Dimond Iii: Image Quality and Dose Management For Digital Radiography

DIMOND III
Image Quality and

Dose Management
For Digital Radiography
Final Report
H.P. Busch
Trier
2004
DIMOND III
Contract: FIGM-CT-2000-00061
Working Group:
Brüderkrankenhaus Trier
Germany
Prof. Dr. H.P. Busch M.D.

Chr. Decker Dipl. Ing
C. Schilz Dipl. Ing
A. Jockenhöfer M.D.
St. Busch cand.med.
M. Anschütz MTRA
Lead Contractor:
Prof. Dr. H.P. Busch
Contents:
Chapter I Image Quality and 5

Dose Management for
Digital Projection Radiography
Chapter II Digital Projection Radiography: 11

Principles, Imaging and Application
Chapter III Diagnostic Requirements 24
Chapter IV Referral Guidelines 52

(with the three level band for image quality)
Chapter V Optimisation and Quality Assurance 66
Chapter VI Image Quality and Dose for 77

(H.P. Busch, S. Busch, Chr. Decker, C. Schilz)
Chapter VII References 93

Annex 1: Final Report
Annex 2: - Janalyser-CDRAD -
Quality Testing for
(C. Schilz)
Annex 3: - Janalyser -
Constancy Testing for
(C. Schilz)
Annex 4: CD1: Quality- and Dose- Management

CD2/3/4: Digital Mammography (Examples)
Annex 5: Acknowledgements
Chapter I
Image Quality and Dose Management for

In the field of projection radiography, conventional film/screen radiography is

continuously being replaced by digital imaging methods (such as storage
phosphor plates and flat detectors). Digital technology has many advantages,
especially for imaging without automatic exposure control, such as bedside
imaging. In recent years, costs have decreased significantly and a greater
range of specific units has become available. The new technology is likely to
lead to increased spatial resolution, decreased dose values, and a faster direct
readout. Storage phosphor plates, for example, can be used with all existing
projection radiography units. Flat detectors have excellent imaging capabilities
and can achieve very high image quality, even at low doses. The immediate
availability of images is another advantage of this technique. The increased
cost of the new technology is compensated by increased patient throughput.
Successful use of this technology in economic terms requires not only
replacement of the previous imaging system, but adaptation of the workflow
and patient management.
Digital imaging differs significantly from film/screen radiography in that the

digital detectors’ imaging capabilities are greater and the individual processes
of image acquisition, image processing and image documentation are
separated from one another.
Choosing suitable technique parameters will often provide additional

diagnostic information, whereas an unsuitable choice can have negative
outcomes right down to misdiagnosis. Optimal use of digital methods in clinical
practise requires optimisation and quality assurance of the acquisition, post-
processing, and image documentation processes.
Imaging methods have to solve diagnostic problems with sufficient quality and
certainty. The full spectrum of imaging methods (from ultrasound to MRI and
CT) has to be included in the discussion of suitable indications for projection
radiography. It must also be taken into account that alterations to the spectrum
of applications can change the value of a given imaging method from that of a
final diagnostic tool to that of a screening method, or a starting point for further
diagnosis.
In recent years, film/screen radiography has been characterised by the search

for optimal image quality, independent of the specific clinical problem. In
striving for such image quality, no effort was made to minimise the dose value,
5
other than ensuring it did not exceed a maximum level. The advantage of
pursuing the strategy of attempting to attain the highest image quality is the
fact that aside from specific diagnostic information, other useful and
unexpected additional information can also be obtained. The use of digital
imaging methods can also mean that dose values can be reduced over a
broad range -- sometimes with decreases in image quality, and sometimes
without. Dose and quality management now uses this potential along the
guidelines of the ALARA principle (As Low As Reasonably Achievable) to
implement the strategy of setting image quality and dose to fit the clinical
situation (“quality as good as necessary, not as good as possible – dose value
as low as possible”).
A clear definition of indications for Digital Projection Radiography should be

the starting point for image quality and dose management. The criteria and
methods of evidence-based medicine should form the basis for the definition of
guidelines. These guidelines must take into account the clinical situation and
the risk to the patient. The aim is to avoid unnecessary, “routine” exposures as
well as reduce the dose per image. Two publications can be considered a
starting point for further discussion here, namely “Referral Guidelines for
Imaging”, published by the European Commission (European Commission –
Radiation protection 118 – ISBN 92-828-9452-5), and the Council Directive
97/43/ Euratom on protecting the health of individuals against the dangers of
ionising radiation in relation to medical exposure. These directives include
purpose and scope, definitions, justification, optimisation and responsibilities.
The choice of specific imaging method and exposure parameters must be

fixed against diagnostic image quality requirements. Examples of these
requirements can be found in “European guidelines on quality criteria for
diagnostic radiographic images” (European Commission – EUR 16260EN), the
results of the European research project DIMOND II and the guidelines of the
German Bundesärztekammer (German Federal Medical Association), for
example.
One parameter that can be varied over a broader range in digital radiography
than in film/screen radiography is the exposure dose value. Past experience
and research demonstrate that for flat detectors, the speed class of 800 is
sufficient for nearly all indications. The dose value of the speed class 200 is no
longer used. For many indications, speed class 1600 is suitable. The storage
phosphor technique’s potential for dose reduction is significantly lower.
Part of the DIMOND III project involved developing a three-level concept of

image quality (high, medium, low) and dose management, which lead to
certain dose values (for example, the speed classes 400, 800, and 1600)
being assigned to certain quality levels, depending on the specific imaging
method.
6
Examples of the high, medium, and low quality classes for skeletal exposures
are:
High: Diagnosis of a fracture/fissure

Medium: Control of the position of fractures
Low: Control of complete metal explantation,
adaptation of hip prostheses,
metric measurements of the spine
Indications for these three dose level bands have been defined corresponding
to the European referral guidelines. Clinical problems have been assigned to
quality levels in “extended” referral guidelines(Chapter IV). Afterwards the
recommended image quality level was transferred to a choice of suitable
parameters for the specific imaging method.
One example resulting out of a comparative study (see chapter VI) is shown in
the following diagram:
Relation between image quality classes and dose for various on imaging
methods (speed class)
Image quality class
High Medium Low

Flat detector (400) Flat detector (800) Flat detector (1600)
Stor.Phos. (200/400) Stor.Phos. (400) Stor.Phos. (800)
Film/screen (200) Film/screen (400) Film/screen (800)
There must be intensive debate on the strategies and methods for optimising
and standardising image quality in the future. Although many individual studies
(see references) describe interesting results, they are missing a methodical
framework.
This new concept consists of three steps:
1. Optimisation (use clinical criteria)

