RADIATION PROTECTION IN DIAGNOSTIC AND INTERVENTIONAL RADIOLOGY

IAEAInternational Atomic Energy Agency

RADIATION PROTECTION INDIAGNOSTIC AND

INTERVENTIONAL RADIOLOGY

Part 18: Optimization of protection in CT scanner

Practical exercise - Quality Control

IAEA Training Material on Radiation Protection in Diagnostic and Interventional Radiology

IAEA 18: Optimization of protection in CT scanner 2

Contents

Quality control tests

• CT accuracy, uniformity, linearity. and noise,

• Low and high contrast resolution

• Z-axis sensitivity

• Alignment, table top travel accuracy

• Gantry tilt measurement

• Dosimetry



Topic 1: Quality Control



Physical parameters of CT image

Image quality• May be expressed in terms of physical parameters such

as uniformity, linearity, noise, spatial resolution, low contrast resolution

• Image quality depends on the technical characteristics of the CT scanner, the exposure factors used and image viewing conditions.

• Quality may be assessed by quantitative measurement using test phantoms, and by the appearance of artifacts.

• Measurements should be conducted regularly


Scanner performance: technical parameters (I)

Test Phantoms:• Test phantoms of a standardized human shape

or test objects of a particular shape, size and structure, are used for the purposes of calibration and evaluation of the performances of CT scanners

• Phantoms should allow for evaluation of CT number; uniformity; noise; spatial resolution; slice thickness; dose; positioning of table top


Scanner performance: technical parameters (II)CT Number Accuracy

• CT number depends on tube voltage, filtration, object thickness• CT number of water is by definition equal to 0 • Measured CT number should be < ± 4 HU in the central ROI

CT Number Linearity• The linear relationship between the calculated CT number and the

linear attenuation coefficient of each element of the object• Deviations from linearity should be < ± 5 HU

CT Number Uniformity• The CT number of each pixel in the image of an homogeneous object

should be the same over the image area• The difference in the CT number between a peripheral and a central

region of an homogeneous test object should be < 8 HU• Differences are largely due to beam hardening


Noise• The local statistical fluctuation (standard deviation) of CT

numbers in a homogeneous Region Of Interest (ROI)• Noise strongly affects the low contrast resolution• Noise is dependent on the radiation dose

• Image noise should be measured over an area of about 10% of the cross-sectional area of the test object.

• Goal— obtain an image with an acceptable level of noise while keeping the patient dose as low as reasonably achievable

Scanner performance: technical parameters (III)

doseNoise 1


Scanner performance: technical parameters (IV)

Spatial Resolution• The high contrast resolution determines the minimum

size of details visualized in the plane of the slice with a contrast >10%.

It is affected by:

• the reconstruction algorithm

•the detector width

• the effective slice thickness

•the object to detector distance

• the x-ray tube focal spot size

•the matrix size.


Scanner performance: technical parameters (V)

Spatial Resolution• The low contrast resolution determines the size

of detail that can be visibly reproduced when there is only a small difference in density relative to the surrounding area• Low contrast resolution is limited by noise.

• The perception threshold in relation to contrast and detail size can be determined by creating contrast-detail curve.


Scanner performance: technical parameters (VI)

Slice Thickness• The slice thickness is determined in

the center of the field of view as the distance between the two points on the sensitivity profile along the axis of rotation at which response has fallen to 50%.

• The use of post-patient collimation to reduce the width of the image slice leads to very significant increases in the patient dose

Positioning of couch• The accuracy of positioning of the

patient couch is evaluated by moving the loaded couch a defined distance and moving it back to the start position


Minimum requirements: CT scanner (I)

• Image noiseThe Standard Deviation of CT numbers in the central 500 mm2 ROI for a water or tissue equivalent phantom should not deviate more than 20% from the baseline.

• CT number valuesThe deviation in the CT number values for water or tissue equivalent material and materials of different densities should <± 20 HU or 5%.

• CT number uniformityThe SD of the CT number over a 500 mm2 region of interest for water or tissue equivalent material at the center and at the periphery of phantom should vary by less than 1.5% of the baseline


Minimum requirements: CT scanner (II)

• Computed tomography dose index (CTDI)

The CTDI for a single slice for each available beam shaping filter and for each available slice thickness should not deviate more than ± 20% from the baseline.

