Académique Documents
Professionnel Documents
Culture Documents
X - ray Phantoms
DESCRIPTION
This set of test objects has been put together in accordance with the Kcare protocols for CR and
DDR QA testing. Copies of the protocols are included in this set, however, please refer to the
KCare website for future revisions http://www.kcare.co.uk/Education/protocols.htm.
Additional information for each of the test objects is given below. Please do not hesitate to
contact Leeds Test Objects for further information.
This resolution test pattern is labelled and comprises test groups up to 10 LP/mm
3. Blurring [MS4]
Test Object TOMS4 consists of a circular plate of diameter 220 mm and thickness 10 mm
containing a precision test mesh (International Test Sieve Standard R565) with approximate
*
spatial frequencies of 1.23 cycles per mm. This mesh is known as Leeds Test Object MS4.
Test Object TO20 consists of a circular plate of diameter 250 mm and thickness 10 mm. The test
details consist of 144 circular details arranged in rows of 12 contrasts; the rows contain details
with the following diameters (mm).
Diameter: 11.1 8.0 5.6 4.0 2.8 2.0 1.4 1.0 0.7 0.5 0.35 0.25
Reference: A B C D E F G H J K L M
Details are presented in a range of 12 contrasts, the values of which at various kVp and copper
pre-filtration are given in the following tables.
*
Nominal aperture size (mm) 0.50
Nominal wire diameter (mm) 0.31
-1
Spatial frequency = [wire diameter + aperture size]
1
CONTRAST DATA
60kV, 1.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0876 0.0586 0.0377 0.0253 0.0183 0.0127 0.0102 0.0078 0.0051 0.0035 0.0025 0.0018
D,E,F 0.1130 0.0876 0.0586 0.0377 0.0253 0.0183 0.0127 0.0102 0.0078 0.0051 0.0035 0.0025
G,H,J 0.3000 0.2130 0.1650 0.1130 0.0876 0.0586 0.0377 0.0253 0.0183 0.0127 0.0102 0.0078
K,L,M 0.9820 0.8140 0.5880 0.4440 0.3000 0.2130 0.1650 0.1130 0.0876 0.0586 0.0377 0.0253
65kV, 1.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0766 0.0512 0.0347 0.0232 0.0168 0.0116 0.0093 0.0071 0.0046 0.0032 0.0023 0.0017
D,E,F 0.0993 0.0766 0.0512 0.0347 0.0232 0.0168 0.0116 0.0093 0.0071 0.0046 0.0032 0.0023
G,H,J 0.2670 0.1880 0.1450 0.0993 0.0766 0.0512 0.0347 0.0232 0.0168 0.0116 0.0093 0.0071
K,L,M 0.9690 0.7690 0.5390 0.3990 0.2670 0.1880 0.1450 0.0993 0.0766 0.0512 0.0347 0.0232
70kV, 1.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0674 0.0450 0.0322 0.0215 0.0155 0.0108 0.0086 0.0065 0.0043 0.0030 0.0021 0.0016
D,E,F 0.0876 0.0674 0.0450 0.0322 0.0215 0.0155 0.0108 0.0086 0.0065 0.0043 0.0030 0.0021
G,H,J 0.2380 0.1670 0.1280 0.0876 0.0674 0.0450 0.0322 0.0215 0.0155 0.0108 0.0086 0.0065
K,L,M 0.9540 0.7260 0.4960 0.3600 0.2380 0.1670 0.1280 0.0876 0.0674 0.0450 0.0322 0.0215
75kV, 1.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0599 0.0399 0.0301 0.0201 0.0145 0.0101 0.0081 0.0062 0.0040 0.0028 0.0020 0.0015
D,E,F 0.0780 0.0599 0.0399 0.0301 0.0201 0.0145 0.0101 0.0081 0.0062 0.0040 0.0028 0.0020
G,H,J 0.2140 0.1490 0.1140 0.0780 0.0599 0.0399 0.0301 0.0201 0.0145 0.0101 0.0081 0.0062
K,L,M 0.9360 0.6860 0.4580 0.3270 0.2140 0.1490 0.1140 0.0780 0.0599 0.0399 0.0301 0.0201
80kV, 1.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0542 0.0360 0.0285 0.0190 0.0138 0.0096 0.0077 0.0059 0.0038 0.0027 0.0019 0.0014
D,E,F 0.0705 0.0542 0.0360 0.0285 0.0190 0.0138 0.0096 0.0077 0.0059 0.0038 0.0027 0.0019
G,H,J 0.1940 0.1350 0.1040 0.0705 0.0542 0.0360 0.0285 0.0190 0.0138 0.0096 0.0077 0.0059
K,L,M 0.9150 0.6490 0.4260 0.2990 0.1940 0.1350 0.1040 0.0705 0.0542 0.0360 0.0285 0.0190
2
60kV, 1.5 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0791 0.0528 0.0353 0.0236 0.0170 0.0119 0.0095 0.0072 0.0047 0.0033 0.0024 0.0017
D,E,F 0.1030 0.0791 0.0528 0.0353 0.0236 0.0170 0.0119 0.0095 0.0072 0.0047 0.0033 0.0024
G,H,J 0.2760 0.1940 0.1500 0.1030 0.0791 0.0528 0.0353 0.0236 0.0170 0.0119 0.0095 0.0072
K,L,M 0.9790 0.7920 0.5590 0.4130 0.2760 0.1940 0.1500 0.1030 0.0791 0.0528 0.0353 0.0236