2. Objectivation (description with phantom exposures)
3. Standardisation (defined bandwidth of image quality)
As a first approach, exposure parameters (focus size, voltage, grid, and so on)
and projection can be chosen as for film/screen exposures. Future research
projects on digital radiography will exhaust the existing potential for further
optimisation.
7
Past experience shows that optimal post-processing is not independent of the
dose range. Therefore, optimisation strategies depend on the organ field, the
clinical problem and the required image quality class. Optimisation must be
carried out according to clinical criteria. Experience of many years shows that
optimisation according to “physical test phantoms” is not possible. However,
test phantom images (such as CDRAD2.0) are well suited for objectivation and
standardisation of image quality.
Standardisation of the various manufacturers’ post-processing algorithms and

unit generations is not possible. Therefore, a “black box” model has to be
taken as a starting point. A “black box” model can be described by the relation
of the output function (image) to the input function (image). A strategy of
quality management leads to necessary optimisation for each organ program
with different units in reference centres. The results can then be objectivated
by phantom exposures. The range of results, which correspond to a suitable
image quality, is defined by these reference centres. It is also possible to
determine the image quality of images produced by unknown equipment by
evaluating phantom exposures under the same exposure conditions.
A digital test pattern should be integrated into a corner of the raw images to
demonstrate the influence of individual post-processing (in additional to
standard post-processing). This test pattern should combine high and low
contrast structures. The same kind of post-processing should be applied to
both the whole image and the test pattern. The results of this individual post-
processing can then be demonstrated on laser film or on a monitor.
An important part of quality assurance is constancy testing, something that

digital radiography offers new ways of performing. Image quality can be
graded by digital parameters (e.g. signal/noise – see chapter V , annex3). A
subjective visual evaluation is no longer necessary. The starting point for the
evaluating the image quality of documentation on laser films or monitors
should be a defined digital test pattern (e.g. SMPTE). Digital imaging units are
often interlinked and offer new potential for central constancy testing using
personal computers.
8
The Future Aspects for “Digital Projection Radiography”
1. There will be an extensive replacement of conventional film/screen

radiography by digital radiographic systems in the near future.
2. Both storage phosphor plates and flat detectors are well suited to the
spectrum of digital projection radiography. Flat detectors have better
imaging capabilities, storage phosphor systems lower costs.
3. The capabilities of digital imaging methods differ significantly from

film/screen radiography. This demands a new choice of parameters and
regulations for equipment standards, as well as guidelines and
recommendations for use.
4. Image quality and dose management strategies must be based on

European and international standards and guidelines. It is anticipated that
national standards and guidelines will soon become obsolete.
5. Image quality and dose management strategies must take into account the
complete diagnostic chain, from the referral indication to the quality of
diagnosis.
6. The application of the ALARA principle means that the image quality
should be as good as necessary, not as good as possible. That means that
the dose value should be as low as possible and consistent with the clinical
objective.
7. No significant differences exist between digital radiography and film/screen

radiography in terms of the definition of diagnostic image criteria.
8. Image optimisation must take place according to clinical criteria. However,

“physical” test phantoms are well suited to objective evaluation and
standardisation of image quality.
9. Economically efficient application of new techniques (e.g. flat detectors)

not only requires the implementation of the new digital detector, but the
adaptation of the workflow and patient management
10. The new technological capability that digital projection radiography

provides demands strategies for optimisation and standardisation.
Discussion of keywords such as radiation protection, dose reduction, and
image quality and dose management can be heard and this demands the
definition of new paradigms.
9
11. The development of new strategies of optimisation and standardisation is a
challenge for the next years.
10
Chapter II

- Principles, Imaging and Application -
The application of the digital image intensifier technique for projection

radiography started at the beginning of the 1980s after the introduction of
digital subtraction angiography (DSA) for fluoroscopic imaging units. This new
technique was used primarily for examining the gastrointestinal tract, for
myelograms, and for arthrographies. In the following years, the clear
advantages of this technique (high image quality, low dose values, direct
availability, easy handling) led to its nearly completely replacing film/screen
technology for fluoroscopic units (1, 2).
In the mid-1980s, the storage phosphor technique came into clinical

application as a new imaging method for exposures at the wall stand, Bucky
table and for bedside imaging. The very high technical requirements and
financial costs, limited image quality and difficult handling without a reduction
of examination time delayed the transfer into routine clinical use, which started
to increase at the beginning of the 1990s. Today, storage phosphor
radiography makes an important contribution to obtaining optimal imaging in all
fields of projection radiography.
In recent years, the progress of technology has allowed direct acquisition of

the image information at the detector (flat detector technology). After first being
successfully tested in manufacturers’ research laboratories, the new
technology made the transition into clinical use (3, 4).
From a radiologist’s point of view, the aim of further developing imaging

systems is to improve diagnostic information while simultaneously decreasing
dose value. Additionally, easy handling, high exposure frequency and a
favourable cost-benefit ratio are also important.
11
New or improved methods of digital projection radiography must fulfil the
following criteria:
1. High image quality (spatial resolution, contrast detectability, dynamic range,

homogeneity)
2. Low dose value (high sensitivity for x-ray quanta)
3. Simple and fast handling
4. Integration into existing x-ray units and workflows
5. Integration into PACS/RIS-systems
6. Favourable cost/benefit relationship
Principles of imaging systems for digital projection radiography
Film/screen imaging is a well-proven system for projection radiography. Its

advantages include the high image quality, “simple” exposure technique and
favourable cost benefit relationship. Its disadvantages are its limited dynamic
range, difficulties in post-processing, relative high dose values, uncomfortable
handling and limited availability of x-ray films. The exposed film is
simultaneously a detector and a medium for archiving and demonstration,
which makes it fundamentally different in principle to digital techniques, which
require separate steps.
The digital imaging chain consists of three independent steps: image

acquisition, image processing and image display. The main differences
between the specific digital projection radiography systems are in detector
technology and signal processing. Optimising image processing and display
on laser films or monitors is a problem common to all systems.
12
In the following, five techniques are described:
1. Image intensifier (II) technique

2. Storage phosphor technique
3. Selenium drum technique
4. CCD (charged coupled device) technique
5. Flat detector technique
5.1. Direct conversion (selenium)
5.2. Indirect conversion (scintillator)
1. Digital image intensifier technique

The image intensifier (II) converts the latent absorption image of x-rays into a
reduced size image at the II exit. The steps involved in this process are the
entrance screen (Caesium iodine) and photocathode steps, which convert x-
ray quanta first into light quanta and then into electrons. An electrical field
leads these electrons to a screen at the exit of the II, upon which the electrons
are converted into light again (1). A television camera records this screen over
tandem optics. If the camera consists of a conventional television tube (e.g.
Vidicon, Saticon) the information is transferred as an analogue signal to an
analogue/digital converter. The image information is recorded directly by a
semiconductor image detector (CCD), digitised, and transferred to a computer
system. CCD detectors have a lower inertia compared to conventional
television tubes and are both more tolerant against burn-out effects and have
very homogenous imaging capabilities. While technical progresses in II
technology have been only slowly forthcoming due to physical and financial
limitations, there have been improvements in pulsed fluoroscopy and
shuttering without radiation dose over the last few years, as well as the
introduction of the matrix size 2048X2028 (5, 6, 7).
2. Storage phosphor radiography

Storage phosphor radiography involves exposing phosphor plates instead of x-
ray films. Within the layers of the storage phosphor plate, electrons are
elevated to a higher energy level by the absorption and deposition of energy
by x-ray quanta. The way these electrons are distributed corresponds to the
intensity distribution of the radiation. A laser beam scans the plates in a
separate readout unit. During this process, electrons drop down to a basic
level and emit light, which is registered by a photomultiplier. After analogue
signals have been converted to digital values, a numerical intensity value is
assigned to every point of the image. Homogenous light irradiation then erases
the remaining image information off the plates.
13
3. Digital radiography with the selenium drum technique
Here, the detector consists of a rotating drum covered with selenium.
Selenium is a semiconductor which becomes conductive by absorbing x-ray
radiation. The selenium layer of the drum is positively charged homogeneously
before exposure. During the exposure, charges are equalised, that is, the
positive charge is reduced. This creates a charge pattern on the surface that is
proportional to the radiation pattern. The charge pattern is scanned during
drum rotation and converted to digital image information (8).
4. Digital CCD technique

