H2 PDF

H2 20-09-2005 20:54 Pagina 57
European protocol for the

quality control of the physical
and technical aspects of
mammography screening
2a Screen-film mammography
2b Digital mammography
E u r o p e a n g u i d e l i n e s f o r q u a l i t y a s s u r a n c e i n b r e a s t c a n c e r s c r e e n i n g a n d d i a g n o s i s Fo u r t h e d i t i o n 57
H2 20-09-2005 20:54 Pagina 58
EUROPEAN PROTOCOL FOR THE QUALITY CONTROL OF THE PHYSICAL AND TECHNICAL ASPECTS OF MAMMOGRAPHY SCREENING
58 E u r o p e a n g u i d e l i n e s f o r q u a l i t y a s s u r a n c e i n b r e a s t c a n c e r s c r e e n i n g a n d d i a g n o s i s Fo u r t h e d i t i o n
H2 20-09-2005 20:54 Pagina 59
Executive summary
A prerequisite for a successful screening project is that the mammograms contain sufficient
diagnostic information to be able to detect breast cancer, using as low a radiation dose as is
reasonably achievable (ALARA). This quality demand holds for every single mammogram. Quality
Control (QC) therefore must ascertain that the equipment performs at a constant high quality
level.
In the framework of ‘Europe Against Cancer’ (EAC), a European approach for mammography
screening is chosen to achieve comparable high quality results for all centres participating in the
mammography screening programme. Within this programme, Quality Assurance (QA) takes into
account the medical, organisational and technical aspects. This section is specifically concerned
with the quality control of physical and technical aspects of medical imaging in mammography
and the dosimetry.
The intention of this part of the guidelines is to indicate the basic test procedures, dose
measurements and their frequencies. The use of these tests and procedures is essential for
ensuring high quality mammography and enables comparison between centres. This document is
intended as a minimum standard for implementation throughout the EC Member States and does
not reduce more comprehensive and refined requirements for QC that are specified in local or
national QA Programmes. Therefore some screening programmes may implement additional
procedures.
Quality Control (QC)

Mammography screening should only be performed using modern dedicated X-ray equipment and
appropriate image receptors.
QC of the physical and technical aspects in mammography screening starts with specification
and purchase of the appropriate equipment, meeting accepted standards of performance. Before
the system is put into clinical use, it must undergo acceptance testing to ensure that the
performance meets these standards. This holds for the mammography X-ray equipment, image
receptor, film processor, viewing device and QC test equipment. After acceptance, the
performance of all equipment must be maintained above the minimum level and at the highest
level possible.
The QC of the physical and technical aspects must guarantee that the following objectives are met:
1. The radiologist is provided with images that have the best possible diagnostic information
obtainable when the appropriate radiographic technique is employed. The images should at
least contain the defined acceptable level of information, necessary to detect the smaller
lesions (see CEC Document EUR 16260).
2. The image quality is stable with respect to information content and optical density and
consistent with that obtained by other participating screening centres.
3. The breast dose is As Low As Reasonably Achievable (ALARA) for the mammographic
information required.
QC Measurements and Frequencies

To attain these objectives, QC measurements should be carried out. Each measurement should
follow a written QC protocol that is adapted to the specific requirements of local or national QA
programmes. The European Protocol for the Quality Control of the Physical and Technical
Aspects of Mammography Screening gives guidance on individual physical, technical and dose
measurements, and their frequencies, that should be performed as part of mammography
screening programmes.
H2 20-09-2005 20:54 Pagina 60
Several measurements can be performed by the local staff. The more elaborate measurements
should be undertaken by medical physicists who are trained and experienced in diagnostic
radiology and specifically trained in mammography QC. Comparability and consistency of the
results from different centres is best achieved if data from all measurements, including those
performed by local technicians or radiographers are collected and analysed centrally.
Image quality and breast dose depend on the equipment used and the radiographic technique
employed. QC should be carried out by monitoring the physical and technical parameters of the
mammographic system and its components. The following components and system parameters
should be monitored:
• X-ray generator and exposure control system

• Bucky and image receptor
• Film processing (for screen-film systems)
• Image processing (for digital systems)
• System properties (including dose)
• Monitors and printers (for digital systems)
• Viewing conditions
The probability of change and the impact of a change on image quality and on breast dose
determine the frequencies at which the parameters should be measured. The protocol gives also
the acceptable and achievable limiting values for some QC parameters. The acceptable values
indicate the minimal performance limits. The achievable values indicate the limits that are
achievable. Limiting values are only indicated when consensus on the measurement method and
parameter values has been obtained. The equipment required for conducting QC tests are listed
together with the appropriate tolerances in Table II.
Methods of dosimetry are described in the ‘European Protocol on Dosimetry in Mammography’
(EUR16263). It provides accepted indicators for breast dose, from both measurements on a
group of women and test objects.
The first (1992) version of this document (REF: EUR 14821) was produced by a Study Group,
selected from the contractors of the CEC Radiation Protection Actions. In the second (1996) and
third (1999) version the test procedures and limiting values have been reviewed critically based
on literature, the experience gained by users of the document and comments from
manufacturers of equipment and screen-film systems. Due to the introduction of digital
mammography an addendum on digital mammography was made in 2003. The current version
incorporates both screen-film and digital mammography and is based on further practical
experience with the protocol, comments from manufacturers and the need to adapt to new
developments in equipment and in the literature.
Corresponding address: Corresponding author:

EUREF office R. van Engen
[email protected] [email protected]
National Expert and Training Centre for

Breast Cancer Screening 451
Radboud University Nijmegen Medical Centre
P.O. Box 9101
6500 HB Nijmegen
The Netherlands
H2 20-09-2005 20:54 Pagina 61

a
Screen-film mammography
Authors
R. van Engen
S. van Woudenberg
H. Bosmans
K. Young
M. Thijssen
H2 20-09-2005 20:54 Pagina 62
S C R E E N - F I L M M A M M O G R A P H Y
EUROPEAN COMMISSION
Authors:
R. van Engen, Nijmegen, the Netherlands
S. van Woudenberg, Nijmegen, the Netherlands
H. Bosmans, Leuven, Belgium
K. Young, Guildford, United Kingdom
M. Thijssen, Nijmegen, the Netherlands
Contributors:
P. Baldelli, Ferrara, Italy
B. Beckers, Nijmegen, the Netherlands
R. Bijkerk, Nijmegen, the Netherlands
M. Borowski, Bremen, Germany
AK. Carton, Leuven, Belgium
D. Dance, London, United Kingdom
T. Deprez, Leuven, Belgium
D. Dierckx, Brussels, Belgium
A. Ferro de Carvalho, Lisbon, Portugal
M. Fitzgerald, London, United Kingdom
A. Flioni Vyza, Athens, Greece
M. Gambaccini, Ferrara, Italy
T. Geertse, Nijmegen, The Netherlands
G. Gennaro, Padua, Italy
N. Gerardy, Brussels, Belgium
A. de Hauwere, Gent, Belgium
P. Heid, Marseille, France
E. van der Kop, Nijmegen, the Netherlands
W. Leitz, Stockholm, Sweden
J. Lindeijer, Nijmegen, the Netherlands
R. van Loon, Brussels, Belgium
C. Maccia, Cachan, France
A. Maidment, Philadelphia, USA
H. Mol, Brussels, Belgium
B. Moores, Liverpool, United Kingdom
L. Oostveen, Nijmegen, the Netherlands
J. Pages, Brussels, Belgium
F. Rogge, Leuven, Belgium
M. Säbel, Erlangen, Germany
H. Schibilla, Brussels, EC
F. Shannoun, Luxembourg, Luxembourg
J. Shekdhar, London, United Kingdom
F. Stieve, Neuherberg, Germany
M. Swinkels, Nijmegen, the Netherlands
A. Taibi, Ferrara, Italy
D. Teunen, Luxembourg, EC
E. Vaño, Madrid, Spain
F. Verdun, Lausanne, Switserland
A. Watt, Edinburgh, United Kingdom
J. Zoetelief, Rijswijk, the Netherlands
Acknowledgements:
Comments have been received from the following manufacturers:
GE, Kodak
Comments have been received from the following group:

DIN working group mammography, IEC MT 31: mammography
H2 20-09-2005 20:54 Pagina 63
S C R E E N - F I L M M A M M O G R A P H Y a
2a.1 Introduction to the measurements

This protocol describes the basic techniques for the quality control (QC) of the physical and
technical aspects of mammography screening. It has been developed from existing protocols
(see section 5, bibliography) and the experience of groups performing QC of mammography
equipment. Since the technique of mammographic imaging and the equipment used are
constantly improving, the protocol is subject to regular updates.
Many measurements are performed using an exposure of a test object. All measurements are
performed under normal working conditions: no special adjustments of the equipment are
necessary.
Two standard types of exposures are specified:
• The reference exposure- which is intended to provide information on the system under defined
conditions, independent of the clinical settings
• The routine exposure- which is intended to provide information on the system under clinical
settings
For the production of the reference or routine exposure, an object is exposed using the machine
settings as follows (unless otherwise mentioned):
Reference exposure: Routine exposure:
test object thickness 45 mm1 45 mm
test object material PMMA PMMA
tube voltage 28 kV as used clinically
target material molybdenum as used clinically
filter material molybdenum as used clinically
compression device in contact with test object in contact with test object
anti scatter grid present present
source-to-image distance matching with focused grid matching with focused grid
phototimer detector in position closest to in position closest to

chest wall chest wall
automatic exposure control on as used clinically
optical density control as leading to the reference as leading to the target

optical density optical density
The optical density (OD) of the processed image is measured at the reference ROI, which lies 60
mm from the chest wall side and laterally centred. The reference optical density is preferably
1.60 ± 0.15 OD.
All measurements should be performed with the same cassette to rule out differences between
screens and cassettes except when testing individual cassettes (as in section 2a.2.2.2).
Limits of acceptable performance are given, but often a better result would be achievable. Both
the acceptable and achievable limits are summarised in section 2a.4, table 1. Occasionally no
limiting value is given, but only a typical value as an indication of what may normally be expected.
The measurement frequencies indicated in the protocol (summarised in section 2a.4) are the
H2 20-09-2005 20:54 Pagina 64
minimum required. When the acceptable limiting value is exceeded the measurement should be
repeated. If necessary, additional measurements should be performed to determine the origin of
the observed problem and appropriate actions should be taken to solve the problem.
For guidance on the specific design and operating criteria of suitable test objects; see the
Proceedings of the CEC Workshop on Test Phantoms (see section 2a.5, Bibliography). Definition
of terms, such as the ‘reference ROI’ and the ‘reference density’ are given in section 2a.1.2. The
evaluation of the results of the QC measurements can be simplified by using the forms for QC
reporting provided in section 2a.6.
2a1.1 Staff and equipment

Several measurements can be performed by the local staff. The more elaborate measurements
should be undertaken by medical physicists who are trained and experienced in diagnostic
radiology and specifically trained in mammography QC. Comparability and consistency of the
results from different centres is best achieved if data from all measurements, including those
performed by local technicians or radiographers are collected and analysed centrally.
The staff conducting the daily/weekly QC-tests will need the following equipment2 at the
screening site:
• Sensitometer • Standard test block3 (45 mm PMMA)

• Densitometer • QC test object
• Thermometer • Reference cassette
• PMMA plates4,5
The medical physics staff conducting the other QC-tests will need the following additional
equipment and may need duplicates of many of the above:
• Dosemeter • Stopwatch
• kVp-meter • Film/screen contact test device
• Exposure time meter • Tape measure
• Light meter • Compression force test device
• QC test objects • Rubber foam
• Aluminium sheets • Lead sheet
• Focal spot test device + stand • Aluminium stepwedge
2a.1.2 Definition of terms

Accuracy Gives how close the measured value of a quantity is to the true
value. In this document accuracy is used to check the
correspondence between nominal and measured values of high
voltage applied to an x-ray tube. The nominal value is taken as true
value. The accuracy is calculated as relative difference between
measured (m) and true (t) value, according to (m/t - 1), or as
percentage, (m/t -1) x 100%.
Air kerma Quotient of dEtr by dm where dEtr is the sum of initial kinetic
energies of all the charged ionising particles liberated by
uncharged ionising particles in a mass of air dm (adapted from
ICRU 1980). The common unit for air kerma is milliGray (mGy).
Air kerma measures, employing a ionization chamber or another
dose detector calibrated in mammography energy range, can be used
to evaluate the entrance dose (Entrance Surface Air Kerma – ESAK).
H2 20-09-2005 20:54 Pagina 65
Antiscatter grid Device positioned close to the entrance surface of an image

receptor for reducing the amount of scattered radiation reaching
the receptor.
Automatic Exposure Operation mode of an X-ray machine by which the tube loading is
Control (AEC) automatically controlled and terminated when a preset radiation
exposure to a dose detector located under the image receptor is
reached. Some more sophisticated equipment also allow the
automatic selection of tube potential (kV), target and filter
materials.
Average glandular dose Reference term (ICRP 1987) for radiation dose estimation from X-
(AGD) ray mammography i.e. the average absorbed dose in the glandular
tissue in a uniformly compressed breast. AGD value depends on X-
ray beam quality (HVL), breast thickness and composition. If
breast thickness and composition are not known, AGD can be
referred to a standard breast.
Baseline value Value of a parameter defined on basis of many repeated

measurements (at least 10), that can be considered typical for a
system. Generally, the baseline value is used when absolute limits
for a parameter do not exist.
Breast compression Application of a compression force to the breast during image

acquisition. This immobilises the breast, which limits motion
artifacts, and reduces breast thickness, which limits scatter
effects and makes breast thickness approximately uniform.
Compression paddle Thin device (few millimetres) rectangular shaped, made of plastic
material (typical PMMA or polycarbonate) that can be positioned
parallel to and above the breast table of a mammography
apparatus.
Contrast threshold Contrast level that produces a just visible difference between an
object and the background.
Dmin Optical density obtained after processing of a non-exposed film.

Dmin is not zero due to the absorption of light in the film support
and emulsion itself. In practice, for QC measurements, the density
of the first step of a sensitometric strip is taken as Dmin. It is taken
as a measure of the ‘base+fog’ value.
Dmax Maximum optical density achievable with an exposed film; usually

the density of the darkest step of a sensitometric strip. It
corresponds to the saturation zone of a film response curve.
Film gradient Index used to evaluate the film contrast.
Grad See Film gradient.
Heel effect Decreasing optical density measurable on a film in the cathode-

anode direction, caused by the non-uniform intensity distribution
of the X-ray beam. It is due to the geometric setup of the X-ray
tube.
Half Value Layer (HVL) Thickness of absorber which attenuates the air kerma of non-
monochromatic X-ray beams by half. The absorber normally used
to evaluate HVL of low energy X-ray beams, such as mammography
beams, is high purity aluminium (≥ 99.9%). It should be noticed
H2 20-09-2005 20:54 Pagina 66
that a correct measurement of HVL requires ‘good geometry’

conditions (proper distances among source, attenuator and image
receptor, collimation and perpendicular incidence at image
receptor entrance), rather far from geometry imposed by
mammography equipment. Thereby, the HVL measurement is a
sort of verification about the compatibility of radiation spectra with
standard values measured with calibrated beams.
Image quality There is not a definition of image quality univocally accepted.

Commonly, it is possible to define quality indices representing the
information content of the image; this is often done by test objects
including details whose visibility can be quantified by means of
proper scoring criteria.
Limiting value Maximum or minimum limit of a possible range, considered

acceptable for a given parameter.
Mean gradient (MGrad) Parameter describing the film contrast in the exposure range,
which contains most diagnostic information. MGrad is calculated
as the slope of the line through the points D0.25 = (Dmin + 0.25) OD
and D2 = (Dmin + 2.00) OD. Since the film curve is constructed from
a limited number of points, D0.25 and D2 must be obtained by
interpolation. Linear interpolation will result in sufficient accuracy.
Measurement error Standard deviation if the number of repeated measurements is

large enough (at least 5); maximum error [(max-min)/2] for few
measures.
Middle gradient (Grad1,2) Parameter describing the film contrast in the middle of the
diagnostic range. Grad1,2 is calculated as the slope of the line
through the points D1 = (Dmin+ 1.00) OD and D2 = (Dmin+ 2.00) OD.
Since the film curve is constructed from a limited number of
points, D1 and D2 must be obtained by interpolation. Linear
interpolation will result in sufficient accuracy.
Net optical density Optical density excluding base and fog. Base+fog value is
determined measuring the optical density into a non-exposed area
of film (see Dmin).
Optical density (OD) Logarithm (base 10) of the ratio between light intensity produced
by a visible light source and perpendicularly incident on a film (Io),
and light intensity transmitted by the film (I): OD = log10 (Io/I)
Optical density is measured by an instrument named
densitometer, that measures transmitted light intensity into an
area of the order of mm2.
Variations in optical density should be measured along a direction
parallel to the major axis of image receptor (perpendicular to
cathode-anode direction), to avoid influences by the angular
distribution of X-ray intensity (heel-effect).
Patient dose Generic term for a variety of radiation dose quantities applied to a
(group of) patient(s).
PMMA The synthetic material polymethylmethacrylate. Trade names

include Lucite, Perspex and Plexiglas.
Precision See Reproducibility.
H2 20-09-2005 20:54 Pagina 67
Quality assurance As defined by the WHO (1982): ‘All those planned and systematic
actions necessary to provide adequate confidence that a
structure, system or component will perform satisfactorily in
service (ISO 6215-1980). Satisfactory performance in service
implies the optimum quality of the entire diagnostic process i.e.,
the consistent production of adequate diagnostic information with
minimum exposure of both patients and personnel.’
Quality control As defined by the WHO (1982): ‘The set of operations

(programming, coordinating, carrying out) intended to maintain or
to improve [ . . . ] (ISO 3534-1977). As applied to a diagnostic
procedure, it covers monitoring, evaluation, and maintenance at
optimum levels of all characteristics of performance that can be
defined, measured, and controlled.’
QC test object Object made of tissue simulating material (typically PMMA) for
image quality evaluation; it generally includes objects simulating
mammographic lesions (microcalcifications, fibers, masses)
and/or resolution patterns and step wedges to measure
parameters such as spatial resolution or contrast, related to
image quality.
Radiation quality See HVL.
Reference cassette Cassette, properly identified, used for QC tests. Using a single
cassette permits to exclude variations in optical density caused by
changes in absorption from different cassettes or individual
screen efficiencies.
Reference exposure Exposure of the standard test object with predetermined values of
parameters to provide an image at reference conditions.
Reference optical density Optical density of (1.6 ± 0.1) OD, measured in the reference ROI.
Reference ROI Considering an image obtained by the standard test block, the
reference ROI is centred 60 mm perpendicular from the chest wall
in the middle of the major film axis.
Relative error Ratio between measurement error and mean value.
Reproducibility Indicates the measurement precision or the reliability of tested

equipment.
Region Of Interest (ROI) Measurement area of optical density whose boundaries can be
virtually defined on an image. ROI size can be around 1 cm2.
Routine exposure Exposure of the standard test block under the conditions that
would normally be used to produce a mammogram having the
routine optical density into the reference ROI. The routine
exposure is used to check optical density and dose stability under
clinical conditions.
Routine optical density Optical density measured in the reference ROI of a standard block
image obtained by a routine exposure. This value is chosen by the
site personnel as optimal value for mean clinical mammograms
provided by a specific imaging chain. The routine net optical
density should be included into the interval [1.4-1.9] OD.
H2 20-09-2005 20:54 Pagina 68
Spatial resolution Describes the smallest detectable detail at a defined contrast

(at high or low contrast) level to a given background. It is commonly evaluated by means of
bar patterns, i.e. test objects constituted by groups of absorbing
lines (typically Pb or Au) alternated to transparent lines of the
same size. Line groups have increasing spatial frequency (typically
expressed in ‘line pairs/mm’); the frequency at which line paires
remain distinguishable is taken as the limiting spatial resolution.
In conditions of ‘high contrast’, that can be obtained by exposing
the bar pattern only, this evaluation provides an estimation of the
limiting spatial resolution of the whole mammography unit. The
spatial resolution test can also be performed for ‘low contrast’
conditions, in order to simulate the degradation of both spatial
resolution and contrast typical of clinical images. Image quality
test objects including among the details one ore more bar patterns
are available on the market.
Speed Synonym of film sensitivity, is a parameter inversely proportional

to dose. Speed is defined as the reciprocal of dose necessary to
produce on a film an optical density equal to 1.00 + Dmin;
conventionally, it has been established that speed 100 means the
film needs 10 µGy to produce 1.00 OD above base+fog, while
speed 400 means the film requires 2.5 µGy to obtain the same
result. If film Speed is higher, the dose necessary to obtain a
same value of optical density is lower. Since in practice the
sensitometric curve is constructed from a limited number of
points, the film speed must be obtained by interpolation. Linear
interpolation will result in sufficient accuracy.
Standard breast Mathematical model generally used for calculations of glandular

dose in Monte Carlo simulations. It consists of a 40 mm thick
central region comprising a certain mixture by weight of adipose
tissue and glandular tissue (dependent on compressed breast
thickness and age) surrounded by a 5 mm thick superficial layer of
adipose tissue (simulating skin absorption). The standard breast
is semicircular with a radius ≥ 80 mm and has a total thickness of
50 mm (40+5+5). It is commonly assumed that a uniform PMMA
block 45 mm thick is equivalent in absorption to a standard
breast. (Note that other definitions of a standard breast have been
used in other protocols. As an example, in the UK the standard
breast has a total thickness of 45 mm with a 35 mm thick central
region.)
Standard test block PMMA test object to simulate the absorption of a standard breast.
Its thickness is (45.0 ± 0.5) mm and the remaining dimensions
can be either rectangular ≥ 150 mm x 100 mm or semi-circular
with a radius of ≥ 100 mm. The standard test block can be used to
check the AEC behaviour or to evaluate a mean value of AGD.
Typical value Value of a parameter found in most facilities in comparable

measurements. The statement of typical value is an indication
about values that could be expected, without imposing any limits
to obtainable results.
Tube loading Product of the X-ray tube current (milliampere, mA) and the
exposure time (seconds, s). It is quantified in units of mAs.
Tube potential The potential difference in units of kilovolt (kV) applied across the
anode and cathode of an X-ray tube during a radiographic
exposure.
H2 20-09-2005 20:54 Pagina 69
Tube yield Ratio between air kerma (mGy) measured without any test object
and the tube loading (mAs), for a known distance between the X-
ray source and the dosimeter and for preset exposure parameters.
X-ray spectrum Distribution of photon energies in an X-ray beam. It depends on

anode and filter material and tube potential, as well on all
attenuators (tube output window, compression device, air gap)
between anode and breast.
2a.2 Description of the measurements

Generally when absolute measurements of dose are performed, make sure that the proper
corrections for temperature and air pressure are applied to the raw values. Use one and the
same box of (fresh) film throughout the tests described in this protocol.
Local basic safety tests should be followed. If no local basic safety tests are available, an
example of such tests can be found in appendix 1.
2a.2.1 X-ray generation

2a.2.1.1 X-ray source
The measurements to determine the focal spot size, source-to-image distance, alignment of X-ray
field and image receptor, radiation leakage and tube output, are described in this section.
2a.2.1.1.1 Focal spot size

The measurement of the focal spot size is intended to determine its physical dimensions at
installation or when resolution has markedly decreased. The focal spot size must be determined
for all available targets of the mammography unit. For routine quality control the evaluation of
spatial resolution is considered adequate.
The focal spot dimensions can be obtained by using one of the

following methods.
• Star pattern method; a convenient method (routine testing)
• Slit camera; a complex, but accurate method for exact
dimensions (acceptance testing)
• Pinhole camera; a complex, but accurate method to
determine the shape (acceptance testing)
• Multi-pinhole test tool, a simple method to determine the size
across the field (routine/acceptance testing)
Some fully automated digital devices to measure focal spot

size are available. If validated they may be used.
When doing focal spot size measurements, it is advised to use
one of the above mentioned methods consistently.
A magnified X-ray image of the test device is produced using a

non-screen cassette. This can be achieved by placing a black
film (OD ≥ 3) between screen and film. Select the focal spot
size required, 28 kV tube voltage and a focal spot charge (mAs)
Fig. 2.1 Focal spot size measurement using the star pattern method
H2 20-09-2005 20:54 Pagina 70
to obtain an optical density between 0.8 and 1.4 OD base and fog excluded (measured in the
central area of the image). The device should be imaged at the reference ROI of the image plane,
which is located at 60 mm from the chest wall side and laterally centred. Remove the
compression device and use the test stand to support the test device. Select about the same
focal spot charge (mAs) that is used to produce the standard image of 45 mm PMMA, which will
result in an optical density of the star pattern image in the range 0.8 to 1.4.
According the IEC/NEMA norm, an 0.3 nominal focal spot is limited to a width of 0.45 mm and a
length of 0.65 mm. A 0.4 nominal focal spot is limited to 0.60 and 0.85 mm respectively. No
specific limiting value is given here, since the measurement of imaging performance of the focal
spot is incorporated in the limits for spatial resolution at high contrast. (see 2a.2.5.2)
Focal spot size: star pattern method

The focal spot dimensions can be estimated from the ‘blurring diameter’ on the image
(magnification 2.5 to 3 times) of the star pattern. The distance between the outermost blurred
regions is measured in two directions: perpendicular and parallel to the tube axis. Position the
cassette on top of the bucky (no grid).
The focal spot is calculated by applying formula 2.1, which can also be found in the completion
form.
(2.1)
where q is the angle of the radiopaque spokes, and dblur is the diameter of the blur.
The magnification factor (mstar) is determined by measuring the diameter of the star pattern on
the acquired image (dimage) and the diameter of the device itself (dstar), directly on the star, and is
calculated by:
mstar=dimage /dstar (2.2)
Limiting value None.

Frequency At acceptance and when resolution has changed.
Equipment Star resolution pattern (spoke angle 1° or 0.5° ) and
appropriate test stand.
Focal spot size: slit camera method

To determine the focal spot dimensions (f) with a slit camera, a 10 mm slit is used. Remove the
compression device and use a test stand to support the slit. Produce two magnified images
(magnification 2.5 to 3 times) of the slit, perpendicular and parallel to the tube axis.
The dimensions of the focal spot are derived by examining and measuring the pair of images
through the magnifying glass. Make a correction for the magnification factor according to
f=F/mslit, where F is the width of the slit image. The magnification factor (mslit) is determined by
measuring the distance from the slit to the plane of the film (dslit-to-film) and the distance from the
focal spot to the plane of the slit (dfocal spot-to-slit). mslit is calculated by:
mslit=dslit-to-film /dfocal spot-to-slit (2.3)
Note: mslit = mimage - 1, and the method requires a higher exposure than the star pattern method.
H2 20-09-2005 20:54 Pagina 71

Equipment Slit camera (10 µm slit) with appropriate test stand and
magnifying glass (5-10x), having a built-in graticule with 0.1 mm
divisions.
Focal spot size: pinhole method

To determine the focal spot dimensions (f) with a pinhole, a µ30 m gold/platinum alloy pinhole is
used. Produce a magnified image (magnification 2.5 to 3 times) of the pinhole.
The dimensions of the focal spot are derived by examining the images through the magnifying
glass and correcting for the magnification factor according to f = F/mpinhole, where F is the size of
the imaged focal spot. The magnification factor (mpinhole) is determined by measuring the distance
from the pinhole to the plane of the film (dpinhole-to-film) and the distance from the focal spot to the
plane of the pinhole (dfocal spot-to-pinhole). mpinhole is calculated by:
mpinhole=dpinhole-to-film /dfocal spot-to-pinhole (2.4)
Note: The method requires a higher exposure than the star pattern method.