• Irradiated slice thickness

The FWHM of the dose profile should not differ more than ± 20% from baseline.

• High contrast resolution (spatial resolution)

The FWHM of the point spread function of a pin, or the edge response function of an edge should not differ more than ± 20% from baseline.

• Low contrast resolution

Polystyrene pins of 0.35 cm diameter inserted in a uniform body water phantom should be visible in the image.



Topic 2: Noise



Imaging performance (Noise)

• Noise is assessed using cylindrical phantoms, which are either filled with water or made of a tissue equivalent material

• Once an axial image of the phantom has

been acquired, noise is measures as the standard deviation in CT number in a region of interest (ROI) placed in the center of the image


Region of Region of interest interest (ROI)(ROI)




• Noise values are given in manufacturers’ specifications are quoted for a specific phantom (e.g., manufacturer’s QA phantom) and for specified scan techniques

• These conditions must be matched exactly for the purposes of the acceptance test

• Manufacturers quote noise at a particular surface dose

• If this is the case, dose for axial scans can be measured by taping an ion chamber to the surface of the phantom



• Baseline noise values should be obtained for several scan protocols that will be used clinically, using the routine QC noise phantom

• To ensure that noise figures are both accurate and representative, it is essential to find the mean value from several scans (10 scans.)



Topic 3: CT uniformity and linearity



CT number uniformity

• CT number uniformity can be assessed at the same time as measuring noise, by placing four additional ROI (N, E, S and W) at positions near the edge of the image of a uniform phantom

• The mean CT number is then noted for these four regions, as well as the center

• The deviation from the central value should be calculated

• The CT number uniformity should be measured for large fields of view


Axial image of Axial image of a homogenous a homogenous phantomphantom




• CT number linearity is assessed using a phantom containing inserts of a number of different materials (materials should cover a wide range of CT numbers)

• One example of a suitable phantom to use at acceptance is the Catphan (The Phantom Laboratory, Salem, NY), which contains four inserts with CT numbers ranging from approximately -1000HU to +1000HU





Topic 4: Low and high contrast resolution



Low contrast resolution

• Low contrast resolution (or low contrast detectability) is often quoted in specification documentation as the smallest visible object at a given contrast for a given dose

• Since this measurement relates directly to imaging performance, it is an important value to verify at acceptance testing

• At least 20 images of the low contrast insert (LCR) should be acquired and then viewed by at least 3 observers under optimal viewing conditions to obtain an average


Typical Typical image of the image of the Catphan LCR Catphan LCR insertinsert

Low contrast resolution


Spatial resolution (high contrast)

• There exist two broad categories of measuring techniques : • analysis of the point spread function, usually by

calculation of the modulation transfer function (MTF) • either objective analysis or visual assessment of images

of a resolution bar phantom.

• The resolution is quoted as the spatial frequency (in line pairs/cm) at which the modulation falls to 50%, 10% or 2% MTF.

• These figures are often given for more than one reconstruction algorithm, e.g., for standard and high-resolution scans.


Spatial resolution (high contrast)

• The number of line pairs per cm just visible in the image is approximately equivalent to the 2% value of the MTF

• This result can then be compared with the 2% MTF, as quoted in the manufacturer’s specifications



Topic 5: Z-axis sensitivity



Z-Sensitivity (Imaged slice width)

Axial measurements • Phantoms used for axial measurements may contain

thin metal plates, wires or arrays of air holes, inclined at an angle to the image plane

• Manufacturers should be able to supply an appropriate phantom or, alternatively, the Catphan contains an insert suitable for this test

Note: to obtain meaningful measurements, the thickness of the plates, wires or holes cannot be greater than the nominal slice width concerned. This may create problems for the sub-millimetre slice widths offered on multi-slice scanners.



