceptable
levels of patient radiation dose. Image quality is not always performed
at maximum resolution because it requires an increase in patient
exposure and patient dose. When it comes to patient dose and dosimetry
it is important that the radiographer understands that dosimetry and
performance standards must comply with Federal Regulation 21CFR
1020.33(C). The Food and Drug Administration (FDA) have established
Specific CT dose indices in this report. The purpose of the report is to
relate radiation dose and image quality of the CT scanner. Often CT
technologists find themselves in discussions concerning patient dose for
a given scan type. In order to properly relate dose information to the
patient or physician it is important to understand how dose is
calculated and the terminology of CT dose, which is slightly different
from general radiography dose.
This example of a daily phantom test
report gives the calculated CTDI and DLP. By monitoring this parameter
daily the radiation safety officer is able to determine that the CT
computer is accurately calculating dose. In order for these
measurements to be valid it is important the radiation physicist
performs the quality assurance testing for dose. It should be pointed
out that QA dose calibrations and the are perform quality by the
radiation physicists. However, technologists should understand how this
test is performed and its relationship to the dose calculated by the
scanner.
The whole purpose of dose calculation is to assess patient
biological risk since dose is related to risk, and highly dependent on
the tissue exposed. In the past, for example in general radiography,
studies such as chest x-rays, fluoroscopy, and so forth were performed
for many years without calculating patient dose. This is because the
effective dose, which is the proper way to express dose, is very
difficult to determine for a plain film x-ray. Now with computed
imaging, especially CT scan, it is easier to calculate effective dose.
What we mean by effective dose is that dose summed from the weighted
dose and radiosensitivities of specific organs or tissues exposed. The
doses are reported in documents published by the National Council on
Radiation Protection (NCRP) and ICRP 60 (International Committee on
radiation Protection, Publication 60). It is not possible to
characterize the specific dose any one individual may receive.
Because it is difficult to determine specific dose of a patient
the CT computer calculates what is called the CT Dose Index (CTDI). The
CTDI is determined from calibrated doses at specific points in a
phantom. A head phantom is used to calculate head dose, and a body
phantom for body dose. The dose is taken in the phantom at center and
peripheral locations. Therefore, the CTDI dose is the dose absorbed in
the phantom material polymethyl methacrylate (PMMA) at a point volume of
+/- 7 contiguous slices adjacent to the point. A calibrated CT scanner
is able to calculate patient dose in the same manner; however, this is a
limited estimate of dose because only 14 slices are used in the
calculation (+/- 7 slices). The calculated dose for CTDI is defined by
U.S. Federal Regulations as:
The formula for CTDI is specified in
U.S. Federal Regulation 2121CFR 1020.33 (C). n=number of image slices
per scan, T + slice width per image, D(z) = Z-axis dose profile
(absorbed in PMMA). What is important for technologists is that CTDI is
a calculated dose based on phantom calibrations to calculate patient
dose. It is a limited dose calculation because it represents the dose
from a single CT slice.
Another term called the CTDI100 provides a better relative index when thin slices are taken on the patient and when helical scanning is used. The CTDI100 is a single reported dose that consists of 2/3 CTDI100 peripheral doses and 1/3 CTDI100
central dose. The dose spans 100mm thickness whereas the CTDI only
covers 14 slices (+/- 7 slices from adjacent selected point). The
combined peripheral and central doses are reported as CTDIw, a relative dose for the entire scan. The CTDIw,
which is specified in the international standard IEC 601-2-44 drafts is
calculated from the CTDI and reported on the patient dose report. You
may be asking why we need both a CTDI and CTDIw calculation.
The answer is that data collected during helical scan length is greater
than the reconstructed image region, and CT irradiation is from a
series of narrow x-ray beams. This is because multiple helical scanning
requires overlap exposure areas. This is more pronounced when the scan
is short, so this overlap is calculated into the CTDIw.
