You are on page 1of 32

Leeds Test Objects

www.leedstestobjects.com CR & DDR


user manual

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.

1. Scaling Errors [M1]

Test Object M1 consists of a matrix of 20mm square spacing with 10 mm diametrical


interpolations, mounted on a circular Perspex plate of diameter 250 mm and thickness 10 mm.
The matrix diameter is 230mm.

2. Limiting Spatial Resolution [Resolution Test Pattern]

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.

4. Threshold Contrast Detail Detectability [TO20]

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

Table 1: TO.20 detail diameters and reference letter.

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

Commissioning and Annual QA Tests

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

Purpose: To assess the level of noise inherent in the system

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’

Fuji: Readout mode – ‘Fixed’


S = 10000
L=1

Kodak: Mode – ‘pattern’

Konica: Readout mode – Fix

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).

d) If possible either archive or print the image for future reference.

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

Figure 1: Set-up for exposure index calibration

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

c) Record the measured dose and repeat twice.

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.

Agfa: S=200, system diagnosis/flat field processing

Kodak: Pattern mode body part.

Fuji: semi-auto, L=1 or 2

Konica: QC S-value, E and F processing turned off

d) Record the detector dose indicator value.

Agfa: SAL (draw an region of interest covering at least 10000 pixels at


the centre of the image

Kodak: Exposure index (EI)

Fuji: Linearity mode (S=200)

Konica: Fix mode

e) Record a pixel value from the centre of the image.

- 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.

f) Repeat for doses of order 4mGy, 12mGy and 50mGy.

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)

where f is some arbitrary function is required.

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.

Manufacturer STP Relationship


Agfa Square root *
Fuji Logarithmic
Kodak Logarithmic
Konica Logarithmic

* 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

1.4 Erasure cycle efficiency

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.

b) Read the plate (the readout parameters are not important).

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.

d) Read the plate using the following parameters.

Agfa: S=200,
examination type - ‘System Diagnosis’
processing – ‘Flat Field’

Kodak: Mode – ‘Pattern’

Fuji: Readout mode – ‘Semi Auto’


L = 1 or 2

Konica: Semi fix


g=1

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.

1.5 Detector dose indicator calibration

Purpose: To assess the accuracy of the plate exposure values calculated using exposure
indicators.

a) Position a 24´30 plate on the table as described for test 1.3

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).

CR system Filtration Tube Voltage


(kVp)
Agfa 1.5mm Cu 75
Kodak 0.5mm Cu 80
+1mm Al
Konica info not known info not known
Fuji none 80

c) Read the plate out as described below

Agfa: no delay between exposure and readout, S=200, system diagnosis/flat


field processing and linear sensitomitry.

Kodak: A 15 minute delay between exposure and readout, readout on Pattern


mode body part.

Fuji: A 10 minute delay between exposure and readout, readout using semi-
auto, L=1 or 2

Konica: Details of the Konica calibration protocol are not available

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)

For Kodak systems the indicated exposure, EKodak , in mGy, is given by

æ EI - 2000 ö
E kodak = 8.7 ´ 10 n , where n = ç ÷ (2)
è 1000 ø

For Fuji systems the indicated exposure, Efuji , in mGy, is given by


1740
E fuji =
S (3)

For Konica systems the equation linking to dose and S value is not available

e) Repeat twice and take a mean value of the indicated exposures.

Tolerance: The indicated exposure should agree with the measured exposure within 20%.

1.6 Detector dose indicator consistency

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)

c) Read the plate as described for test 1.5.

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).

c) Read the plate as described for test 1.3.

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

Fig 2: Positions of the ROI’s for uniformity tests

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%.

1.8 Scaling errors

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.

b) Expose at 50-60 kVp with no filtration and 10mAs.

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.

c) Read out plate using processing as for test 1.3.

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

Purpose: To test for any localised distortion or blurring of the image.

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.

c) Repeat for at least two other plates.

d) Repeat with a fine mesh if available.