65kV, 1.5 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0687 0.0458 0.0325 0.0217 0.0156 0.0109 0.0087 0.0066 0.0043 0.0030 0.0022 0.0016
D,E,F 0.0895 0.0687 0.0458 0.0325 0.0217 0.0156 0.0109 0.0087 0.0066 0.0043 0.0030 0.0022
G,H,J 0.2430 0.1700 0.1310 0.0895 0.0687 0.0458 0.0325 0.0217 0.0156 0.0109 0.0087 0.0066
K,L,M 0.9630 0.7430 0.5090 0.3680 0.2430 0.1700 0.1310 0.0895 0.0687 0.0458 0.0325 0.0217
70kV, 1.5 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0601 0.0400 0.0301 0.0201 0.0145 0.0101 0.0081 0.0062 0.0040 0.0028 0.0020 0.0015
D,E,F 0.0782 0.0601 0.0400 0.0301 0.0201 0.0145 0.0101 0.0081 0.0062 0.0040 0.0028 0.0020
G,H,J 0.2150 0.1500 0.1150 0.0782 0.0601 0.0400 0.0301 0.0201 0.0145 0.0101 0.0081 0.0062
K,L,M 0.9450 0.6970 0.4650 0.3300 0.2150 0.1500 0.1150 0.0782 0.0601 0.0400 0.0301 0.0201
75kV, 1.5 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0532 0.0354 0.0282 0.0188 0.0136 0.0095 0.0076 0.0058 0.0038 0.0026 0.0019 0.0014
D,E,F 0.0693 0.0532 0.0354 0.0282 0.0188 0.0136 0.0095 0.0076 0.0058 0.0038 0.0026 0.0019
G,H,J 0.1920 0.1340 0.1020 0.0693 0.0532 0.0354 0.0282 0.0188 0.0136 0.0095 0.0076 0.0058
K,L,M 0.9240 0.6550 0.4270 0.2980 0.1920 0.1340 0.1020 0.0693 0.0532 0.0354 0.0282 0.0188
80kV, 1.5 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0480 0.0319 0.0268 0.0179 0.0129 0.0090 0.0072 0.0055 0.0036 0.0025 0.0018 0.0013
D,E,F 0.0626 0.0480 0.0319 0.0268 0.0179 0.0129 0.0090 0.0072 0.0055 0.0036 0.0025 0.0018
G,H,J 0.1750 0.1210 0.0922 0.0626 0.0480 0.0319 0.0268 0.0179 0.0129 0.0090 0.0072 0.0055
K,L,M 0.9000 0.6160 0.3950 0.2720 0.1750 0.1210 0.0922 0.0626 0.0480 0.0319 0.0268 0.0179
3
60kV, 2.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0742 0.0495 0.0339 0.0226 0.0164 0.0114 0.0091 0.0069 0.0045 0.0031 0.0023 0.0016
D,E,F 0.0964 0.0742 0.0495 0.0339 0.0226 0.0164 0.0114 0.0091 0.0069 0.0045 0.0031 0.0023
G,H,J 0.2610 0.1830 0.1410 0.0964 0.0742 0.0495 0.0339 0.0226 0.0164 0.0114 0.0091 0.0069
K,L,M 0.9750 0.7760 0.5400 0.3940 0.2610 0.1830 0.1410 0.0964 0.0742 0.0495 0.0339 0.0226
65kV, 2.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0641 0.0427 0.0312 0.0208 0.0150 0.0105 0.0084 0.0064 0.0042 0.0029 0.0021 0.0015
D,E,F 0.0834 0.0641 0.0427 0.0312 0.0208 0.0150 0.0105 0.0084 0.0064 0.0042 0.0029 0.0021
G,H,J 0.2280 0.1600 0.1220 0.0834 0.0641 0.0427 0.0312 0.0208 0.0150 0.0105 0.0084 0.0064
K,L,M 0.9570 0.7240 0.4890 0.3490 0.2280 0.1600 0.1220 0.0834 0.0641 0.0427 0.0312 0.0208
70kV, 2.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0557 0.0370 0.0289 0.0193 0.0139 0.0097 0.0078 0.0059 0.0039 0.0027 0.0019 0.0014
D,E,F 0.0726 0.0557 0.0370 0.0289 0.0193 0.0139 0.0097 0.0078 0.0059 0.0039 0.0027 0.0019
G,H,J 0.2010 0.1400 0.1070 0.0726 0.0557 0.0370 0.0289 0.0193 0.0139 0.0097 0.0078 0.0059
K,L,M 0.9370 0.6760 0.4440 0.3110 0.2010 0.1400 0.1070 0.0726 0.0557 0.0370 0.0289 0.0193
75kV, 2.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0491 0.0326 0.0271 0.0181 0.0130 0.0091 0.0073 0.0055 0.0036 0.0025 0.0018 0.0013
D,E,F 0.0640 0.0491 0.0326 0.0271 0.0181 0.0130 0.0091 0.0073 0.0055 0.0036 0.0025 0.0018
G,H,J 0.1790 0.1240 0.0946 0.0640 0.0491 0.0326 0.0271 0.0181 0.0130 0.0091 0.0073 0.0055
K,L,M 0.9140 0.6320 0.4060 0.2790 0.1790 0.1240 0.0946 0.0640 0.0491 0.0326 0.0271 0.0181
80kV, 2.0 mm Cu
position
1 2 3 4 5 6 7 8 9 10 11 12
A,B,C 0.0442 0.0293 0.0258 0.0172 0.0124 0.0086 0.0069 0.0052 0.0034 0.0024 0.0017 0.0012
D,E,F 0.0577 0.0442 0.0293 0.0258 0.0172 0.0124 0.0086 0.0069 0.0052 0.0034 0.0024 0.0017
G,H,J 0.1620 0.1120 0.0852 0.0577 0.0442 0.0293 0.0258 0.0172 0.0124 0.0086 0.0069 0.0052
K,L,M 0.8880 0.5930 0.3740 0.2540 0.1620 0.1120 0.0852 0.0577 0.0442 0.0293 0.0258 0.0172
4
Figure 1: TO.20 detail layout, showing size references A to M
(see table 2 and 3 for detail diameters and contrasts)
5
Protocol for the QA of Computed Radiography Systems
This document describes a series of tests to assess CR plate and reader performance. The tests
are intended to detect artefacts and monitor image quality and sensitivity. The tests are split into