The mathematical formulas for CTDI100 and CTDIw
are shown above. What is important to know is that a body or head
phantom is used to calibrate the scanner to calculate these doses. The
scanner calculates the CTDIw for a patient scan. The
technologist does not need to perform these calculations, but should
understand their relationship to calculated patient dose.
The dose length product (DLP) is the final calculation
performed by the scanner’s computer. This is a raw calculation that
must be converted using a tissue-weighted factor to arrive at the
effective dose. The effective dose is the best estimate of patient risk
because we can compare this dose to other x-ray procedures. The dose
length product is the volume CTDI (CTDIvol) multiplied by the scan length (slice thickness × number of slices) in centimeters. The volume CTDI is defined as CTDIw divided by the beam pitch factor, which is the most commonly cited index for modern MDCT equipment.15
DLP is expressed in mGyCm (milliGray Centimeters). DLP displayed at
the end of a scan is computed for each group in the scan and for the sum
of the groups. While DLP is a good measure for managing patient dose a
minor downside is that it is nonspecific for what is actually scanned.
It does not matter if the patient is a 10-pound infant or a 200-pound
adult, or if the scan is of the brain vs. chest or abdomen. So DLP data
is used retrospectively to estimate effective dose equivalence (ED).
To more accurately calculate effective dose a weighted factor for organ
dose is applied.
A classical scenario encountered in CT imaging is that the
patient or patient’s physician inquires about dose prior to ordering a
CT scan. For example, the patient is 5 months pregnant, presents with
new onset chest pain and swollen right leg. The physician wishes to
perform a pulmonary CT angiogram to rule out pulmonary embolus.
However, the question is what is the estimated patient dose and risk for
the scan? Average broad estimates of effective dose (E) for common CT
procedures should be available at the scanner. These estimates may be
derived from values of DLP for an examination using appropriately
normalized (weighted tissue) coefficients:
The mathematical formulas for CTDI100 and CTDIw
are shown above. What is important to know is that a body or head
phantom is used to calibrate the scanner to calculate these doses. The
scanner calculates the CTDIw for a patient scan. The
technologist does not need to perform these calculations, but should
understand their relationship to calculated patient dose.
For a CT procedure in the chest, one would then multiply the DLP for that procedure (in mGy-cm) by the factor 0.017 mSv mGy-1 cm-1 to get an effective dose in mSv.
Now let’s get back to our reading as calculated by the CT
scanner for the daily phantom test. In the example below we see that
the scanner calculates the DLP, but does not calculate the effective
dose equivalency. ED is not calculated routinely, but is a
retrospective calculation should you need to explain risk to the patient
or their physician. As a CT technologist you should know how to read
the DLP dose report produced at the completion of a scan. The dose from
the scout images is not reported since there is currently no standard
to follow. Generally the scout images are not made for diagnostic
reading purposes, just for alignment and are low dose. Scouts are a
small part of the total patient dose, which is considered to have no
practical dose significance.
Note the volume CT dose index
(CTDIvol) (arrow) and dose length product (DLP) (asterisk). Scan
parameters: mAs 150, kVp 120, rotation time 0.6 sec. Effective dose in
mSv = DLP × conversion factor: 193.14 mGy-cm × 0.02185 mSv/mGy-cm-1 (conversion factor) = 4.23 mSv. (Conversion factor ICRP Report 103).
The dose from the scout images is not
reported since there is currently no standard to follow. This is also
true for routine computed radiography exams too. Generally, the scout
images are made with substantially low dose for image alignment.
Because of the low exposure technique and low dose these images are not
used for diagnostic reading. The DLP calculated for this head CT is
based on accuracy of the calculation-performed daily as in the previous
example.