X-ray quanta is converted into light on a fluorescent entrance screen. The
emitted light is then transferred to four CCD-cameras (9), which convert the
optical image into digital information. A CCD detector uses the property of
selenium to convert light into moveable charge carriers. The light image is
stored as a charge image in a number of pixel-elements. After exposure, the
“charge” image is read as a sequence of signals and, after conversion from
analog to digital, is stored as digital information.
5.1. Flat detectors with direct conversion (selenium)

Flat detectors allow the direct conversion of x-ray quanta into charges
(electrons). A voltage of 6 kV is passed through a 500 µm thick layer of
selenium. A matrix of image elements (pixels) lies behind the layer. Every pixel
contains an electrode for charge detection, a charge capacitor and a transistor.
In the selenium layer, x-ray quanta are converted to charges, transferred to the
electrode, and stored in the capacitor. If the element is selected, the charge is
transferred to an analogue/digital converter. A digital image is then generated
by line-by-line readout (10, 11, 12, 13).
5.2. Flat detector with indirect conversion (scintillator)
A glass plate is coated with a layer of amorphous silicon (thickness 500 µm).
The structure is as a matrix of silicon photodiodes similar to the CCD detector.
A switching transistor selected by a control line is connected to the element
(pixel). A scintillator layer with caesium iodine is in front of the silicon
elements. Within this layer the x-ray quanta are converted to light quanta. The
needle-structured caesium iodine crystals focus the light quanta to the detector
elements. The charge of the specific photodiode is read out and transferred to
an analogue/digital converter (14, 15).
14
Imaging capabilities of systems for digital projection radiography
The imaging capabilities of digital radiography are determined by the

characteristics of detectors, signal processing (amplification, digitalisation),
digital post-processing and documentation (16, 17, 18).
The parameters characterising digital systems are:
Spatial resolution
Dynamic range
Quantum efficiency (DQE)
Modular transfer function (MTF)
Spatial resolution:
The parameter spatial resolution describes the detectability of small, high-
contrast objects. Spatial resolution is usually demonstrated by a lead bar
pattern and measured as linepairs per millimetre (lp/mm). Scattering inside the
detector has a significant influence on resolution, similar to film/screen
exposures. For analogue systems, the maximum resolution is determined by
the evaluation of the image of the lead bar pattern. The cut-off resolution is
defined as 4% of the maximal amplitude of the modulation transfer function
(MTF). For digital systems, there exists an additional limitation by pixel size.
The scanning formula means that for a pixel size of a, the maximal transfer
frequency can be 1/2a. A pixel size of 0.25 mm, therefore, would lead to a cut-
off frequency (Nyquist cut-off) of 2 lp/mm. If this limitation is exceeded by a
lead bar pattern image, an aliasing pattern will result, with a reduction of image
quality.
Dynamic range:
The dynamic range is determined by the dose range, which can be imaged
without the effects of underexposure or overexposure. Digital systems have a
linear correlation between dose value and signal amplitude. Conventional
film/screen systems have only a narrow dynamic range for imaging that is
characterized by the linear perform of s-shaped density curve. A large dynamic
range is connected to a broad range for the exposure parameter dose (e.g.
bedside imaging (reduction of the number of retakes!)). It allows simultaneous
imaging of large absorption differences (e.g. bone/soft tissue:
mediastinum/parenchyma of the lung) within only one exposure (reduction of
dose value!).
15
Detective quantum efficiency (DQE):
DQE describes the efficiency of x-ray quanta-to-signal conversion. The DQE
depends on the dose value and the spatial resolution (object size) and is
influenced by the quantum noise and the noise of the detector. An ideal
detector has a DQE=100%. A high DQE is correlated with a high effective use
of x-ray quanta. This results in a reduction of dose value without loss of image
quality.
Modulation transfer function (MTF):

Imaging means demonstrating different sized objects with different attenuation
of x-rays. A description of imaging of contrast and object size can be done
using the modular transfer function. The MTF shows how the contrast of
different sized objects are demonstrated by the image. The MTF should be
high in the range of object size between 0 and 2 lp/mm, because this range
contains the medically-relevant details. The MTF’s special parameters are cut-
off resolution (resolution at 4% of the maximal amplitude) and specific
modulation at fixed resolution (e.g. 60% for 1 lp/mm). The reference point is
100% for 0 lp/mm (that is, a homogenous background). Characteristic
modulation underlines that the shape of the MTF is more important in the
range of 0 – 2 lp/mm than the cut-off frequency because, again, this range
frequently contains structures relevant to diagnosis.
Image quality:
The image quality at the end of the imaging chain (laser film, monitor) is
determined by the parameters already described, with the addition of digital
post-processing. Post-processing can lead to increased image quality and
additional diagnostic information. Image processing should result in improved
detail contrast, edge enhancement, reduction of dynamic range and noise
(19).
Digital Image Intensifier Technique
The diameter of the entrance surface is between 15 cm and 40 cm. With a

matrix size of 1024X1024 (or 2048X2048), the pixel size is between
0.15(0.07)mm and 0.4(0.2)mm. The II entrance dose for a diameter of 17cm is
about 2,2 µGy, comparable to film/screen speed class 400, and for a 40cm
entrance screen the dose is about 0.4 µGy. The dynamic range (1:100) is
larger than for film/screen (1:30). The DQE is 60% (70kV). Alongside single
exposures, image intensifier technology can also be used for fluoroscopy.
Fluoroscopy should guarantee sufficient image quality for controlling catheter
16
position, projections and vessels. The II entrance dose for fluoroscopy is 0.02
µGy/sec (3 pulses/sec, II 27cm) and 0.17 µGy/sec (continuous beam, II 27
cm). Pulsed fluoroscopy, image integration (recursive filtering) and collimation
without radiation is the major potential of dose reduction.
Digital storage phosphor radiography
In storage phosphor radiography, pixel size and spatial resolution depend on

the entrance size. A format of 35X43cm (matrix 1769x2140) has a pixel size of
0.2 mm and a cut off resolution of 2.5 lp/mm, whereas a format of 18x24cm
(1770x2370) results a pixel size of 0.1 mm and a cut-off frequency of 5.0
lp/mm. The dynamic range is 1:40000 and the quantum efficiency 25% (70 kV,
0 lp/mm). This means that, in comparison to film/screen radiography for
storage phosphor radiography, no significant reduce of dose value is possible.
The large dynamic range, the high contrast amplification and a limited spatial
resolution are characteristic of storage phosphor radiography. The method of
post-processing also has significant role in attaining high image quality of
storage phosphor radiography (19, 20).
Digital radiography with the selenium drum technique
The selenium drum technique is used in dedicated systems for the lung.
(Philips: Thoravision). For an image format of 35x43 cm (matrix: 2166x2448),
the pixel size is 0.2mm and the corresponding cut-off frequency 2.5 lp/mm.
The dynamic range is higher than 1:10.000. The DQE of 60% (60 kV) points to
the possibility of dose reduction from that of film/screen (DQE 20%).
Digital CCD technique
The number of pixels of a CCD camera is 1024x1280. The sum of 4 CCD