Equipment Pinhole (diameter 30 µm) with appropriate test stand and
magnifying glass (5-10x), having a built-in graticule with 0.1 mm
divisions.
The multi-pinhole device is used similarly. It allows an estimate of the focal spot size at any
position in the x-ray field. This method is not suitable for measuring the dimension of fine focus
because of the relatively large size of the pin-holes.
2a.2.1.1.2 Source-to-image distance

Measure the distance between the focal spot indication mark on the tube housing and the top
surface of the bucky. Add distance between bucky surface and the top of the image receptor.
The source-to-image distance can be determined more accurately by imaging an object with
known dimensions a (≥ 10 cm) positioned on
the breast support table and positioned at a
distance d (≥ 20 cm) above the breast support
table. Measure the dimensions of the imaged
object on image 1 (object on breast support
table) and image 2 (object at distance d above
the breast support table). Using formula 2.5 the
source-to-image distance can be calculated.
f d
f =
⎛ 1 1 ⎞
object a * ⎜⎜ − ⎟ (2.5)
⎟
⎝ p1 p 2 ⎠
f = source-to-image distance
d
d = distance between the object in position 1
and 2
breast support table ob ject a = size of the imaged object
image receptor p1 = size of the object on image 1 (object on
p1 the breast support table)
p2 p2 = size of the object on image 2 (object at
a distance d above the breast support
Fig. 2.2 Source-to-image distance measurement table)
H2 20-09-2005 20:54 Pagina 72
Limiting value Manufacturers specification, typical 600-650 mm.

Frequency At acceptance, if the source-to-image distance is adjustable:
every six months.
Equipment Tape measure, arbitrary test object.
2a.2.1.1.3 Alignment of X-ray field/image receptor

The alignment of the X-ray field and image receptor at the chest
wall side can be determined with two loaded cassettes and two
X-ray absorbers, e.g. coins.
Place one cassette in the bucky tray and the other on top of the
breast support table. Make sure the second cassette has a
film loaded with the emulsion side away from the screen. It
must extend beyond the chest wall side about 30 mm. Mark
the chest wall side of the bucky by placing the absorbers on top
of the cassette. Automatic exposure will result in sufficient
optical densities. Reposition the films on a light box using the
imaged absorbers as a reference. The alignment between the
film, X-ray field and chest wall edge of the bucky should be
measured.
Fig. 2.3 Alignment of X-ray field/image receptor measurement
Note 1: The lateral edges of the X-ray field should at least expose the image receptor. A slight
extension beyond any edge of the image receptor is acceptable.
Note 2: If more than one field size or target is used, the measurement should be repeated for
each.
Limiting value For all focal spots:

All sides: X-rays must cover the film by no more than 5 mm
outside the film.
On chest wall edge: distance between film edge and edge of the
bucky must be ≤ 5 mm.
Frequency Yearly.
Equipment X-ray absorbers - e.g. coins, rulers, iron balls, tape measure.
2a.2.1.1.4 Radiation leakage

The measurement of leakage radiation comprises two parts; firstly the location of leakage and
secondly, the measurement of its intensity.
Position a beam stopper (e.g. lead sheet) over the end of the diaphragm assembly such that no
primary radiation is emitted. Enclose the tube housing with loaded cassettes and expose to the
maximum tube voltage and a high tube current (several exposures). Process the films and pin-
point any excessive leakage. Next, quantify the amount of radiation at the ‘hot-spots’ at a
distance of 50 mm of the tube with a suitable detector. Correct the readings to air kerma rate in
mGy/h (free in air) at the distance of 1 m from the focal spot at the maximum rating of the tube.
Limiting value Not more than 1 mGy in 1 hour at 1 m from the focus at the
maximum rating of the tube averaged over an area not exceeding
100 cm2, and according to local regulations.
Frequency At acceptance and after intervention on the tube housing.
Equipment Dose meter and appropriate detector.
2a.2.1.1.5 Tube output

The specific tube output (µGy/mAs) and the output rate (mGy/s) should both be measured using
a Molybdenum-Molybdenum target-filter combination at 28 kVp on a line passing through the
H2 20-09-2005 20:54 Pagina 73
focal spot and the reference ROI, in the absence of scatter material and attenuation (e.g. due to
the compression plate). A tube load (mAs) similar to that required for the reference exposure
should be used for the measurement. Correct for the distance from the focal spot to the detector
and calculate the specific output at 1 metre and the output rate at a distance equal to the focus-
to-film distance (FFD).
If the measurements are used for dosimetry, tube output measurements should be performed at
all relevant spectra with the compression plate in position.
Limiting values Acceptable: > 30 µGy/mAs at 1 metre, achievable > 40 µGy/ mAs
at 1 metre > 70% of value at acceptance for all target-filter
combinations.
Frequency Every six months and when problems occur.
Equipment Dose-meter, exposure timer.
Note: A high output is desirable for a number of reasons e.g. it results in shorter exposure
times, minimising the effects of patient movement and ensures adequate penetration of
large/dense breasts within the setting of the guard timer. In addition any marked changes
in output require investigation.
2a.2.1.2 Tube voltage and beam quality

The radiation quality of the emitted X-ray beam is determined by tube voltage, anode material and
filtration. Tube voltage and Half Value Layer (i.e. beam quality assessment) can be assessed by
the measurements described below.
2a.2.1.2.1 Reproducibility and accuracy

A number of tube voltages should be checked, which cover the range of clinically used settings.
The reproducibility is measured by repeated exposures at one fixed tube voltage that is normally
used clinically (e.g. 28 kVp).
Note: Consult the instruction manual of the kVp-meter for the correct positioning.
Limiting value Accuracy for the range of clinically used tube voltages: < ± 1 kV,
reproducibility < ± 0.5 kV.
Frequency Every six months.
Equipment kVp-meter.
2a.2.1.2.2 Half Value Layer (HVL)

The Half Value Layer can be assessed by adding thin aluminium (Al) filters to the X-ray beam and
measuring the attenuation.
Position the exposure detector at the reference ROI (since the HVL is position dependent) on top
of the bucky. Place the compression device halfway between focal spot and detector. Select a
molybdenum/molybdenum target/filter combination, 28 kV tube voltage and an adequate tube
loading (mAs-setting), and expose the detector directly. The filters can be positioned on the
compression device and must intercept the whole radiation field. Use the same tube load (mAs)
setting and expose the detector through each filter. For higher accuracy (about 2%) a diaphragm,
positioned on the compression paddle, limiting the exposure to the area of the detector may be
used (see European Protocol on Dosimetry in Mammography, ISBN 92-827-7289-6). At
acceptance the measurements should be repeated for all relevant spectra for average glandular
dose calculations. The HVL is calculated by applying formula 2.5.
(2.6)
H2 20-09-2005 20:54 Pagina 74
The direct exposure reading is denoted as Y0; Y1 and Y2 are the exposure readings with added
aluminium thickness of X1 and X2 respectively.
Note 1: The purity of the aluminium ≥ 99.9% is required. The thickness of the aluminium sheets
should be measured with an accuracy of 1%.
Note 2: For this measurement the output of the X-ray machine needs to be stable.
Note 3: The HVL for other (clinical) tube voltages and other target materials and filters may also
be measured for assessment of the average glandular dose (see appendix 5 and the
European Protocol on Dosimetry in Mammography, ISBN 92-827-7289-6).
Note 4: Alternatively a digital HVL-meter can be used, but correct these readings under extra
filtration following the manufacturers’ manual.
Limiting value For 28 kV Mo/Mo the HVL must be over 0.30 mm Al equivalent,
and is typically < 0.40 mm Al. Typical values of HVL for relevant
target-filter combinations and tube voltages, are shown in
appendix 5, table A5.3.
Frequency Yearly.
Equipment Dosemeter, aluminium sheets: 0.30, 0.40, 0.50, 0.60 mm.
2a.2.1.3 AEC-system
The performance of the Automatic Exposure Control (AEC) system can be described by the
reproducibility and accuracy of the automatic optical density control under varying conditions, like
different object thickness and tube voltages. Essential prerequisites for these measurements
are a stable operating film-processor and the use of the reference cassette. If more than one
breast support table, with a different AEC detector attached, is used then each system must be
assessed separately.
2a.2.1.3.1 Optical density control setting: central value and difference per step
To compensate for the long term variations in mean density due to system variations the central
optical density setting and the difference per step of the selector are assessed. To verify the
adjustment of the optical density control, produce exposures in the clinically used AEC mode of
the standard test object with varying settings of the optical density control selector.
A target value for the mean optical density at the reference ROI should be established according
to local preference, in the range: 1.4 – 1.9 OD, base and fog included.
Limiting value The optical density (base and fog included) of the step used
clinically at the reference ROI should remain within ± 0.15 OD of
the target value.
The change produced by each step in the optical density control
should be about 0.10 OD.
Step-sizes within the range 0.05 to 0.20 OD are acceptable.
The acceptable value for the range covered by full adjustment of
the density control is > 1.0 OD.
Frequency Step-size and adjustable range: every six months.
Density and mAs-value for clinically used AEC setting: daily.
Equipment Standard test block, densitometer.
2a.2.1.3.2 Back-up timer and security cut-off

The AEC system should also be equipped with a back-up timer or security cut-off which will
terminate the exposure in case of malfunctioning of the AEC system or when the required
exposure is not possible. Record the mAs-value at which the system terminates the exposure
e.g. when using increasing thickness of PMMA plates.
Warning: An incorrect functioning of the back-up timer or security cut-off could damage the tube.
To avoid excessive tube load consult the manual for maximum permitted exposure time.
H2 20-09-2005 20:54 Pagina 75
Limiting value The back-up timer and/or security cut-off should function properly.
Frequency Yearly.
Equipment PMMA plates or sheet of lead covering the detector.
2a.2.1.3.3 Short term reproducibility

Position the dosemeter in the x-ray beam but without covering the AEC-detector. The short term
reproducibility of the AEC system is calculated by the deviation of the exposure meter reading of
ten routine exposures (45 mm PMMA).
If it is noticed that the system switches between two spectra, release the compression paddle
and compress again or use another PMMA thickness (add for example 0.5 cm PMMA) to force
the choice of one single spectrum and repeat the measurement.
Limiting value Deviations from the mean value of exposures < ± 5%, achievable
< ± 2%.
Equipment Standard test block, dosemeter.
Note: For the assessment of the reproducibility, also compare these results from the short term
reproducibility with the results from the thickness and tube voltage compensation and
from the optical density control setting at 45 mm PMMA at identical settings. Any problem
will be indicated by a mismatch between those figures.
2a.2.1.3.4 Long term reproducibility

The long term reproducibility can be assessed from the measurement of optical density and tube
load (mAs) resulting from the exposures of a PMMA-block or the QC test object in the daily quality
control. Causes of deviations can be found by comparison of the daily sensitometry data and
tube load (mAs) recordings (see 2a.2.3.2).
Limiting value The variation from the target value must be within < ± 0.20 OD;
achievable < ± 0.15 OD.
Frequency Daily.
Equipment Standard test block or QC test object, densitometer.
2a.2.1.3.5 Object thickness and tube voltage compensation

Compensation for object thickness and tube voltage should be measured by exposures of PMMA
plates in the thickness range 20 to 70 mm in the clinically used AEC mode. If the system only
incorporates a semi-automatic exposure control, spectrum should be manually increased with
thickness, see appendix 4. At acceptance all AEC modes should be checked. Record the
spectrum, which is chosen by the AEC at all thicknesses. Record the value of the thickness
indicator at all thicknesses. Measure the optical density in the reference ROI.
Limiting value All optical density variations from the chosen target optical
density must be within ± 0.15 OD. Achievable: ± 0.10 OD.
Typical spectra for each PMMA thickness can be found in
appendix 4. The value of the thickness indicator must be within
± 0.5 cm of the thickness of the PMMA plates.
Frequency Every six months: full test.
Weekly: 20, 45, 65 mm PMMA exposed as for clinical setting.
Equipment PMMA: plates 10x180x240 mm3, densitometer.
2a.2.1.3.6 Correspondence between AEC sensors

Some mammography systems incorporate several independent AEC sensors. For these systems
it should be checked whether the optical density of images made with different sensors
correspond and if the correct sensor is chosen by the system.
To test correspondence, images of a homogeneous PMMA plate (45 mm thick) should be made
H2 20-09-2005 20:54 Pagina 76
with each AEC sensor. Choose the sensors manually. The optical density at the position of the
AEC sensor, which was used for that particular image, should be measured.
To test whether the correct AEC sensor is chosen, extra attenuation material (for example: 2 or 3
aluminium sheets used for HVL measurements) should be positioned above one AEC sensor
position. The markers on the compression paddle can be used as guidance. The whole sensor
should be covered and adjacent sensors should not be covered. The sensor, above which the
extra attenuation has been placed, must be chosen automatically by the system. If another
sensor is chosen, increase the amount of attenuating material until the correct sensor is chosen
or until it is beyond any doubt that the sensor does not work properly. This procedure must be
repeated for all sensor positions.
Note: If the Heel effect is large, it may be necessary to add extra attenuating material for sensor
positions near nipple side. The marker on the compression paddle may not always
completely coincide with the real position of the sensor.
Limiting value The variation in optical density between all AEC sensors should
be within 0.20 OD. The correct AEC sensor must be chosen.
Frequency Every six months: full test.
Equipment Standard test block, densitometer.
2a.2.1.4 Compression
The compression of the breast tissue should be firm but tolerable. There is no optimal value
known for the force, but attention should be given to the applied compression and the accuracy
of the indication. All units must have motorised compression.
2a.2.1.4.1 Compression force

The compression force can be adequately
measured with a compression force test device
or a bathroom scale (use compressible material
e.g. a tennis ball to protect the bucky and
compression device). The compression device
should be examined for possible cracks (which
might only be clearly visible under compression)
and sharp edges.
Fig. 2.4 Compression force measurement
When compression force is indicated on the console, it should be verified whether the figure
corresponds with the measured value. It should also be verified whether the applied
compression force is maintained over a period of 1 minute. A loss of force over this time may be
explained, for example, by a leakage in the pneumatic system.
Limiting value Maximum automatically applied force: 130 - 200 N. (~ 13-20 kg),
and must be maintained unchanged for at least 1 minute. The
indicated compression force should be within ± 20 N of the
measured value.
The compression device should not contain any cracks or sharp
edges.
Frequency Yearly.
Equipment Compression force test device.
H2 20-09-2005 20:54 Pagina 77
2a.2.1.4.2 Compression plate alignment

The alignment of the compression device at maximum force can be visualised and measured
when a piece of foam-rubber is compressed. Measure the distance between bucky surface and
compression device on each corner. Normally, those four distances are equal. Misalignment
normal to the chest wall side is less disturbing than in the parallel direction, as it compensates
for the heel effect. The upright edge of the device must be projected outside the receptor area
and optimally within the chest wall side of the bucky.
compression plate
bucky
foam rubber
Fig. 2.5 Compression plate alignment measurement, symmetrical load
Limiting value Minimal misalignment is allowed, the difference between the

measured distances at the left and the right side of the
compression paddle should be ≤ 5 mm for symmetrical load.
Frequency Yearly.
Equipment Foam rubber (specific mass: about 30 mg/cm3), tape measure.
2a.2.2 Bucky and image receptor

If more than one bucky and image receptor system is attached to the imaging chain than each
system must be assessed separately.
2a.2.2.1 Anti scatter grid

The anti scatter grid is composed of strips of lead and low density interspace material and is
designed to absorb scattered photons. The grid system is composed of the grid, a cassette
holder, a breast support table and a mechanism for moving the grid.
2a.2.2.1.1 Grid system factor

The grid system factor can be determined by dose measurements. Produce two images, one with
and one without the grid system. Use manual exposure control to obtain images of about
reference optical density. The first image is made with the cassette in the bucky tray (imaged
using the grid system) and PMMA on top of the bucky. The second with the cassette on top of the
bucky (imaging not using the grid system) and PMMA on top of the cassette. The grid system
factor is calculated by dividing the dose meter readings, corrected for the inverse square law and
optical density differences.
Note: Not correcting the doses for the inverse square law will result in an over estimation of 5%.
Typical value < 3.

Frequency At acceptance and when dose or exposure time increases suddenly.
Equipment Dosemeter, standard test block and densitometer.
H2 20-09-2005 20:54 Pagina 78
2a.2.2.1.2 Grid imaging

To assess the homogeneity of the grid in case of suspected damage or looking for the origin of
artefacts, the grid may be imaged by automatic exposure of the bucky at the lowest position of
the AEC-selector, without any added PMMA. This in general gives a good image of the gridlines.
Remark: For some systems it is not possible to image the grid due to the minimum required
exposure time.
Limiting value No significant non uniformity.

Frequency Yearly.
Equipment None.
2a.2.2.2 Screen-film
The current image receptor in screen-film mammography consists of a cassette with one
intensifying screen in close contact with a single emulsion film. The performance of the stock of
cassettes is described by the inter cassette sensitivity variation and screen-film contact.
2a.2.2.2.1 Inter cassette sensitivity and attenuation variation and optical density range
The differences between cassettes can be assessed with the reference exposure (section 2a.1).
Select an AEC setting (should be the normal position and using a fixed tube voltage, target and
filter) to produce an image having about the clinically used mean optical density on the
processed film. Repeat for each cassette using films from the same box or batch. Make sure the
cassettes are identified properly. Measure the exposure (in terms of mGy or mAs) and the
corresponding optical densities on each film at the reference ROI. To ensure that the cassette
tests are valid the AEC system in the mammography unit needs to be sufficiently stable. It will be
sufficient if the variation in repeated exposures selected by the AEC for a single cassette is (in
terms of mGy and mAs) < ± 2%.
Limiting value The exposure, in terms of mGy (or mAs), must be within ± 5% of
the mean for all cassettes.
The maximum difference in optical density between all
cassettes: ± 0.10 OD is acceptable, ± 0.08 OD is achievable.
Frequency Yearly, and after introducing new screens.
Equipment Standard test object, dosemeter, densitometer.
2a.2.2.2.2 Screen-film contact

Clean the inside of the cassette and the screen. Wait for at least 5 minutes to allow air between
the screen and film to escape. Place the mammography contact test device (about 40 metal
wires/inch, 1.5 wires/mm) on top of the cassette and make a non grid exposure to produce a
film with an average optical density of about 2 OD at the reference ROI. Regions of poor contact
will be blurred and appear as dark spots in the image. Reject cassettes only when they show the
same spots when the test is repeated after cleaning. View at a distance of 1 meter. Additionally
the screen resolution may be measured by imaging a resolution pattern placed directly on top of
a cassette.
Limiting value No significant areas (i.e. > 1 cm2) of poor contact are allowed in
the diagnostically relevant part of the film.
Frequency Every six months and after introducing new screens.
Equipment Mammography screen-film contact test device, densitometer and
viewbox.
H2 20-09-2005 20:54 Pagina 79
2a.2.3 Film processing

The performance of the film processing greatly affects image quality. The best way to measure
the performance is by sensitometry. Measurements of temperature and processing time are
performed to establish the baseline performance.
2a.2.3.1 Baseline performance of the processor

2a.2.3.1.1 Temperature verification and baseline
To establish a baseline performance of the automatic processor, the temperature of developer
and fixer are measured. Take care that the temperature is measured at a fixed point, as
recommended by the manufacturers. The measured values can be used as background
information when malfunction is suspected. Do not use a glass thermometer because of the
contamination risk in the event of breakage.
Limiting value Compliance with the manufacturer’s recommendations.

Equipment Electronic thermometer.
2a.2.3.1.2 Processing time

The total processing time can be measured with a stopwatch. Insert the film into the processor
and start the timer when the signal is given by the processor. When the processed film is
available, stop the timer. When malfunction of the processor is suspected, measure this
processing time exactly the same way again and check to see if there is any difference.
Limiting value Compliance with the manufacturer’s recommendations.

Frequency At acceptance and when problems occur.
Equipment Stopwatch.
2a.2.3.2 Film and processor

The films used in mammography should be specially designed for that purpose. Light
sensitometry is a suitable method to measure the performance of the processor. Disturbing
processor artefacts should not be present on the processed image.
2a.2.3.2.1 Sensitometry
Use a sensitometer to expose a film with light and insert the exposed side into the processor
first. Before measuring the optical densities of the step-wedge, a visual comparison can be made
with a reference strip to rule out a procedure fault, like exposure with a different colour of light or
exposure of the base instead of the emulsion side.
From the characteristic curve (the graph of measured optical density against the logarithm of
exposure by light) the values of base and fog, maximum density, speed and film gradients can be
derived. These parameters characterise the processing performance. A detailed description of
these ANSI-parameters and their clinical relevance can be found in appendix 2, film parameters.
Typical values: base and fog: 0.15 – 0.25 OD

contrast: MGrad: 3.0 - 4.0 see note
contrast: Grad1-2: 3.5 – 5.0
Frequency Daily.
Equipment Sensitometer, densitometer.
Note: There is no clear evidence for the optimal value of film gradient; the ranges quoted are
based on what is typical of current practice and are dependent on the film, which is used.
At the top end of these ranges the high film gradient may lead to under- and over exposure
of parts of the image for some types of breast, thereby reducing the information content.
H2 20-09-2005 20:54 Pagina 80
A further complication of using a very high film contrast is that stable conditions with very
low variability of the parameters are required to achieve any benefit in terms of overall
image quality (See appendix 3).
2a.2.3.2.2 Daily performance

The daily performance of the processor is assessed by sensitometry. After the processor has
been used for about one hour each morning, perform the sensitometry as described above. The
variability of the parameters can be calculated over a period of time e.g. one month (see
calculation of film parameters in appendix 2).
Limiting value See table below.

Frequency Daily and more often when problems occur.
Equipment Sensitometer, densitometer.
The assessment of variations can be found in the use of the following table, where the values are
expressed as a range (Max value - Min value). Acceptable and achievable ranges are quoted in
the table below. For centres where computer facilities for calculating speed and film gradient
(Mgrad and Grad1,2) are not available, speed and contrast indices are given. However, this
approach is less satisfactory as these indices are not pure measures of speed and contrast.
Assessment of variations acceptable achievable
base and fog < 0.03 < 0.02 OD
speed < 0.05 < 0.03
mean gradient (Mgrad) < 10% of baseline value < 5% of baseline value
mid gradient (Grad1,2) < 0.40 < 0.20
speed index < 0.30 < 0.20 OD
contrast index < 0.30 < 0.20 OD
2a.2.3.2.3 Artefacts
An image of the standard test block obtained daily, using a routine exposure, should be
inspected. This should show a homogeneous density, without significant scratches, shades or
other marks indicating artefacts.
Limiting value No artefacts.

Frequency Daily.
Equipment Standard test block or PMMA plates 40-60 mm and area
18X24 cm, viewing box.
2a.2.3.3 Darkroom
Light tightness of the darkroom should be verified. It is reported, that about half of darkrooms
are found to be unacceptable. Extra fogging by the safelights must be within given limits.
2a.2.3.3.1 Light leakage

Remain in the darkroom for a minimum of five minutes with all the lights, including the safelights,
turned off. Ensure that adjacent rooms are fully illuminated. Inspect all those areas likely to be a
source of light leakage. To measure the extra fog as a result of any light leakage or other light
sources, a pre-exposed film of about 1.2 OD is needed. This film can be obtained by a reference
H2 20-09-2005 20:54 Pagina 81
exposure of a uniform PMMA block. Always

measure the optical density differences in a line
perpendicular to the tube axis to avoid influence
of the heel effect.
Fig. 2.6 Line of measurement when

performing the light leakage measurement
Open the cassette with pre-exposed film and position the film (emulsion up) on the (appropriate
part of the) workbench. Cover half the film and expose for two minutes. Position the cover
parallel to the tube axis to avoid the influence of the heel effect in the measurements. Measure
the optical density difference of the background (Dbg) and the fogged area (Dfogged). The extra fog
(ΔD) equals:
ΔD = Dfogged - Dbg (2.7)
Limiting value Extra fog: ΔD ≤ 0.02 OD in 2 minutes.

Frequency Yearly and when light leakage is suspected.
Equipment Film cover, densitometer.
2a.2.3.3.2 Safelights
Perform a visual check that all safelights are in good working order (filters not cracked). To
measure the extra fog as a result of the safelights, repeat the procedure for light leakage but with
the safelights on. Make sure that the safelights were on for more than 5 minutes to avoid start-
up effects.
Limiting value Extra fog: ΔD ≤ 0.10 OD in 2 minutes.

Frequency Yearly and every time the darkroom environment has changed.
Equipment Film cover, densitometer.
2a.2.4 Viewing conditions
Since good viewing conditions are important for the correct interpretation of the diagnostic
images, they must be optimised. Although the need for relatively bright light boxes is generally
appreciated, the level of ambient lighting is also very important and should be kept low. In
addition it is imperative that glare is minimised by masking the film.
The procedures for photometric measurements and the values required for optimum
mammographic viewing are not well established. However there is general agreement on the
parameters that are important. The two main measurements in photometry are luminance and
illuminance. The luminance of viewing boxes is the amount of light emitted from a surface
measured in candela/m2. Illuminance is the amount of light falling on a surface and is measured
in lux (lumen/m2). The illuminance that is of concern here is the light falling on the viewing box,
i.e. the ambient light level. (An alternative approach is to measure the light falling on the film
reader’s eye by pointing the light detector at the viewing box from a suitable distance with the
viewing box off.) Whether one is measuring luminance or illuminance one requires a detector and
a photometric filter. This combination is designed to provide a spectral sensitivity similar to the
human eye. The collection geometry and calibration of the instrument is different for luminance
H2 20-09-2005 20:54 Pagina 82
and illuminance. To measure luminance a lens or fibre-optic probe is used, whereas a cosine
diffuser is fitted when measuring illuminance. Where the only instrument available is an
illuminance meter calibrated in lux it is common practice to measure luminance by placing the
light detector in contact facing the surface of the viewing box and converting from lux to cd/m2 by
dividing by π. Since this approach makes assumptions about the collection geometry, a correctly
calibrated luminance detector is preferred.
There is no clear consensus on what luminance is required for viewing boxes. It is generally
thought that viewing boxes for mammography need to be higher than for general radiography. In a
review of 20 viewing boxes used in mammographic screening in the UK, luminance averaged
4500 cd/m2 and ranged from 2300 to 6700 cd/m2. In the USA the ACR recommend a minimum
of 3500 cd/m2 for mammography. However some experts have suggested that the viewing box
luminance need not be very high provided the ambient light is sufficiently low and that the level
of ambient light is the most critical factor. The limiting values suggested here represent a
compromise position until clearer evidence is available.
2a.2.4.1 Viewing box

2a.2.4.1.1 Luminance
The tendency to use a high optical density for mammography means that one must ensure that
the luminance of the viewbox is adequate. Measure the luminance in the centre of each viewing
panel using a luminance meter calibrated in cd/m2. An upper limit is included to minimise glare
where films are imperfectly masked.
Limiting value Luminance should be in the range 3000-6000 cd/m2. The

deviation of the luminance between the centres of all panels of a
viewing box < ± 15% from the mean value of all panels.
Frequency Yearly.
Equipment Luminance meter.
2a.2.4.1.2 Homogeneity
The homogeneity of a single viewing box is measured by multiple readings of luminance over the
surface of the illuminator, compared with the luminance in the middle of the viewing panel.
Readings very near the edges (e.g. within 5 cm) of the viewing box should be avoided. Gross
mismatch between viewing boxes or between viewing conditions used by the radiologist and
those used by the radiographer should be avoided. If a colour mismatch exists, check to see that
all lamps are of the same brand, type and age. The local personnel has to make sure that all
tubes are changed at the same time. To avoid inhomogeneities as a result of dust, clean the light
boxes should be regularly cleaned inside and outside.
Limiting value The luminance across each panel should be within 30% of the
luminance in the centre of the panel.
Frequency Yearly.
Equipment Luminance meter.
2a.2.4.2 Ambient light

When measuring the ambient light level (illuminance), the viewing box should be switched off.
Place the detector against the viewing area and rotate away from the surface to obtain a maximal
reading. This value is denoted as the ambient light level.
Limiting value Ambient light level < 50 lux.