the following categories,
- commissioning tests
- annual QA tests.
All the tests described should be performed on all available reader systems.
KCARE have data from performing tests on all manufacturers CR systems and are available for
advice and information.
1 Commissioning Tests
List of equipment
· Tape measure
· Adhesive tape
· 1.0 mm Copper filtration (>10 x 10 cm)
· 1.5 mm Copper filtration (>10 x 10 cm) - for Agfa only
· 0.5 mm Copper and 1mm Aluminium filtration (>10 x 10 cm) - for Kodak only
· TO20 threshold contrast test object
· Resolution test object (e.g. Huttner 18)
· M1 geometry test object or lead ruler
· Contact mesh
· Ionisation chamber
· Small lead or Copper block (~5 x 5 cm)
· Steel ruler
In all tests described the unique plate identification code should be recorded.
These tests should be performed using an x-ray unit that has recently passed QC tests. In
particular, the accuracy of the kVp selected for detector dose indicator consistency and
calibration should be tested.
It may be possible to undertake some of these tests on a review workstation depending on the
processing tools available (which will be dependent on the manufacturer). However certain tests
require the use of the higher quality reporting workstation.
6
1.1 Dark Noise
a) Erase an image plate and without making an exposure read it using the following
parameters
Agfa: S=800,
examination type - ‘System Diagnosis’
processing – ‘Flat Field’
b) Examine the images visually for uniformity and record the detector dose indicator value
for Agfa (SAL – at centre of plate) and Kodak (Exposure Index).
c) Record a mean pixel value using region of interest analysis (for systems not offering ROI
analysis see appendix for details of how to measure a mean pixel value).
Tolerance: For Agfa and Fuji systems a uniform artefact free image should be expected. Kodak
systems add a collector profile to the image to compensate for non uniform collection efficiency
across the place. This results in series of bands appearing across the image. Agfa systems
should have an SAL < 100. For Fuji the pixel value should be <280. For Kodak the EI value
should be <80 for GP plates and <380 for GP plates. For Konica a pixel value >3975 should be
expected.
1.2 Dosimetry
Purpose: To measure receptor doses required for later tests 1.3,1.5, 1.6, 1.7 and 1.11
a) Position an ion chamber at ~1.2 m from the focus (see figure 1) and at least 30 cm
above the table (record the actual distances). Collimate to the ion chamber.
7
focus
added filter
> 150 cm
chamber
CR plate
> 30cm
b) Expose the chamber at 70kVp with 1.0mm of copper in filtration at the tube head. The
mAs should, by ‘trial and error’, be set such that the inverse square law corrected dose to
the table level is approximately 10mGy
d) Under the same beam conditions determine the mAs required to deliver a receptor
entrance air kerma of 1mGy, 4mGy 12mGy and 50mGy
e) If a Fuji system is being tested determine the mAs to deliver a receptor entrance air
kerma of 10mGy at 80kVp with no filtration in the beam.
If a Kodak system is being tested determine the mAs to deliver a receptor entrance air
kerma of 10mGy at 80kVp with 1mm Al and 0.5mm Cu filtration at the tube head.
8
1.3 Linearity and System transfer properties
Purpose: To establish the relationship between receptor dose and pixel value so that this
relationship can be corrected for in tests 1.4 and 1.7. Also to establish that the indicated exposure
(calculated from the detector dose indicator) responds linearly to increases in dose.
a) Place a 24 x 30 cm cassette on the table at ~1.50m (as described for test 1.2). Set the field to
just cover the cassette. Mark the corners of the cassette on the table with transpore, so that the
cassette can be easily repositioned.
b) Expose a plate at 70kVp with 1.0mm copper at the tube head to deliver a dose of order 1mGy
as measured in test 1.2.
c) After a minimal fixed time delay (e.g.1 to 5 mins), read the plate as described below.