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.

c) Readout the plate using the following parameters

Agfa: S=100,
examination type - ‘System Diagnosis’
processing – ‘Flat Field’

Fuji: Readout mode – ‘fix’ with L=2 and S=200

Kodak: Mode – ‘Pattern’ with raw data and no edge enhancement

Konica: QC S-value, E and F processing off

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).

g) If possible either archive or print the image for future reference.

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’

Fuji: Readout mode – ‘semi-auto’ with test/sensitivity GA=1

Kodak: Mode – ‘Pattern’ with raw data and no edge enhancement

Konica: QC S-value, E and F processing off

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.

c) Calculate an image quality factor, IQF,

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.

e) If possible either archive or print the images for future reference.

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

Purpose: To assess laser beam scanline integrity and jitter

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.

1.13 Moiré Patterns

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.

c) Read the plate as described for test 1.3.

d) Visually inspect the image for Moiré line pattern artefacts.

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

The following routine QA tests should be performed approximately annually

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.

Test Remedial Level


detector dose indicator consistency baseline ± 20% exposure equivalent
(sensitivity)
limiting resolution baseline ± 20%
TCDD (Quality Index) baseline ± 30%

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

[3] IPEM draft CR QC protocol

[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

Commissioning and Annual QA Tests

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.

It may be possible to undertake some of these tests on the acquisition/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.

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

Figure 1: Set-up for dosimetry

b) Position a chamber at least 30 cm in front of the detector (i.e. away from


backscatter). If possible either point the tube away from the detector or shield the
detector with a lead apron.

c) Collimate to the chamber.

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.

e) Record the measured dose and repeat twice.

f) Under the same beam conditions determine the mAs required to deliver 1mGy,
4mGy 12mGy and 50mGy

22
1.2Dark Noise

Purpose: To assess the level of noise inherent in the system

a) If possible, remove the grid from the system


.
b) Close the collimators and cover the detector with a lead apron. Set a low exposure
e.g. 50kVp and 0.5mAs. This will give an effectively zero dose, or ‘dark noise’ image.

c) Record the detector dose indicator value, and pixel value.

Tolerance: This test is used to set a baseline for future QA tests.

1.3 Linearity, system transfer properties and dark 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).

a) If possible, remove the grid from the system

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).

c) Record the detector dose indicator value.

d) Repeat for doses of order 4mGy, 12mGy and 50mGy.

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

dose =f(pixel value)

where f is some arbitrary function is required.

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.

b) Set the focus to detector distance (FDD) to be approximately 180 cm.

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

Figure 2: Regions of interest for image retention

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.

a) If possible, remove the grid from the system.

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.

e) Repeat steps a to c at least 3 times

f) Also repeat for 1µGy and 12µGy (1 image for each).

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

Fig 3: Positions of the ROI’s for uniformity tests

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.

a) Ensure the grid is removed from the system, if possible.

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.

Figure 4: Scaling errors test

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.

d) Visually inspect the image for blurring and stitching artefacts.

e) Repeat with a finer mesh if available.

Tolerance: No blurring should be present. If stitching artefacts are present there should be no
loss of information.

1.9 Limiting Spatial Resolution

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.

a) Ensure the grid is removed from the system, if possible.

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.

c) Set 70 kVp and the appropriate mAs to deliver ~4 mGy.

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.

e) Calculate an image quality factor, IQF,

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.

f) Repeat this test for exposures of ~1 mGy and ~12 mGy.

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

The following QA tests should be performed approximately annually:

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:

Test Remedial Level


Dark noise baseline± 50% exposure equivalent

Detector dose indicator consistency baseline ± 20% exposure equivalent

Limiting resolution baseline ± 20%

TCDD (Quality Index) baseline ± 30%

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

LEEDS TEST OBJECTS LTD


Miro House Phone +44(0)1423 321102
Becklands Close Fax +44(0)1423 322672
Boroughbridge Email info@leedstestobjects.com
North Yorkshire
YO51 9NR Web www.leedstestobjects.com