Frequency Yearly.
Equipment Illuminance meter.
H2 20-09-2005 20:54 Pagina 83
2a.2.5 System properties

The success of a screening programme is dependent on the proper information transfer and
therefore on the image quality of the mammogram. Decreasing the dose per image for reasons
of radiation protection is only justified when the information content of the image remains
sufficient to achieve the aims of the breast cancer screening programme.
2a.2.5.1 Dosimetry
Image PMMA plates of 20 mm thickness in clinical settings. Record the entrance surface air
kerma and the exposure factors chosen by the AEC. Repeat this measurement for 30, 40, 45,
50, 60 and 70 mm PMMA thickness. Calculate the average glandular dose for a breast
equivalent to each PMMA thickness. A detailed description of the calculation of the average
glandular dose can be found in appendix 5.
Limiting value A maximum average glandular dose is set per PMMA thickness:
Thickness Equivalent Maximum average glandular

of PMMA breast thickness dose to equivalent breasts
acceptable level achievable level
[cm] [cm] [mGy] [mGy]
2.0 2.1 < 1.0 < 0.6
3.0 3.2 < 1.5 < 1.0
4.0 4.5 < 2.0 < 1.6
4.5 5.3 < 2.5 < 2.0
5.0 6.0 < 3.0 < 2.4
6.0 7.5 < 4.5 < 3.6
7.0 9.0 < 6.5 < 5.1

Equipment Dose meter, standard test block, densitometer.
2a.2.5.2 Image Quality

The information content of an image may best be defined in terms of just visible contrasts and
details, characterised by its contrast-detail curve. The basic conditions for good performance and
the constancy of a system can be assessed by measurement of the following: resolution,
contrast visibility, threshold contrast and exposure time.
H2 20-09-2005 20:54 Pagina 84
2a.2.5.2.1 Spatial resolution

One of the parameters which determine image quality is the system spatial resolution. It can be
adequately measured by imaging two resolution lead bar patterns, up to 20 line pairs per mm
(lp/mm) each. They should be placed on top of PMMA plates with a total thickness of 45 mm.
Image the patterns at the reference ROI both parallel and perpendicular to the tube axis, and
determine these resolutions.
Note: If the resolution is measured at different heights between 25 and 50 mm from the
tabletop it can differ by as much as 4 lp/mm. The distance from the chest wall edge is
critical, but the position parallel to the chest wall side is not critical within ± 5 cm from the
reference ROI. Resolution is generally worse parallel to the tube axis due to the
asymmetrical shape of the focal spot.
Limiting value Acceptable: > 12 lp/mm, achievable: > 15 lp/mm at the

reference ROI in both directions.
Frequency Weekly.
Equipment PMMA plates 180x240 mm, resolution pattern(s) up to
20 lp/mm, densitometer.
2a.2.5.2.2 Image contrast

Since image contrast is affected by various parameters (like tube voltage, film contrast etc.) this
measurement is an effective method to detect a range of system faults. Make a reference
exposure of an aluminium or PMMA stepwedge and measure the optical density of each step in
the stepwedge. Draw a graph of the readings at each step against the stepnumber. The graph
gives an impression of the contrast. Since this graph includes the processing conditions, the film
curve has to be excluded to find the radiation contrast, see Appendix 3.
Remark: The value for image contrast is dependent on the whole imaging chain, therefore no
absolute limits are given. Ideally the object is part of, or placed on top of, the daily
quality control test object.
Limiting value Acceptable: ± 10%, achievable: ± 5%.

Frequency Weekly, and when problems occur.
Equipment PMMA or aluminium stepwedge, densitometer.
2a.2.5.2.3 Threshold contrast visibility

Extensive test: Threshold contrast visibility is determined for circular details with diameters in
the range from 0.1 to 2 mm. The details are imaged on a background object with a thickness
equivalent (in terms of attenuation) to 50 mm of PMMA. The details must be positioned at a
height of 20 to 25 mm above the breast support table6. Use the exposure factors that would be
selected clinically. Make two images. Three experienced observers should determine the
minimal contrast visible on both images. The detail diameter must cover the range from 0.1 to 2
mm. In this range minimal contrast visible for a large number of detail diameter must be
determined at acceptance and at least 5 detail diameters in subsequent tests.
The threshold contrast performance specified here relates to the nominal contrast calculated for
the details for a 28 kV tube voltage with a molybdenum target and filter materials as explained in
appendix 6. This nominal contrast depends on the thickness and materials used to manufacture
the test object, and is independent of the actual spectrum used to form the image, which should
be that used clinically. It does not include the effects of scatter. The average nominal threshold
contrasts should be compared with the limiting values below.
Weekly a simple test should give an indication of the lowest detectable contrast of ‘large’ objects
(diameter > 5 mm). Therefore a selection of low contrast objects have to be embedded in a
PMMA test object to mimic clinical exposures. There should be at least two visible and two non-
visible objects. Note, that the result is dependent on the mean OD of the image and on noise.
Produce a routine exposure and let two or three observers examine the low contrast objects. The
number of visible objects is recorded. Ideally the object is part of, or placed on top of, the daily
quality control test object.
H2 20-09-2005 20:54 Pagina 85
Limiting value Extensive test: results at acceptance are used as reference.

Weekly test: minimum detectable contrast for a 5-6 mm detail
< 1.5%.
Frequency Yearly (extensive test), weekly (simple test).
Equipment Test object with low contrast details plus PMMA plates, to a
thickness of 45 mm, densitometer.
2a.2.5.2.4 Exposure time
Long exposure times can give rise to motion unsharpness. Exposure time may be measured by
some designs of kVp- and output meters. Otherwise a dedicated exposure timer has to be used.
The time for a routine exposure is measured.
Limiting value Acceptable: < 2 sec.; achievable: < 1.5 sec.

Frequency Yearly and when problems occur.
Equipment Exposure time meter, standard test block.
2a.3 Daily and weekly QC tests

There are a number of tests that should be conducted daily or weekly. For this purpose, a
dedicated QC-test object or set of test objects are convenient. The actual frequencies
recommended for each measurement are specified in section 2a.2.3.2.2 and summarised in
Table 2a.4.1. The procedure must facilitate the measurement of some essential physical
quantities, and it should be designed to evaluate:
• AEC reproducibility
• Tube output stability
• Reference optical density
• Spatial resolution
• Image contrast
• Threshold contrast visibility
• Homogeneity, artefacts
• Sensitometry (speed, contrast, fog)
Practical considerations:
• Ideally the sensitometric stepwedge should be on the same film as the image of the test
object, to be able to correct optimally for the processing conditions.
• To improve the accuracy of the daily measurement, the test object should be designed in such
a way that it can be positioned reproducibly on the bucky.
• The shape of the test object does not have to be breast-like. To be able to perform a good
homogeneity check, the test object should cover the normally imaged area on the image
receptor (180x240 mm).
• For testing the AEC reproducibility, the PMMA test object may comprise several layers of
PMMA, 10 or 20 mm thick. It is important to use the same PMMA blocks since variations in
thickness of the PMMA plates will influence the tube load (mAs) read-out. Sufficient blocks
are required to make up a thickness in the range 20-70 mm to adequately simulate the range
of breast thickness found clinically.
H2 20-09-2005 20:54 Pagina 86
2a.4 Tables
Table 2a.4.1 Frequency of quality control, measured and limiting values
2a.2.1 X-ray generation and control frequency typical limiting value unit
value acceptable achievable
X-ray source
- focal spot size i 0.3 IEC/NEMA - -
- source-to-image distance i ≥ 600 - - mm
- alignment of x-ray field/ 12 - <5 <5 mm
- image receptor
- film/bucky edge 12 - ≤5 ≤5 mm
- radiation leakage i - <1 <1 mGy/hr
* output 6 - > 30% > 40 µGy/mAs
> 70% of baseline
tube voltage
- reproducibility 6 - < ± 0.5 < ± 0.5 kV
- accuracy (25 – 31 kV) 6 - < ± 1.0 < ± 1.0 kV
- HVL 12 - See appendix 5 See appendix 5
table A5.3 table A5.3
AEC
* central opt. dens 6 - < ± 0.15 of target value - OD
- control setting
- target opt. dens. control setting 6 - 1.4 -1.9 OD
- opt. dens. control step 6 - 0.05 - 0.20 0.05 - 0.10 OD
- adjustable range 6 - > 1.0z a > 1.0z a OD
* short term reproducibility 6 - < ± 5%a < ± 2%zz mGy
* long term reproducibility d - < ± 0.20 < ± 0.15 OD
- object thickness w - < ± 0.15 < ± 0.10 OD
- and tube voltage compensation 6 - < ± 0.15 < ± 0.10 OD
- spectra 6 See appendix 4
- correspondence between AEC sensors 6 - < 0.20 OD
compression
- compression force 12 - 130 - 200 - N
- maintain force for 1 minute 12 - 1 1 min
- compression force indicator 12 - < ± 20 < ± 20 N
- compression plate alignment, 12 - ≤5 ≤5 mm
- symmetric
2a.2.2 Bucky and image receptor frequency typical limiting value unit
anti scatter grid

* grid system factor i <3 - - -
screen-film
* inter cassette sensitivity 12 - < ± 5% < ± 5% mGy
- - variation (mAs)
* inter cassette sensitivity 12 - < ± 0.10 < ± 0.08 OD
- - variation (OD range)
- screen-film contact 12 - No significant areas - -
of poor contact
H2 20-09-2005 20:54 Pagina 87
Table 2a.4.1 continued
2a.2.3 Film processing frequency typical limiting value unit

processor
- temperature i 34 - 36 - - °C
- processing time i 90 - 120 - - s
film
- sensitometry: base and fog d 0.15 - 0.251 - - OD
speed d - - - -
Contrast Mgrad: d 3.0 - 4.02 - - -
Grad1,2 d 3.5 - 5.0 - - -
- daily performance d - See 2a.2.3.2 See 2a.2.3.2 -
- artefacts d - No disturbing artefacts - -
darkroom
- light leakage (extra fog 12 - ≤ + 0.02 ≤ + 0.02 OD
- in 2 minutes)
- safelights (extra fog 12 - ≤ + 0.10 ≤ + 0.10 OD
- in 2 minutes)
2a.2.4 Viewing conditions frequency typical limiting value unit

viewing box
- luminance 12 - 3000 - 6000 3000 - 6000 cd/m2
- homogeneity 12 - < ± 30% < ± 30% cd/m2
- luminance difference 12 - < ± 15% < ± 15% cd/m2
between panels
environment
- ambient light level 12 - < 50 < 50 lux
2a.2.5 System properties frequency typical limiting value unit

dosimetry
* Glandular dose 6
- PMMA thickness (cm)
2.0 < 1.0 < 0.6 mGy
3.0 < 1.5 < 1.0 mGy
4.0 < 2.0 < 1.6 mGy
4.5 < 2.5 < 2.0 mGy
5.0 < 3.0 < 2.4 mGy
6.0 < 4.5 < 3.6 mGy
7.0 < 6.5 < 5.1 mGy
image quality --
* spatial resolution, w - > 12 > 15 lp/mm
- - reference ROI
* threshold contrast visibility w - < 1.5% < 1.5% -
* exposure time 12 - <2 < 1.5 s
End of table 2a.4.1 i = At acceptance; d = daily; w = weekly; 6 = every 6 months; 12 = every 12 months * standard measurement conditions
1. For standard blue based films only 2. Depend on the film which is used
H2 20-09-2005 20:54 Pagina 88
Table 2a4.2 QC equipment specifications
QC equipment accuracy reproducibility unit
sensitometer - ± 2% OD
densitometer ± 2% at 1.0 ± 1% OD
dosemeter ± 5% ± 1% mGy
thermometer ± 0.3 ± 0.1 °C
kVp-meter for mammographic use ± 2% ± 1% kV
exposure time meter ± 5% ± 1% s
luminance meter ± 10% ± 5% Cd.m-2
illuminance meter ± 10% ± 5% klux
test objects, PMMA ± 2% - mm
compression force test device ± 10% ± 5% N
aluminium filters (purity ≥ 99.9%)
aluminium stepwedge
resolution pattern
focal spot test device
stopwatch
film/screen contact test tool
tape measure
rubber foam for compression plate alignment
lead sheet
H2 20-09-2005 20:54 Pagina 89
2a.5 Bibliography
CEC-Reports
1. Technical and Physical Parameters for Quality Assurance in Medical Diagnostic Radiology;
Tolerances, Limiting Values and Appropriate Measuring Methods.
1989: British Institute of Radiology; BIR-Report 18, CEC-Report EUR 11620.
2. Optimisation of Image Quality and Patient Exposure in Diagnostic Radiology.
1989: British Institute of Radiology; BIR-Report 20, CEC-Report EUR 11842.
3. Dosimetry in Diagnostic Radiology.
Proceedings of a Seminar held in Luxembourg, March 19-21, 1991.
1992: Rad. Prot. Dosimetry vol 43, nr 1-4, CEC-Report EUR 14180.
4. Test Objects and Optimisation in Diagnostic Radiology and Nuclear Medicine.
Proceedings of a Discussion Workshop held in Würtzburg (FRG), June 15-17, 1992
1993: Rad. Prot. Dosimetry vol 49, nr 1-3; CEC-Report EUR 14767.
5. Quality Control and Radiation Protection of the Patient in Diagnostic Radiology and Nuclear
Medicine.
1995: Rad. Prot. Dosimetry vol 57, nr 1-4, CEC-Report EUR 15257.
6. European Guidelines on Quality Criteria for Diagnostic Radiographic Images.
1996: CEC-Report EUR 16260.
Protocols
1. The European Protocol for the Quality Control of the Technical Aspects of Mammography
Screening.
2. European Protocol on Dosimetry in Mammography.
3. Protocol acceptance inspection of screening units for breast cancer screening, version April
2002.(in Dutch)
National Expert and Training Centre for Breast Cancer Screening, University Hospital
Nijmegen (NL) 2002.
4. LNETI/DPSR: Protocol of quality control in mammography (in English) 1991.
5. ISS: Controllo di Qualità in Mammografia: aspetti technici e clinici.
Instituto superiore de sanità (in Italian),
1995: ISTASAN 95/12.
6. IPSM: Commissioning and Routine testing of Mammographic X-Ray Systems - second edition
The Institute of Physical Sciences in Medicine, York.
1994: Report no. 59/2.
7. American College of Radiology (ACR), Committee on Quality Assurance in Mammography:
Mammography quality control.
1994, revised edition.
8. American Association of Physicists in Medicine (AAPM): Equipment requirements and quality
control for mammography.
1990: report No. 29.
9. Quality Control in Mammography,
1995: Physics consulting group Ontario Breast Screening Programme.
10. QARAD/LUCK: Belgisch Protocol voor de kwaliteitszorg van de fysische en technishe
aspecten bij mammografische screening (in Dutch) 1999.
11. Protocollo italiano per il controllo di qualità degli aspetti fisici e tecnici in mammografia,
2004: AIFM report n. l, http://www.aifm.it/report/
Publications
1. Chakraborty D.P.: Quantitative versus subjective evaluation of mammography accreditation
test object images.
1995: Med. Phys. 22(2):133-143.
2. Wagner A.J.: Quantitative mammography contrast threshold test tool.
1995: Med. Phys. 22(2):127-132.
European guidelines for quality assurance in mammography screening Four th edition 89

H2 20-09-2005 20:54 Pagina 90
3. Widmer J.H.: Identifying and correcting processing artefacts.

Technical and scientific monograph.
Health Sciences Division.
Eastman Kodak Company, Rochester, New York, 1994.
4. Caldwell C.B.: Evaluation of mammographic image quality: pilot study comparing five
methods.
1992: AJR 159:295-301.
5. Wu X.: Spectral dependence of glandular tissue dose in screen-film mammography.
1991: Radiology 179:143-148.
6. Hendrick R.E.: Standardization of image quality and radiation dose in mammography.
1990: Radiology 174(3):648-654.
7. Baines C.J.: Canadian national breast screening study: assessment of technical quality by
external review.
1990: AJR 155:743-747.
8. Jacobson D.R.: Simple devices for the determination of mammography dose or radiographic
exposure.
1994: Z. Med. Phys. 4:91-93.
9. Conway B.J.: National survey of mammographic facilities in 1985, 1988 and 1992.
1994: Radiology 191:323-330.
10. Farria D.M.: Mammography quality assurance from A to Z.
1994: Radiographics 14: 371-385.
11. Sickles E.A.: Latent image fading in screen-film mammography: lack of clinical relevance for
batch-processed films.
1995: Radiology 194:389-392.
12. Sullivan D.C.: Measurement of force applied during mammography.
1991: Radiology 181:355-357.
13. Russell D.G.: Pressures in a simulated breast subjected to compression forces comparable
to those of mammography.
1995: Radiology 194:383-387.
14. Faulkner K.: Technical note: perspex blocks for estimation of dose to a standard breast -
effect of variation in block thickness.
1995: Br. J. Radiol. 68:194-196.
15. Faulkner K.: An investigation into variations in the estimation of mean glandular dose in
mammography.
1995: Radiat. Prot. Dosimet. 57:405-407.
16. K.C. Young, M.G. Wallis, M. L. Ramsdale: Mammographic Film Density and Detection of
Small Breast cancers, 1994 Clin. Radiol (49) 461-465.
17. Tang S.: Slit camera focal spot measurement errors in mammography.
1995: Med. Phys. 22:1803-1814.
18. Hartmann E.: Quality control of radiographic illuminators and associated viewing equipment.
Retrieval and viewing conditions.
1989: BIR report 18:135-137.
19. Haus A.G.: Technologic improvements in screen-film mammography.
1990: Radiology 174(3):628-637.
20. L.K. Wagner, B.R. Archer, F. Cerra; On the measurement of half-value layer in screen-film
mammography.
1990: Med. Phys. (17):989-997.
21. J.D. Everson, J.E. Gray: Focal-Spot Measurement: Comparison of Slit, Pinhole, and Star
Resolution Pattern Techniques.
1987: Radiology (165):261-264.
22. J. Law: The measurement and routine checking of mammography X-ray tube kV.
1991: Phys.Med.Biol. (36):1133-1139.
23. J. Law: Measurements of focal spot size in mammography X-ray tubes.
1993: Brit. J. Of Radiology (66):44-50.
24. M. Thijssen et al: A definition of image quality: the image quality figure.
1989: Brit. Inst. Radiology, BIR-report 20: 29-34.
25. R.L. Tanner: Simple test pattern for mammographic screen-film contact measurement.
1991: Radiology (178):883-884.
H2 20-09-2005 20:54 Pagina 91
26. K.C. Young, M.L. Ramsdale, A. Rust: Mammographic dose and image quality in the UK
breast screening programme. 1998, NHSBSP report 35.
27. J. Zaers, S. van Woudenberg, G. Brix: Qualitätssicherung in der Röntgenmammographie
1997: Der Radiologe (37):617-620.
28. J. Law: Checking the consistency of sensitometers and film processors in a mammographic
screening.
programme. 1996 Brit. J. Of Radiology (69) 143-147.
29. K.J. Robson, C.J. Kotre, K. Faulkner : The use of a contrast-detail test object in the
optimization of optical density in mammography, 1995 Brit. J. Of Radiology (68) 277-282.
30. J.A. Terry, R.G. Waggener, M.A. Miller Blough: Half-value layer and intensity variations as a
function of position in the radiation field for film screen mammography 1999.
Med. Phys. 26 259-266.
31. S. Tang, G.T. Barnes, R.L. Tanner: Slit camera focal spot measurement errors in
mammography.
1995 Med. Phys. (22) 1803-1814.
32. J. Coletti et al.: Comparison of exposure standards in the mammography x-ray region, 1997.
Med. Phys (8) 1263-1267.
33. C. Kimme Smith et al. Mammography film processor replenishment rate: Bromide level
monitoring.
1997 Med. Phys. (3) 369-372.
35. M. Goodsitt, H. Chan, B. Liu: Investigation of the line-pair method for evaluating
mammographic focal spot performance, 1997. Med. Phys. (1) 11-15.
34. A. Krol et al. Scatter reduction in mammography with air gap.
1996 Med. Phys. (7) 1263-1270.
35. D. McLean, J. Gray: K-characteristic photon absorption from intensifying screens and other
materials:
Theoretical calculations and measurements.
1996 Med. Phys. (7) 1253-1261.
36. P. Rezentes, A de Almeida, G. Barnes: Mammographic Grid Performance.
1999 Radiology 210:227-232.
37. J. Hogge et al. Quality assurance in Mammography: Artifact Analysis.
1999 Radiographics 19:503-522.
38. J. Byng et al.: Analysis of Mammographic Density and Breast Cancer Risk from Digitized
Mammograms.
1998 Radiographics 18:1587-1598.
39. D.R. Dance et al.: Additional factors for the estimation of mean glandular breast dose using
the UK mammography dosimetry protocol, 2000 Phys. Med. Biol. (45) 3225-3240.
Other reports
1. International Electrotechnical Commission (IEC), Geneva, Switzerland: Characteristics of
focal spots in diagnostic X-ray tube assemblies for medical use.
1982: IEC-Publication 336.
2. Quality assurance in mammography - quality control of performance and constancy
1990: Series of Nordic Reports on radiation Safety No. 1, Denmark, Finland, Iceland, Norway
and Sweden.
3. Société française des physiciens d’hôpital, Nancy: Contrôle de qualité et mesure de dose
en mammographie - aspects théoriques et pratiques (in French).
1991.
4. Department Health & Social Security, Supplies Technology Division (DHSS): Guidance notes
for health authorities on mammographic equipment requirements for breast cancer
screening.
1987: STD
5. Department of Radiodiagnostic Radiology, University of Lund, Sweden: Quality Assurance in
Mammography.
1989.
6. Sicherung der Bildqualität in röntgendiagnostischen Betrieben - Filmverarbeitung.
1996: DIN 6868-2: Beuth Verlag GmbH, Berlin.
H2 20-09-2005 20:54 Pagina 92
7. American Association of Physicists in Medicine (AAPM): Basic quality control in diagnostic

radiology.
1978: report No. 4.
8. ECRI: Special issue: Mammography Units.
1989: Health Devices:Vol.18:No.1:Plymouth Meeting (PA).
9. ECRI: Double issue: Mammography Units.
1990: Health Devices:Vol.19:No.5-6:Plymouth Meeting (PA).
10. Siemens Medical Systems Inc., New Jersey: Mammography QA - Doc.# 54780/up
1990.
11. ANSI: Determination of ISO speed and average gradient.
American National Standards Institute (ANSI).
1983: Nr. PH2.50.
12. Sicherung der Bildqualität in röntgendiagnostischen Betrieben - Konstantzprüfung an
Röntgen- einrichtungen für Mammographie.
2004: DIN-1:6868-7:Beuth Verlag GmbH, Berlin.
13. Sicherung der Bildqualität in röntgendiagnostischen Betrieben - Abnahmeprüfung an
Röntgen-Einrichtungen für Mammographie.
2003: DIN-2:6868 teil 152: Beuth Verlag GmbH, Berlin.
14. ICRP Publication 52, including the Statement from the Como Meeting of the ICRP.
1987: Annals of the ICRP 17 (4), i-v, Pergamon Press, Oxford, UK.
15. ICRP Publication 60, 1990 Recommendations of the ICRP.
(Adopted by the Commission in November 1990).
1991: Annals of the ICRP 21 (1-3), Pergamon Press, Oxford, UK.
16. Bewertung und routinemäßige Prüfung in Abteilungen für medizinische Bildgebung -
Abnahmeprüfungen – Abbildungsqualität von Röntgen-Einrichtungen für die Mammographie
2001: DIN EN 61223-3-2: Beuth Verlag GmbH, Berlin.
17. International Electrotechnical Commission (IEC), Geneva, Switzerland: Evaluation and
routine testing in medical imaging departments, part 3-2: Acceptance tests – Imaging
performance of mammographic X-ray equipment.
2004: IEC 61223-3-2 Ed. 2.
H2 20-09-2005 20:54 Pagina 93
2a.6 Completion forms for QC reporting
QC report
based on
The European protocol for the quality control of the physical

and technical aspects of mammography screening
Fourth edition
Date:
Contact:
Institute:
Address:
Telephone:
Conducted by:
H2 20-09-2005 20:54 Pagina 94
2a.2.1 X-ray generation and control
2a.2.1.1 X-ray source

2a.2.1.1.1 Focal spot size
Class (large) focal spot: (IEC)
* star pattern method

diameter star pattern dstar mm
spoke angle θ θ °
diameter magnified star image dmag mm
diameter first MTF zero ⊥ AC axis dblur, ⊥ mm
diameter first MTF zero // AC axis dblur, // mm
* slit camera method

width slit mm
distance slit-to-film dslit-film mm
distance focus-to-slit dfocus-slit mm
width slit image ⊥ AC axis F⊥ mm
width slit image // AC axis F // mm
* pinhole method
diameter pinhole µm
distance pinhole-to-film dpinhole-film mm
distance focus-to-pinhole dfocus-pinhole mm
diameter pinhole ⊥ AC axis f⊥ mm
diameter pinhole // AC axis f // mm
Focal spot size f⊥ = mm
f// = mm
Accepted: yes / no
2a.2.1.1.2 Source-to-image distance

Nominal value: mm
Measured value :
- Focus indication to bucky: mm
- Bucky to cassette: mm
Source-to-image distance: mm
H2 20-09-2005 20:54 Pagina 95
2a.2.1.1.3 Alignment of X-ray field / image receptor