- For Agfa systems the SAL values obtained from ROI analysis on the review workstation should
be used.
- For Fuji, Konica and Kodak systems the images should be transferred to reporting workstations
to use ROI analysis tools if available.
g) Plot a graph of pixel value versus receptor dose using a graph plotting package (e.g. Microsoft
excel). Obtain the equation of the trend-line for this graph (i.e. the pixel value as a function of
receptor dose). This equation is the system transfer properties (STP) equation and is used for
making corrections in tests 1.4 and 1.7. An equation of the form
9
dose =f(pixel value)
Tolerance: For all images the ratio, k, of indicated exposure to exposure should not differ by
2
greater than ±10% from the mean k value. The trend-line plotted in excel should have an R fit
value >0.95. There is no tolerance for the STP equation. However the pixel value to dose
relationship should be a simple relationship (e.g. log, linear or square root). For systems
evaluated by KCARE the following has been found.
* For the Agfa system there is a square root relationship between SAL values and dose. The
relationship between raw data pixel values and dose however was logarithmic for systems
evaluated by KCARE
Purpose: To test that minimal residual signal (ghosting) remains on a plate after readout and
erasure.
a) Position a plate on the table at ~1.5 m. Set a 10 cm x 10 cm field and position a piece of
attenuating material (e.g. Copper or lead) at the centre of the CR plate. Expose at 80kVp,
25mAs with no filtration.
c) Re-expose the plate with a 9 cm x 9 cm field centred on the same point on the plate with
no attenuating material in place, using 80kVp, 0.5mAs and no filtration.
Agfa: S=200,
examination type - ‘System Diagnosis’
processing – ‘Flat Field’
10
e) e) Set a very narrow window and adjust the level. Visually inspect the image for any remnant of
the previous image (look for both the attenuating material and the position of the collimators).
If a remnant is visible, use region of interest analysis to quantify the difference in pixel value
between the ghosted and unghosted areas.
f)
¨ For Agfa systems the SAL values obtained from ROI analysis on the review
workstation should be used.
¨ For Fuji, Konica and Kodak systems the images should be transferred to
reporting workstations to use ROI analysis tools if available.
The ROI values should be used to calculate indicated receptor doses using the STP
equation established in test 1.3.
Tolerance: If no evidence of ghosting is found from visual inspection of the images then the test
is passed and there is no need to perform ROI analysis. There should be <1% (remedial)
difference between the STP corrected pixel values in the ghosted region and the surrounding
areas. A suspension level of <5% is set.
Purpose: To assess the accuracy of the plate exposure values calculated using exposure
indicators.
b) Expose the plate to a known dose of ~10 mGy using the kVp and filtration listed below
(use mAs found in test 1.2).
Fuji: A 10 minute delay between exposure and readout, readout using semi-
auto, L=1 or 2
11
d) Record the detector dose indicator, and calculate the indicated exposure using the equations
given below.
For Agfa systems the indicated exposure, EAgfa , in mGy, for a 200 speed readout is given by
–6 2
EAgfa = 5.90 ´ 10 ´ SAL (1)
æ EI - 2000 ö
E kodak = 8.7 ´ 10 n , where n = ç ÷ (2)
è 1000 ø
For Konica systems the equation linking to dose and S value is not available
Tolerance: The indicated exposure should agree with the measured exposure within 20%.
Purpose: To assess the variation of sensitivity between plates, and set a baseline for monitoring
system sensitivity for future QA testing
a) Place a 24 x 30 cm CR cassette on the couch and set up as described for test 1.2/1.3
(see figure 1) and with 1.0mm Cu filtration.
b) Expose the plate at 70kVp to give a known dose of ~10 mGy. The dose to the plate
calculated from inverse square law corrected ion chamber measurements should be
recorded (see test 1.2)
d) Record the detector dose indicator, and calculate the indicated exposure using equations
1-3. Repeat twice for the same plate.
12
e) Calculate the indicated exposure using equations given in test 1.5
f) Repeat this test for all plates for acceptance testing(making only one exposure to each
plate). It is helpful at this point to identify a plate that has a detector dose indicator in the
middle of the range for future QA.
Tolerance: The variation in the calculated indicated exposures should not differ by greater than
20% between plates. The measurements repeated on the same plate should be used to lay down
a baseline for future QA tests. Al images should be inspected for gross artefacts
1.7 Uniformity
Purpose: To assess the uniformity of the recorded signal from a uniformly exposed plate. A non-
uniform response could affect clinical image quality.
a) Expose a plate as described for test 1.6 but with half the mAs.
o
b) Rotate the plate through 180 about the vertical axis and re-expose using the same
parameters (this should largely cancel out the non uniformities due to the anode heal
affect).
d) Visually inspect all images obtained in test 1.3, 1.5 and 1.6 for uniformity and artefacts.
Likely artefacts include dust on the plate or readout optics, and scratches on plates.
e) The uniformity of the image obtained in 1.7b should be assessed using region of interest
analysis (ROI ) if available, to measure the mean pixel values in positions a-e, as
indicated in figure 2 below (i.e.at the centre of the image, and at the centre of each of the
four quadrants of the image). The size of ROI should be of order 10000 pixels.