targets is therefore 4096x5120 for an entrance field of 34 x 43 cm. The pixel
size is 169 µm and the cut-off resolution 3 lp/mm. CCD systems have a
dynamic range of about 1:4000 and a DQE of 40%. Decreased signal-to-noise
ratio is attributed to the transfer of light from the scintillator surface to the CCD
detector.
Flat detector with direct conversion (selenium)
For a detector area of 35x43 cm (matrix 2560x3072), the pixel size is 0.139
mm and the cut-off resolution 3.6 lp/mm. The DQE is 43% (70kV, 0 lp/mm)
and the dynamic range >1:10.000.
17
Flat detector with indirect conversion (scintillator)
Flat detectors manufactured by Trixel (a joint venture by Thomson, Siemens,

and Philips) have a detector area of 43x43 cm (matrix 3120x3120). The pixel
size is 0.143 mm and the cut-off resolution 3.5 lp/mm. The dynamic range is
>1:10.000 and the DQE 60% (70kV, 0 lp/mm). The readout time is 1.25 sec.
The General Electric silicon detector has a pixel size of 0.2 mm (cut-off
resolution 2.5mm) and a DQE of 80% (70 kV, 0 lp/mm). The first successful
fluoroscopy applications for silicon detectors are now been described. With an
entrance field of 20cm, (Matrix 1024x1024) the pixel size is 0.2mm. The
detector has an image frequency of 25 images/second in continuous mode.
For medium and high dose levels, image quality is considered to be higher in
comparison to II radiography. For low dose values, however, there is lower
signal-to-noise ratio when compared to II radiography.
Clinical applications of digital projection radiography
Today, image intensifier, storage phosphor and selenium drum radiography all
play an important role in projection radiography as it is used every day.
The advantages of digital imaging techniques are:
1. Large dynamic range (no retakes, imaging of great differences of

attenuation)
2. High contrast detectability
3. The potential for dose reduction
4. Increased image quality through post-processing
5. Immediate availability of image information
6. Integration into PACS/RIS systems (storage, transfer, availability)
The digital image intensifier technique
The digital imaging technique has completely replaced film/screen radiography

for fluoroscopic units. This is because the digital technique has several
advantages over film/screen radiography, despite the lower detail resolution
when compared to film/screen radiography (21). The future is likely to bring
only limited innovation because of the boundaries of physics and financial
resources. The increase of matrix size from 1024 to 2048 is advantageous, but
it is difficult to prove this in every-day clinical use.
18
There are similar problems associated with the increase in matrix size from
2000 to 4000 for storage phosphor radiography (22). Increasing the matrix size
makes sense only if it results in increased diagnostic information. The image
quality of state-of-the-art II units should be considered to be sufficient because
of its being able to be applied to all existing clinical situations. Further
innovation is likely to focus on reducing the dose (pulsed fluoroscopy,
collimation without radiation) and improving handling rather than improving the
image quality itself.
Because the digital image intensifier technique can already be integrated into
PACS systems, there is no strong need for new detectors in fluoroscopic units
when compared to projection radiography as a whole. But the first applications
of flat detectors for single exposures and fluoroscopy show promising results.
Flat detector fluoroscopy’s potential of improving image quality is greater than
its potential to reduce dose. The signal-to-noise ratio is still lower in the low
dose range in comparison to II units (23). It is, however, likely that image
intensifiers will be replaced by this flat detector technique in the future.
Storage phosphor technique
As can be read in current literature (24, 25) and the results of a consensus
conference (26), the storage phosphor technique can be applied to all clinical
situations requiring projection radiography. One excellent field of application is
bedside imaging without automatic exposure control. The large dynamic range
guaranties a constantly high image quality. However, early expectations for a
significant reduction of dose value (in comparison to film/screen) didn’t hold.
The necessary dose level is between the speed class 200 and 400. Handling
of the storage phosphor technique is disadvantageous, because exposed
cassettes have to be brought to a central reading unit and erased cassettes
have to be carried back to the exposure room. Storage phosphor radiography
is the only method for bedside imaging because all other detector techniques
are restricted to a fixed examination room. Currently, the only way to conduct
horizontal beam imaging at Bucky tables is the storage phosphor technique.
An alternative to this would be C-arm units, which are not marketed by all
manufacturers. A flat detector which can be read out like a mobile cassette is
expected in the near future. However, innovation to the detective material of
storage phosphor plates (with needle configuration) and a faster readout by
detection of the information line by line could result in improved image quality
and handling. First attempts to read out the plate just at the exposure unit
could lead to competition with flat detectors (that is, in terms of cost).
19
Selenium drum technique
Clinical studies have shown that, based on the better imaging capabilities of
selenium detectors, the selenium drum technique is more favourable for lung
units than the storage phosphor technique (22). The literature describes the
ways in which doses can be reduced. One disadvantage of the selenium drum
technique is that the units for selenium technique can be used only for thoracic
imaging in an upright position, requiring a high frequency of these
examinations. It is assumed that flat detectors for Bucky tables and wall stands
will play a greater role in the future and will probably replace the selenium
drum technique.
CCD technology
Until now, only a limited number of reports of clinical results for this technology
have been published. Based on its imaging capabilities, CCD technology
probably produces similar image quality to storage phosphor radiography for
the same dose value. A significant reduction of dose cannot be expected
because of the limited DQE. CCD technology does, however, have the
advantageous of no cassettes having to be carried.
Flat detector technique
Flat detectors with amorphous selenium have the advantage of a direct