Distance at chest wall side film: inside/outside image receptor:
position
- left: mm, in / out
- nipple: mm, in / out
- right : mm, in / out
- chest : mm, in / out
Distance between film edge and bucky edge: mm
Accepted: yes / no
2a.2.1.1.4 Radiation leakage

Description of position of ‘hot spots’
1
2
3
detector surface area: mm2
measured: calculated for

distance from tube: 50 mm 1000 mm,
surface area: mm2 100 cm2:
nr:
1. mGy/hr
2. mGy/hr
3. mGy/hr
Accepted: yes / no
2a.2.1.1.5 Tube output

Focus to detector distance: mm
Surface air kerma: mGy
Focal spot charge: mAs
Specific tube output at 1 m µGy/mAs

Output rate at FFD mGy/s
Accepted: yes / no
2a.2.1.2 Tube voltage

2a.2.1.2.1 Reproducibility and accuracy
Pre-set tube load: mAs
Clinically most relevant kV: kV
Accuracy at clinical tube voltage settings

Setting 28 kV
Measured kV
Deviation kV
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 96
Reproducibility at the clinically most relevant tube voltage setting

Set tube voltage kV
Measured value: 1. 2. 3. 4. 5. kV
Reproducibility (max difference from the mean): kV
Accepted: yes / no
2a.2.1.2.2 Half Value Layer

Target/filter: Mo/Mo
Measured tube voltage: 28 kV
Filtration: 0.0 0.30 0.40 mm Al
Exposure: Y0 Y1 Y2
1. mGy
2. mGy
3. mGy
Average exposure: mGy
= mm Al
Deviation exposure at 0 mm Al: %

Accepted: yes / no
Half Value Layer for average Glandular Dose calculations

Target/filter: /
Measured tube voltage: kV
Filtration: 0.0 mm Al
Exposure: Y0 Y1 Y2
1. mGy
2. mGy
3. mGy
Average exposure: mGy
HVL: mm Al
Deviation exposure at 0 mm Al : %
H2 20-09-2005 20:54 Pagina 97
2a.2.1.3 AEC-system
2a.2.1.3.1 Optical density control setting: central value and difference per step
Target density value: OD
Setting Exposure Tube load Density Density incr.

mGy mAs OD OD
-3
-2
-1
-0
-1
-2
-3
Accepted: yes / no
Adjustable range: OD
Accepted: yes / no
Optical density control setting for reference density:

Optical density control setting for target density:
2a.2.1.3.2 Back-up timer and security cut-off

Exposure terminates by exposure limit : yes/no
Alarm or error code: yes/no
Exposure: mGy
Tube load: mAs
2a.2.1.3.3 Short term reproducibility

Optical density control setting:
Exp. # Exposure (mGy) Tube load (mAs)

1
2
3
4
5
6
7
8
9
10
Deviation in tube load: % (= 100 x (max-min)/mean )
Accepted: yes / no
2a.2.1.3.4 Long term reproducibility: Forms should be made to suit the local preferences.
H2 20-09-2005 20:54 Pagina 98
2a.2.1.3.5 Object thickness and tube voltage compensation

Optical density control setting:
Mode name:
PMMA Target/ Tube voltage Optical Thickness

thickness filter (kV) Density (OD) Indication (mm)
10 mm
20 mm
30 mm
40 mm
50 mm
60 mm
70 mm
Variation in optical density: OD

Accepted: yes / no
2a.2.1.3.6 Correspondence between AEC sensors

AEC sensor position Tube load Optical
(mAs) Density
OD
OD
OD
OD
OD
OD
Difference in Optical Density OD
AEC sensor position Position extra Tube load Chosen AEC

attenuation (mAs) sensor position
Accepted: yes / no
2a.2.1.4 Compression
2a.2.1.4.1 Compression force

Force indication: N
Measured compression force: N
Compression force after 1 min: N
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 99
2a.2.1.4.2 Compression plate alignment

Attachment compression plate: in order/out of order
Symmetric load
Thickness indication: cm
Height of compression plate above the bucky at full compression:
left right difference(l/r)

Rear : cm
Front : cm
Difference(r/f) cm
Accepted: yes / no
2a.2.2 Bucky and image receptor

2a.2.2.1 Anti scatter grid
2a.2.2.1.1 Grid system factor
Exposure Tube load Density
[mGy] [mAs] [OD]
Present:
Absent:
Grid system factor:
Accepted: yes / no
2a.2.2.1.2 Grid imaging

Additional grid images made:
# Added PMMA Description of artefacts
1. yes/no
2. yes/no
3. yes/no
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 100
2a.2.2.2 Screen-film
2a.2.2.2.1 Inter cassette sensitivity and attenuation variation and optical density range
AEC setting:
Cassette id Exposure Tube load Density

[mGy] [mAs] [OD]
1
2
3
4
5
6
7
8
9
10
11
12
Average values: mAs OD

Max. deviation: % mAs OD
Reference cassette:
Accepted: yes / no
2a.2.2.2.2 Screen-film contact

Cassette id: Description of artefacts:
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 101
2a.2.3 Film processing

2a.2.3.1 Baseline performance of the processor
2a.2.3.1.1 Temperature
Point of measurement in bath:
Developer Fixer
Reference/nominal:
Thermometer
Reference:
Local:
Console:
2a.2.3.1.2 Process time

Time from processor signal to film available: s
2a.2.3.2 Film and processor

2a.2.3.2.1 Sensitometry, daily performance, artefacts:
Forms should be made to suit the local preferences
2a.2.3.3 Darkroom
2a.2.3.3.1 Light leakage
Fog (after 2 min.) of a pre-exposed film on the workbench:
point: 1 2 3 4 5
D(fogged): OD
D(background): OD
Difference: OD
Average difference: OD
Accepted: yes / no
Positions of light sources and leaks in the darkroom:

-
-
2a.2.3.3.2 Safelights
Type of lighting: direct/indirect

Height : ± meter above workbench
Setting:
Filter condition : good / insufficient / absent / not checked
Fog (after 2 min.) of a pre-exposed film on the workbench:

point: 1 2 3 4 5
D(fogged) OD
D(background): OD
Difference: OD
Average difference: OD
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 102
2a.2.4 Viewing conditions

2a.2.4.1 Viewing box
2a.2.4.1.1 Viewing box luminance and 2a.2.4.1.2 Homogeneity
Viewing panel 1 2 3 4 5
Luminance (cd/m2)
Centre
Top left
Top right
Bottom left
Bottom right
Difference in
luminance
between the
centres (%)
Maximum deviation
in luminance compared
to the luminance in the
centre (%)
Accepted: yes / no
2a.2.4.2 Ambient light level

Reading from the illuminance meter (detector at the image plane, box is off): lux
Accepted: yes / no
2a.2.5 System properties

2a.2.5.1 Dosimetry
PMMA thickness (mm) Average glandular dose (mGy)
20
30
40
50
60
70
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 103
2a.2.5.2 Image quality

2a.2.5.2.1 Spatial resolution
Position of the centre of the pattern:
Height above the bucky surface: mm
Distance from thorax side of the bucky: mm
Distance from AC axis: mm
Resolution R⊥ AC-axis R//AC-axis
image 1
image 2
image 3
image 4
Accepted: yes / no
2a.2.5.2.2 Image contrast

image mAs #1 #2 #3 #4 #5 #6 #7 #8 #9 #10
1
2
3
4
5
Graph(s) attached
2a.2.5.2.3 Threshold contrast visibility

Observer # objects identified
1
Accepted: yes / no
Diameter disc Threshold contrast

(mm) (%)
0,1
0,2
0,5
1,0
2,0
H2 20-09-2005 20:54 Pagina 104
2a.2.5.2.4 Exposure time

AEC setting for a routine image:
Tube load obtained: mAs
Exposure time: s
Accepted: yes / no
H2 20-09-2005 20:54 Pagina 105

b
Digital mammography
Authors
R. van Engen
K. Young
H. Bosmans
M. Thijssen
H2 20-09-2005 20:54 Pagina 106
D I G I T A L M A M M O G R A P H Y
Authors:
R. van Engen, Nijmegen, the Netherlands
K. Young, Guildford, United Kingdom
H. Bosmans, Leuven, Belgium
M. Thijssen, Nijmegen, the Netherlands (project leader)
Co-authors:
R. Visser, Nijmegen, the Netherlands
L. Oostveen, Nijmegen, the Netherlands
T. Geertse, Nijmegen, the Netherlands
R. Bijkerk, Nijmegen, the Netherlands
P. Heid, Marseille, France
Contributors:
P. Baldelli, Ferrara, Italy A. Noel, Nancy, France
B. Beckers, Nijmegen, the Netherlands K. Pedersen, Oslo, Norway
A. Bloomquist, Toronto, Canada N. Phelan, Dublin, Ireland
M. Borowski, Bremen, Germany F. Rogge, Leuven, Belgium
AK. Carton, Leuven, Belgium M. Säbel, Erlangen, Germany
M. Chevalier, Madrid, Spain M. Schutten, Nijmegen, the Netherlands
M. Gambaccini, Ferrara, Italy F. Shannoun, Luxembourg, Luxembourg
S. von Gehlen, Oldenburg, Germany A. Stargardt, Aachen, Germany
G. Gennaro, Padua, Italy M. Swinkels, Nijmegen, the Netherlands
A. de Hauwere, Gent, Belgium A. Taibi, Ferrara, Italy
B. Johnson, Guildford, United Kingdom J. Teubner, Heidelberg, Germany
B. Lazzari, Florence, Italy H. Thierens, Gent, Belgium
A. Lisbona, Nantes, France P. Torbica, Innsbruck, Austria
A. Maidment, Philadelphia, USA F. van der Meer, Rotterdam, the Netherlands
N. Marshall, London, United Kingdom F.R. Verdun, Lausanne, Switzerland
G. Mawdsley, Toronto, Canada A. Workman, Belfast, United Kingdom
P. Moran, Madrid, Spain
Acknowledgements:
Comments have been received from the following manufacturers:
Agfa, Barco, Eizo, Fischer, Fuji, GE, Hologic, IMS, Instrumentarium, Kodak, Mevis, Sectra,
Siemens, X-counter.
Comments have been received from the following groups:

DIN working group mammography, IEC MT 31: mammography.
The authors acknowledge the valuable contributions of numerous scientists as

well as manufacturers in the preparation of this document.
Furthermore the financial support of the European Commission through the European Breast
Cancer Network contracts 2000 (SI2.307923(2000CVG2-031)), 2001 (SI2.328176(2001CVG2-
013)) and 2002 (SPC2002482) is acknowledged.
This protocol is written as part of two projects:

• the European Guidelines and Certification Project of the European Reference Organisation for
Quality Assured Breast Screening and Diagnostic Services (EUREF) and
• the Digital Mammography Project of the Leuven Universitary Center of Cancer Prevention
(LUCK).
Both projects are partners of the European Breast Cancer Network (EBCN).
The publication has been coordinated by the EUREF office in Nijmegen.
H2 20-09-2005 20:54 Pagina 107
D I G I T A L M A M M O G R A P H Y b
Foreword
The ‘European guidelines for quality assurance in mammography screening’ (European
Guidelines, 2001) include as chapter 3 the ‘European protocol for the quality control of the
physical and technical aspects of mammography screening’. In this protocol the requirements for
an adequate screen-film imaging system are defined. In recent years, the image detection
technology used in mammography has extended to include the use of digital detector systems.
This technology is different in so many ways, that it is necessary to set new quality standards and
test procedures specifically for digital systems.
This document is based on the addendum to chapter 3 of the European guidelines (3rd edition),
which was released in November 2003 (Addendum, 2003). The approach to quality assessment
and control in this protocol is comparable in the sense, that the measurement and evaluation of
performance are in principle independent of the type and brand of the system used. The
measurements are generally based on parameters that are extracted from the images that are
produced when a phantom with known physical properties is exposed under defined conditions.
The limiting values are based upon the quality that is achieved by screen-film systems, which
fulfil the demands of the European guidelines.
To fulfil the European guidelines in mammography screening, the digital x-ray system must pass
all relevant tests at the acceptable level. The achievable level reflects the state of the art for the
individual parameter.
This protocol for digital mammography is work-in-progress and subject to improvements as more
experience in digital mammography is obtained and new types of digital mammography
equipment are developed. Changes in measuring techniques or limiting values will lead to a new
version number, changes in wording or added comments will change the sub-number. Updates on
the current version will be made available on the EUREF website (www.euref.org). It is
recommended that users check the website for updates before testing digital mammography
equipment.
In the text some lines are printed in parentheses [like these]. This text is a remark.
Text in a box like this needs further evaluation.
H2 20-09-2005 20:54 Pagina 108
2b.1 Introduction to the measurements

To produce images with adequate quality, each part of the imaging chain must function within the
limits of performance given. Experience with some digital systems shows, that non-compliance
results in a seriously diminished information transfer to the observer. This can be expected to
result in a lower detection rate for microcalcifications and/or for low contrast lesions.
To facilitate the relevant quality control, the user must be able to evaluate the status of the
acquisition system including detector, the processing system and the display system (see fig. 1).
This protocol follows recommendations according to the DICOM standard (NEMA, Digital Imaging
and Communications in Medicine). The equipment therefore must be able to transmit and
receive digital mammogram images as IOD’s (Information Object Definitions) of modality ‘MG’
(mammography) or ‘CR’ (computed radiography), in compliance with parts 3 (IOD definition), 5
(data structures and encoding), 6 (data dictionary) and 7 (message exchange) of the DICOM
standard. Modality MG is preferred over modality CR (for example because MG includes exposure
parameters and the terms ‘for processing’ and ‘for presentation’ are used to distinguish
unprocessed and processed images).
The general principles for testing the three main parts of the imaging chain, illustrated in figure 1,
are discussed below.
Acquisition system including image receptor

The acquisition system (fig. 1, A) can be evaluated:
• By inspection of a recent ‘bad pixel map’. This map (either an image or a table) defines the
position of all pixels of which the pixel value is not based on its own del reading (see
2b.2.2.3.2). It must be accessible to the user at any time and be usable independently of a
given equipment and manufacturers permission.
• By the assessment of the relation between X-ray exposure parameters, dose to the image
receptor and pixel values. An ‘unprocessed image’ (DICOM defines such an image as ‘for
processing’), presenting a linear or other known mathematical relationship between del dose
and pixel value, must be accessible. This image type must also be available for CAD (computer
aided detection) or other processing software.
• By an indication of the nominal sensitivity setting of the system in every image.
Since image quality increases with dose, the preference for higher system dose can be
expected. This leads to a higher mean glandular dose and consequent higher radiation risk to
the women screened. In screen-film systems the dose to the image receptor is linked to the
mean optical density of each film, given the speed class of the system (speed class 100
corresponds roughly to an air kerma of 10 µGy at the place of the image receptor). An
indication, comparable to speed class, must be provided for digital systems to keep the
radiologist informed on the average doses delivered. It is recommended that manufacturers
provide sufficient information in the header of the file to allow calculation of the average
glandular dose for each individual patient. A working group of DICOM is drafting the definitions.
• For the evaluation of the acquisition system this protocol follows some draft procedures of the
American Association of Physicists in Medicine (AAPM) Task Group 10 (Samei, 2001) and of
preliminary results of the American College of Radiology Imaging Network Dmist trial (ACRIN Dmist).
Processing system
• In future the processing system (fig 1, B) may be evaluated by the inspection and scoring of a
test set of images (either mammograms or phantom images), which have been processed in
the available standard processing algorithm.
- These images are to be inserted by the user as ‘unprocessed images’ (DICOM: ‘for
processing’) and processed by the software of the manufacturer before displaying.
- The manufacturer must provide information in general terms on the processing applied.
• The processing algorithms are built to enhance the visibility of specific image details. At this
moment little experience and literature on the effects is available. These algorithms therefore
are not addressed in the present protocol. The observer is urged to convince himself of the
value of the algorithms provided.
H2 20-09-2005 20:54 Pagina 109
• Evaluation of processing algorithms and CAD (computer aided detection) will be addressed in
a future version of this protocol.
Display system
• The display system (fig 1, C) can be evaluated by the inspection on the display system (printer
or monitor) of synthetic test images, produced in DICOM format and independent of the
phantom images delivered by the manufacturer. The user must be able to insert these images as
‘processed images’ (DICOM: ‘for presentation’). They are not processed further before displaying.
Evaluation of such images is necessary to confirm compliance to quality standards other than
those of the manufacturer. It must be possible to load and display these phantom images
using the imaging system under evaluation.
• For the evaluation of the display system this protocol follows the advice of AAPM Task Group
18 (Samei, 2004) and of preliminary results of the ACRIN Dmist trial.
The measurements in the protocol are in principle chosen and described to be generally
applicable. Where the tests are similar to those required for screen-film mammography, a
reference to the relevant part of the European guidelines is given. When necessary, different test
procedures are given for CR (computed radiology, i.e. photo-stimulable phosphor type) systems
and DR (direct radiology, i.e. solid state type, including scanning slot) systems separately.
Many measurements are performed by an exposure of a test object. All measurements are
performed under normal working conditions: no special adjustment of the equipment is
necessary. Since the available settings in the different systems vary in spectrum and X-ray
quantity for the different breast thicknesses, no common standard exposure can be indicated.
Therefore dose calculations for the comparison of systems are based on the AGD (average
glandular dose) to the breast (or simulated breast) rather than on entrance surface air kerma.
To evaluate the clinical use of a system, a standard type of exposure is specified: the routine
exposure, which is intended to provide information on the system under clinical settings.
For the production of the routine exposure, a test object is exposed using the machine settings
as follows (unless stated otherwise):
Routine exposure:
test object thickness: 45 mm
test object material: PMMA
tube voltage: as used clinically
target material: as used clinically
filter material: as used clinically
compression device: in contact with test object
anti scatter grid: as used clinically
source-to-image distance: as used clinically
photo timer detector (for CR): in position closest to chest wall
automatic exposure control: as used clinically
exposure control step: as used clinically
exposure-to-read-time (for CR): 1 minute7
image processing: off
H2 20-09-2005 20:54 Pagina 110
Mean pixel values and their standard deviation are measured in a standard region of interest
(ROI), which has an area of 4 cm2 and is positioned 60 mm from the chest wall side and laterally
centred.
Limits of acceptable performance for image quality and dose are based on the limits of
acceptable performance of screen-film mammography systems. The relation between dose and
limits of visibility of details for a certain contrast are based on the performance of a large number
of screen-film systems in the UK, the Netherlands, Germany, Belgium and France. These
acceptable limits are given, but often a better result is achievable. When applicable the
achievable values are also given. Both the acceptable and achievable values are summarised in
Appendix 7. Occasionally no limiting value is given, but only a typical value as an indication of
what may normally be expected. The measurement frequencies indicated in the protocol
(appendix 6) are the minimum required. When the acceptable limiting value is exceeded the
measurement should be repeated. If necessary, additional measurements should be performed
to determine the origin of the observed problem and appropriate actions that should be taken to
solve the problem.
For some tests the limiting values are provisional, this means that the limiting value needs
further evaluation and may be changed in the future. Check the EUREF website for updates. In
some cases further remarks about the limiting values can be found in a box.
Guidance on the specific design and operating criteria of suitable test objects will be given by a
separate project group of the European Breast Cancer Network (EBCN). Definition of terms, such
as the reference ROI and signal-to-noise-ratio are given in section 2b.1.5. The evaluation of the
results of the QC measurements can be simplified by using the forms for QC reporting that are
provided on the EUREF homepage (www.euref.org).
2b.1.1 Staff and equipment

The local staff can perform several measurements. The more elaborate measurements should
be undertaken by medical physicists who are trained and experienced in diagnostic radiology and
specifically trained in mammography QC. Comparability and consistency of the results from
different centres is best achieved if data from all measurements, including those performed by
local technicians or radiographers, are collected and analysed centrally.
The staff conducting the daily/weekly QC-tests will need the following equipment 8 at the
screening site:
• Standard test block9 (45 mm PMMA10) • Digital QC test images

• Reference cassette (CR systems) • PMMA plates11
The medical physics staff conducting the other QC-tests will need the following additional
equipment and may need duplicates of some of the above3:
• Dose meter • Aluminium sheets

• Tube voltage meter • Focal spot test device + stand
• Exposure time meter • Screen-film contact test device
• Telescopic luminance meter • Tape ruler
• Illuminance meter • Compression force test device
• QC test objects • Rubber foam
• Digital QC test images • Lead sheet
• Contrast-detail test object • Expanded polymer spacers
• Densitometer
H2 20-09-2005 20:54 Pagina 111
2b.1.2 System demands

Accessibility
It must be possible to access and insert DICOM images as ‘for processing’ and ‘for
presentation’ to allow evaluation of the image receptor, image processing and image
presentation separately.
AEC
The As Low As Reasonably Achievable (ALARA) principle on dose administered to the patient
necessitates the use of an automated exposure control (AEC) system to ensure the optimal
exposure of the image receptor compensating for breast thickness and composition. The use of
a look up table, only based on the measured thickness of the compressed breast, increases the
mean dose to the patients. This is due to the necessary margin in exposure to avoid increased
noise by underexposure in dense breasts and to compensate for the incorrect reading of the
thickness.
Image receptor
The required physical size of the image receptor and the amount of missed tissue at the short
sides and especially at the chest wall side are important for an optimal imaging of the breast
tissues. An upper limit is given for the amount of missed tissue at chest wall side, but the
acceptance of other margins remains the responsibility of the radiologist.
Display system
Optimal transfer of the information in digital mammograms will be reached, when every pixel in
the matrix is projected to at least one pixel on the display system and when the pixel size on the
display system is sufficiently small to show details that coincide with the maximum sensitivity of
the eye of the observer (1-3 lp/mm at a viewing distance of 30 cm). In screening the monitor
should allow for the inspection of the image at full size in full resolution, since the number of
images read does not allow time consuming procedures like roaming or zooming. Normally two
images are viewed at the same time, and with the current technology it is therefore
recommended that diagnostic workstations with two large (45-50 cm diagonal (19-21II)) high
quality, 5 megapixel monitors are used.
On the acquisition unit it may be acceptable to use a monitor with lower specification, depending
on the tasks of the radiographer.
Further research is needed to demonstrate whether cheaper solutions (e.g. 3 megapixel
monitors) may be sufficient in clinical situations.
Viewing conditions
Since the maximum intensity on the monitor (300-800 cd/m2) is much lower than that of a
viewing box with unexposed and developed film (3000-6000 cd/m2) and due to the reflection
characteristics of the monitor, the amount of ambient light might seriously diminish the visible
dynamic range and the visibility of low contrast lesions. The ambient light level therefore should
be low (less than 10 lux) to allow maximal extension to the lower part of the range. Although this
level has proven to be acceptable, a short time to adapt to this level might be necessary.
Computed Radiography (CR) system

Measurements should be performed with the same phosphor screen to rule out differences
between screens except when testing individual screens as in section 2b.2.2.4 and when testing
contrast threshold visibility as in section 2b.2.4.1. The exposure-to-read-time is standardised to
minimize differences caused by varying time delays (i.e. fading of the latent image).
The DICOM standard allows both IOD’s of ‘CR’ and ‘MG’ to be used for CR images. This may lead
to improper hanging of the images by different display systems.
H2 20-09-2005 20:54 Pagina 112
Direct Radiography (DR) detector

When measurements are performed for which no image is required (e.g. HVL or tube voltage), the
detector should be covered sufficiently to prevent ghost images appearing on subsequent use of
the system.
When the absorbers in the QC test object lead to automatic exposure values other than those
that would be obtained with homogeneous PMMA, set the system manually to these values.
Printer
The pixel size of the printer should be in the same order of magnitude as (or less than) the pixel
size of the image and should be < 100 micron.
2b.1.3 Order of the measurements

It is advisable to perform measurements such as homogeneity, NPS, linearity, MTF first and
ghosting last to prevent the influence of possible ghost images. After the ghosting measurement
it is advised to make some additional images with a homogeneous block of PMMA covering the
whole detector to make sure that ghosts do not appear on clinical images.
Fig.1
Fig.1A
H2 20-09-2005 20:54 Pagina 113
Fig.1C (monitor)
Fig.1C (printer)
2b.1.4 Philosophy
Introduction
The primary scope of this document is setting standards for mammography screening, however
similar standards are expected for diagnostic mammography.
The imaging chain in digital mammography can be divided into three independent parts:
1. Image acquisition, which includes the X-ray generation, the image receptor and (for some
systems) image receptor corrections.
2. Image processing, which includes the image processing software.
3. Image presentation, including monitor, image presentation software, printer and viewing box.
In the European protocol for the quality control of the physical and technical aspects of
mammography screening these parts of the imaging chain are evaluated separately. This is a
practical approach because each requires very different evaluation techniques and it allows the
use of equipment and software from different vendors. In the present version of the protocol
(version 4) only image acquisition and image presentation are considered. Due to the large
number of processing techniques and the shortcomings of classical test objects with regard to
the evaluation of post processing as histogram and texture based processing, evaluation of the
image processing part of the imaging chain has not been addressed (yet). However, manufacturers
have to specify in general terms, which image processing techniques are applied and it is
advised to evaluate image processing by comparing mammograms to images from the previous
screening round by experienced readers.
The digital section of this chapter of the European guidelines should not be considered as a
guide for the optimal working point of a particular system or as a guide to optimise image quality.
Different research groups are studying these issues and manufacturers are still working on the
optimizing of current systems and the development of new techniques. We urge the reader of this
document to keep track of all new developments in this rapidly changing technology. Updates of
this protocol will be available at www.euref.org.
H2 20-09-2005 20:54 Pagina 114
2b.1.4.1 Methods of testing

The tests as described in the present text on image acquisition are based on the expertise of the
different European groups in digital mammography, the American College of Radiology Imaging
Network Dmist trial (ACRIN Dmist ) QC protocol (and experiences with this protocol which were
shared generously by the QC team of the trial), manufacturers QC tests and the publication from
American Association of Physicists in Medicine (AAPM) Task Group 10 concerning CR systems.
The tests in the image presentation section are based on the testing methods and test images
of AAPM Task Group 18. This includes conformation to the DICOM standard for presentations.
Before publication the test methods have been evaluated using a number of different types of
digital systems. For some types of systems only a small number of evaluation tests have been
performed due to limited accessibility. Due to the rapidly changing technologies, new methods of
testing may be necessary in the future. Check for updates on the EUREF website.
2b.1.4.2 Limiting values

Limiting values have been derived as much as possible from practice using screen-film
mammography: it is assumed to be a requirement that digital mammography should perform at
least as well as screen-film mammography.
For some test items the limiting values need more evaluation. In these cases, the limiting values
have been made provisional. For some requirements, we depend on the provision of additional
features by the manufacturers. In these cases, a date is given by which the items should be
made available.
To remain up-to-date with the latest insights, the protocol will be updated continuously. Latest
versions will be made available on the EUREF website (www.euref.org).
The philosophy of important QC tests and remarks are explained in the following paragraphs.
2b.1.4.3 Image acquisition

The X-ray generation part of the protocol is essentially identical to that of screen-film
mammography. Therefore it will not be discussed in this section.
Automatic exposure control