Now that we have examined dose data and how the CT scanner
calculates dose, we have to ask, “what do we do with effective dose
data?” The answer is that this data permits us to compare patient EDs
from CT with those of other x-ray based imaging examinations, as well as
with benchmark values, including natural background radiation and
regulatory dose limits. Since effective dose is measured in mSv it is
possible to compare dose from different modalities. Table 1 below lists
of representative diagnostic procedures and associated doses.
| Table I. - Radiation Dose Comparison
|
| Diagnostic Procedure
|
Typical
Effective Dose
(mSv)
|
Number of
Chest
X rays
(PA film)
Equivalent
Effective Dose
|
Time Period for
Equivalent Effective
Dose from Natural
Background
Radiation
|
Chest x ray
(PA film)
|
0.02
|
1
|
2.4 days
|
| Skull x-ray
|
0.1
|
5
|
12 days
|
| Lumbar spine
|
1.5
|
75
|
182 days
|
| I.V. urogram
|
3
|
150
|
1 year
|
| Upper G.I. exam
|
6
|
300
|
2 years
|
| Barium enema
|
8
|
400
|
2.7 years
|
| CT head
|
2
|
100
|
243 days
|
| CT abdomen
|
8
|
400
|
2.7 years
|
Average effective dose in
millisieverts (mSv) as compiled by Fred A. Mettler, Jr., et al.,
"Effective Doses in Radiology and Diagnostic Nuclear Medicine: A
Catalog," Radiology Vol. 248, No. 1, pp. 254-263, July 2008. This chart
compares the average “effective dose” of selected imaging studies
compared to the average effective dose from PA chest x-ray of 0.02 mSv.
The assumed average "effective dose" from natural background radiation
in the United States is 3 mSv per year.
A CT examination with an effective dose of 10 millisieverts
(abbreviated 10 mSv; 1 mSv = 1 mGy in the case of x rays.) may be
associated with an increase in the possibility of fatal cancer of
approximately 1 chance in 2000.7. This increase in the
possibility of a fatal cancer from radiation can be compared to the
natural incidence of fatal cancer in the U.S. population, about 1 chance
in 5. In other words, for any one person the risk of radiation-induced
cancer is much smaller than the natural risk of cancer. Nevertheless,
this small increase in radiation-associated cancer risk for an
individual can become a public health concern if large numbers of the
population undergo increased numbers of CT screening procedures of
uncertain benefit18.
A CT examination with an effective dose of 10 millisieverts
(abbreviated 10 mSv; 1 mSv = 1 mGy in the case of x rays.) may be
associated with an increase in the possibility of fatal cancer of
approximately 1 chance in 20007. This increase in the possibility of a
fatal cancer from radiation can be compared to the natural incidence of
fatal cancer in the U.S. population, about 1 chance in 5. In other
words, for any one person the risk of radiation-induced cancer is much
smaller than the natural risk of cancer. Nevertheless, this small
increase in radiation-associated cancer risk for an individual can
become a public health concern if large numbers of the population
undergo increased numbers of CT screening procedures of uncertain
benefit18.
When the radiation safety officer or radiation physicist
performs QA dose measurements either a head and body technique and head
or body phantom is used. A pencil ionization chamber attached to an
electrometer is placed within drilled 1 cm holes at various locations of
an acrylic phantom and measured. This is the easiest and most accurate
way to measure dose. Dosage is measured in mR (millirad) and is
specified as the CTDI or CT dose index. The ionization chamber is
placed anterior, posterior, in the center, and around the perimeter of
the phantom to determine various simulated patient doses. Holes that are
not utilized during testing should be filled with an acrylic plug. The
typical positions tested are anterior center and at surface max which
is usually at the anterior 12:00 position. Results should not vary more
than 10% from one assessment to the next. This test is usually done
annually and upon acceptance.
This is a picture of a pencil
ionization chamber that is inserted into a head or body phantom for
patient dose calibration testing. A head phantom of 16 cm diameter is
used for head dose calculation, and a 32 cm diameter body phantom is
used for calibrating dose to the torso. Phantom material must be made
of polymethyl methacrylate (PMMA) with holes greater than 14 cm thick.