Axial measurements • Phantoms manufacturers may quote the tolerance

for each nominal slice width setting in their specification documentation

• Z-sensitivity measurements in axial mode can be used to check that imaged slice widths are within the tolerances given

• This can also be used in conjunction with irradiated slice width measurements to assess the accuracy of post patient collimation and to calculate the geometric efficiency for the scanner


Plan view of a test object used to measure imaged slice widths for axial scans




Topic 6: Alignment



Alignment of indicating lights with scan, coronal and sagittal planes

• Several methods can be used to perform these tests

• The techniques described here are straightforward to implement and require minimal test equipment.


Agreement between internal and external scan plane lights

• Use an envelope-wrapped film (therapy verificaiton film) for the measurement. However, a piece of paper or card can also be used

• The film is placed flat on the table and illuminated by the external scan plane light

• The position of the light is marked on the film envelope and the table is moved automatically to the scan plane

• If the distance between the internal and external lights is correct, the internal light should now coincide with the mark on the film envelope.


Co-incidence of internal scan plane lights and scan plane

• Pin pricks are made in a piece of therapy verification film (or similar) along the line of the internal scan plane light

• The film is wrapped around a phantom and scanned with a narrow axial and developed

• Coincidence between the pin pricks and the x-ray beam exposure indicates alignment between the alignment lights and the scan plane



• For multi-slice scanners, pin pricks (and, thus, the internal alignment light) are usually found to coincide with the center of the four slices

• To plan a scan so that the x-ray beam is centered over the internal scan plane lights at zero, it will be necessary to center the first and last slices symmetrically around zero (e.g. four slices from –7.5 mm to +7.5 mm on a 4 x 5 mm scan)


Z

X

Pin pricks made in film at position of scan plane light

X-ray beam exposure



Coronal and Sagital plane lights

• A long, thin object, with a high CT number relative to air, such as the ‘lead’ in a pencil or a straightened paper clip, can be used as a marker to perform this test

• The marker is supported above the patient table and aligned, using the indicating lights, so that it is positioned at the isocentre, parallel to the z-axis and perpendicular to the scan plane

• If indicating lights are accurately aligned to the coronal and sagital planes, the marker should appear as a dot at exactly x = 0, y = 0 on the axial image.



Topic 7: Table travel accuracy



Table travel accuracy

• A ruler or tape measure placed alongside the table, can be used to check that the amount of table movement indicated on the gantry agrees with the actual distance moved.

• A load of approximately 70- 80 kg should be placed on the table in order to simulate the weight of a patient.

• The test should be performed twice: by driving the table top both away from and towards the gantry.


Assessment of distance indicator accuracy

Table travel accuracy


Axial scan incrementation accuracy

• Verification of incrementation accuracy between successive axial slices can be achieved by placing therapy verification film on the couch (in the isocentre plane) and exposing it to an axial scan sequence

• Narrow slices separated by a couch increment greater than 1 slice width can be used, and the distance between the lines on the film measured


Table travel accuracy for helical scans

• In helical scanning, it is not sufficient to use a simple mechanical test because the distance imaged depends on couch speed and scanner software

• One method of assessing distance accuracy is to use an acrylic test object containing two small radio-opaque markers, separated by a fixed distance (e.g., 20 cm)

• The test object is scanned in Scan Projection Radiography (SPR) mode and a helical scan is planned to start at the first marker and to end at a distance x from the first marker

• If couch travel is accurate during the helical scan, the markers should be clearly seen on the first and final images of the series.


Couch travel accuracy for helical scans



Topic 8: Gantry tilt measurement



Gantry Tilt

• The accuracy of displayed gantry tilt can be assessed by supporting therapy verification film at the gantry end of the patient table

• The film must be held vertically (e.g., by taping to an acrylic block), so that it is parallel to the sagital plane and intersects scan and coronal planes at right angles


Gantry Tilt

• Three axial exposures are made using the same film:• one for the maximum superior gantry tilt,

• one for the maximum inferior gantry tilt

• one at 0º gantry tilt

• The three scan planes should then be visible on the developed film

• The angles + and - between scan planes at maximum tilt relative to that at 0º tilt should equal tilt angles displayed on the gantry.