For Agfa systems the SAL values obtained from ROI analysis on the review
workstation should be used.
For Fuji, Konica and Kodak systems the images should be transferred
to reporting workstations to use ROI analysis tools if available.
a b
d e
13
f) The five values obtained from ROI analysis should be used to calculate five indicated
receptor dose values using the STP equation obtained in test 1.3
Tolerance: The images should not have obvious artefacts. The ratio of the standard deviation of
the 5 STP corrected ROI values to their mean (the coefficient of variation) should be less than
10%.
Purpose: To assess the accuracy of software distance indicators and check for distortion.
a) Position the M1 test object directly on the centre of a CR cassette at > 150 FDD.
N.B. A lead ruler could be used in place of the M1 test object. If so 2 exposures should be
made with the ruler placed in first the scan direction then the subscan direction.
d) Using the distance measuring software tools measure the dimensions (x and y) of five
central squares in both fast and slow scan directions. Calculate the aspect ratio x/y. For
Agfa systems the review workstation software can be used. For Fuji, Kodak and Konica
systems the images should be transferred to the reporting workstation to use distance
measuring software tools if available. If images are reported from film then they should be
printed at full size. Distances can then be measured with a ruler.
e) Reposition the test object over the edge of the plate as indicated in figure 3 and repeat
steps b and c
scan direction
sub-
scan
direct
ion
Figure 3
f) Along the edge of the plate measure the horizontal (x1) and vertical (y1) sizes of two
squares as indicated in figure 4. Calculate the aspect ratio x1/y1.
14
g) If possible download the image as a DICOM file. Open the image using a DICOM viewer
such as Santeviewer. Hold the curser over a corner of a square in the grid. Record the
position within the image (i.e. the x and y coordinates). Move the curser to the corner of
the square of the grid 10cm from the first corner in the x direction. Record the coordinates
again. Calculate the pixel pitch, p(mm)=100/n, where n =number of pixels covering 10cm
of the grid. Repeat for the y direction. This test is only necessary on commissioning.
Compare the pixel pitch to that stated by the manufacturer. The difference should be no
greater than the estimated measurement error.
y1
Figure 4
x1
Tolerance: The measured distances x and y should agree within 3% of the actual distances at the
centre of the plate and 5% at the edge. All calculated aspect ratios should be within 1.00 ± 0.03 at
the centre of the plate, or 1.00 ± 0.05 at the edge.
1.9 Blurring
a) With the contact mesh in placed on the cassette at >150cm FDD, expose at 50-60 kVp,
fine focus, with no filtration and 10mAs.Read the plate as described for test 1.3.
b) Visually inspect the image for distortions. If distortion occurs clean the plate and repeat.
Tolerance: No blurring should be present. If blurring is present on all plates this suggests the
reader is at fault, whilst imperfections in individual plates may also lead to blurring. If blurring
remains on a region of a plate after cleaning it should not be used clinically.
15
1.10 Limiting Spatial Resolution
Purpose: To test the high contrast limit of the systems ability to resolve details.
a) Place a general purpose cassette on the couch with the Huttner test object positioned at
o
its centre aligned at 45 to its edges.
-1
NB A Huttner test object with line spacings up to 8 lp.mm may be required.
b) Set 50-60 kVp, fine focus, and expose the cassette using 10mAs.
Agfa: S=100,
examination type - ‘System Diagnosis’
processing – ‘Flat Field’
d) Adjust the window level and magnification to optimise the resolution. Score the number of
resolvable groups of lines from the screen. Look up the corresponding resolution. The
image should be scored at a magnification of order x 5. If this facility is not available on
the review workstation then images should be transferred to the reporting workstation for
scoring.
e) Repeat the measurement twice with the resolution test object placed at a slight angle to
the first the lateral or then the longitudinal axis.
f) Repeat this process for all available image pixel pitches (nb different plate sizes will often
default to being scanned at different pixel pitches).
Tolerance: These measurements should be used to set a baseline for future QA tests. Print or
save the images for future reference, if possible. Comparable images are available for most
systems through KCARE.
o
N.B. The limiting resolution should be expected to approach the Nyquist limit. At 45 the Nyquist
frequency is defined by Ö2/2p where p is the pixel pitch. The measured limiting resolution may be
limited by display when scoring from a review workstation, particularly if no zoom or limited zoom
facilities are available.
16
1.11 Threshold Contrast Detail Detectability
Purpose: To monitor image quality by assessing the visibility of low contrast details.
a) With the tube, plate, and 1.0mm copper filtration in the same positions as for the
sensitivity tests, place the TO20 (or equivalent) test object on the plate. Collimate down
to the size of the test object.
b) Set 70 kVp and an mAs to deliver ~4 mGy. Read the plate using the following parameters.
Agfa: S=200,
examination type - ‘System Diagnosis’
processing – ‘Flat Field’
Ascertain whether clinical images are most commonly viewed soft or hard copy. If they
view hardcopy, adjust the window to optimise the visibility of the details, ensuring that
background noise is perceptible, and print the image out on the largest film size. View the
image on a masked light box, and score each detail size using fixed distance viewing
(<1m). If images are viewed softcopy, score them on a reporting workstation optimising
window and level settings for each detail size.