The pencil ionization chamber is
inserted into the center hole (A) and peripheral holes that are 1 cm
from the surface (A through E) in the 12 o’clock, 3 o’clock, 6 o’clock,
and 9 o’clock positions to determine various simulated patient doses.
Positions B through E measure peripheral doses and center dose is
calculated at position A. Holes that are not utilized during testing
should be filled with an acrylic plug.
For a list of complete tests or for those not mentioned, please see
the equipment manufacturers� specification manual for your specific
model or unit. The Quality Control survey from each facility's medical
physicist will also outline all required tests at the state and federal
levels along with data outlining past test results for comparison.
Conclusions
As CT has evolved through the years, various phantoms and test
tools have been used to test CT scanners and equipment performance.
These tests reveal a great deal of information to medical physicists,
patients, physicians, and quality assurance personnel. Conventional CT
phantoms for single slice scanners are not adequate for newer helical
multi-slice technology. The Food and Drug Administration has established
standards for phantoms and specific tests for CT scanners.
Manufacturer and model specific phantoms that meet these standards are
widely available. New phantoms are constantly being developed to
evaluate problem areas in maintaining quality images. Proper use of
quality assurance phantoms for daily QA testing, recording daily test
data, as well as understanding tests described in this module remain a
duty of the CT technologist.
CT is considered a high-dose procedure and accounts for 35% of total patient dose.2 The dose from CT is approximately that of an average fluoroscopic study.2
Currently, there are no dose limits for patients receiving a CT scan.
However, the need for the exam must outweigh the risks to the patient.
Refer to maximum permissible dose (MPD), and the ALARA principle
(as low as reasonably achievable) in the standards and goals of the
(NCRP) National Council on Radiation Protection and Measurements.2 The general non-patient population is limited to 5 mSv/yr (500 mrem). Occupational dose limit is 50mSv/yr (5000 mrem).2,15 Each state and facility may have different limits set forth by their own governing agencies.
The measured dose from a spiral scanner can vary from
manufacturer and model, to the age of the equipment. It is important to
base all dose results on the specific equipment for your facility. You
may, of course, choose to compare your results with other facilities and
their quality assurance programs. A good quality assurance program
containing accurate documentation, along with the newest testing
devices, are necessary in maintaining the highest standard of quality
and the minimum quantities of radiation exposure possible to our
patients. Because of the excellent diagnostic information that CT
provides, it will continue to be an essential tool for medical
diagnosis.
For a list of complete tests or for those not mentioned, please
see the equipment manufacturer specification manual for your specific
model or unit. The Quality Control survey from each facility's medical
physicist will also outline all required tests at the state and federal
levels along with data outlining past test results for comparison.
Summary Points!
Under the Radiation Control for Health and Safety Act, all CT
equipment is subject to an equipment performance standard. Within this
act are minimum radiation safety requirements and manufacturer
requirements regarding safety standards and performance.
Quality assurance is the component that measures CT scanner
performance to assure operation is at an acceptable level. Quality
control takes action to correct inadequacies before they are
problematic. These measures assure high imaging performance standards,
and reduce the risk of patient harm due to change in equipment
performance characteristics.
There are three basic tenets of an acceptable quality control
program. QA must be performed on a regular basis, there must be prompt
interpretation of test results, and the third tenet is accurate
bookkeeping.
There are two broad categories of QA test phantoms: calibration
phantoms and imaging phantoms. A calibration phantom is used for
testing detectors and correcting quantitative information obtained from
digital images.8 Imaging phantoms assess image quality and are usually
further classified as head, body, standard, or reference phantoms.
QA testing is performed upon receiving new equipment or if in use equipment is relocated, or the x-ray tube is replaced.
Scan data is based on penetration and attenuation measurements of
photons as they traverse matter. This raw data must then be converted
to digital data and displayed as a CT image. Each pixel in the CT image
is assigned a number collectively referred to as a CT number. CT
numbers are referred to as Hounsfield Unit (HU) named after the inventor
of computed tomography.