conversion of x-ray quanta to electrical charge. This can lead to very high
image quality. Selenium flat detectors offer the potential for high image quality
and/or significant dose reduction.
Flat detectors with scintillators demonstrated higher image quality and a lower
dose level compared to film/screen and storage phosphor radiography (27, 28,
29, 30).
Digital radiography - a new challenge
The new potential of digital projection radiography can be discussed from the
standpoint of a single stand-alone unit without integration into a digital system
or from the standpoint of a necessary component for PACS (picture archiving
and communication system) systems. For the near future, the view from a
stand alone unit will remain more common. The option of integration into a fully
digital system, with its many advantages (transfer, archiving, post-processing)
is an important aspect of this new digital technique. X-ray units for projection
20
radiography are part of radiologists’ tools; image quality thus has to fulfil the
diagnostic demands – not more and not less. An increase of image quality
beyond these clinical demands is useless as it costs money and also requires
higher doses. New discussion of the quality of new imaging methods must be
based on the 100 years of experience with film/screen radiography, 25 years
of experience with digital image intensifier radiography and 20 years of
experience with storage phosphor radiography that the industry has. A suitable
spectrum of indications has been discussed in a large number of publications
and a consensus conference. Image quality is defined by physical
measurements (spatial resolution, MTF, DQE, homogeneity) and clinical
studies comparing different methods. Initial results of the new flat detector
systems demonstrate the potential to increase image quality and to decrease
dose values. It cannot be foreseen whether this will be connected to a
broadening of the spectrum of referral indications.
There is a close correlation between dose and image quality. The commonly
known image quality and radiation protection requirements are leading to new
strategies for dose reduction. One way of achieving this could be a higher
image quality with the same dose; another way the same image quality with a
lower dose value. In the past, the focus has been more on image quality;
today, however, dose reduction, easy handling and a favourable cost/benefit
relation are important decision criteria. Better image quality is only useful if it
has consequences on the outcome of therapy and follow-up and reduces the
number of additional examinations.
Discussion of the financial aspects is difficult in light of the high purchase and
service costs and the need for specific IT infrastructure. The costs for a flat
detector system are three times as high as a film/screen Bucky table with wall
stand. For a stand alone unit, the potential of cost reduction lies in time and
personnel costs (increase of throughput, reduction of personnel costs) and in a
reduction of film costs (smaller film size, no retakes, lower number of films).
The reduction of patient risk cannot, however, be calculated in terms of
economics. The cost/benefit relation of a PACS system is also very difficult to
calculate and largely depends on the requirements placed on the examinations
themselves, the system and personnel.
Further clinical experience must now show whether this new potential will lead
to a new examination workflow and to a new spectrum of referral indications.
Higher costs demand an intensive discussion of cost/benefit relation. The
physical imaging capabilities of flat detectors point to the possibility of a
successful introduction into every-day clinical world in the near future.
21
References:
[1] Busch HP, Lehmann KJ, Freund MC, Georgi (1991). Digitale
Projektionsradiographie. Röntgenpraxis 44: 329-335
[2] Busch HP, Lehmann KJ, Freund MC, Georgi M (1992). Digitale
Projektionsradiographie: Klinische Anwendungsmöglichkeiten.
Röntgenpraxis 45: 35-43
[3] Frija G (1999) Flat Panel Sensors: Questions and Answers. Medical
Imaging Techhnology, Vol 17, 2: 99 – 104
[4] Chotas HG, Dobbins JT, Ravin CE(1999). Principles of digital
radiography with large-area, electronically readout detectors: A review of
the basics. Radiology 210: 595 – 599
[5] Vetter S, Heckmann H, Strecker EP, Busch HP, Kamm KF,
Allmendinger H(1998). Klinische Aspekte zu Bildqualität und Dosis bei
gittergesteuerter gepulster Durchleuchtung. Akt.Radiol. 8: 191 – 195
[6] Vetter S, Faulkner K, Strecker EP, Busch HP (1998) Dose reduction and
image quality in pulsed fluoroscopy. Radiation Protection Dosimetry. 80:
299 – 301
[7] Busch HP, Hoffmann HG, Kruppert H, Mörsdorf M(1997). Digitale BV-
Radiography – Eine Methode hat sich durchgesetzt. Electromedica 65:
62 – 64
[8] Neitzel U (1993). Selenium: a new image detector for digital chest
radiography. Medica Mundi 38: 89 – 93
[9] Gotsch K (1998) Röntgendiagnostik ohne Film. F&M 106: 448-450
[10] Lee DL, Cheung LK, Jeromin LS, Palecki E (1996). Imaging
performance of a direct digital radiographic detector using selenium and
a thin film transistor array. In: Lemke HU(ed) Computer assisted
Radiology. Elesevier: 41 – 46
[11] Lee DL, Cheung LK, Jeromin LS (1995). A new digital detector for
projection radiography. SPIE 2432: 237 – 249
[12] Shaber GS, Maidment ADA, Bell J, Jeromin LS, Lee DL, Powell GF
(1997) Full field digital projection radiography system: Principles and
image evaluaten. In: Lemke HU (ed) Computer assisted radiology and
surgery. Elsevier: 1 – 7
[13] Yaffee MJ, Rowlands JA (1997) X-ray detectors for digital projection
radiography. Phys Med Biol: 1 - 39
[14] Chaussat Ch, Chabbal J, Ducourant Th, Spinnler V, Vieux G, Neyret R
(1998) New CsJ/a-Si X-ray flat panel detector provides superior
detectivity and immediate direct digital output for General Radiography
systems.
SPIE 3336: 45 – 56
[15] Neitzel U(1999) Integrated digital radiography with a flat –panel sensor.
Medical Imaging Technology 17,2: 123-129
[16] Neitzel U (1998) Grundlagen der digitalen Bildgebung. In Ewen K(Hrsg)
Moderne Bildgebung. Georg Thieme: 71 – 76
22
[17] Neitzel U(1998) Systeme für die digitale Bildgebung. In: Ewen K (Hrsg)
[18] Kamm KF (1998) Grundlagen der Röntgenabbildung. In: Ewen E (Hrsg)
[19] Prokop M, Schaefer-Prokop CM(1997) Digital image processing. Eur
Radiol 7(Suppl3): 73 – 82
[20] Vuysteke P, Schoeters E (1994) Multiscale image contrast amplification
(MUSICA). SPIE 2167: 551 – 560
[21] Lehmann KJ, Busch HP, Georgi M (1992) Digitale Bildverstärker
Radiographie – welche Aufnahmedosis für welche Fragestellung? Akt
Radiol 2:11 – 15
[22] Schaefer-Prokop CM, Prokop M (1996) Digitale Radiographie des
Thorax – der Selendetektor im Vergleich zu anderen
Abbildungssystemen.
Kontraste: 14 – 22
[23] Bruijns TJC, Alving PL, Bury R, Cowen AR, Jung N, Luijendijk HA,
Meulenbrugge HJ, Stouten HJ (1998) Technical and clinical results of an
experimental FLat Dynamic (digital) X-ray image Detector (FDXD)
system with real-time corrections. SPIE 3336: 33 – 44
[24] Busch HP (1997) Digital radiography for clinical application. Eur Radiol 7
(Suppl 3): 66 – 72
[25] Leitlinien der Bundesärztekammer zur Qualitätssicherung in der
Röntgendiagnostik (1995) Deutsche Ärzteblatt 92, 49: 1691 – 1703
[26] Busch HP, Klose KJ, Braunschweig R, Neugebauer E (1999) Digitale
Radiographie – Ergebnisse einer Anwendungsumfrage und einer
Konsensuskonferenz. Akt.Radiol. 7: 56 – 63
[27] Hamers S, Freyschmidt J (1998) Digital radiography with an electronic
flat-panel detector: First clinical experience in skeletal diagnostics.
Medica Mundi 42,3: 2 – 6
[28] Reiff KJ (1999) Flat panel detectors – closing the (digital) gap in Chest
and skeletal radiology. European Journal of Radiology (in press)
[29] Strotzer M, Gmeinwieser J, Völk M, Fründ R, Feuerbach S (1999)
Digitale Flachbilddetektortechnik basierend auf Cäsiumjodid und
amorphem Silizium: Experimentelle Untersuchungen und erste klinische
Ergebnisse. Fortschr Röntgenstr 170: 66 – 72
[30] Strotzer M, Gmeinwieser J, Völk M, Fründ R, Seitz J, Manke C, Albricht
H, Feuerbach S (1998) Clinical application of a flat-panel X-ray detector,
based on amorphous silicon technology: image quality and potential for
radiation dose reduction in skeletal radiography. Amer J Roentgenol
171: 23 – 27
This article was translated from a publication in German language:

H.P.Busch: Digitale Projektionsradiographie - Technische Grundlagen,
Abbildungseigenschaften und Anwendungsmöglichkeiten.
Radiologe 1999 – 39: 710 – 724.
23
Chapter III
Diagnostic Requirements for Digital Projection