Assessment of accuracy of gantry tilt



Topic 9: Dosimetry



Dosimetry - CTDI in air

• The Computed Tomography Dose Index (CTDI) in air can be measured using a 10 cm pencil ionization chamber, located in the scan plane at the isocentre

• The ion chamber can be supported using a laboratory test stand and clamp


Ionization chamber

Table




• It is important to measure the CTDI in air under the following conditions When commissioning a CT scanner:• For all beam shaping filters

• For all nominal slice widths

• For all clinical kV settings

• For a range of scan times

• For a range of mA settings






Axial slice positionsAxial slice positions

Helical scan (pitch 1)Helical scan (pitch 1)

Dosimetry - CTDI in air (helical)


Dosimetry - CTDI in Acrylic Phantoms

• The body phantom is placed on the patient table and the head phantom is supported in the head rest

• Phantoms are aligned and centred at the scan isocentre

• The ion chamber is inserted into either the central or one of the peripheral holes of the phantom (all other holes being filled with acrylic rods)


Dosimetry - CTDI in Acrylic Phantoms

• Dose measurements at the center are used to calculate the central CTDI

• Peripheral CTDI is measured in at least four positions around the phantom, so as to achieve a true average

Note: If gantry rotation is initiated from different angular positions for successive scans it may be necessary to take a number of measurements at each position in order to get a representative mean dose


Head phantomHead phantomBody phantom Body phantom (or annulus (or annulus

to fit over head phantom)to fit over head phantom)

Insert to plug holesInsert to plug holes

Dosimetry - CTDI in Perspex Phantoms


• Central and peripheral CTDI’s are used to calculate weighted CTDI, CTDIw:

• CTDIws can be compared against diagnostic reference levels for standard scan examinations

Dosimetry - CTDI in Perspex Phantoms

)( CTDI3

2 + CTDI

3

1

C

1 = CTDI p100,c100,wn


Example of CTDI Values for some CT units

MODEL KV FAD S nCDTI (mGy/mAs) Air weighted

Head Body Siemens HIQ

133 700 10 0.195 0.161 (0.83) 0.093 (0.48)

Siemens Plus S

120 700 10 0.128 0.110 (0.86) 0.062 (0.48)

Siemens Plus S

137 700 8 0.161 0.082 (0.51)

GE Pace 120 525 10 0.344 0.200 (0.58) 0.094 (0.27)

Philips LX

120 606 10 0.200 0.160 (0.80) 0.081 (0.41)

FAD(mm): Focus to Axis distanceS (mm): Slice thickness


Dose Profiles (Irradiated slice width)

• Measurement of irradiated slice widths, for all nominal slice width settings, provides a direct test of pre-patient beam collimation and allows geometric efficiencies to be calculated for the scanner

• Geometric efficiency (GE) is defined as:GE = width slice irradiated x 100 %

width slice imaged• It is recommended that GE value is displayed on

the console if it is less than 70%



• First method : expose therapy verification film, supported in air at the isocentre, at each of the slice width settings

• Once developed, optical density profiles may be plotted using a digitizer or the width measured using a ruler

• To provide an accurate dose profile, a calibration curve can be applied to convert optical density profiles into dose profiles, from which irradiated slice widths (FWHM of dose profiles) may be derived

• Alternative technique: measure dose profiles using TLD’s.


Measurement of Measurement of irradiated slice irradiated slice

widths for a range widths for a range of nominal slice of nominal slice width settingswidth settings



Quality control program

Example of a quality control program for a CT unit:

Baseline ± 20%H1 yBHigh contrast resolution

> ± 20% or 1 mmM1 yBImage slice thickness

> ± 20% or 1 mmH1 yBIrradiated slice thickness

Baseline ± 20%H1 yACTDI

> ± 1.5% H1 yBCT number uniformity

Water: ± 10 HU

Other materials ± 20 HU

*more scans/parameters

H

H

D/W

1 Y*

A

B

CT number values

Baseline ± 20%

*more scans/parameters

H

H

D/W

1 y*

A

B

Image noise

Remedial levelPriorityFrequencyLevel of expertise

Physical parameter

Date post:	12-Jan-2016
Category:	Documents
Upload:	neviah
View:	29 times
Download:	1 times

RADIATION PROTECTION IN DIAGNOSTIC AND INTERVENTIONAL RADIOLOGY

Documents