Some digital mammography equipment on the market, are still ‘in development’. One of the
features not yet incorporated in some systems, is an automatic exposure control device. This
has a number of disadvantages:
1. In the case of completely manual settings, mistakes in exposure settings may occur and lead
to under- or overexposure and leading to insufficient image quality or unnecessary patient
dose. Contrary to screen-film mammography, in which underexposure is immediately
recognized from a change in the optical density of the film, underexposure in digital
mammography is not easily recognized by the radiographer or the radiologist. This may lead
to insufficient image quality.
2. The system might not be able to handle the high throughput necessary in mammography
screening.
3. Due to the unknown breast content, exposure factors must be tuned to dense breasts to
guarantee a sufficiently high image quality. This leads, however, to unnecessarily high
exposures for other women and does not comply with the As Low As Reasonably Achievable
(ALARA) principle. Some manufacturers try to compensate for this by providing exposure
tables for several types of breast composition. However, it is not always clear how these
tables were set-up and how the categories of breast content are defined or anticipated. The
problem comes down to the user who has to choose the right exposure table. This is difficult
since breast content may not be known until the breast is imaged.
Therefore, the authors have stated that systems used for mammography screening should
incorporate an AEC. Manufacturers of equipment without Automatic Exposure Control (AEC), are
H2 20-09-2005 20:54 Pagina 115
urged to implement such a device in their systems before January 2006. For the time being
systems which incorporate an exposure table in the software that account only for compressed
breast thickness, will be allowed. We advise against systems in which both spectrum and dose
have to be set completely manually.
2b.1.4.4 Image quality evaluation

Image quality is evaluated in terms of threshold contrast visibility at a standard simulated breast
thickness. This provides a measure of image quality for an average breast. As this test is rather
time consuming, the evaluation is restricted to this standard thickness. The image quality of
other thicknesses is related to the image quality at the standard thickness using simpler
parameters, which describe the image quality relative to the image quality at the standard
thickness.
2b.1.4.4.1 Image quality at standard thickness

Image quality is expressed in terms of threshold contrast visibility using clinical exposure
settings. This allows evaluation of the image quality of a digital image receptor in relationship to
the spectrum and dose levels, which are used clinically on that particular system.
For the evaluation of threshold contrast visibility ‘unprocessed’ phantom images must be used.
In this way, only the image acquisition part of the system is included and the image quality
evaluation cannot be considered as a ‘whole-system’ test.
It is acknowledged that it is not possible to get ‘unprocessed’ images from all systems yet. For
these systems threshold contrast visibility has to be determined on images with the least
possible image processing. This processing may introduce artefacts due to histogram or texture
based processing techniques. Therefore care needs to be taken in interpretation of these
processed contrast-detail (CD) images.
To increase reliability at least six phantom images are required. To reduce the (in-) visibility of
small disks due to the accidental relative position of the disks and the dels of the detector the
phantom has to be repositioned slightly after acquisition of each image. Extensive window
levelling and zooming must be performed to optimize the visibility of the dots in each section of
the phantom image. This prevents the monitor from being the limiting factor for the threshold
contrast evaluation instead of the quality of the unprocessed image. At least three readers
should score two different images each.
A problem with scoring CD images is the inter- and intra-reader variability. Therefore CDMAM
images with scores are available on the EUREF website for reference purposes. In future, the
threshold contrast visibility test may be performed using computer readout of the phantom
images. This will solve problems with inter- and intra-observer variability Allowance may need to
be made for differences between human and machine measurements of threshold contrast.
At this moment image quality is evaluated using a total attenuation equivalent to 50 mm PMMA
thickness. This has been chosen because image quality information was available for this
thickness. In the future, the image quality evaluation may be performed at the thickness of 45
mm PMMA, which has been chosen as the standard thickness for other tests in the European
Guidelines.
Two kinds of limiting values have been set: acceptable and achievable limiting values. The
acceptable limiting values have been derived from screen-film mammography, the achievable
limiting values have been derived from current full-field digital mammography systems.
The acceptable limiting values have been derived by stating that image quality of digital
mammography must be (at least) comparable to screen-film mammography (Young, 2004). For
this purpose the image quality of a large number of screen-film mammography systems in
different screening programmes has been determined using CD analysis. The CDMAM phantom
has been used for these measurements. It was chosen that the image quality limiting values for
digital mammography should be such that 97.5% of the screen-film systems in the UK would
comply. This means that the image quality limits are not very demanding and it must be realized
that a system just complying to the acceptable limits would probably not be considered
equivalent to top quality screen-film systems.
H2 20-09-2005 20:54 Pagina 116
The resulting limiting values have been checked with the image quality levels found in the Dutch
screening and in some of the German screening projects (of which data was available) and were
found to be realistic.
Furthermore the limiting values have been checked with the CD curves from some hospitals in
which it was established (by radiologists) that image quality of the digital system was too low for
mammography. (The visibility of microcalcifications was regarded insufficient). Threshold
contrast visibility for small diameters did not meet the acceptable limiting values in these
hospitals. The image quality of a system is only acceptable if contrast threshold values for all
diameters comply with the limiting values.
The achievable limiting values have been derived as averages from a number of established
digital mammography systems. At the EUREF website CDMAM images with scores can be found
for reference purposes.
2b.1.4.4.2 Image quality at other PMMA thicknesses

In version 1.0 of the protocol for digital mammography image quality at thicknesses other than
standard thickness is related to the image quality at the standard thickness using Signal-to-
Noise Ratio (SNR) and Contrast-to-Noise Ratio (CNR) requirements. The absolute values of SNR
and CNR are system dependent (they are dependent on for example pixel size), therefore limiting
values need to be expressed in terms of variation in SNR over the whole range of simulated
breast thicknesses and percentage of CNR at standard thickness respectively.
However there are difficulties with this measurement. At this moment three kinds of parameters
are used by manufacturers to control image quality in AEC systems: dose to the detector (pixel
value), SNR and CNR.
Screen-film mammography systems and some digital systems keep dose to the AEC detector
constant over the whole range of breast thicknesses (for digital systems this means that pixel
value is kept constant), some other systems keep SNR constant and recently a system has been
introduced which tries to keep CNR constant.
In the view of the authors CNR would be the right measure to quantify image quality at
thicknesses other than standard thickness. CNR should not necessarily be equal across the
whole range of breast thicknesses. However, problems arise when setting the CNR value at
standard thickness as reference for other thicknesses using the method described in version 1.0
of the Protocol for digital mammography. If the CNR at standard thickness is high, CNR at other
thicknesses may fail, not because image quality is too low, but because image quality at
standard thickness is relatively high. So the method of testing and limiting values needed to be
revised.
In this fourth edition of the guidelines the value of CNR at standard thickness is estimated which
would be obtained on a system if this system just complied with the acceptable limiting values of
threshold contrast visibility. In the calculation of this minimum CNR level it is assumed that
quantum noise is the main source of noise in the system. The calculation is based on the Rose
theory, from which can be derived that threshold contrast visibility is inversely related to CNR.
The calculated CNR at the acceptable limiting value of threshold contrast is the lower limit of
CNR at standard thickness. Lower limits of CNR at other thickness are related to this value
providing sufficient CNR for the whole range of breast thickness.
2b.1.4.5 Glandular dose

It is assumed that average glandular dose levels in digital mammography systems should be no
more than for screen-film systems. To ensure this, the limiting dose values have been changed
compared to the third edition of the European guidelines for quality assurance in mammography
screening in three aspects:
In the present version of the protocol the clinical spectrum is used for dose measurements
instead of a standard spectrum, the dose limits have been made independent of optical density
and a limiting dose value per PMMA thickness is introduced. The reasons for these changes will
be explained in the next paragraphs.
H2 20-09-2005 20:54 Pagina 117
In the third edition of the protocol for screen-film systems, limiting dose values were measured
using a standard spectrum. This requirement of the third edition cannot be fulfilled by some
digital mammography systems due to the available spectra. For example: scanning slot systems
use tungsten instead of molybdenum targets due to the required tube loading. Furthermore using
clinical spectra in dose measurements is closer to clinical practice.
In the third edition, Entrance Surface Air Kerma (ESAK) limits at standard thickness have to be
measured for a given optical density. Practical ESAK values are found to be far below the limiting
value, even at the clinically used optical density. In digital mammography the link between
limiting dose values and OD is non-existent. Therefore a choice had to be made what limiting
dose value would be appropriate for digital mammography. In the view of the authors, inspired by
the ALARA principle, dose should not increase substantially when changing to digital
mammography. Data from the Dutch (Beckers 2003), Swedish (Leitz 2001), Norwegian (Pedersen
2000) and UK (NHSBSP 2000, 2003) screening programmes show that average glandular dose
levels in screen-film mammography systems are between 0.8 and 2.5 mGy for 4.5 cm PMMA in
clinical settings (corrected for difference in standard PMMA thickness in the UK and the
Netherlands). Therefore an average glandular dose limit of 2.5 mGy at standard thickness in
clinical settings has been chosen to ensure that dose levels in digital mammography will not
exceed those of screen-film mammography. This limiting value is comparable to the objective of
the NHSBSP in the UK to have average glandular dose levels of 2 mGy or less (for 4.0 cm PMMA)
and the limiting average glandular dose value for the Dutch screening programme (3 mGy for 5.0
cm PMMA). In the present version of the protocol limiting dose values for a range of PMMA
thickness have been introduced. This has been done because in some non-AEC systems it was
noticed that manufacturers decreased dose at standard thicknesses to comply with the limiting
value at standard thickness while dose levels at other thickness were found to be (much) higher
than found in screen-film mammography. Besides this it has been found that some systems did
use much lower tube voltages than in screen-film mammography (thus increasing patient dose
substantially). In measurements performed by some of the authors, these very low values proved
unnecessary for image quality, therefore the use of these tube voltages does not comply with the
ALARA principle. Setting limiting dose levels per PMMA thickness prevents this situation. The
limiting values for PMMA thicknesses other than standard thickness have been obtained by
averaging all measured glandular dose levels per PMMA thickness from all X-ray units of the
Dutch screening programme and some German screening trials. The resulting average glandular
dose against PMMA thickness curve has been scaled to the limiting value at standard thickness.
The results have been compared with the dose values per PMMA thickness found in the UK and
some of the German screening projects (screen-film mammography). The limiting values were
found to be reasonable.
2b.1.4.6 Exposure time

The exposure time should be sufficiently short to avoid motion unsharpness. For scanning slot
systems a distinction has to be made between the time in which each individual part of the
breast is exposed and the total scanning time. The first is important for motion unsharpness, the
latter for the time during which the breast of a woman is compressed.
For most systems exposure time increases rapidly with breast thickness and content. Depending
on the screen-film combination and the clinically used spectra this range may vary from 0.2 to 3
seconds. For some scanning slot systems however, scanning time and exposure time are fairly
constant for the whole range of breast thickness and content. Due to this design, these systems
may not comply with the limiting value of 2 seconds at standard thickness. Ideally exposure time
should be below a certain limiting value even for very thick and dense breasts, so the limiting
value at standard thickness may not be the right measure to prevent motion unsharpness for all
breasts. Because this worst case liming value has not been determined yet, the value of 2
seconds at standard thickness is maintained, with the exception that scanning slot systems for
which exposure time is only slightly dependent on breast thickness and content do not have to
comply. For these systems clinical results will have to show that motion unsharpness is not a
problem.
H2 20-09-2005 20:54 Pagina 118
2b.1.4.7 Image receptor

2b.1.4.7.1 Response function
The response function should comply with a specification. The response function may be linear
or logarithmic or fulfil some other mathematical relationship. The response function should be
monotonous increasing (or decreasing). In some systems manufacturers certain value is added
to the pixel value of all pixels to prevent negative values. When calculating SNR this offset must
be taken into account. The response function of current CR systems is not linear but may be
logarithmic. For these systems the response function needs to be linearised before SNR and
CNR calculations are performed.
2b.1.4.7.2 Noise evaluation

Noise is evaluated by plotting SNR squared against entrance surface air kerma (ESAK) for
systems with a linear response (such as current DR systems) and Standard Deviation squared
against one over ESAK for systems with a logarithmic response (such as current CR systems).
The non-linearity and the offset of the curve are indications of the presence of additional noise.
At acceptance a reference curve is measured. At subsequent QC tests the results should be
compared to the reference curve.
2b.1.4.7.3 Missed tissue at chest wall side

The limiting value on the amount of missed tissue at chest wall side is based on characteristic
values found in screen-film mammography systems. In November 2003 (date of publication of
version 1.0 of the protocol for digital mammography), some specific designs of digital
mammography system did not comply with the limiting value of 5 mm. Because it is stated that
digital mammography should be at least equal to screen-film mammography, the manufacturers
of systems, that do not comply with the limiting value, have been urged to reduce the amount of
missed tissue at chest wall side for their system(s). A number of responding manufacturers have
stated that they will comply.
2b.1.4.7.4 Detector element failure (DR)

It is very important to check the number and position of defective detector elements (dels). At
this moment manufacturers are reluctant to provide this kind of information to users, but buyers
of equipment have the right to know the extent to which the images on their systems are
reconstructed. Therefore this information should be made available to the user.
It is demanded that a bad pixel map (either an image or a table with the position of all pixels of
which the pixel value is not based on its own del reading) is incorporated and that this map is
accessible to the user at any time and in such a format that it can be used independent of the
equipment of the manufacturer.
Limiting values on detector element failure should firstly (and most importantly) be based on
clinical relevance. At this moment there is not much information available on this subject. It is
expected that the loss of individual microcalcifications will not influence diagnostic decisions, so
(reconstructed) individual defective dels can be allowed. If a large number of dels are defective
within a certain area, this might influence diagnostic decisions. The difficulty is where to draw the
line.
Secondly, the correction algorithm, which is used on a particular X-ray unit, must be considered.
If an algorithm cannot handle the reconstruction of certain defective del values, this might lead
to unwanted artefacts on the image, even if the area is sufficiently small not to influence
diagnostics. For both reasons it is currently advisable to refer to the specifications of the
manufacturer for the number of defective detector elements, which can be allowed on a
particular detector.
2b.1.4.7.5 Image receptor homogeneity

For DR detectors, detector corrections are applied. In this correction the pixel value of defective
dels is reconstructed from the readout of neighbouring dels and corrections for differences in
electronic gain of the read-out and individual detector element sensitivity variations are
H2 20-09-2005 20:54 Pagina 119
performed. For some systems, the latter correction has to be performed by the user. If the user
has to perform this calibration, sufficient time must have passed since the last images were
made to prevent the influence of possible ghost images on the calibration. These corrections can
be checked in a homogeneity test.
Strictly speaking the correction for differences in sensitivity is only valid at the spectrum and
simulated breast thickness at which calibration is performed. Therefore it is advised to perform
the homogeneity test at several clinically relevant spectra and thickness at acceptance. At
acceptance, a baseline is established for the homogeneity test. For some types of DR detectors
homogeneity changes relatively quickly over time. Therefore it is advised to check image
homogeneity regularly (weekly) and check the results with the baseline. The frequency may
change in future and may be dependent on the type of digital system. For CR systems the
usefulness of the homogeneity test needs to be established. For the moment it is also
recommended to perform this test on CR systems.
Problems may occur if the Heel effect and geometric effects are relatively large. These effects
might influence the results of the image receptor homogeneity measurement. If a specific
system does not comply with the provisional limiting values it is advised to check whether
geometry or the Heel effect causes this deviation or any malfunction in the system. It is
recommended that the images are checked visually for artefacts.
2b.1.4.7.6 Fading of latent image (CR)

At acceptance it is advised that the fading of the latent image on the phosphor screens is
measured. With the results of this test the importance of using the same exposure-to-processing
time (in clinical practice and during quality control tests) can be determined.
2b.1.4.7.7 Ghosting
Several reports on ghosting in DR systems have been published (for example: Siewerdsen,
1999). In CR systems ghosts may occur if the erasure of the screens is not performed optimally.
This ghosting is quantified by comparing the pixel value of an induced ghost image to a known
contrast in the image (contrast of an aluminium sheet). After the ghosting measurement it is
advised to make some additional images with a homogeneous block of PMMA covering the whole
detector to make sure that ghosts will not appear on clinical images.
For scanning slot systems ghosts will not show with the proposed method of testing, but any
ghosting is included in MTF measurements.
2b.1.4.8 Image presentation

The whole image presentation section of this protocol is based on the work of AAPM Task Group
18. Only measurements which differ from the recommendations of Task Group 18 and limiting
values for which systems do not comply are mentioned below.
2b.1.4.8.1 Monitors
2b.1.4.8.1.1 Ambient light

AAPM Task Group 18 does not have specific limiting values on ambient light. In fact maximum
ambient light levels are dependent on the minimum luminance and reflection characteristics of
the monitor. For reasons of simplicity a single ambient light limiting value has been set.
2b.1.4.8.1.2 Grayscale display function

The grayscale display function of the monitor is checked against the DICOM Greyscale Standard
Display Function (GSDF). It is noticed that a number of display systems do not comply to the
GSDF. Manufacturers are urged to comply with this part of the DICOM standard. Test image
TG18-QC seems to be a good and quick (daily) test for the display on the monitor.
H2 20-09-2005 20:54 Pagina 120
2b.1.4.3.2 Printers
2b.1.4.3.2.1 Greyscale display function

The suggestions for QC made by AAPM Task Group 18 have been adapted slightly. Task Group 18
measurements are based on measuring luminances of a printed test pattern on a viewing box.
These measurements should be performed for all combinations of printers and viewing boxes.
From a quality control point of view this is impractical, therefore a standard viewing box has been
defined (luminance of the viewing box without film: 4000 cd/m2, luminance contribution due to
ambient illuminance reflecting of the printout: 1 cd/m2. The optical densities of the test pattern
should be such that the printout in combination with this virtual viewing box would comply with
the GSDF. The luminance of the viewing boxes is controlled by the tests described in the screen-
film section of the European guidelines.
2b.1.4.3.2.2 Pixel size

To be able to print images with sufficient resolution, the pixel size of the printer should be in the
same order of magnitude as (or less than) the pixel size of the image and should always be
< 100 micron.
2b.1.5 Definition of terms

The definitions given here specify the meaning of the terms used in this document.
Active display area The part of the display used for displaying images, applications
and the desktop.
Bad pixel map A map (either an image or a table) which defines the position of all
pixels of which the pixel value is not based on its own del reading.
Bit-depth Number of values which can be assigned to a pixel in a certain

digital system, expressed in bits.
Computer Aided Detection Software to aid the radiologists’ detection of suspect areas in the
(CAD) breast image.
Computed Radiography Digital radiology technology using photostimulable phosphor plates.

(CR)
Contrast to Noise Ratio The CNR is calculated as follows for a specific test object (e.g. 0.2
(CNR) mm Al thickness on 45 mm PMMA).
mean pixel value (signal) - mean pixel value (background)

CNR =
Standarddeviation
Standard (signal)2 2++SStandard
deviation(signal) tan dard deviation
deviation((background)
2
background ) 2
2
Del Discrete element in a DR detector.
Detective Quantum Function which describes the transfer of SNR as function of spatial
Efficiency (DQE) frequency when recording an X-ray image. The DQE gives the
efficiency with which the device uses the available quanta.
Detector corrections Correction in DR systems whereby the pixel value of defective

detector elements are reconstructed and pixel values are
corrected for individual detector element sensitivity variations and
electronic gain of the read-out.
H2 20-09-2005 20:54 Pagina 121
Direct Radiography (DR) Digital radiology technology using sealed units mounted on a
radiography system, which captures X-rays and produces a digital
image by sampling the X-ray image.
Digital Driving Level (DDL) Digital value which is the input for a display system.
Exposure indicator Number ascribed to an image related to the exposure.
Exposure time The time between the first and last moment that primary X-rays
reach an individual part of an imaged object.
Ghost image Residuals of previous images visible on the current image.
Modulation Transfer Function, which describes how the contrast of image components
Function (MTF) is transmitted as a function of their spatial frequency content.
Noise Fluctuations in pixel values which are unrelated to the imaged

object. The standard deviation in a ROI in the output image is
taken as measure of noise.
Noise power spectrum Function which describes image noise as a function of spatial
(NPS) frequency.
P-value See presentation value.
Pixel Picture element, the smallest unit in the image.
Pixel pitch Physical distance between the centres of adjacent pixels. In the
DICOM tags pixel pitch is called imager pixel spacing and is
generally equal to detector element spacing.
Pixel value Discrete value assigned to a pixel, in mammography systems the

number of pixel values range from 1024 (10-bits) to 16384 (14
bits), depending on the detector.
Pixel value offset Constant value that is added to the values of all pixels.
Presentation value Pixel value after Value Of Interest Look-Up-Table (VOI LUT) or
window width and window level settings have been applied.
Primary class display A display device used for the interpretation of medical images
device (also referred to in the text as ‘diagnostic display device’).
Processed image The image after image processing, ready for presentation on the
monitor or print-out. In the DICOM file the value of tag Pixel
Intensity Relationship (0028,1040) is ‘for presentation’.
Raw image See unprocessed image.
Reference region-of- The region-of-interest (≈ 4 cm2, either circular or square) in which

interest (ROI) mean pixel values and standard deviation are measured. The
centre of the region-of-interest is positioned 60 mm perpendicular
to the chest wall edge of the table and centred laterally.
Secondary class A display device used for viewing the images, but not for
display device diagnosis.
H2 20-09-2005 20:54 Pagina 122
(Nominal) sensitivity Indication of the sensitivity setting of the system, comparable to

setting the speed class in screen-film systems. The practical method to
implement a (nominal) sensitivity setting will be discussed with
manufacturers.
Screen processing Image processing applied in a CR system during read-out of the

imaging plate.
Signal to Noise Ratio The SNR is calculated as follows for a specific ROI:
(SNR)
mean pixel value - pixel value offset
SNR =
standard deviation in pixel value
Standard test block PMMA test object to represent approximately the average breast
(although not an exact tissue-substitute) so that the X-ray machine
operates correctly under automatic exposure control and the dose
meter readings may be converted into dose to glandular tissue.
The thickness is 45 ± 0.5 mm. The standard test block covers the
whole detector.
Threshold contrast The smallest detectable contrast for a given detail size that can be
shown by the imaging system with different intensity (density) over
the whole dynamic range. The threshold contrast is a measure for
imaging of low-contrast structures.
Uncorrected image The image in a DR system before any image processing, including
detector corrections and flat-fielding, is performed.
Unprocessed image The image of a DR system after flat-fielding and detector

corrections but before other image processing has been applied.
In the unprocessed image the pixel value is in general linear to
pixel exposure. In the DICOM file the value of tag Pixel Intensity
Relationship (0028,1040) is ‘for processing’. International
Electrotechnical Commission (IEC) Maintenance Team (MT) 31
refers to the unprocessed image as ‘raw data’.
maximum value - minmum value

Variation Variation = x 100%
mean value
VOI LUT Value of interest lookup table, defines the (non-linear)

transformation of pixel values into values meaningful for
presentation (presentation values).
Window centre Setting defining (together with window width) a linear relationship
between modality pixel values and pixel values meaningful for
Window width Setting defining (together with window centre) a linear relationship
between modality pixel values and pixel values meaningful for
H2 20-09-2005 20:54 Pagina 123
2b.2 Image acquisition

2b.2.1 X-ray generation
2b.2.1.1 X-ray source
The measurements to determine the focal spot size, source-to-image distance, alignment of X-ray
field and image receptor, radiation leakage and tube output are described in this section.
To prevent ghosting artefacts, it is advised to cover the detector with a lead sheet during
all tests for which no image is required and use the non-imaging mode (if available) on the
X-ray unit.
2b.2.1.1.1 Focal spot size

Use the methods and limiting values described in section 2a.2.1.1.1 of the screen-film part of
the European guidelines. Either film or the digital detector may be used, but beware of detector
saturation.
2b.2.1.1.2 Source-to-image distance

Use the method and limiting values described in section 2a.2.1.1.2 of the screen-film part of the
European guidelines. The distance on the digital images may be obtained by multiplying distance
in number of pixels with the pixel pitch.
2b.2.1.1.3 Alignment of X-ray field/image area

For CR systems use the method and limiting values described in section 2a.2.1.1.3 of the
screen-film part of the European guidelines. (Currently the most convenient method for DR
systems is with screen-film cassettes or CR cassettes. In future these facilities might not be
available. If cassettes and film processor are unavailable at the test site, use cassettes that can
be read-out or processed elsewhere or use self developing such as Polaroid Type 57 or
Gafchromic XR Type T film12).
2b.2.1.1.4 Radiation leakage

For CR systems use the method and limiting values described in section 2a.2.1.1.4 of the
screen-film part of the European guidelines. (Currently the most convenient method for DR
systems is with screen-film or CR cassettes. In future this might be a problem. If cassettes and
film processor are unavailable at the test site, use cassettes that can be read-out or processed
elsewhere or use self developing such as Polaroid Type 57 or Gafchromic XR Type T film12).
2b.2.1.1.5 Tube output

Use the measurement method described in section 2a.2.1.1.5 of the screen-film part of the
European guidelines. Tube output measurements should be performed at all clinically used
target-filter combinations for dose calculations (if necessary). Measurements should be
performed with compression paddle in place. To calculate the transmission factor of the
compression paddle, which may be needed for glandular dose estimates, tube output
measurements should also be performed without compression paddle. The transmission factor
should be calculated as the measured air kerma in presence of the compression paddle, divided
by the measured air kerma in absence of the compression paddle.
H2 20-09-2005 20:54 Pagina 124
2b.2.1.2 Tube voltage and beam quality

The beam quality of the emitted X-ray beam is determined by tube voltage, anode material and
filtration. Tube voltage and beam quality can be assessed by the measurements described
below.
2b.2.1.2.1 Tube voltage

Both the accuracy and reproducibility of the tube voltage are measured. Use the method and
limiting values described in section 2a.2.1.2.1 of the screen-film part of the European guidelines.
2b.2.1.2.2 Half Value Layer (HVL)

Use the method described in section 2a.2.1.2.2 of the screen-film part of the European
guidelines.
2b.2.1.3 AEC-system
It is generally recommended that systems used for mammography screening incorporate an AEC.
The performance of the AEC system should be tested in terms of reproducibility and accuracy
under varying conditions (object thickness and beam quality). The AEC system should adjust
target, filter and tube voltage such that image quality is sufficient and dose is within an
acceptable range. Semi-automated systems that start from a user defined target, filter and tube
voltage but adapt dose according to breast transparency, are also acceptable.
The use of a look-up-table (LUT) for the determination of target, filter, tube voltage and dose
based on compressed breast thickness can only be allowed if this LUT is programmed into the X-
ray unit. However, it must be realized that these systems do not take breast composition into
account and therefore cannot be fully optimized with respect to image quality and dose. For this
kind of system some guidance for QC measurements is given in appendix 8.
For dose measurements it is essential that the dosimeter is positioned outside the region in
which the exposure settings are determined. Alternatively, dose can be calculated using tube
loading (mAs) and tube output.
Manufacturers of equipment, which do not incorporate an AEC, are urged to implement an

AEC in their mammography X-ray units before January 2006.
The authors advise against the use of mammography X-ray units on which the exposure
settings have to be set completely manually.
2b.2.1.3.1 Exposure control steps: central value and difference per step (if applicable)
This test item only applies to mammography units with exposure control steps. Image the
standard test block at the different exposure control steps (or a relevant subset). Record
entrance dose (or tube loading). Calculate exposure steps in entrance dose (or tube loading).
Remark: If it is noticed that the system switches between two spectra, release the compression
paddle and compress again or use another PMMA thickness (add for example 0.5 cm
PMMA) to force the choice of one single spectrum and repeat the measurement.
The central setting is the standard setting. In this setting image quality must be sufficient, this is
determined by contrast threshold visibility measurements, see section 2b.2.4.1.
Typical value 5 - 15% increase in exposure per step13.