Radiography
Based on:
1. European Guidelines on Quality Criteria

for Diagnostic images (EC)
2. Diagnostic Requirements for
Digital Radiographic Procedures (DIMOND II )
3. Leitlinien der Bundesärztekammer zur
Qualitätssicherung in der Röntgendiagnostik – Qualitätskriterien
röntgendiagnostischer Untersuchungen (the German Federal Medical
Association's Quality Assurance Guidelines for Radiological Diagnostics:
Quality criteria for radiological examinations)
24
Content:
1) Chest/lungs and heart PA projection, wall stand
2) Chest/lungs and heart Lateral projection, wall stand
3) Chest/lungs and heart AP projection, bedside
4) Cervical spine AP projection, grid table or vertical stand
5) Cervical spine Lateral projection, grid table or vertical stand
6) Thoracic spine AP projection, grid table or vertical stand
7) Thoracic spine Lateral projection, grid table or vertical stand
8) Lumbar spine AP projection, grid table or vertical stand
9) Lumbar spine Lateral projection, grid table or vertical stand
10) Pelvis AP projection, grid table or vertical stand
11) Skull AP projection, grid table or vertical stand
12) Skull Various projections, grid table or vertical stand
13) Extremity I Various projections, grid table or vertical stand
Hip
Thigh
14) Extremities II Various projections, grid table or vertical stand
Shoulder
Upper arm
Ribs
Sternum
Knee joint
Lower leg
15) Extremities III Various projections, grid table or vertical stand
Elbow
Forearm
Ankle
Tarsus
16) Extremities IV Various projections, grid table or vertical stand
Hand
Fingers
Foot
Toes
17) Abdomen AP projection, grid table or vertical stand
25
Chest/lungs and heart - PA projection, wall stand
Image criteria
Performed at full inspiration (as assessed by the position of the ribs above the
diaphragm - either 6 anteriorly or 10 posteriorly) and with suspended
respiration
Symmetrical imaging of the thorax, as shown by the central position of a
spinous process between the medical ends of the clavicles
Medical border of the scapulae to be outside the lung fields
Visualisation of the whole lung including the costophrenic angles
Clear visualisation of the lung structure and vascular pattern throughout the
lung fields including the retrocardiac area
Clear delineation of vertebral disc spaces
Suggested radiographic technique

Radiographic device: Vertical stand with moving grid
Nominal focal spot value: ≤ 1.3 mm
Total filtration: ≥ 3.0 mm Al
Anti-scatter grid: r = 10; 40/cm
FFD: 180 cm (140-200)
Radiographic voltage: 125 kV
Automatic exposure control: Both lateral chambers
Exposure time: < 20 ms
Imaging technique: Flat detector, storage phosphor

radiography
26
Chest/Lungs and Heart - lateral projection, wall stand
Image criteria
Performed at full inspiration and with suspended respiration
Arms should be raised clear of the thorax
Superimposition of the posterior lung borders
Visualisation of the trachea
Visualisation of costophrenic angles
Visually sharp reproduction of the posterior border of the heart, aortic arch,
mediastinum, diaphragm, sternum and thoracic spine

Radiographic device: Vertical stand with moving grid
Total filtration; ≥ 3.0 mm Al
FFD: 180 cm (140-200)
Radiographic voltage: 125 kV
Automatic exposure control: Central chamber

radiography
27
Chest/ Lungs and Heart - AP projection, bedside
Image criteria
Imaging of the whole rib cage above the diaphragm and both hemidiaphragms
at full inspiration (if possible)
The mediastinum should be sufficiently penetrated to visualize trachea and
major bronchi.
Visualisation of venous catheters, draining catheters… including the tip
Visually sharp reproduction of the vascular pattern

Radiographic device: Mobile unit
FFD: 100 cm
Radiographic voltage: >100 kV
Automatic exposure control: No
Alternative
Anti-scatter grid: No
FFD: 100 cm
Radiographic voltage: 80 - 90 kV
Imaging technique: Storage phosphor radiography
28
Cervical Spine - AP projection
Image criteria
Complete imaging of the cervical spine, including the upper cervical spine and
the 7th vertebra, if necessary with additional spotfilm
Visually sharp imaging, as a single line, of the upper and lower-plate surface in
the centred beam area
Visualisation of the intervertebral joints and the spinous processes
Visually sharp imaging of the cortical and trabecular structures

Radiographic device: Grid table or vertical stand with moving grid
FFD: 115 cm (100-150)
Radiographic voltage: 65-75 kV

radiography
29
Cervical Spine - lateral projection
Image criteria
Complete imaging of the cervical spine, including the upper cervical spine and
the 7th vertebra
Visualisation of the intervertebral spaces, intervertebral joints and spinous
processes
Visualisation of the soft tissues, particularly the retrotracheal space

FFD: 115 cm (100-150)

radiography
30
Thoracic Spine - AP projection
Image criteria
Complete imaging of the thoracic spine, including Th1
Visually sharp imaging of the pedicle, spinous processes and costovertebral
joints

Radiographic device: Table or vertical stand with moving grid
FFD: 115 cm (100-150)

radiography
31
Thoracic Spine - lateral projection
Image criteria
Complete imaging of the thoracic spine from T2 down to the thoracolumbar
junction
Visualisation of the intervertebral spaces and intervertebral joints in the
centred beam area

FFD: 115 cm (100-150)

radiography
32
Lumbar Spine - AP projection
Image criteria
Complete visualisation of the lumbar spine and sacrum
Visually sharp imaging, as a single line, of the upper and lower-plate surfaces
in the centred beam area
Visualisation of the intervertebral spaces in the centred beam area
Visually sharp imaging of the pedicles, transverse processes, spinous
processes and intervertebral joints
Visualisation of the sacroiliac joints

FFD: 115 cm (100-150)

radiography
33
Lumbar Spine - lateral projection
Image criteria
Complete visualisation of the lumbar spine and lumbosacral junction
Visualisation of intervertebral spaces, intervertebral joints and spinous
processes

FFD: 115 cm (100-150)

radiography
34
Pelvis - PA projection
Image criteria
Symmetrical imaging of both sides of the pelvis
Visually sharp imaging of the necks of the femora – should not be distorted by
foreshortening or rotation
Visually sharp imaging of the pubic and ischial rami and the sacroiliac joints
Visually sharp imaging of the cortial and trabecular structures including the
trochanters
Visually sharp imaging of the sacrum and its intervertebral foramina

FFD: 115 cm (100-150)

radiography
35
Skull - PA projection
Image criteria
Symmetrical reproduction of the skull, particularly cranial vault, orbits and
petrous bones
Projection of the apex of the petrous temporal bone into the centre of the orbits
Visually sharp reproduction of the frontal sinus, ethmoid cells and apex of the
petrous temporal bones and the internal auditory canals
Visually sharp reproduction of the outer and inner lamina of the cranial vault

Nominal focal spot value: 0,6 mm
Total filtration: ≥ 2,5 mm Al
FFD: 115 cm (100-150)

radiography
36
Skull - lateral projection
Image criteria
Visually sharp reproduction of the outer and inner lamina of the cranial vault,
the floor of the sella, and the apex of the petrous temporal bone
Superimposition respectively of the contours of the frontal cranial fossa, the
lesser wing of the sphenoid bone, the clinoid processes and the external
auditory canals
Visually sharp reproduction of the vascular channels, the vertex of the skull
and the trabecular structure of the cranium
Superimposition of the mandibular angles and ascending rami