The Hounsfield scale (H scale) is a calculation from the linear
attenuation coefficients of tissues that make up an image slice. CT
numbers are relative to the attenuation of water, which is assigned a
value of 0 HU.
To perform the CT number test a water phantom test tool (this is a
water-filled cylinder with a 20cm diameter for head, or 32 cm diameter
for body) is used. The phantom is scanned using a head technique and a
20 cm FOV. Using a 2-3 cm ROI containing about 200-300 pixels and
measure the average CT number. Air measures -1000 HU; water should be
close to 0 (zero). Water should not exceed +/-3 HU and no more than +/- 5
HU from center to periphery. The CT number for water is one of the
daily tests performed on the CT scanner.
Noise in an image affects low contrast visibility. Noise level can
be stated as a percentage of image contrast in CT numbers. The stand
deviation for noise should be in the range of +/- 3. Since CT numbers
range from +/-1000 HU, noise should be less than 0.3%. The maximum
standard deviation between the center ROI and any peripheral ROI is less
than +/- 5 HU.
<.li> Spatial resolution is important in detecting the edges of
structures, margins of tumors, small foreign bodies, and small bony
structures. High (spatial) resolution test measures spatial resolution
by measuring the high-resolution pattern in a phantom image.
The MTF is the most common method of describing spatial resolution
in CT, digital radiography, and film-screen radiography systems. The
MTF analysis allows us to compare system performance on a day to day
basis, or to compare a system’s performance against another CT system.
Modulation transfer function is expressed in line pairs per centimeter
(lp/cm). When counting a line pair, one line and its adjacent space are
called a line pair.
An object’s size in a given length is also known as its spatial
frequency. What this means is that how frequently an object fits into a
given space is it spatial frequency. Generally speaking, the smaller
an object is the higher spatial frequency and the more difficult it is
to be displayed accurately. Likewise, a large object will have low
spatial frequency and will be more accurately displayed.
If an object is displayed accurately as it is, then the MFT is
given a value of 1.0. The modulation transfer function scale is from
zero to 1. A MTF value of 0 would mean the image is blank and contains
no information about the object scanned. Scanned objects will have
values between 0 and 1; however, the closer to 1 an object is the better
the MTF of the scanner.
There must be at least a 10% density difference between structures
to be detected with conventional x-ray imaging. This is why we use a
15% increase or decrease in the kVp to change the scale of contrast in
conventional radiography. CT is able to resolve density differences as
little as 0.1%. Most of these subtle densities are in the low contrast
scale of the image.
Low-resolution contrast is determined as the differenced in HU of
objects and background. High noise in the image will cause a decrease in
low-contrast resolution. To visualize small holes with differing
contrast embedded in a polystyrene membrane test tool we subtract the CT
number for water from the membrane measurement. The derived HU number
is used to set contrast level for viewing the number of visualized
holes.
The purpose for testing slice thickness (sensitivity profile) is to
determine if the collimators which shape the x-ray beam are correctly
open to the appropriate size set at the console. The accuracy of
collimators is important for limiting patient dose as well as image
quality assurance.
Most modern CT scanners provide dose information in terms of CT
Dose Index (CTDI) and its relative, dose length product (DLP). CT dose
is measured in units of mGy (milli Gray). Former units of measurement
was the radiation absorbed dose or Rad (1 rad equals 10 mGy). Exposure
is measured in units of coulombs/kilogram (C/kg).
Dosimetry and CT image quality must comply with Federal
Regulation 21CFR 1020.33(C). The Food and Drug Administration (FDA)
have established Specific CT dose indices in this report. The purpose of
the report is to relate radiation dose and image quality of the CT
scanner.
The whole purpose of dose calculation is to assess patient
biological risk since dose is related to risk, and highly dependent on
the tissue exposed.