Equipment Standard test block, dose meter.
H2 20-09-2005 20:54 Pagina 125
2b.2.1.3.2 Back-up timer and security cut-off

European guidelines. Make sure that the detector is completely covered, or tape some lead
plates to the tube window.
Warning: An incorrect functioning of the back-up timer could damage the tube. To avoid excessive
tube load, consult the manual for maximum permitted exposure time.
2b.2.1.3.3 Short term reproducibility

Use the method and limiting values described in section 2a.2.1.3.3 of the screen-film section of
the European guidelines.
Remark: If it is noticed that the system switches between two spectra, release the compression
paddle and compress again or use another PMMA thickness (add for example 0.5 cm
PMMA) to force the choice of one single spectrum and repeat the measurement.
2b.2.1.3.4 Long term reproducibility

Use the weekly homogeneity check (see section 2b.2.2.3.1) for long term reproducibility.
Limiting value The variation of SNR in the reference ROI and dose < ± 10%.
Frequency Weekly.
Equipment Standard test block.
2b.2.1.3.5 Object thickness and tube voltage compensation

Compensation for object thickness should be measured by exposures of PMMA plates in the
thickness range from 20 to 70 mm (steps of 10 mm), using the clinical AEC settings (tube
voltage, target, filter and mode). The compression paddle must be in contact with the PMMA
plates.
Image PMMA plates of 20 mm thickness, with an aluminium object of 0.2 mm thickness on top,
if necessary in manual mode and with settings as close as possible to the clinical AEC settings
(if manual mode is used, substract the pre-exposure from the settings). Position the aluminium
object as shown in figure 2.1. Measure the mean pixel value and standard deviation in a ROI (4
cm2) with (position 2) and without (position 1) aluminium object. Calculate CNR. Repeat this
measurement for 30, 40, 45, 50, 60 and 70 mm PMMA thickness.
Fig. 2.1 Position of the aluminium filter for the CNR measurement
Image quality is evaluated for one thickness (at the equivalent of 5.0 cm PMMA) using contrast
threshold measurements (section 2.4.1). At other PMMA thicknesses CNRlimiting value is related to
the CNRlimiting value at 5.0 cm PMMA to ensure image quality at other thicknesses14.
The following formula is used to calculate the limiting value of CNR at standard thickness:
Threshold contrastmeasured * CNRmeasured = Threshold contrastlimiting value * CNRlimiting value
H2 20-09-2005 20:54 Pagina 126
The value of CNR at 5.0 cm thickness is related to the measured threshold contrast visibility in
section 2b.2.4.1. Using the formula above the limiting value of CNR at standard thickness can
be estimated using the measured threshold contrast in section 2b.2.4.1 and the (acceptable)
limiting value of value of the 0.1 mm diameter disc. The calculated CNRlimiting value should be used
as the 100% level mentioned in the limiting values below.
Limiting value CNR per PMMA thickness, see table for provisional limiting
values; Compare CNR values with results at acceptance
PMMA CNR15
Thickness (relative to 5.0 cm PMMA)
[cm] [%]
2.0 > 115
3.0 > 110
4.0 > 105
4.5 > 103
5.0 > 100
6.0 > 95
7.0 > 90

Equipment PMMA: a set of 10 mm thick PMMA plates covering the complete
detector area, 0.2 mm thick Al object (for example: the filters
which are used for the HVL measurement).
2b.2.1.4 Compression
Use the method and limiting values described in section 2a.2.1.4 of the screen-film part of the
European guidelines.
2b.2.1.5 Anti scatter grid

The anti scatter grid is designed to absorb scattered photons. The tests in this section only apply
to mammography units with (removable) grid. Some digital mammography systems do not
incorporate anti scatter grids (e.g. scanning systems).
2b.2.1.5.1 Grid system factor

Image the standard test block in clinical setting with grid. Record entrance dose and measure the
mean pixel value in the reference ROI. Expose two images without grid with mean pixel values
respectively below and above the value of the image with grid. Interpolate the pixel values to
obtain the entrance dose for which the pixel value is similar to the image with grid. Calculate the
grid system factor by dividing the entrance dose with grid by the entrance dose without grid.
Limiting value Manufacturers specification, typical value < 3.

Frequency At acceptance.
H2 20-09-2005 20:54 Pagina 127
2b.2.1.5.2 Grid imaging

European guidelines. The imaging of the grid is not possible for some grids due to minimum
required exposure times.
2b.2.2 Image receptor

This section describes measurements applicable to both DR and CR systems i.e. the image
receptor response and missed tissue at chest wall side. Other measurements apply to DR or CR
systems only. For a DR system detector element failure is determined. The performance of the
imaging plates of a CR system can be described by the CR plate sensitivity and the sensitivity to
other sources of radiation.
2b.2.2.1 Image receptor response

The measurement of the response is performed to check compliance with manufacturers
specifications, pixel value offset and the presence of additional noise sources beside quantum
noise.

The response function of the detector can be assessed by imaging a standard test block with
different entrance doses (tube loading) at the clinically used beam quality. Use the manual mode
for this measurement. Use at least 10 different tube loadings (mAs values). The range of mAs
values should be chosen such that the linearity measurement includes a wide range of entrance
surface air kerma (for example: 1/10 to 5 times16 the entrance surface air kerma for a routine
exposure).
For systems with a linear response, such as currently available DR systems, measure the mean
pixel value and standard deviation in the reference ROI on the unprocessed image. Plot the mean
pixel value against entrance surface air kerma. Determine linearity by plotting a best fit through
all measured points and determine the zero crossing to check presence of a pixel value offset.
Calculate the square of the correlation coefficient (R2). Compare the results to previous
measurements.
For systems with a non-linear response, such as currently available CR systems, plot mean pixel
value against . log relative entrance surface air kerma. Refer to the information provided by the
manufacturer whether pixel value should be linear or logarithmic versus entrance surface air
kerma at the applied screen processing. Post processing should be turned off. The screen
processing should be turned off as much as possible (see appendix 7). Determine linearity by
plotting a best fit through all measured points. Calculate the square of the correlation coefficient
(R2). Compare the results to previous measurements.
Appendix 7 provides information about the relation between entrance surface air kerma and
exposure indicator for some CR systems and screen processing modes.
Limiting value R2 > 0.99, results at acceptance are used as reference.

Frequency Every six months. At acceptance: additional measurements at
minimum and maximum tube voltage used in clinical practice at
every target-filter combination.
H2 20-09-2005 20:54 Pagina 128

Measure the mean pixel value and standard deviation in the reference ROI on the unprocessed
images of the response function measurement (2b.2.2.1.1). For systems with a linear response,
calculate the SNR and plot SNR2 against entrance surface air kerma. Determine linearity by
plotting a best fit through all measured points. Calculate the square of the correlation coefficient
(R2). Repeat this measurement for all available target-filter combinations used in clinical
practice. Non-linearity is an indication for the presence of additional noise sources besides
quantum noise. (At acceptance: additional measurements at minimum and maximum tube
voltage used in clinical practice for each target-filter combination). Compare the results to
previous measurements.
For systems with a logarithmic response plot standard deviation squared against 1/entrance
surface air kerma. Determine linearity by plotting a best fit through all measured points. Calculate
the square of the correlation coefficient (R2). The offset is an indication for the presence of
additional noise sources besides quantum noise. Compare the results to previous measurements.
For CR systems: No post processing should be applied, the screen processing should be turned
off as much as possible (see appendix 7).
Limiting value Results at acceptance are used as reference

Frequency Every six months. At acceptance: additional measurements at
minimum and maximum tube voltage used in clinical practice at
every target-filter combination
Equipment Standard test block, dose meter
2b.2.2.2 Missed tissue at chest wall side

Determine the width of tissue not imaged between the edge of the breast support table and the
imaged area. This can be done by several methods. In some phantoms markers at a fixed
distance from chest wall side are incorporated. The position of these markers on the image can
be used to determine the missed tissue at chest wall side. For CR systems, this measurement
should be repeated 5 times to check whether the insertion of the plate in the cassette is
reproducible.
Limiting value Width of missed tissue at chest wall side ≤ 5 mm.

Frequency At acceptance.
Equipment Phantom with markers positioned close to the bucky surface.
2b.2.2.3 Image receptor homogeneity and stability

2b.2.2.3.1 Image receptor homogeneity
The homogeneity of the image receptor can be obtained by exposing at clinical settings a
standard test block covering the complete detector. Record the exposure settings and tube
loading. Evaluate the unprocessed image by calculating the mean pixel value and standard
deviation in a ROI (a square with an area of 1 cm2). Move the ROI over the whole image.
Determine the mean pixel value in the whole image and the mean SNR in all ROI’s. Compare the
mean pixel value and the SNR of each ROI to the overall mean pixel value and the mean SNR.
Compare the SNR to previous homogeneity tests. Software for determining detector homogeneity
is available on: www.euref.org.
To exclude failure due to inhomogeneities in the standard block, rotate the standard test block
180° and repeat the measurement.
Check the homogeneity visually. The window width should be set at 10% of the mean pixel value.
Perform this measurement at acceptance also at other PMMA thickness (for example with PMMA
blocks of 20 and 70 mm thickness). For all measurements clinical settings should be used.
H2 20-09-2005 20:54 Pagina 129
It is acknowledged that the Heel effect and geometry effects influences the results of the
homogeneity measurement. If a specific system does not comply with the provisional
limiting values it is advised to check whether geometry or the Heel effect causes this
deviation or some malfunction in the system. For CR systems an additional homogeneity
image can be obtained by exposing a cassette using half dose under normal conditions
and half dose with the cassette rotated 180° in the bucky to minimize the Heel effect and
geometric effects.
Limiting value (provisional) Maximum deviation in mean pixel value < ± 15% of
mean pixel value in whole image, maximum deviation in SNR
< ± 15% of mean SNR in all ROI’s, maximum variation of the
mean SNR between weekly images < ± 10%, entrance surface air
kerma (or tube loading) between weekly images < ± 10%.
Frequency Weekly and after maintenance, at acceptance also at 20 and
70 mm PMMA thickness.
Equipment Standard test block covering the complete detector, at acceptance
also PMMA blocks of 20 and 70 mm thickness covering the complete
detector, software for determining detector homogeneity.
2b.2.2.3.2 Detector element failure (DR systems)

Inspect the most recent ‘bad pixel map’ of the manufacturer. This map (either an image or a
table) defines the position of all pixels of which the pixel value is not based on its own del
reading. This bad pixel map must be accessible by the user at any time and must be usable
independent of the equipment of that manufacturer.
Evaluate the up to date information on bad columns and bad dels from the manufacturer and
compare the position and number of defective dels to previous maps. Large clusters of defective
dels and dels from which the reading is influenced by neighbouring defective dels may become
visible in the image of a screen-film contact tool.
Limiting value At this moment no limits have been established. In future

versions of this protocol limits will be set and probably the
number of defective dels/columns will (also) be limited by the
percentage of a certain area, which is defective. At this moment
it is advised to refer to the limits of the manufacturer.
Equipment Bad pixel map.
2b.2.2.3.3 Uncorrected defective detector elements (DR systems)

To determine the number and position of defective detector elements not corrected by the
manufacturer, an image of the standard test block made at clinical settings should be evaluated
by calculating the mean pixel value in ROIs (squares with an area of 1 cm2). Move the ROI over
the whole image. Determine the pixels deviating more than 20% from the mean pixel value in a
ROI. To increase reliability deviating pixels can be determined on four images. Pixels, which
deviate more than 20% on several images, are potentially bad pixels. If the deviating pixels are in
one column, it is likely to be a bad column. Software for determining the number of uncorrected
defective detector elements is available on: www.euref.org.
Limiting value No limits have been set yet on the number of uncorrected
defective detector elements.
Frequency Weekly.
Equipment Standard test block covering the complete detector, at acceptance
also PMMA blocks of 20 and 70 mm thickness covering the
complete detector.
H2 20-09-2005 20:54 Pagina 130
2b.2.2.4 Inter plate sensitivity variations (CR systems)

Image the standard test block using the AEC exposure setting that is normally used clinically.
Record the entrance surface air kerma (or tube loading). Process the plate. The screen
processing should be turned off as much as possible (see appendix 7). No post processing
should be applied. Measure the mean pixel value and standard deviation in the reference ROI.
Calculate SNR. Repeat this measurement for all imaging plates. Evaluate the homogeneity of
each image.
Limiting values SNR variation in the reference ROI between all imaging plates
< ± 15%, variation in entrance surface air kerma (or tube loading)
< ± 10%, no major inhomogeneities on the images.
Frequency Yearly and after introducing new imaging plates.
2b.2.2.5 Influence of other sources of radiation (CR systems)

Erase a single imaging plate. Tape two different coins, one on each side of the cassette. Store
the imaging plate in the storage area during a maximal time period, for example during the
complete acceptance test. Process the plate. The screen processing should be turned off as
much as possible (see appendix 7). No post processing should be applied. Evaluate the visibility
of the coins on the resulting image.
Limiting value The coins should not be visible.

Frequency At acceptance and when changes in storage of the cassettes
have occurred.
Equipment Two coins of different size (for example a one and a two Euro
coin).
2b.2.2.6 Fading of latent image (CR systems)

Image the standard test block using one fixed exposure that is normally used clinically. Process
the plate after 1 minute. Measure the mean pixel value in the reference ROI. Repeat the
measurement with different time periods before read-out (2, 5, 10, 30 minutes).
Limiting value Results at acceptance are used as reference.

Frequency At acceptance and when image quality problems are suspected.
2b.2.3 Dosimetry
Use the method and limiting values described in paragraph 2.5.1 of the screen-film part of the
European guidelines. The PMMA plates should cover the whole detector. For dose measurements
it is essential that the dose probe is positioned outside the region in which the exposure settings
are determined. Alternatively, dose can be calculated using tube loading (mAs) and tube output.
2b.2.4 Image Quality

2b.2.4.1 Threshold contrast visibility
Threshold contrast visibility is determined for circular details with diameters in the range from
0.1 to 2 mm. The details are imaged on a background object with a thickness equivalent (in
terms of attenuation) to 50 mm of PMMA. The details must be positioned at a height of 20 to 25
mm above the breast support table17. Use the exposure factors that would be selected clinically.
Make six images of the details and move the details slightly between the images to obtain
130 European guidelines for quality assurance in mammography screening Four th edition
H2 20-09-2005 20:54 Pagina 131
images with different relative position of the details and the detector elements. Three
experienced observers should determine the minimal contrast visible on two images. Every
observer must score two different images. The whole detail diameter range specified in the table
below must be covered. In this range minimal contrast visible for a large number of detail
diameter must be determined at acceptance and at least 5 detail diameters in subsequent tests.
This evaluation should be done on unprocessed images. The window width and level and zoom
facilities must be adjusted to maximise the visibility of the details on the displayed images.
It is acknowledged that at present it is not possible to get unprocessed images from some
systems. For these systems threshold contrast visibility evaluation should be done on
processed images. The image processing may introduce artefacts on phantom images and
may be different from image processing for mammograms due to histogram or local
texture based processing techniques. Therefore care needs to be taken in interpretation
of these processed images.
The threshold contrast performance specified here relates to the nominal contrast calculated for
the details for a 28 kV tube voltage with molybdenum target and filter materials as explained in
appendix 6. This nominal contrast depends on the thickness and materials used to manufacture
the test object, and is independent of the actual spectrum used to form the image, which should
be that used clinically. It does not include the effects of scatter. The average nominal threshold
contrasts should be compared with the limiting values below.
off as much as possible (see appendix 7). If the screens comply with the limiting values of
section 2b.2.2.4 inter plate sensitivity variations, it is not necessary to use the same screen in
the threshold contrast visibility measurement.
Limiting value See table
Threshold contrast
Acceptable value Achievable value
Diameter Radiation contrast Equivalent gold Radiation contrast Equivalent gold

of detail using Mo/Mo 28 kV thickness18 using Mo/Mo 28 kV thickness11
[mm] [%] [µm] [%] [µm]
5* < 0.85 0.056 < 0.45 0.032
2 < 1.05 0.069 < 0.55 0.038
1 < 1.40 0.091 < 0.85 0.056
0.5 < 2.35 0.150 < 1.60 0.103
0.25 < 5.45 0.352 < 3.80 0.244
0.1 < 23.0 1.68 < 15.8 1.10

* This diameter size is optional
Frequency Yearly.
Equipment Contrast detail phantom.
The threshold contrast standards defined in the table above are chosen to ensure that digital
mammography systems perform at least as well as screen-film systems (Young, 2004). They
have been derived from measurements on screen-film and digital mammography systems using
the Nijmegen CDMAM contrast detail phantom version 3.4 (see section 2b.1.4). However it is
intended that they are sufficiently flexible to allow testing by other designs and makes of test
H2 20-09-2005 20:54 Pagina 132
objects. The values quoted form a smooth curve and may be interpolated for other detail
diameters. It is expected that a new design of test object will be developed that will simplify the
testing against these standards on a routine basis.
On the EUREF website (www.euref.org) CDMAM images and scores are available for reference
purposes.
2b.2.4.2 Modulation Transfer Function (MTF) and Noise Power Spectrum

(NPS) [optional]
Image an MTF test tool. Determine the MTF of the detector by using appropriate software tools.
Image a NPS phantom, or the standard test block. Determine the NPS of the detector by using
appropriate software. Use the resulting MTF and NPS of the acceptance test as reference. The
measurement can be repeated when in doubt about the quality of the detector.
Limiting value Results at acceptance are used as reference.

Frequency At acceptance and when image quality problems are suspected.
Equipment MTF test tool, software to calculate MTF, NPS phantom [standard
test block], software to calculate NPS.
2b.2.4.3 Exposure time

Long exposure times can give rise to motion unsharpness. Exposure time is defined as the time
during which primary X-rays reach each individual part of an imaged object. Exposure time may
be measured by some designs of tube voltage and output meters. Otherwise a dedicated
exposure timer has to be used. The time for a routine exposure in all clinical AEC modes is
measured at standard PMMA thickness. For scanning slot systems, also measure the scanning
time.
Remark: For most systems exposure time increases rapidly with breast thickness and content.
Depending on the screen-film combination and the clinically used spectra this range
may vary from 0.2 to 3 seconds. For some scanning slot systems however, scanning
time and exposure time are fairly constant for the whole range of breast thickness and
content. Due to this design, these systems may not comply with the limiting value of 2
seconds at standard thickness. Ideally exposure time should be below a certain limiting
value even for very thick and dense breasts, so the limiting value at standard thickness
may not be the right measure to prevent motion unsharpness for all breasts. Because
this worst case liming value has not been determined yet, the value of 2 seconds at
standard thickness is maintained, with the exception that scanning slot systems for
which exposure time is only slightly dependent on breast thickness and content do not
have to comply. For these systems clinical results will have to show that motion
unsharpness is not a problem.
Limiting value Exposure time: acceptable: < 2 s19; achievable: <1.5 s; scanning
time: values at acceptance are used as reference, typical value:
5 - 8 s.
Frequency Yearly.
Equipment Exposure time meter, standard test block.
2b.2.4.4 Geometric distortion and artefact evaluation

Evaluate geometric distortion by measuring distances (with digital distance measuring tools) on
an image of a phantom with straight lines (CDMAM, Toronto geometric distortion phantom etc.).
Image a wire mesh (e.g. mammography screen-film contact test device) at the standard AEC
setting. For CR systems: process the plate. The screen processing should be turned off as much
as possible (see appendix 7). No post processing should be applied. Evaluate the grid pattern on
the resulting image.
132 European guidelines for quality assurance in mammography screening Four th edition
H2 20-09-2005 20:54 Pagina 133
For the different digital systems, different types of artefacts can occur. Inspect all test images for
artefacts.
Limiting value No disturbing artefacts, no visible distortion.

Equipment Test object with horizontal, vertical and diagonal lines, wire mesh.
2b.2.4.5 Ghost image/erasure thoroughness

A ghost image is the residue of a previous image on the present image. In this measurement an
induced ghost image is related to the contrast of 0.1 mm Al at clinical setting.
In manual mode an image of the standard test block is made using clinical settings. The block is
positioned such that half of the detector is covered and half of the detector is not covered. For
the second image (at clinical settings) the standard test block covers the whole detector and the
aluminium object is placed exactly centred on top of the standard block (see figure 2.2). The time
between both images should be approximately one minute.
Repeat the ghost image measurement a number of times during testing.
Fig. 2.2 Ghost image / erasure thoroughness measurement
Measure the mean pixel value (PV) in the ROI (area: 4 cm2) on the locations shown in the figure
above (on the second image) and calculate the ‘ghost image’-factor.
mean pixel value (region 3) - mean pixel value (region 2)

Ghost image factor =
mean pixel value (region 1) - mean pixel value (region 2)
If the system fails to meet the limiting value, check the homogeneity of the image. If the Heel
effect is large regions 1 to 3 should be chosen on a line parallel to chest wall side.
If the ghost image test is performed last, it is advised to make a number of images of a
homogeneous block PMMA covering the whole detector afterwards to get rid of possible
ghosts.
Limiting value ‘Ghost image’-factor < 0.3 (provisional).

Frequency Yearly.
Equipment Standard test block, aluminium object of 0.1 mm thickness (for
example: the filters which are used for the HVL measurement).
H2 20-09-2005 20:54 Pagina 134
2b.3 Image processing

Image processing will not be considered in this version of the protocol. Manufacturers have to
specify in general terms which image processing is applied. It is advised that image processing
is evaluated clinically by comparing the image quality of mammograms (for example: a set of 50
mammograms) to mammograms of previous screening rounds by experienced readers. Special
attention should be given to the visualization of microcalcifications and subtle structures.
2b.4 Image presentation

The tests in this section are based upon the work of AAPM TG18 (American Association of
Physicists in Medicine, Task Group 18). The TG18 test patterns described in this section should
be obtained independently from the manufacturer, and can be downloaded from the TG18
website (2k versions should be used when available): http://deckard.mc.duke.edu/~samei/
tg18. Some mammography display systems need adjusted versions of the test patterns, these
will be available from the EUREF website.
Some general remarks:

• The test patterns have to be displayed at full resolution (exactly one display pixel for each pixel
in the digital image) or printed at full size, contrast and brightness of the images may not be
adjusted.
• For the tests in this chapter, the use of the display (primary class (diagnostic) or secondary
class display device) often determines the limiting values.
• Some of the tests in this chapter are for Cathode Ray Tube (CRT) displays or Liquid Crystal
Displays (LCDs) only.
• A magnifying glass may be used in the evaluation of printed images.
• The monitors should be tested as used clinically (e.g. third monitor on, viewing boxes on
covered with films).
2b.4.1 Monitors
2b.4.1.1 Ambient light
Most of the quality tests in this chapter are highly sensitive to ambient light, therefore all of them
should be performed under clinical conditions (room lights, light boxes and other display devices
should be at the same luminance level as under clinical conditions). The ambient light should be
measured at the centre of the display with the light detector facing outwards and the display
switched off.
Limiting value Ambient light should be less than 10 lux for primary display
devices. [The maximum ambient light actually depends on the
reflection characteristics and minimum luminance of the monitor,
but for reasons of simplicity this is ignored here.]
Frequency Every six months. (Every time the system is used, it has to be
made sure that ambient light conditions have not changed.)
Equipment Illuminance meter.
H2 20-09-2005 20:54 Pagina 135
2b.4.1.2 Geometrical distortion (CRT displays)

Visually check whether the TG18-QC image (fig. 4.1) is displayed without geometrical distortion.
To do so, inspect the lines and borders of the test pattern.
Fig. 4.1 TG18-QC test pattern
Limiting value Borders should be completely visible, lines should be straight,

the active display area should be centred on the screen.
Frequency Daily.
Equipment TG18-QC test pattern.
2b.4.1.3 Contrast visibility

The TG18-QC test pattern contains several items for evaluating the contrast visibility of a display.
Each of the sixteen luminance patches located approximately equidistant from the centre of the
image, contains four corner squares at equal low contrast steps to the patch (fig 4.2). The two
patches in the bottom with minimum and maximum pixel value, surrounding the test pattern
name, contain a centre square with a pixel value of 5% and 95% of the maximal grey level
respectively. The letters ‘QUALITY CONTROL’ in the three rectangles below these patches are
displayed with decreasing contrast to the background. The visible part of the letters should be
written down and checked with the visibility at acceptance, in order to keep track of contrast
degradation. If contrast visibility is not sufficient, it may help to dim the room lights. If this is
done however, the lights should also be dimmed while using the displaying system clinically. The
appearance of the TG18-QC test pattern also depends on the mapping of pixel values to
luminance. Therefore if this test has failed, the tests in sections 2b.4.1.6 and 2b.4.1.7 should
be performed.
Remark: It should be kept in mind that the luminance of LCD monitors depends on the viewing
angle. When large viewing angles are used, contrast visibility may not comply with the
limiting values.
H2 20-09-2005 20:54 Pagina 136
Fig. 4.2 Contrast visibility test item in TG18-QC test image
Limiting value All corner patches should be visible, the 5% and 95% pixel value
squares should be clearly visible.
Frequency Daily.
2b.4.1.4 Resolution
Evaluate horizontal and vertical line patterns to check display resolution visually.
AAPM Task Group 18 provides 6 line patterns at different background luminance levels.
(Horizontal line patterns TG18-LPH10, -LPH50 and -LPH89; Vertical line patterns TG18-LPV10,
-LPV50 and -LPV89.)
Fig. 4.3 Zoomed versions of the TG18-LPH50 pattern
H2 20-09-2005 20:54 Pagina 137
Limiting value All line patterns should be discernible.

Frequency Every 6 months.
Equipment 2kx2k TG18-LPH10, TG18-LPH50, TG18-LPH89, TG18-LPV10,
TG18-LPV50 and TG18-LPV89 test patterns.
2b.4.1.5 Display artefacts

The TG18-QC test pattern also contains some elements, which can be used for recognising
display artefacts. The image should be carefully checked for defect pixels (LCD only), steps in the
black-to-white and white-to-black ramp bars (this can reveal an insufficient bit depth), and
artefacts near the black-to-white and white-to-black transitions (video card). Also pay attention to
temporal instability (flicker) and spatial instability (jitter).
Limiting Values No disturbing artefacts should be visible.