0.5
1 n H T (Ai ) é Dref ù
IQF = å ref ê ú (4)
n i =1 H T (Ai )ë D û
where
HT(A) = threshold contrast detail index values calculated from the image,
ref
HT (A) = threshold contrast detail index values calculated from a reference image of a
system known to be in good adjustment
D = the dose to the image plate
Dref = the dose to the image plate for the reference image
n = the number of details in the test object.
d) Repeat this test for two other imaging plates and also for a single plate at exposures of
~1mGy and ~12mGy.
Tolerance: The results of this test are used to set a baseline for future QA tests. Results could be
compared to those from other similar systems if available.
17
1.12 Laser beam function
a) Place a steel ruler slightly angled to the subscan direction on a large cassette.
b) Expose at ~70 kVp, 150cm FSD and an mAs to deliver an incident exposure of
~50mGy. Read the plate as described for test 1.3
c) Using the software magnify the image x10. This will usually require the image to be
viewed from a reporting monitor. Select a narrow window width such that the image
appears largely polarised to black or white. This should allow the edge to be easily
differentiated from the background. Laser beam jitter can be evaluated by examining the
edge of the ruler on the image.
Tolerance: The edge should be continuous across the full length of the image. Stair step
characteristics should be uniform across the length of the image. Regions of over or undershoot
of the scan lines indicate a timer or laser beam modulation problem.
Purpose: To test for the presence of Moiré pattern artefacts caused by grids.
a) Place a CR cassette in the bucky such that the scan lines are vertical to the gridlines.
The cassette should be 1.5m from the focus, and the collimation should cover the whole
plate.
b) Expose at 70 kVp using the AEC with 1.0 mm of copper in the beam, and the grid in
place.
e) Repeat with the CR cassette positioned in the bucky such that the scan lines are
horizontal to the gridlines.
f) Repeat for all buckies and grids that may be used with the CR system, including any
grids used in mobile radiography.
Tolerance: No Moiré patterns should be visible. If Moiré patterns are visible with a particular grid,
it should not be used with the CR plates. The cause of Moire patterns may be the failure of the
motion of moving grids or insufficient grid density.
18
2 Annual QA tests
1.2 Dosimetry (only for 4mGy and 10mGy with 70kVp and 1.0mmCu)
1.4 Erasure cycle efficiency
1.6 Detector dose indicator consistency/sensitivity (for 1 plate of each size)
1.7 Uniformity
1.8 Scaling errors
1.9 Blurring
o
1.10 Limiting resolution (45 only)
1.11 TCDD (only 4mGy).
The tests should be performed as described in the previous section. Table 1 below summarises
the relevant remedial levels where these are different to those described for commissioning.
It should be noted that TCDD and limiting resolution are subjective measures. Some effort should
therefore be made to train scorers to score to similar thresholds.
19
Appendix A – Measuring a mean pixel value using a Fuji CR system
Reduce the window width to 1, so that the image has only pixels appear as either one of just two
levels, black or white. Adjust the level until approximately half the pixels are black and half are
white. This level value is the mean pixel value of the image.
References
[1] Draft Report of Task Group #10, American Association of Physicists in Medicine, Acceptance
Testing and Quality Control of Photostimulable Storage Phosphor Imaging Systems, August 1998
[2] British Institute of Radiology, ‘Assurance of the quality in the diagnostic imaging department’,
2001, ISBN 0-905749-48-0
[4] Samei E, Seibert JA, Willis CE, Flynn MJ, Mah E, Junck KL, ‘Performance evaluation of
computed radiography systems’, 2001, Med.Phys. Vol28 (3) p361-371.
20
Protocol for the QA of Direct Digital Radiography Systems
This document describes a series of tests to assess digital detector performance. The tests are
intended to detect artefacts and test image quality and sensitivity. The tests are split into the
following categories,
- commissioning tests
- annual QA tests.
All images should be acquired with a minimal amount of pre-processing and a linear Look Up
Table (LUT) applied, unless otherwise stated. Some of the images should be assessed on either
a reporting quality monitor or printer. These devices should be monitored as part of a Quality
Assurance program. Ambient lighting conditions should be consistent and appropriate.
KCARE have data from performing tests on many manufacturers DDR systems and are available
for advice and information regarding expected results.
1 Commissioning Tests
List of equipment
· Tape measure
· Adhesive tape
· 1.0 mm Copper filtration (>10 ´ 10 cm)
· TO20 threshold contrast test object
· Resolution test object (e.g. Huttner 18)
· M1 geometry test object
· Contact mesh
· Ionisation chamber
· Small Lead or Copper block (~5 ´ 5 cm).
These tests should be performed using an x-ray unit that has recently passed QC tests. In
particular, the accuracy of the kVp selected for detector dose indicator consistency and
calibration should be tested.
Unless otherwise stated, a image processing mode should be selected such that there is a direct
relationship between pixel value and dose.
21
1.1 Dosimetry
Purpose: To measure entrance detector doses required for later test 1.4, 1.5, 1.6 and 1.10
a) Set an FDD of at least 150 cm (see figure 1). Record this distance.