Nominal focal spot value: 0,6 mm
FFD: 115 cm (100-150)

radiography
37
Extremities I - Various projections, grid table or vertical stand
Hips
Thighs
Image criteria
Visualisation of typical structures of compacta and spongiosa
Imaging of the joints in typical projections
Visually sharp reproduction of the cortical joint surface

Nominal focal spot value: ≤ 1,3 mm
FFD: 115 cm (100-150)

radiography
38
Extremities II - Various projections, grid table or vertical stand
Shoulder
Upper arm
Ribs
Sternum
Knee joint
Lower leg
Image criteria

FFD: 115 cm (100-150)

radiography
39
Extremities III - Various projections, grid table or vertical stand
Elbow
Forearm
Ankle
Tarsus
Image criteria

Radiographic device: Grid table
FFD: 105 cm (100-150)
Automatic exposure control: Off

radiography
40
Extremities IV - Various projections, grid table or vertical stand
Hands
Fingers
Feet
Toes
Image criteria

Radiographic device: Table
FFD: 105 cm (100-150)
Automatic exposure control: Off

radiography
41
Abdomen - PA projections, grid table or vertical stand
Image criteria
Reproduction of the area of the whole abdomen from the diaphragm to the
base of the bladder
Reproduction of the kidney outlines
Visualisation of the psoas outlines
Visually sharp reproduction of the bones

Radiographic device: Grid table
FFD: 115 cm (100-150)
Automatic exposure control: Chamber selected – central or lateral

radiography
42
Guidelines and user-defined recommendations for the
selection of imaging parameters
Based on:
Recommendations of the EU(EU)

German Guidelines (RiLiBÄ)
Krankenhaus der Barmherzigen Brüder Trier (BKT-Trier)
University Hospital, Innsbruck (Innsbruck)
Athens Hospital (Athens)
San Carlos Hospital, Madrid (San Carlos/Madrid)
Limerick Hospital, Ireland (Limerick)
43
44
45
46
47
48
49
50
51
Chapter IV
Referral Guidelines with the Three Level Band

for Image Quality
In recent years, film/screen radiography has been characterised by the search

for optimal image quality, independent of the specific clinical problem. The only
restricting parameter was the dose value, which could not exceed a maximum
value. The advantage of pursuing a strategy of trying to obtain the highest
level of image quality is the fact that a range of unexpected, yet
complementary diagnostic information can be obtained, over and above the
information normally obtained at a lower dose. By implementing digital imaging
methods, however, doses can be reduced over a broad range – which
sometimes does decrease image quality, but often does not. Dose and quality
management realises this new potential using the ALARA principle (As Low As
Reasonably Achievable), adapting image quality and dose to fit the clinical
situation (“quality as good as necessary, not as good as possible – dose value
as low as possible”).
Quality and dose management should start with establishing a clear definition
of indications for Digital Projection Radiography. These guidelines should be
based on the criteria and methods of evidence-based medicine, taking into
account the clinical situation and the risk for to the patient. The aim is to avoid
unnecessary, routine exposures as well as to reduce the dose per image. Two
documents can be considered a starting point here, namely “Referral
Guidelines for Imaging”, published by the European Commission (European
Commission – Radiation protection 118 – ISBN 92-828-9452-5). and the
Council Directive 97/43/ Euratom on health protection of individuals against
dangers of ionizing radiation in relation to medical exposure. The latter
publication includes sections on purpose and scope, definitions, justification,
optimisation and responsibilities.
One parameter that can be varied over a broader range in digital radiography
than in film/screen radiography is the exposure dose value. Past experience
and research demonstrate that for flat detectors, the speed class of 800 is
sufficient for nearly all indications. The dose value of the speed class 200 is no
longer used. For many indications speed class, 1600 is suitable. The storage
phosphor technique’s potential for dose reduction is significantly lower.
Part of the DIMOND III project involved developing a three-level concept of

image quality (high, medium, low) for quality and dose management. Certain
dose values (for example, the speed classes 400, 800, and 1600) have been
assigned to certain quality levels, depending on the specific imaging method.
52
One example resulting out of a comparative study (see chapter VI) is shown in
the following diagram:
Image Quality classes

( ) Speed class
High Medium Low
Flat detector (400) Flat detector (800) Flat detector (1600)
Storage Phosphor Storage Phosphor Storage Phosphor

(200/400) (400) (800)
Film (200) Film (400) Film (800)
There must be intensive discussion on the strategies and methods for

optimising and standardising image quality in the future. Although many
individual studies (see references) describe interesting results, they are
missing a methodical framework.
53
Examples of quality classes high, medium and low:
Referral Guidelines for Imaging: European Commission – Radiation protection

118 – ISBN 92-828-9452-5
The column “Quality Class” has been added as a result of the DIMOND III
project
INVESTIGATION {DOSE}
XR Plain radiography one or more films
CXR Chest radiograph
Typical effective
Class Examples
Dose (mSv)
0 0 US, MRI
I <1 CXR, limb XR, pelvis XR
IVU, lumbar spine XR, NM
II 1–5
(e.g. skeletal scintigram), CT head & neck
III 5–10 CT chest and abdomen, NM (e.g. cardiac)
IV >10 Some NM studies (e.g. PET)
QUALITY CLASS
H High
M Middle
L Low
RECOMMENDATION {GRADE}
(A) Randomised controlled trials (RCTs), metaanalyses, systematic reviews; or
(B) Robust experimental or observational studies; or
(C) Other evidence where the advice relies on expert opinion and has the
endorsement of respected authorities.
54
QUALITY - DOSE
- LEVEL RECOM-
INVESTI-
CLINICAL MEN-
GATION COMMENT
Madrid
Dublin
Athen
PROBLEM DATION
Trier
{DOSE}
{GRADE}
A. Head
Especially for those who have worked with
Orbits
XR orbits (I) H H H H Indicated (B) metallic materials, power tools, etc. Some centres
Metallic FB (before MRI)
use CT. (see Trauma Section K for acute injury).
XR skull, sinus C Not indicated Radiography of little use in the absence of focal
Headache: chronic M M M M
spine (I) routinely (B) signs/symptoms. See A13 below.
Pituitary and Not indicated

SXR (I) . M H H M Patients who require investigation need MRI or CT
juxta-sellar problems routinely (C)
Hydrocephalus XR L L L L Indicated (C) XR can demonstrate whole valve system.
Not indicated Thickened mucosa is a non-specific finding and

Sinus disease Sinus XR (I) M M M-H M
routinely (B) may occur in asymptomatic patients.
Dementia and memory

Not indicated Consider investigation if clinical course unusual or
disorders, first onset SXR (I) M M L-M M
routinely (B) in younger patient.
psychosis
Especially for those who have worked with

Orbits
XR orbits (I) H H H H Indicated (B) metallic materials, power tools, etc. Some centres
Metallic FB (before MRI)
use CT. (see Trauma Section K for acute injury).
Not indicated Plain XRs rarely contributory. Specialists may

Visual disturbances SXR (I) M M M M
routinely (C) require CT or MRI.
Evaluation requires specialist expertise. Late