The CT computer calculates CTDI and dose length product (DLP),
which are the final dose calculations performed by the scanner’s
computer after the patient’s scan. These are raw calculations that must
be converted using a tissue-weighted factor to arrive at the effective
dose. The effective dose (E) is the best estimate of patient risk
because we can compare this dose to other x-ray procedures. Effective
dose is reported in units of mSv.
References:
- Seeram, E. Computed Tomography-Physical Principles, Clinical
Applications, and Quality Control. Pennsylvania: W.B. Saunders Co.;
1994.
- Bushong, SC. Radiologic Science for Technologists- Physics,
Biology, and Protection, 7th ed. Missouri: Mosby; 2001: 392-439,
520-523, 533-555.
- (No authors listed) Compliance Guidelines for Computed Tomography
Quality Control. New Jersey Department of Environmental Protection;
Bureau of Radiological Health. Jan 11, 2001:5-43. Website: http://
www.state.nj.us/dep/rpp. Accessed June22, 2004.
- Hsieh J. Investigation of an image artefact induced by projection
noise inhomogeneity in multi-slice helical computed tomography. Phys Med
Biol. 2003 Feb 7; 48(3): 341-56.
- 5. Suess C., Kalender WA. Coman JM. New low-contrast phantoms
for computed tomography. Med Phys. 1999 Feb; 26(2): 296-302. Available
at: http://www.ncbi.nlm.nih.gov/entrez/query. Accessed May 25, 2004.
- U.S. Food and Drug Administration. How does FDA regulate CT
systems? 2002. Available at: http://
www.fda.gov/cdrh/ct/regulatory.html. Accessed May 30, 2004.
- AAPM Fact Sheet. AAPM online. The website of the American
Association of Physicists in Medicine. 2004. Available at: http://
www.aapm.org. Accessed July 19,2004.
- Phantom Applications and Technology Overview. CIRS- Computerized
Imaging Reference Systems website 2001. Available at: http://
www.cirsinc.com/overview.html. Accessed June 22, 2004.
- ACR Technical Standard for Diagnostic Medical Physics Performance
Monitoring of Computed Tomography (CT) Equipment. ACR-American College
of Radiology web site revised 2002 (res. 21) Effective 1/1/03. Available
at: http:// www.acr.org. Accessed June 8, 2004.
- Ballinger, PW. Frank, ED. Merrill’s Atlas of Radiographic Positions
& Radiologic Procedures. Vol3. 10th Ed. Missouri: Mosby;
2003:330-371.
- Morin, RL. PhD, FACR. Computed Tomography Quality Control; Department of Radiology. Mayo Clinic. Jacksonville. 2004:3-23.
- Spiral/Helical CT Phantom. CIRS- Computerized Imaging Reference
Systems 2001. Available at: http://
www.cirsinc.com/products/model061.html. Accessed May 26, 2004.
- CT-SSP (Slice Sensitivity Profile) Point Response Phantom. Cardinal Health. Nuclear Associates Sales Manual.2004.157.
- Davros WJ, Herts BR, Walmsley JJ, Obuchowski NA. Determination of
Spiral CT Slice Sensitivity Profiles Using a Point Response Phantom
Journal of Computer Assisted Tomography. 1995;Vol. 19, No. 5:838-843.
- Yoshizumi TT, Nelson RC. Radiation issues with multidetector row helical.Crit Rev Comput Tomogr. 2003;44:95-117.
- U.S. Food and Drug Administration. What Are the Radiation Risks
from CT? 2002. Available at: http:// www.fda.gov/cdrh/ct/risks.html.
Accessed May 30, 2004.
- ACR Position Statement. ACR Practice Guidelines and Technical
Standards. ACR- American College of Radiology. Website. Revised 2002
(res.9). Available at: http:// www.acr.org. Accessed June 8, 2004 and
January 11, 2010.
- http://www.fda.gov/Radiation
EmittingProducts/RadiationEmittingProductsandProcedures/MedicalImaging/MedicalX-Rays/ucm115329.htm
Accessed January 10, 2010
|
No comments:
Post a Comment