Focus
Added filter
> 150 cm
Ionisation
chamber
Detector
> 30cm
d) Expose the chamber such that the inverse square law corrected receptor entrance
air kerma is approximately 10 mGy, using 70 kVp and 1.0 mmCu filtration.
f) Under the same beam conditions determine the mAs required to deliver 1mGy,
4mGy 12mGy and 50mGy
22
1.2Dark Noise
Purpose: To establish the relationship between receptor dose and pixel value so that this
relationship can be corrected for in tests 1.4 and 1.6. Also to establish that the indicated exposure
(calculated from the detector dose indicator responds linearly to increases in dose).
b) Open the collimators and expose the entire area of the detector at 70kVp with 1.0mm
Cu at the tube head. Set a mAs and FDD to deliver a dose of order 1mGy (as determined
in test 1.1).
e) Record a pixel value from the centre of each image from the acquisition workstation. If
ROI analysis is not available at the acquisition workstation the images should be
transferred to the reporting workstation to perform this task
f) Plot a graph of pixel value versus receptor dose using a graph plotting package (e.g.
Microsoft excel). A zero dose point can be obtained using the result of test 1.2. Obtain
the equation of the trendline for this graph (i.e. the pixel value as a function of receptor
dose). This equation is the system transfer properties (STP) equation and is used for
making corrections in tests 1.4 and 1.6. An equation of the form
2
Tolerance: The trend-line plotted in excel should have an R fit value >0.95. There is no tolerance
for the STP equation. However the pixel value to dose relationship should be a simple
relationship (e.g. log, linear or square root). There is also no tolerance for the dark noise image,
however the data may be useful as a baseline when re-checking the dark noise in the event of
other image quality problems.
NB KCARE have tested many of the digital detectors on the market. Please contact us for
information regarding the expected form of the STP and the relationship between detector dose
indicator and dose.
23
1.4 Image retention
Purpose: To test that any detectable residual signal (ghosting) that remains in subsequent
images is minimal.
a) Ensure the grid is removed from the system if possible and there is no attenuation in
the beam.
c) Close the collimators and cover the detector with a lead apron. Set a low exposure
e.g. 50kVp and 0.5mAs
d) Open the collimators and place the attenuating material (copper or lead 5´5 cm) on
the detector such that it covers part of the field. Make an exposure at 70kVp and an
mAs to deliver a receptor dose of order 4mGy.
e) Obtain another blank image as described in step c. This exposure should be made 1
minute after the previous one.
f) Set a very narrow window and adjust the level. Visually inspect the image for any
remnant of the previous image. If a remnant is visible, use region of interest analysis
to quantify the difference in pixel value between the ghosted and unghosted areas. If
ROI analysis is not available the image should be transferred to a reporting
workstation. The ROI values should be used to calculate indicated receptor doses
using the STP equation established in test 1.3 (see later).
Tolerance: If no evidence of ghosting is found from visual inspection of the images then the
test is passed and there is no need to perform ROI analysis. There should be <5% (remedial)
difference between the STP corrected pixel values in the ghosted region and the surrounding
areas (this tolerance is under review by KCARE).
attenuated
region
ROI 1 ROI 2
24
1.5 Detector dose indicator consistency
This test can only be performed if the unit has a form of detector dose indicator.
Purpose: To assess the variation of sensitivity between exposures, and set a baseline for
monitoring system sensitivity for future QA testing.
b) Set a field size to cover the entire detector and a FDD as for the dosimetry.
c) Expose the detector to a known dose of 10 mGy at 70kVp with 1.0mm Cu at the tube
head. Set an mAs as established from test 1.1
d) Record the organ program, LUT name and detector dose indicator, without changing
the window and levelling.
Tolerance: The indicated sensitivity indices should not differ by greater than 20% of equivalent
exposure, between exposures. The measurements should be used to set a baseline for future QA
tests.
NB The relationship between detector dose indicator and dose may not be linear or even a
simple relationship. As far as possible the index should be converted to an exposure for
comparison (see test 1.4).
25
1.6 Uniformity
Purpose: To assess the uniformity of the recorded signal from a uniformly exposed detector. A
non-uniform response could affect clinical image quality.
a) Visually inspect all images obtained in test 1.5 for uniformity and artefacts.
b) The uniformity one of images should be assessed using region of interest (ROI)
analysis if available; to measure the mean and standard deviation of the pixel values
in position a-e, as indicated in figure 2 below (i.e. the centre of the image an the
centre of the four quadrants). For detectors that are tiled detectors an ROI should be
drawn at the centre of all tiles. The ROIs should be of order 10000 pixels. If ROI
analysis is not available at the acquisition workstation then images should be
transferred to a reporting workstation. If uniformity is poor in the direction of the
anode cathode axis this is likely to be a result of the anode heel affect. To confirm
o
this the test should be repeated with the tube rotated through 90 .
c) The five values obtained from ROI analysis should be used to calculate five indicated
receptor dose values using the STP equation obtained in test 1.4
a b
d e
Tolerance: The images should not have obvious artefacts. The ratio of the standard deviation of
the 5 STP corrected ROI values to their mean (the coefficient of variation) should be less than
10%.