Frequency Daily.
Equipment 2kx2k TG18-QC test pattern.
2b.4.1.6 Luminance range

Measure the maximum and minimum luminance of the display device. Test patterns TG18-LN12-
01 and TG18-LN12-18 can be used.
The ratio of maximum and minimum display luminance, in the presence of ambient light, is an
indicator of luminance contrast response capabilities of the monitor (under the current
environmental conditions). Both luminances should be measured using a telescopic luminance
meter, to include the influence of ambient light.
The ratio can be increased by reducing ambient light or by display adjustments. DICOM GSDF
conformance (section 2b.4.1.7) makes sure the available contrast is spread out in an
appropriate and standard manner over the full greyscale range of the monitor.
angle. When large viewing angles are used, the luminance range may not comply with
the limiting values.
Limiting Values The maximum to minimum luminance ratio should be at least

250 for primary display devices, or 100 for secondary display
devices. The difference of maximum luminances between
displays belonging to one displaying station should not exceed
5% of the lowest.
Frequency Every six months or when contrast visibility has changed.
Equipment Telescopic luminance meter, TG18-LN12-01 and TG18-LN12-18
test patterns.
2b.4.1.7 Greyscale Display Function

To make sure a mammogram will appear similarly on different viewing stations and on printed
film, the mapping of greyscale values to display luminance or optical density should be
consistent. In this measurement it is determined whether a display conforms to the DICOM
Greyscale Standard Display Function (GSDF).
The greyscale display function (GDF) can be determined by measuring the luminance of the 18
AAPM luminance test patterns (TG18-LN12-01 through TG18-LN12-18). The test patterns should
be displayed full screen and the luminance has to be measured at the centre of the screen. The
shape of the GDF depends on the ambient light in the room. Therefore room lights, light boxes
and other display devices should be at the same luminance level as when the system is used
clinically. A telescopic luminance meter should be used to include the influence of ambient light.
H2 20-09-2005 20:54 Pagina 138
The measured values can be inserted into a spreadsheet (available on the EUREF website:
www.euref.org) to automatically determine GSDF conformance.
After doing this measurement, the amount of ambient light may not be increased anymore,
otherwise the contrast response has to be measured again!
Remark: This test only applies to primary and secondary display systems. The acquisition
workstation monitor is excluded from this test. Due to the required ambient light levels
in the mammography room the acquisition workstation monitor will not comply with the
limiting values of primary and secondary displays. Therefore this monitor should only be
used to check positioning techniques, not for diagnosis and image quality checks.
It is acknowledged that some displaying systems do not comply with the DICOM Greyscale
Standard Function. Manufacturers are urged to comply with this standard.
angle. When large viewing angles are used, the display on a monitor may not comply with
the GSDF.
Limiting value The calculated contrast response should fall within ± 10% of the
GSDF contrast response for primary class displays (± 20% for
secondary class displays).
Frequency Every six months and when contrast visibility has changed.
Equipment Telescopic luminance meter, TG18-LN12-01 through TG18-LN12-
18 test patterns.
2b.4.1.8 Luminance uniformity

When the display has been tested for DICOM conformance at the centre of the monitor, this does
not mean contrast visibility is optimal at every position on the monitor. One could test the GDF
for several locations on the monitor, but it is more convenient to check display uniformity.
Measure the display luminance at five locations for each monitor. The test patterns TG18-UNL10
and TG18-UNL80 can be used (fig. 4.4).
Fig. 4.4 TG18-UNL10 an TG18-UNL80 test pattern
H2 20-09-2005 20:54 Pagina 139
Limiting value Maximum luminance deviation of a display device should be less

than 30% for CRT displays and LCD displays ((Lmax-Lmin)/
Lcentre < 0.3).
Equipment Luminance meter (telescopic luminance meters should be
equipped with a cone or baffle for this measurement),
TG18-UNL10 and TG18-UNL80 test patterns.
2b.4.2 Printers
2b.4.2.1 Geometrical distortion
Print the TG18-QC test pattern (fig. 4.1) and check visually if the image is printed without
geometrical distortion. Only the lines and borders of the test pattern are used to do this.
Limiting value Borders should be completely visible, lines should be straight.

Frequency Daily.
2b.4.2.2 Contrast visibility

Print the TG18-QC test pattern (see fig. 4.1). Check the visibility of the several items for
evaluating the contrast visibility (see fig. 4.2). Be sure that the viewing box, on which the test
pattern is checked, has sufficient luminance.
If contrast visibility is not sufficient, it may help to use diaphragms (if clinically used) or dim the
room lights. If this is done however, the lights should also be dimmed while using the displaying
system clinically. The appearance of the TG18-QC test pattern also depends on the mapping of
pixel values to densities. Therefore if this test has failed, the tests in sections 2b.4.2.5 and
2b.4.2.6 should be performed.
Limiting value All corner patches should be visible, the 5% and 95% pixel value
squares should be clearly visible.
Frequency Daily.
2b.4.2.3 Resolution
Evaluate horizontal and vertical line patterns to
check the resolution of a print-out.
The fine detail horizontal and vertical line
patterns in the TG18-PQC test pattern (fig 4.5)
can be used.
Limiting value All line patterns should be

discernible20.
Frequency At acceptance and when
decreased resolution is
suspected.
Equipment TG18-PQC test pattern.
Fig. 4.5 TG18-PQC test pattern
H2 20-09-2005 20:54 Pagina 140
2b.4.2.4 Printer artefacts

Print the TG18-QC, -PQC, -UN80 and -UN10 test patterns. Check the image for printer artefacts,
for example banding and streaking artefacts, pick-off artefacts, etc.
Limiting Values No disturbing artefacts should be visible.

Frequency Daily.
Equipment TG18-QC, TG18-PQC, TG18-UN10 and TG18-UN80 test patterns.
2b.4.2.5 Optical Density Range (optional)

Print the TG18-QC test pattern. Measure Dmin and Dmax on this image.
Limiting value Dmin < 0.25 OD, Dmax > 3.40 OD21 (provisional).
Equipment Densitometer, TG18-QC test pattern.
2b.4.2.6 Greyscale Display Function

To make sure a mammogram will appear similarly on different viewing stations and on printed
film, the mapping of greyscale values to display luminance or optical density should be
consistent. In this measurement it is determined whether a printer conforms to the DICOM
Greyscale Standard Display Function (GSDF).
The greyscale display function (GDF) can be determined by printing the TG18-PQC test pattern
and measuring the optical density of marked regions of the 18 bars. The GDF is determined by
the luminance corresponding with the optical density. The relationship between the luminance (L)
and the optical density (D) of the printed bars is:
L = La + L0 * 10-D
where: La is the luminance contribution due to ambient illuminance reflected off the film, and
L0 is the luminance of the light box with no film present
Printed mammograms may be viewed on different viewing boxes and under a variety of viewing
conditions. It is not desirable to repeat this measurement for each viewing box. Assuming each
viewing box, on which printed mammograms will be diagnosed, complies with the limiting values,
a standard viewing box is defined. For this standard viewing box La is 1 cd/m2 and L0 is 4000
cd/m2.
The measured values can be inserted into an spreadsheet (available on the EUREF website:
www.euref.org) to automatically determine GSDF conformance.
Limiting value The calculated contrast response should fall within ± 10% of the
GSDF contrast response.
Equipment Densitometer, TG18-PQC test pattern.
2b.4.2.7 Density uniformity

Print the test patterns TG18-UNL10 and TG18-UNL80. Measure the optical density at the five
marked locations.
Limiting value Maximum optical density deviation should be less than 10%
((Dmax-Dmin)/Dcentre < 0.1).
Equipment Densitometer, TG18-UNL10 and TG18-UNL80 test patterns.
H2 20-09-2005 20:54 Pagina 141
2b.4.3 Viewing boxes

If mammograms are read on printed images, check the viewing boxes using the method and
limiting values described in the European guidelines for quality assurance in mammography
screening, third edition (page 87).
2b.5 CAD software

May be considered in future versions of this protocol.
2b.6 References and Bibliography

2b.6.1 References
ACRIN Dmist ACRIN Dmist trial results, personal communication with G.
Mawdsley and A. Bloomquist.
Addendum, 2003 R. van Engen, K. Young, H. Bosmans, M. Thijssen: Addendum on
digital mammography to chapter 3 of the: European Guidelines
for Quality Assurance in Mammography Screening, version 1.0,
November 2003.
Beckers, 2003 Results of technical quality control in the Dutch breast cancer
screening programme (2001-2002), S.W. Beckers et al.,
Nijmegen 2003.
Dance, 2000 D.R. Dance et al., Additional factors for the estimation of mean
glandular dose using the UK mammography protocol,
Phys.Med.Biol., 2000, Vol. 45, 3225-3240.
European Guidelines, 2001 N. Perry, M. Broeders, C. de Wolf, S. Törnberg (ed.): European
Guidelines for Quality Assurance in Mammography Screening,
third edition, European Communities, 2001.
NEMA NEMA, Digital Imaging and Communications in Medicine,
http://medical.nema.org.
NHSBSP, 2000 Performance of mammographic equipment in the UK breast
cancer screening programme in 1998/99, 2000 (NHSBSP report
no 45).
NHSBSP, 2003 Review of radiation risk in breast screening, 2003 (NHSBSP
report no 54).
Leitz, 2001 Leitz et al., Patientdoser från röntgenundersökningar i Sverige,
Statens strålskyddsinstitut 2001.
Pedersen, 2000 K. Pedersen et al., Mammografiscreening, teknisk
kvalitetskontroll – resultater og evaluering etter fire års
prøveprosjekt,, Statens strålevern 2000.
Samei, 2001 E. Samei, J.A. Seibert, C.E. Willis, M.J. Flynn, E. Mah, K.L. Junck:
Performance evaluation of computed radiography systems. 2001:
Med. Phys 28(3): 361-371.
Samei, 2004 E. Samei, A. Badano, D. Chakraborty, K. Compton, C. Cornelius,
K. Corrigan, M.J. Flynn, B. Hemminger, N. Hangiandreou, J.
Johnson, M. Moxley, W. Pavlicek, H. Roehrig, L. Rutz, J. Shepard,
R. Uzenoff, J. Wang, C. Willis, Assessment of Display
Performance for Medical Imaging Systems. Draft Report of the
American Association of Physicists in Medicine (AAPM) Task
Group 18, Version 10.0, August 2004.
H2 20-09-2005 20:54 Pagina 142
Siewerdsen, 1999 J. H. Siewerdsen, D. A. Jaffray, A ghost story: Spatio-temporal

response characteristics of an indirect-detection flat-panel
imager, 1999: Med. Phys. 26 (8), 1624-1641.
Young, 2004 K.C. Young, B. Johnson, H. Bosmans, R.E. van Engen,
Development of minimum standards for image quality and dose
in digital mammography (to be published in IWDM 2004
proceedings).
2b.6.2 Bibliography
Protocols
1. European Guidelines for Quality Assurance in Mammography Screening, third edition,
European Communities, 2001.
2. Quality Control Procedures for Full-field Digital Mammography, 2002: ACRIN # 6652, Digital
Mammography Imaging Screening Trial, rev. 2.06, April 2002.
3. Meetprotocol Digitale mammografie: Acceptatietest van Screeningseenheden voor
Bevolkingsonderzoek op Borstkanker, versie 2002, National Expert and Training Centre for
Breast Cancer Screening, University Medical Centre Nijmegen.
4. Recommandations pour un programme d’assurance de qualité en mammographie
numérique. A. Noel, J. Stinès (red.). J. Radiol 2003; 84: 723-9.
5. European protocol on dosimetry in mammography, European Communities, 1996.
6. IPSM: Commissioning and routine testing of mammography X-ray systems – second edition,
The Institute of Physical Sciences in Medicine, York, 1994 (report no. 59/2).
7. Guidelines on quality assurance visits, NHSBSP, Second edition October 2000 (NHSBSP
Publication No 40).
8. Addendum on digital mammography to chapter 3 of the: European Guidelines for Quality
Assurance in Mammography Screening, version 1.0, November 2003.
9. Assessment of display performance for medical imaging systems, pre-final draft (version
8.1), AAPM, Task Group 18, E. Samei (chairman) et al.
2002: AAPM TG 18.
11. Performance of mammographic equipment in the UK breast cancer screening programme in
1998/99, 2000 (NHSBSP report no 45).
12. Review of radiation risk in breast screening, 2003 (NHSBSP report no 54).
13. Results of technical quality control in the Dutch breast cancer screening programme (2001-
2002), S.W. Beckers et al., Nijmegen 2003.
14. Patientdoser från röntgenundersökningar i Sverige, W. Leitz et al. Statens strålskyddsinstitut
2001.
15. Mammografiscreening, teknisk kvalitetskontroll – resultater og evaluering etter fire års
prøveprosjekt, K. Pedersen et al., Statens strålevern 2000.
Publications
1. E. Samei, J.A. Seibert, C.E. Willis, M.J. Flynn, E. Mah, K.L. Junck: Performance evaluation of
computed radiography systems.
2001: Med. Phys 28(3): 361-371.
2. C. Kimme-Smith, C. Lewis, M. Beifuss, L.W.Bassett: Establishing minimum performance
standards, calibration intervals and optimal exposure values for a whole breast digital
mammography unit.
1998: Med. Phys. 25 (12): December .
3. H. Fujita, D.Tsai, T. Itoh, K. Doi, J. Morishita, K. Ueda, A. Ohtsuka: A simple method for
determining the modulation transfer function in digital radiography.
1992: IEEE Transactions on Medical Imaging 11(1): 34-39.
4. E. Samei, M.J. Flynn, D.A. Reimann: A method for measuring the presampled MTF of digital
radiographic systems using an edge test device.
1998: Med. Phys. 25(1): 102-113.
5. A. Noël, J. Stines: Contrôle de qualité: du conventionnel au numérique.
2002: J. Le Sein 12(1-2): 17-21.
H2 20-09-2005 20:54 Pagina 143
6. K.G. Lisk: SMPTE test pattern for certification of medical diagnostic display devices
1984: Proc. of SPIE 486: 79-82.
7. W.Huda, A.M. Sajewicz, K.M. Ogden, E.M. Scalzetti, D.R. Dance: How good is the ACR
accreditation phantom for assessing image quality in digital mammography.
2002: Acad. Radiol. 9: 764-772.
8. J.T. Dobbins: Effects of undersampling on the proper interpretation of modulation transfer
function, noise power spectra and noise equivalent quanta of digital imaging systems
Med. Phys. 22: 171-181.
9. J.P. Moy, B. Bosset: How does real offset and gain corrections affect the DQE in images from
X-ray Flat detectors.
1999: proc. of SPIE 3659.
10. C.K. Ly: Softcopy Display Quality Assurance Program at Texas Children’s Hospital.
2002: Journal Of Digital Imaging, online publication: March.
11. D.R. Dance et al., Monte Carlo calculation of conversion factors for the estimation of mean
glandular breast dose, Phys.Med.Biol., 1990, Vol. 35, 1211-1219.
12. D.R. Dance et al., Additional factors for the estimation of mean glandular dose using the UK
mammography protocol, Phys.Med.Biol., 2000, Vol. 45, 3225-3240.
13. Z.F. Lu et al., Monthly monitoring program on DryviewTM laser imager: One year experience
on five Imation units, Med. Phys. 26 (9), 1817-1821.
14. H. Jung et al., Assessment of flat panel LCD primary class display performance based on
AAPM TG18 acceptance protocol, Med. Phys. 31 (7), 2155-2164.
15. J. H. Siewerdsen, D. A. Jaffray, A ghost story: Spatio-temporal response characteristics of
an indirect-detection flat-panel imager Med. Phys. 26 (8), 1624-1641.
16. Development of minimum standards for image quality and dose in digital mammography,
K.C. Young, B. Johnson, H. Bosmans, R.E. van Engen (to be published in IWDM 2004
proceedings).
17. M. Thijssen, W. Veldkamp, R. van Engen, M. Swinkels, N. Karssemeijer, J. Hendriks,
Comparison of the detectability of small details in a film-screen and a digital mammography
system by the imageing of a new CDMAM phantom, in M.J. Yaffe (ed). Digital mammography
IWDM 2000, Medical Physics Publishing 2001, 666-672.
18. AK. Carton, Development and application of methods for the assessment of image quality
and detector performance in digital mammography. PhD thesis at the KU Leuven, July 2004.
Other reports
1. Samei E, Badano A, Chakraborty D, Compton K, Cornelius C, Corrigan K, Flynn MJ,
Hemminger B, Hangiandreou N, Johnson J, Moxley M, Pavlicek W, Roehrig H, Rutz L, Shepard
J, Uzenoff R, Wang J, and Willis C. Assessment of Display Performance for Medical Imaging
Systems. Draft Report of the American Association of Physicists in Medicine (AAPM) Task
Group 18, Version 10.0, August 2004.
2004: AAPM TG 18
2. Mammography – recent technical developments and their clinical potential, B. Hemdal et al.
2002: Statens strålskyddinstitut, Swedish Radiation Protection Authority.
3. Quality Assurance, meeting the challenge in the Digital Medical Enterprise, B.I. Reiner, E.L.
Siegel, J.A. Carrino (ed.).
2002: Society for Computer Applications in Radiology.
4. Qualitätssicherung an Bildwiedergabegeräten, ed. D. Richter.
2002: ZVEI-Fachverband Elektromedizinische Technik.
5. DIN 6868-13:2003-02 Sicherung der Bildqualität in röntgendiagnostischen Betrieben – Teil
13: Konstanzprüfungbei Projektionsradiographie mit digitalen Bildempfänger-Systemen.
6. DIN V 6868-58:2001-01 Sicherung der Bildqualität in röntgendiagnostischen Betrieben
– Teil 58: Abnahmeprüfung an medizinischen Röntgeneinrichtungen der Projektions-
radiographie mit digitalen Bildempfängersystemen.
7. International Electrotechnical Commission (IEC), Geneva, Switzerland: Evaluation and
routine testing in medical imaging departments, part 3-2: Acceptance tests – Imaging
performance of mammographic X-ray equipment.
2004: IEC 61223-3-2 Ed. 2.
8. Handbook of Medical Imaging vol 1, J. Beutel, H.L. Kundel, R.L. Van Metter
2000: SPIE Press.
H2 20-09-2005 20:54 Pagina 144
9. Breast dose surveys in the NHSBSP, software and instruction manual, K.C. Young., NHS
Cancer Screening Programmes, 2001 (NHSBSP report no 01/10).
10. IPEM report 78 Catalogue of diagnostic X-ray spectra & other data, K. Cranley et al., 1997.
Internet
1. EUREF website: www.euref.org
2. AAPM Task group 18 (test patterns monitor QC): http://deckard.mc.duke.edu/~samei/tg18
3. DICOM standard: http://medical.nema.org
H2 20-09-2005 20:54 Pagina 145
Table 2b.1: Frequencies of Quality Control

This protocol is work-in-progress and subject to improvements as more experience in digital
mammography is obtained and new types of digital mammography equipment are developed.
Therefore the frequencies of quality control may change in future. Updates will be made available
on the EUREF website (www.euref.org). It is recommended that users check the website for
updates before testing digital mammography equipment.
Table 2b.1.1 Frequencies of Quality Control
2b.2 Image acquisition
test-item acceptance yearly six weekly daily

and on monthly
indication
2b.1.1 X-ray source
2b.1.1.1 Focal spot size X
2b.1.1.2 Source-to– X if adjustable

image distance
2b.2.1.1.3 Alignment of X-ray X X

field/image area
2b.2.1.1.4 Radiation leakage X
2b.2.1.1.5 Radiation output X X
2b.2.1.2 Tube voltage and

beam quality
2b.2.1.2.1 Tube voltage X X
2b.2.1.2.2 Half Value Layer X
2b.2.1.3 AEC-system
2b.2.1.3.1 Exposure control X X

steps
2b.2.1.3.2 Back-up timer and X X

security cut-off
2b.2.1.3.3 Short term X X

reproducibility
2b.2.1.3.4 Long term X X

reproducibility
2b.2.1.3.5 Object thickness X X

and tube voltage
compensation
O: optional test, X: required test => This table is continued on the next page
H2 20-09-2005 20:54 Pagina 146
Table 2b.1.1 continued

and on monthly
indication
2b.2.1.4 Compression X X
2b.2.1.5 Anti scatter grid
2b.2.1.5.1 Grid system factor X

(if present)
2b.2.1.5.2 Grid imaging O O
2b.2.2 Image receptor
2b.2.2.1 Image receptor

response
2b.2.2.1.1 Response function X X
2b.2.2.1.2 Noise evaluation X X
2b.2.2.2 Missed tissue at X

chest wall side
2b.2.2.3 Detector homogeneity

and stability
2b.2.2.3.1 Detector homogeneity X X
2b.2.2.3.2 Detector element

failure (DR) X X
2b.2.2.3.3 Uncorrected X X
defective DELs (DR)
2b.2.2.4 Inter plate sensitivity X X

variations (CR)
2b.2.2.5 Influence of other X

sources of radiation (CR)
2b.2.2.6 Fading of latent X

image (CR)
2b.2.3 Dosimetry X X
2b.2.4 Image quality
2b.2.4.1 Threshold contrast X X

visibility
2b.2.4.2 MTF and NPS O
2b.2.4.3 Exposure time X X

O: optional test, X: required test => This table is continued on the next page
H2 20-09-2005 20:54 Pagina 147

and on monthly
indication
2b.2.4.4 Geometric distortion X X

and artefact evaluation
2b.2.4.5 Ghost image / X X

erasure thoroughness
2b.4 Image presentation
2b.4.1 Monitors
2b.4.1.1 Ambient light X X
2b.4.1.2 Geometrical X X
distortion (CRT)
2b.4.1.3 Contrast visibility X X
2b.4.1.4 Resolution X X
2b.4.1.5 Displaying artefacts X X
2b.4.1.6 Luminance range X X
2b.4.1.7 DICOM Greyscale X X

Standard Display Function
2b.4.1.8 Luminance uniformity X X
2b.4.2 Printers
2b.4.2.1 Geometrical distortion X X
2b.4.2.2 Contrast visibility X X
2b.4.2.3 Resolution X
2b.4.2.4 Printer artefacts X X
2b.4.2.5 Optical Density range O O
2b.4.2.6 DICOM GSDF X X
2b.4.2.7 Density uniformity X X
2b.4.3 Viewing boxes X X

O: optional test, X: required test
H2 20-09-2005 20:54 Pagina 148
Table 2b.2 Limiting values
Table 2b.2.1 Limiting values

2b.2. Image acquisition typical limiting value unit
X-ray source
See European Guidelines, part A, table 4.1.
tube voltage
AEC
- exposure contol steps 5 - 15% mGy or mAs
- back-up timer and security cut-off - function properly
- short-term reproducibility - < ± 5% < ± 2% mGy
- long-term reproducibility
variation in SNR - < ± 10% mGy
variation in dose - < ± 10% mGy
- object thickness and tube voltage compensation
CNR per PMMA thickness
2.0 cm - > 115%
3.0 cm - > 110%
4.0 cm - > 105%
4.5 cm - > 103%
5.0 cm - > 100%
6.0 cm - > 95%
7.0 cm - > 90%
compression
anti scatter grid

2b.2.2 Image receptor typical limiting value unit

response function
- linearity - R 2 > 0.99 - -
- noise evaluation - - - -
missed tissue at chest wall side

detector homogeneity - ≤5 - mm
- variation in mean pixel value (on image) - < ± 15% - -
- variation in SNR (on image) - < ± 15% - -
- variation in mean SNR (between images) - < ± 15% - -
- variation in dose (between images) - < ± 10% - mGy
detector element failure

- number of defective dels - not yet established not yet established -
- position of defective dels - not yet established not yet established -
=> This table is continued on the next page
H2 20-09-2005 20:54 Pagina 149
2b.2.2 Image receptor typical limiting value unit

uncorrected dels
- number of uncorrected defective dels - not yet established not yet established -
- position of uncorrected defective dels - not yet established not yet established -
inter plate sensitivity variations

- variation in SNR - < ± 15% - -
- variation in dose - < ± 10% - -
influence of other sources of radiation - coin not visible - -
fading of latent image - - - -
2b.2.3 Dosimetry typical limiting value unit

- glandular dose per PMMA thickness

2.0 cm - < 1.0 < 0.6 mGy
3.0 cm - < 1.5 < 1.0 mGy
4.0 cm - < 2.0 < 1.6 mGy
4.5 cm - < 2.5 < 2.0 mGy
5.0 cm - < 3.0 < 2.4 mGy
6.0 cm - < 4.5 < 3.6 mGy
7.0 cm - < 6.5 < 5.1 mGy
2b.2.4 Image quality typical limiting value unit

threshold contrast visibility

- detail
5.0 mm (optional) - < 0.85% < 0.45% -
2.0 mm - < 1.05% < 0.55% -
1.0 mm - < 1.40% < 0.85% -
0.5 mm - < 2.35% < 1.60% -
0.25 mm - < 5.45% < 3.80% -
0.10 mm - < 23.0% < 15.8% -
MTF and NPS

- MTF (optional) - - - -
- NPS (optional) - - - -
exposure time - < 2.0 < 1.5 s
scanning time 5 to 8 s
geometric distortion and artefact evaluation

- geometric distortion - no distortions - -
- artefact evaluation - no disturbing artefacts - -
ghost image factor - 0.3 - -
=> This table is continued on the next page
H2 20-09-2005 20:54 Pagina 150
2b.4 Image presentation typical limiting value unit

2b.4.1 monitors
- ambient light - < 10 - lux
- geometrical distortion - straight lines
- contrast visibility - corner patches visible -
squares visible -
- resolution - line pattern discernible -
- display artefacts - no disturbing artefacts -
- luminance range
* ratio maximum/minimum luminance - 250 -
* difference in luminance left and right monitor - 5% - Cd/m2
- DICOM greyscale standard display function - ± 10% of GSDF -
- luminance uniformity
* deviation in luminance (CRT display) - 30% - Cd/m2
2b.4.2 printers
- geometrical distortion - straight lines -
- contrast visibility - corner patches visible -
squares visible
- resolution - line pattern discernible -
- printer artefacts - no disturbing artefacts -
- optical density range (optional) - Dmin < 0.251, Dmax > 3.41 - OD
- DICOM greyscale standard display function - ± 10% of GSDF -
- density uniformity
* deviation in optical density - < 10% - OD
2b.4.3 viewing boxes

1
Provisional limiting values
H2 20-09-2005 20:54 Pagina 151
European protocol for the quality control of the physical and technical aspects
of mammography screening
Appendices
Appendix 1: Mechanical and electrical safety checks
Introduction
Basic mechanical and electrical safety tests should be performed according to local regulations.
If such regulations do not exist this appendix gives an example of such tests based on the UK
protocol.
Mechanical Function and Safety checks