26
1.7 Scaling errors
Purpose: To assess the accuracy of software distance indicators and check for distortion.
b) Position the M1 test object directly onto the detector with an FDD of 150 cm.
c) Expose the detector at 50-60 kV with no attenuation in the beam and 10 mAs.
N.B. A lead ruler could be used in place of the M1 test object. If so 2 exposures should
be made with the ruler placed in first the scan direction then the subscan direction.
d) Using the distance measuring software tools measure the dimensions (x and y) of
five central squares in both the horizontal and vertical directions. Calculate the aspect
ratio x/y.
e) Select any corner of the image and measure the horizontal (a) and vertical (b) sizes
of two squares as indicated in figure 3. Calculate the aspect ratio a/b. Repeat this for
one other corner.
f) If possible download the image as a DICOM file. Open the image using a DICOM
viewer such as Santeviewer. Hold the curser over a corner of a square in the grid.
Record the position within the image (i.e. the x and y coordinates). Move the curser
to the corner of the square of the grid 10cm from the first corner in the x direction.
Record the coordinates again. Calculate the pixel pitch, p(mm)=100/n, where n
=number of pixels covering 10cm of the grid. Repeat for the y direction. This test is
only necessary on commissioning. Compare the pixel pitch to that stated by the
manufacturer. The difference should be no greater than the estimated measurement
error.
Tolerance: The measured distances x and y should agree within 3% of the actual distances at
the centre or 5% at the corners. All calculated aspect ratios should be within 1.00 ± 0.03 at the
centre or 5% at the corners.
27
1.8 Blurring and stitching artefacts
Purpose: To test for any localised distortion or blurring and to highlight any stitching artefacts if
the system is formed from more than one detector element.
a) The test should be made with the grid both in and out of the detector.
b) Ensure there is no attenuation in the beam and that the FDD is set as large as
possible.
c) With a contact mesh on the detector, make an exposure at 50 - 60 kVp and 10 mAs
using fine focus. An MS4 or MS5 test object is appropriate.
Tolerance: No blurring should be present. If stitching artefacts are present there should be no
loss of information.
Purpose: To test the high contrast limit of the systems ability to resolve details.
a) Ensure the grid is removed from the system, there is no attenuation in the beam and
the FDD is set as large as possible.
o
b) Place the resolution test object onto the detector aligned at 45 to its edges.
-1
N.B. A Huttner test object with line spacings up to 8 lp.mm may be required
c) Set 50-60 kV and expose the cassette using 10mAs on fine focus.
d) Adjust the window level and magnification to optimise the resolution. Score the
number of resolvable groups of lines from the screen. The image should be scored at
a magnification of order x 5. If this facility is not available on the review workstation
then images should be transferred to the reporting workstation for scoring. Look up
the corresponding resolution.
e) Repeat the measurement twice with the resolution test object placed at a slight angle
to the lateral or longitudinal axis.
f) If the system has more than one detector element, measurements at 45° should be
made for each element.
Tolerance: These measurements should be used to set a baseline for future QA tests. Print or
save the images for future reference, if possible.
o
N.B. The limiting resolution should be expected to approach the Nyquist limit. At 45 the Nyquist
frequency is defined by Ö2/2p where p is the pixel pitch. The measured limiting resolution may be
limited by display when scoring from a review workstation, particularly if no zoom or limited zoom
facilities are available
28
1.10 Threshold Contrast Detail Detectability
Purpose: To monitor image quality by assessing the visibility of low contrast details.
b) With the tube, detector, and 1.0 mmCu filtration in the same positions as for the
sensitivity tests (1.5), place the TO20 (or equivalent) test object on the detector.
Collimate down to the size of the test object.
d) Ascertain whether clinical images are most commonly viewed soft or hard copy. If
images are viewed softcopy, score them on a reporting workstation optimising
window and level settings for each detail size. If they view hardcopy, adjust the
window to optimise the visibility of the details, ensuring that background noise is
perceptible, and print the image out on the largest film size. View the image on a
masked light box.
0.5
1 n H T (Ai ) é Dref ù
IQF = å ref ê ú
n i =1 H T (Ai )ë D û
where:
HT(A) = threshold contrast detail index values calculated from the image,
ref
HT (A) = threshold contrast detail index values calculated from a reference
image of a system known to be in good adjustment,
D = the dose to the image plate,
Dref = the dose to the image plate for the reference image
n = the number of details in the test object.
Tolerance: The results of this test are used to set a baseline for future QA tests. Results could be
compared to those from other similar systems if available. Print or save the images for future
reference, if possible.
29
2 Annual QA tests
1.1 Dosimetry
1.2 Dark Noise
1.5 Detector dose indicator consistency
1.6 Uniformity
1.8 Blurring and stitching artefacts
1.9 Limiting spatial resolution (in one quadrant at 45° only)
1.10 TCDD (only 4mGy)
The tests should be performed as described in the previous section. The table below
summarises the relevant remedial levels where these are different to those tolerances for
commissioning:
It should be noted that TCDD and limiting resolution are subjective measures. Some effort
should therefore be made to train scorers to score to similar thresholds. Retention of images
during annual QA is preferred but not essential.
30
Man 091b 12/07/11