The following features of the equipment should be checked:
• All movements should operate smoothly and be free running. The force needed to move any
part should be less than 30 N.
• All mechanical/electromechanical brakes should function properly.
• All scales/indications on linear/rotational movements and focus film distance (FFD) (if
adjustable) should be clearly marked.
• All beam limiting diaphragms should be marked with their field sizes at the relevant FFD.
• Power driven vertical movement of the U-arm should be possible with the patient leaning
against the breast support platform (without compression applied).
• Vertical and rotational movement of the U-arm should be prevented when compression is
applied.
• All foot switches should operate correctly.
• All attachments should locate correctly and their locks should function properly.
• It should be possible to move the AEC detector properly into the pre-set positions.
• The bucky assembly should provide firm retention of the cassette (with the U-arm both vertical
and horizontal) but allow easy insertion and removal.
• The interlock to prevent exposure when the cassette is not correctly positioned should operate
correctly.
• The light intensity from the x-ray field light should be adequate.
• The movement of the compression device should be smooth.
• When compression is applied, it should not be possible to move the U-arm.
• The automatic release of the compression plate after an exposure should function correctly.
The override of this automatic release should also function correctly.
• An emergency release of compression should be available and function properly.
• The compression paddle and breast support platform should be smooth and must not have
any sharp edges or surfaces, etc. which may injure the patient.
• The edges of the radiation protection screen should be clearly defined so that the operator is
aware of the outline.
• The restraining devices use (for X-ray unit, radiation protective screen, etc.) provided on mobile
units should be effective in use.
Markings and labelling

The following should be clearly marked or indicated:
• The focal spot size and position.
• The amount of inherent, added and total filtration (usually in mm of aluminium) including that
of alterable or removable filters.
• The position of AEC detectors.
• The function of all controls.
H2 20-09-2005 20:54 Pagina 152
Radiation safety
The following checks relate to the safe operation of the x-ray unit:
• A mains isolator, accessible from the normal operating position should be provided.
• A visible indication must be provided on the control panel to show that the mains are switched
on.
• The visible exposure warning indication must function correctly.
• The total filtration must be equivalent to at least 0.5 mm Al or 0.03 mm Mo.
• If the added filtration is removable or interchangeable, an interlock must be provided to
prevent exposure if the filter is removed or incorrectly inserted.
• If the field-limiting diaphragm can be removed, an interlock should be provided to prevent
exposure unless the diaphragm is properly aligned.
• The exposure must terminate if the exposure control is released prematurely.
• The location of the exposure control should confine the operator to the protected area during
exposure.
• The exposure control should be designed to prevent inadvertent production of x-rays.
• The design of the exposure control should prevent further exposure unless pressure on the
control is first released.
Integral radiation protection screen

A radiation protection screen should be provided to afford protection equivalent to at least 0.1
mm of lead at 50 kV and should allow good visibility of the patient by the operator and vice versa.
The lead equivalence of the radiation protection screen should be marked (on both the glass and
the panel where appropriate) at a specified voltage. If the lead equivalence is not marked and is
not shown in the accompanying documentation, it will need to be measured.
X-ray room
• Room warning lights should be provided at all entrances to the x-ray room. These should
indicate when x-rays are being or are about to be generated.
• A check on the room shielding, either visually, against the local requirements at the planning
stage, or by transmission measurements, should be undertaken at or prior to installation.
H2 20-09-2005 20:54 Pagina 153
Appendix 2: Film-parameters
The film curve can be characterised by a few parameters. Most important items are contrast,
sensitivity and base and fog. There are different methods to calculate the film parameters.
Existing normalisation’s differ so much that the following method is suggested, derived from the
Dutch protocol (1991), which is based on the ANSI (1983) norm.
Very high contrast can be a problem because of an associated reduction in dynamic range which
may result in dense breast tissue being imaged in relatively low film densities where the film
performance is relatively poor. To some extent this can be compensated for by setting relatively
high average film densities, but even then a lower film contrast may better image local areas of
dense tissue. Conversely a very low overall film contrast may indicate an inadequately processed
film and subtle details may be missed by the radiologist.
Research has shown that film gradient measured by light sensitometry correlates well with film
gradient measured by x-ray sensitometry using a fixed kV and target filter combination. One must
bear in mind that film emulsions may respond slightly differently to the light from a sensitometer
as opposed to the light from the screen used for imaging.
Dmin Base and fog; the optical density of a non exposed film after
developing. The minimum optical density can be visualised by
fixation only of an unexposed film. The extra fog is a result of
developing the (unexposed) emulsion.
Dmax The maximum density achievable with an exposed film; i.e. the
highest density step.
MGrad Mean Gradient; the property which expresses the filmcontrast in

the diagnostic range. MGrad is calculated as the slope of the line
through the points D1=Dmin+0.25 OD and D2=Dmin+2.00 OD. Since
the film curve is constructed from a limited number of points, D1
and D2 must be interpolated. Linear interpolation of the
construction points of the film curve will result in sufficient
accuracy.
Grad1,2 Middle Gradient; the property which expresses the filmcontrast in

the diagnostic range. Grad1,2 is calculated as the slope of the line
through the points D1=Dmin+1.00 OD and D2=Dmin+2.00 OD. Since
the film curve is constructed from a limited number of points, D1
and D2 must be interpolated. Linear interpolation of the
construction points of the film curve will result in sufficient
accuracy.
Gradgland The glandular tissue gradient can be defined as an alternatively.

This is the gradient at glandular densities 0.8 – 1.2 OD. This
gradient is used in combination with the Gradfat.
Gradfat The alternative fat gradient is defined between densities of 2.0

and 2.4 OD. This gradient is used in combination with the Gradfat.
Speed Sensitivity; the property of the film emulsion directly related to the
dose. The Speed is calculated as the x-axis cut-off at optical
density 1.00+Dmin, also called ‘Speedpoint’. The higher the figure
for Speed, the more dose is needed to obtain the right optical
density. Since the film curve is constructed from a limited number
of points, the Speed must be interpolated. Linear interpolation
will result in sufficient accuracy.
Since these parameters are derived from the characteristic curve by
interpolation they are not very practical if a computer is not available.
A simpler procedure is to use the parameters below which are based
on density measurements of particular sensitometric steps.
H2 20-09-2005 20:54 Pagina 154
Speed Index The density of the step near to the speedpoint density 1.0 OD,
base and fog excluded. Usually this is the density of step 11 of
the sensitometric stepwedge.
Contrast Index 1 The difference in density found between the step nearest to the
speedpoint density (1.0 OD, base and fog excluded) and the one
with a 0.6 log E (factor 4) higher light exposure (normally
4 density steps) (ACR).
Contrast Index 2 The difference in density steps found between the step nearest
to the speedpoint and the step nearest to a density at 2.0 OD,
base and fog excluded (IPSM, see bibliography).
H2 20-09-2005 20:54 Pagina 155
Appendix 3: A method to discriminate between processing and

exposure variations by correction for the film-curve
The optical density of a film is the result of X-ray exposure and processing. The film is mainly
exposed by light emitted by the intensifying screen. The light-emission of the screen is
proportional with the incident X-ray exposure. Primary X-rays only contribute up to 5% of the total
exposure. The developing process determines the optical density of the exposed area.
When an optical density in any given film is measured, the corresponding exposure is unknown.
However, the film curve (measured with light-sensitometry) describes the relation between light-
exposure and optical density. Any measured optical density can be converted into a relative log
(light-exposure) or log (I’) by interpolation of the film curve. This figure log (I’) is a relative value
and strongly depends on the sensitometer used. But still it is a useful value, closely related with
the radiation dose applied and is therefore suitable to calculate the mass attenuation coefficient
of an arbitrary X-ray step wedge.
Note that recently available films, using a different type of sensitizing and grains, in some cases
show a discrepancy between the gradient as a result of light and by X-rays.
When the optical density of several images, taken under identical conditions, are measured,
there will be a range of optical densities. This can either be the result of a change in exposure or
a change in developing conditions. By calculating the relative figure log (I’) we are able to
distinguish between processor faults and tube malfunctions.
Approximation of X-ray contrast

To assess the X-ray contrast, correct the OD-readings of an Al-stepwedge for the processing
conditions by converting the optical densities into a fictional ‘exposure’, log (I’), according the
film curve. Now, a graph of the stepwedge number against ‘exposure’ will result in an almost
straight line. The slope of this line is a measure for the X-ray contrast.
H2 20-09-2005 20:54 Pagina 156
Appendix 4: Typical spectra per PMMA thickness in screen-film

mammography
Changing X-ray spectrum influences both glandular dose and image quality. The choice of
spectrum should be based on the optimization between both effects. In general the X-ray
spectrum should be harder when (simulated) breast thickness is increased. In the table below
some typical spectra, which are used in mammography, and which do not reduce contrast by
more than 10% compared to an image made with Mo-Mo 28 kV are given. The results should be
taken as typical values, not limiting values. When using the newly introduced high contrast films
(like the Kodak EV film), the values in the table below may need adaptation.
A4.1: Typical spectra per PMMA thickness
Spectrum
PMMA thickness (cm) Mo-Mo Mo-Rh Rh-Rh W-Rh
2 25, 26 kV
3 25-27 kV
4 26-28 kV 26, 27 kV
5 27-29 kV 26, 27 kV
6 28-30 kV 27-30 kV 27-30 kV
7 30, 31 kV 29-31 kV 29-31 kV 27-29 kV
H2 20-09-2005 20:54 Pagina 157
Appendix 5: Procedure for determination of average glandular dose

A5.1 Dose to typical breasts simulated with PMMA
The doses to a range of typical breasts can be assessed using blocks of PMMA as breast
substitutes. This method relies on the equivalence in attenuation between different thicknesses
of PMMA and typical breasts [Dance et al, 2000] as listed in tables A5.1 and A5.2. It should be
noted that since PMMA is generally denser than breast tissue any automatic selection of kV,
target or filter may be slightly different from real breasts. This can be corrected by adding
expanded polystyrene blocks to the PMMA as a spacer to make up a total thickness equal to the
equivalent breast. On systems that determine the exposure factors primarily on attenuation such
as the GE 2000D this should not be necessary. The average glandular dose (D) to a typical
breast of thickness and composition equivalent to the thickness of PMMA tested is calculated by
applying the following formula.
D = Kgcs (A5.1)
where K is the entrance surface air kerma (without backscatter) calculated at the upper surface
of the PMMA. The factor g, corresponds to a glandularity of 50%, and is derived from the values
calculated by Dance et al 2000 and is shown in table A5.1 for a range of HVL. The c-factor
corrects for the difference in composition of typical breasts from 50% glandularity [Dance et al
2000] and is given here for typical breasts in the age range 50 to 64 in table A5.2. Note that the
c and g-factors applied are those for the corresponding thickness of typical breast rather than the
thickness of PMMA block used. Where necessary interpolation may be made for different values
of HVL. Typical values of HVL for various spectra are given in table A5.3. The factor s shown in
table A5.4 corrects for differences due to the choice of X-ray spectrum (Dance et al 2000).
The dose should be determined using the usual clinically selected exposure factors including any
automatic selection of kV and target/filter combination.
A5.2 Clinical breast doses

It is also possible to measure the average glandular doses for a series of breast examinations
on each mammography system. To do this, the breast thickness under compression is
measured, and the tube voltage, and tube loading delivered are recorded.
From a knowledge of the output of the X-ray set for the kV and target and filter material used, this
tube loading may be used to estimate average glandular dose using the following formula:
D = Kgcs (A5.2)
where K is the entrance surface air kerma calculated (in the absence of scatter) at the upper
surface of the breast. The factor g, corresponds to a glandularity of 50%, and is shown in table
A5.5 (Dance et al 2000). The factor c corrects for any difference in breast composition from 50%
glandularity. C-factors for typical breast compositions in the age range 50 to 64 and 40 to 49 are
shown in tables A5.6 and A5.7. The factor s corrects for differences due to the choice of X-ray
spectrum as noted earlier. Measurement of compressed breast thickness for this purpose is
performed by the radiographer, by reading the displayed compressed thickness on the X-ray set.
The accuracy of the displayed thickness should be verified by applying a typical force (e.g. 100
N) to rigid material of known thickness. It may be necessary to apply correction factors if the
displayed values are in error. An accuracy of better than ± 2 mm is required. Software for making
such dose calculations has been published by the UK Breast Screening Programme (Young,
2001).
H2 20-09-2005 20:54 Pagina 158
Table A5.1: g-factors for breasts simulated with PMMA
PMMA Equivalent g-factors (mGy/mGy)

thickness breast
(mm) thickness HVL (mm Al)
(mm)
0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60
20 21 0.329 0.378 0.421 0.460 0.496 0.529 0.559 0.585
30 32 0.222 0.261 0.294 0.326 0.357 0.388 0.419 0.448
40 45 0.155 0.183 0.208 0.232 0.258 0.285 0.311 0.339
45 53 0.130 0.155 0.177 0.198 0.220 0.245 0.272 0.295
50 60 0.112 0.135 0.154 0.172 0.192 0.214 0.236 0.261
60 75 0.088 0.106 0.121 0.136 0.152 0.166 0.189 0.210
70 90 0.086 0.098 0.111 0.123 0.136 0.154 0.172
80 103 0.074 0.085 0.096 0.106 0.117 0.133 0.149
Table A5.2: c-factors for breasts simulated with PMMA
PMMA Equivalent Glandularity c-factors

thickness breast of equivalent
(mm) thickness breast HVL (mm Al)
(mm)
0.30 0.35 0.40 0.45 0.50 0.55 0.60
20 21 97 0.889 0.895 0.903 0.908 0.912 0.917 0.921
30 32 67 0.940 0.943 0.945 0.946 0.949 0.952 0.953
40 45 41 1.043 1.041 1.040 1.039 1.037 1.035 1.034
45 53 29 1.109 1.105 1.102 1.099 1.096 1.091 1.088
50 60 20 1.164 1.160 1.151 1.150 1.144 1.139 1.134
60 75 9 1.254 1.245 1.235 1.231 1.225 1.217 1.207
70 90 4 1.299 1.292 1.282 1.275 1.270 1.260 1.249
80 103 3 1.307 1.299 1.292 1.287 1.283 1.273 1.262
H2 20-09-2005 20:54 Pagina 159
Table A5.3: Typical HVL measurements for different tube voltage and target filter
combinations. (Data includes the effect on measured HVL of attenuation by a
PMMA compression plate*.)
HVL (mm Al) for target filter combination
kV Mo + 30 µm Mo Mo +25 µm Rh Rh +25 µm Rh W +50 µm Rh W +0.45 µm Al22
25 0.33 ± .02 0.40 ± .02 0.38 ± .02 0.52 ± .03 0.31 ± .03
28 0.36 ± .02 0.42 ± .02 0.43 ± .02 0.54 ± .03 0.37 ± .03
31 0.39 ± .02 0.44 ± .02 0.48 ± .02 0.56 ± .03 0.42 ± .03
34 0.47 ± .02 0.59 ± .03 0.47 ± .03
37 0.50 ± .02 0.51 ± .03
* Some compression paddles are made of Lexan, the HVL values with this type of compression
plate are 0.01 mm Al lower compared with the values in the table.
Table A5.4: s-factors for clinically used spectra [Dance et al. 2000]
Spectrum s-factor
Mo/Mo 1.000
Mo/Rh 1.017
Rh/Rh 1.061
Rh/Al 1.044
W/Rh 1.042
W/Al 1.05*
* This value is not given in the paper of Dance et al. The value in the table has been estimated
using the S-values of other spectra.
H2 20-09-2005 20:54 Pagina 160
Table A5.5: g-factors (mGy/mGy) for breast thicknesses of 2-11 cm and the HVL range 0.30-
0.60 mm Al. The g-factors for breast thicknesses of 2-8 cm are taken from Dance
(1990), and for 9-11 cm from Dance et al. (2000)
Breast g-factors (mGy/mGy)

Thickness
(cm) HVL (mm Al)
0.30 0.35 0.40 0.45 0.50 0.55 0.60
2 0.390 0.433 0.473 0.509 0.543 0.573 0.587
3 0.274 0.309 0.342 0.374 0.406 0.437 0.466
4 0.207 0.235 0.261 0.289 0.318 0.346 0.374
4.5 0.183 0.208 0.232 0.258 0.285 0.311 0.339
5 0.164 0.187 0.209 0.232 0.258 0.287 0.310
6 0.135 0.154 0.172 0.192 0.214 0.236 0.261
7 0.114 0.130 0.145 0.163 0.177 0.202 0.224
8 0.098 0.112 0.126 0.140 0.154 0.175 0.195
9 0.0859 0.0981 0.1106 0.1233 0.1357 0.1543 0.1723
10 0.0763 0.0873 0.0986 0.1096 0.1207 0.1375 0.1540
11 0.0687 0.0786 0.0887 0.0988 0.1088 0.1240 0.1385
H2 20-09-2005 20:54 Pagina 161
Table A5.6: c-factors for average breasts for women in age group 50 to 64 (Dance et al. 2000)
Breast c-factors
Thickness
(cm) HVL (mm Al)
0.30 0.35 0.40 0.45 0.50 0.55 0.60
2 0.885 0.891 0.900 0.905 0.910 0.914 0.919
3 0.925 0.929 0.931 0.933 0.937 0.940 0.941
4 1.000 1.000 1.000 1.000 1.000 1.000 1.000
5 1.086 1.082 1.081 1.078 1.075 1.071 1.069
6 1.164 1.160 1.151 1.150 1.144 1.139 1.134
7 1.232 1.225 1.214 1.208 1.204 1.196 1.188
8 1.275 1.265 1.257 1.254 1.247 1.237 1.227
9 1.299 1.292 1.282 1.275 1.270 1.260 1.249
10 1.307 1.298 1.290 1.286 1.283 1.272 1.261
11 1.306 1.301 1.294 1.291 1.283 1.274 1.266
Table A5.7: c-factors for average breasts for women in age group 40 to 49 (Dance et al. 2000)
Breast c-factors
Thickness
(cm) HVL (mm Al)
0.30 0.35 0.40 0.45 0.50 0.55 0.60
2 0.885 0.891 0.900 0.905 0.910 0.914 0.919
3 0.894 0.898 0.903 0.906 0.911 0.915 0.918
4 0.940 0.943 0.945 0.947 0.948 0.952 0.955
5 1.005 1.005 1.005 1.004 1.004 1.004 1.004
6 1.080 1.078 1.074 1.074 1.071 1.068 1.066
7 1.152 1.147 1.141 1.138 1.135 1.130 1.127
8 1.220 1.213 1.206 1.205 1.199 1.190 1.183
9 1.270 1.264 1.254 1.248 1.244 1.235 1.225
10 1.295 1.287 1.279 1.275 1.272 1.262 1.251
11 1.294 1.290 1.283 1.281 1.273 1.264 1.256
H2 20-09-2005 20:54 Pagina 162
Appendix 6: Calculation of contrast for details in a contrast-detail

test object
The minimum and achievable standards in section 2b.2.4.1 depend on the calculation of
nominal contrast for the details involved. To allow different designs of test object the standard is
specified in terms of radiation contrast for a typical spectrum using a tube voltage of 28 kV, a
molybdenum target material and a 30 mm thick molybdenum filter. (The spectrum was derived
from IPEM Report 78). The contrast of the discs and the threshold limiting values have been
determined using the CDMAM phantom with a 2 cm thickness of PMMA above and 2 cm
thickness below the test object. The CDMAM phantom includes an aluminium base which is
approximately equivalent to 1cm of PMMA in terms of attenuation. In the European guidelines
third edition however 4.5 cm has been chosen as the standard thickness of PMMA. Therefore in
future threshold contrast might be determined at a total thickness equivalent to 4.5 cm PMMA.
Calculated contrast for various thicknesses of gold are shown in Table A6.1. The corresponding
contrast calculated for the use of a CDMAM phantom with 4 cm of PMMA and for gold details on
4.5 cm PMMA is shown. In both cases the effect of scatter is not included in the calculation.
Table A6.1: Calculated radiation contrast for various gold thickness on the standard test
object
Thickness of Radiation contrast (%) Radiation contrast (%)

gold (µm) for gold disc for CDMAM
on 4.5 cm PMMA with 4 cm PMMA
0.1 1.63 1.57
0.5 7.83 7.60
1.0 15.02 14.55
1.5 21.57 20.92
2.0 27.56 26.76
H2 20-09-2005 20:54 Pagina 163
Appendix 7: Computed Radiography screen processing modes

For all test-items the following screen processing settings must be chosen except for the test-
items listed below. If a specific system or screen processing mode is not mentioned below, it is
advised to refer to the manual of the manufacturer:
Fuji systems Use FIXED EDR screen processing, suggested: S = 120, L = 2
Kodak systems Use Pattern screen processing
Agfa systems Use System diagnostics/flat field screen processing
Remark: For all measurements on the Fuji system an L-value of 2 is advised (which resembles
the L-value in clinical practice). If clipping occurs with an S-value of 120, another S-value
should be chosen.

The following relations between pixel value (sensitivity/exposure index) and entrance surface air
kerma should be linear (If a screen processing mode is not mentioned below, it is advised to refer
to the manual of the manufacturer):
Fuji systems: Linear relations:

Fixed EDR screen Plot the mean pixel value in the reference ROI versus log
processingsuggested: entrance surface air kerma
S = 120, L = 2
Semi EDR screen processing Plot sensitivity index versus inverse entrance surface air kerma
Kodak systems:
Pattern screen processing Plot the mean pixel value in the reference ROI versus log
entrance surface air kerma
Agfa systems:
System diagnostics/flat Plot the mean pixel value in the reference ROI versus log
field screen processing entrance surface air kerma
Fuji systems FIXED EDR screen processing, suggested: S = 120, L = 2
H2 20-09-2005 20:54 Pagina 164
2b.2.4.1 Threshold contrast visibility
Fuji systems Use FIXED EDR screen processing. The S and L value must be
chosen such that they are typical for the clinical situation. These
values may differ from site to site. Typical values (according to
Fuji): S = 40 to 100, L = 1.8 to 2.6.
2b.2.4.5 Ghost image / erasure thoroughness
Fuji systems Use FIXED EDR screen processing. The S and L value must be
chosen such that they are typical for the clinical situation. These
values may differ from site to site. Typical values (according to
Fuji): S = 40 to 100, L = 1.8 to 2.6.
H2 20-09-2005 20:54 Pagina 165
Notes
1. This is the PMMA thickness most commonly used, but others may be specified in parts of this
protocol.
2. The specifications of the listed equipment are given, where appropriate, in section 4, table 2.
3. The standard test block may be composed of several PMMA plates.
4. PMMA (polymethylmethacrylate) is commercially available under several brandnames, e.g.
Lucite, Plexiglas and perspex.
5. 150 X 100 mm or semi-circular with a radius of ≥µ100 mm, and covering a total thickness
range from 20 to 70 mm PMMA.
6. In future the PMMA thickness may change to the ‘standard thickness’ of 45 mm with the
details positioned at a height of 40 to 45 mm above the breast support table. This may mean
that the limiting values need slight adjustment.
7. If the exposure-to-read-time other than one minute is more relevant for practical reasons, that
other time should be chosen.
8. The specifications of the listed equipment are given, where appropriate, in chapter 3.5, table
2 of the European Guidelines, third edition.
9. The standard test block, covering the whole imaging area, may be composed of several PMMA
plates.
10. PMMA (polymethylmethacrylate) is commercially available under several brand names, e.g.
Lucite, Plexiglas and Perspex.
11. Covering the whole imaging area, and covering a total thickness range from 20 to 70 mm
PMMA (Normally PMMA of 180 X 240mm is available).
12. These films have been reported as suitable for use in collimation assessment by Beideck and
Gingold at the AAPM 2004 annual meeting.
13. These values are derived from screen-film mammography. At this moment no limiting values
on exposure increase per step for digital mammography have been set, but they should be
approximately uniform.
14. In future the contrast threshold visibility may be determined at the standard PMMA thickness
of 45 mm, so CNR limits will also be relative to 45 mm in future.
15. These values are provisional, it is advised to check the EUREF website for alterations
16. For some scanning slot systems only a limited range of mA or mAs settings are available, for
these systems images should be made at all settings.
17. In future the PMMA thickness may change to the ‘standard thickness’ of 45 mm with the
details positioned at a height of 40 to 45 mm above the breast support table. This may mean
that the limiting values need slight adjustment.
18. CDMAM phantom with a 4 cm thickness of PMMA, see appendix 6.
19. For some scanning slot systems, see the remark above.
20. Aliasing problems may occur due to the difference in pixel size of the printer and test pattern.
21. Further research is necessary to investigate whether the Dmin and Dmax limiting values are
appropriate.
22. Data partly based on: Bengt Hemdal, Lars Herrnsdorf, Ingvar Andersson, Gert Bengtsson, Boel
Heddson and Magnus Olsson, Average glandular dose in routine mammography screening with
Sectra MicroDose Mammography, MDM, poster at: Medicinska Riksstämman, Göteborg,
Sweden 2004.
H2 20-09-2005 20:54 Pagina 166

H2 PDF

Uploaded by

Copyright:

Available Formats

H2 PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

H2 PDF

Uploaded by

Copyright:

Available Formats

H2 20-09-2005 20:54 Pagina 57

European protocol for the

Quality Control (QC)

QC Measurements and Frequencies

• X-ray generator and exposure control system

Corresponding address: Corresponding author:

National Expert and Training Centre for

European protocol for the

Comments have been received from the following group:

2a.1 Introduction to the measurements

Reference exposure: Routine exposure:

test object thickness 45 mm1 45 mm

test object material PMMA PMMA

tube voltage 28 kV as used clinically

target material molybdenum as used clinically

filter material molybdenum as used clinically

anti scatter grid present present

phototimer detector in position closest to in position closest to

automatic exposure control on as used clinically

optical density control as leading to the reference as leading to the target

2a1.1 Staff and equipment

• Sensitometer • Standard test block3 (45 mm PMMA)

2a.1.2 Definition of terms

Antiscatter grid Device positioned close to the entrance surface of an image

Baseline value Value of a parameter defined on basis of many repeated

Breast compression Application of a compression force to the breast during image

Dmin Optical density obtained after processing of a non-exposed film.

Dmax Maximum optical density achievable with an exposed film; usually

Film gradient Index used to evaluate the film contrast.

Grad See Film gradient.

Heel effect Decreasing optical density measurable on a film in the cathode-

that a correct measurement of HVL requires ‘good geometry’

Image quality There is not a definition of image quality univocally accepted.

Limiting value Maximum or minimum limit of a possible range, considered

Measurement error Standard deviation if the number of repeated measurements is

PMMA The synthetic material polymethylmethacrylate. Trade names

Precision See Reproducibility.

Quality control As defined by the WHO (1982): ‘The set of operations

Radiation quality See HVL.

Relative error Ratio between measurement error and mean value.

Reproducibility Indicates the measurement precision or the reliability of tested

Spatial resolution Describes the smallest detectable detail at a defined contrast

Speed Synonym of film sensitivity, is a parameter inversely proportional

Standard breast Mathematical model generally used for calculations of glandular

Typical value Value of a parameter found in most facilities in comparable

X-ray spectrum Distribution of photon energies in an X-ray beam. It depends on

2a.2 Description of the measurements

2a.2.1 X-ray generation

2a.2.1.1.1 Focal spot size

The focal spot dimensions can be obtained by using one of the

Some fully automated digital devices to measure focal spot

A magnified X-ray image of the test device is produced using a

Focal spot size: star pattern method

mstar=dimage /dstar (2.2)

Limiting value None.

Focal spot size: slit camera method

mslit=dslit-to-film /dfocal spot-to-slit (2.3)

Limiting value None.