Not indicated onset seizures should normally be investigated
Epilepsy (adult) SXR (I) M M M M
routinely (B) but imaging may be unnecessary if clearly
alcohol-related.
B. Neck (for the spine see Sections C [The spine] and K [Trauma])
Soft tissues
Specialised Radiographs will demonstrate bony abnormalities,
Temporo-mandibular joint
XR (I) M H M-H M investigation but are normal in great majority, as problems are
dysfunction
(B) usually related to articular disk dysfunction.
C. The spine
General (for trauma see Section K)
Specialised
e.g. Full-length standing radiograph for scoliosis.
Congenital disorders XR (I) L L L L investigation
See Section M for back pain (M10).
(C)
Cervical spine
A single lateral cervical spine XR with the patient
in subluxation supervised comfortable flexion
should reveal any significant subluxation in
Possible atlanto-axial
XR (I) M M M-H M Indicated (C) patients with rheumatoid arthritis, Down’s
subluxation
Syndrome, etc. MRI (flexion/extension) shows
effect on cord when XR positive or neurological
signs present.
55
Degenerative changes begin in early middle-age
and are often unrelated to symptoms which are
Neck pain, brachalgia, Not indicated
XR (I) M M M M usually due to disk/ligamentous changes
degenerative change routinely (B)
undetectable on plain XR. MRI increasingly being
used, especially when brachalgia is present.
Thoracic spine
Degenerative changes are invariable from middle-
age. Examination rarely useful in the absence of
Not indicated neurological signs or pointers to metastases or
Pain without trauma:
XR (I) M M M M routinely (B) infection. Consider more urgent referral in elderly
degenerative disease
onwards. patients with sudden pain to show osteoporotic
collapse or other forms of bone destruction.
Consider NM for possible metastatic lesions.
Lumbar spine
Degenerative changes are common and non-
Chronic back pain with no
Not indicated specific. Main value in younger patients (e.g.
pointers to infection or XR (II) M M M M
routinely (C) younger than 20, spondylolisthesis, ankylosing
neoplasm
spondylitis, etc.) or in older patients e.g. >55.
Acute back pain is usually due to conditions which

cannot be diagnosed on plain XR (osteoporotic
Acute back pain: disk
collapse an exception). ‘Normal’ plain XRs may be
herniation; sciatica with Not indicated
XR (II) M M M M falsely reassuring. Demonstration of disk
no adverse features (see routinely (C)
herniation requires MRI or CT and should be
above).
considered immediately after failed conservative
management.
D. Musculoskeletal system
The 2–3 phase skeletal scintigram is more
sensitive than XR. However, findings are not
XR (I) + NM (II) specific and further specialised NM with
Osteomyelitis M H M M Indicated (B)
or MRI (0) alternative agents may be needed. Fat-
suppressed MRI is becoming regarded as the
optimal investigation.
Primary bone tumour XR (I) H H H H Indicated (B) XR may characterise the lesion.
Known primary tumour. Skeletal survey

H H H H Not indicated routinely (C)
Skeletal metastases (II)
Bone pain XR (I) M M M M Indicated (B) Local view of symptomatic areas only.
For staging and identifying lesions which may

Skeletal survey
Myeloma M M M M Indicated (C) benefit from radiotherapy. Survey can be very
(II)
limited for follow-up.
Biochemical tests usually suffice. If needed, this

Skeletal survey Not indicated should be limited (e.g. hands, CXR, pelvis and
Metabolic bone disease H H H H
(II) routinely (C) lateral lumbar spine). Bone densitometry may be
needed. (see D9).
Localised XR to establish cause of local pain or

Osteomalacia XR (0) H H H H Indicated (B)
equivocal lesion on NM.
Lateral views will demonstrate compression

fractures. NM or MRI more useful in distinguishing
between recent and old fractures and can help
XR (II) lateral
Pain — osteoporotic exclude pathological fractures. Bone densitometry
thoracic and M M M M Indicated (B)
collapse (dual energy XR absorptiometry (DEXA) or
lumbar spine
quantitative CT) provides objective measurements
of bone mineral content; can also be used for
metabolic bone disease (see D7, D8).
XR (I) affected May be helpful to determine cause although

Arthropathy, presentation H H H H Indicated (C)
joint erosions are a relatively late feature.
56
In patients with suspected rheumatoid arthritis,
XR (I) hands/feet H H H H Indicated (C) XR feet may show erosions even when
symptomatic hand(s) appear normal.
XR (II) multiple Not indicated

H H H H
joint(s) routinely (C)
Not indicated XRs needed by specialists to assist management

Arthropathy, follow-up XR (I) M M M M
routinely (C) decisions.
Degenerative changes in the acromio-clavicular

Not indicated
Painful shoulder joint XR (I) M M M M joints and rotator cuff are common. Earlier XR if
routinely (C)
soft tissue calcification is expected.
A normal NM study excludes most late

Painful prosthesis XR (I) + NM (II) M M M-H M Indicated (B) complications. Further specialised NM studies can
help distinguish loosening from infection.
May help in investigation of sero-negative

SI joint lesion XR SI joints (II) H H H H Indicated (B) arthropathy. SI joints usually adequately
demonstrated on AP lumbar spine.
XR only if symptoms and signs persist or complex

Not indicated
Hip pain: full movement XR pelvis (I) M M M M history (e.g. chance of avascular necrosis, see
routinely (C)
D20)
Symptoms often transient. XR if hip replacement

Hip pain: limited Not indicated might be considered or symptoms persist. PET
XR pelvis (I) M M M M
movement routinely (C) may be helpful, if XR, MRI standard NM all
normal.
Hip pain: avascular

XR Pelvis (I) M H M M Indicated (B) Abnormal in established disease.
necrosis
Symptoms frequently arise from soft tissues and

Knee pain: without
Not indicated these will not be demonstrated on XR. OA
locking or restriction in XR (I) M M M M
routinely (C) changes common. XRs needed when considering
movement
surgery.
Knee pain: with locking,

restricted movement or XR (I) M M M M Indicated (C) To identify radio-opaque loose bodies.
effusion (loose body)
Specialised
Hallux valgus XR (I) L L L L investigation For assessment before surgery.
(C)
Plantar spurs are common incidental findings. The

cause of the pain is seldom detectable on XR. US,
Plantar fasciitis — Not indicated
XR (I) M M M-H M NM and MRI are more sensitive in showing
calcaneal spur routinely (B)
inflammatory change but the majority of patients
can be managed without imaging.
E. Cardiovascular system
CXR must not delay admission to a specialised
unit. CXR can assess heart size, pulmonary
oedema, etc. and can exclude other causes.
Central chest pain Department film preferable. Subsequent imaging
CXR (I) M M M M Indicated (B)
myocardial infarction involves specialised investigations (NM, coronary
angiography, etc.) and depend on local policy. NM
offers myocardial perfusion and ventriculography
data. Increasing E1 interest in MRI.
Chest pain: aortic

CXR (I) M M M M Indicated (B) Mainly to exclude other causes; rarely diagnostic.
dissection: acute
Pericarditis — pericardial May be normal; effusion volume/effect not

effusion determined.
Suspected valvular CXR (I) and Used for initial assessment and when there is a
M M M M Indicated (B)
cardiac disease cardiac US (0) change in the clinical picture.
57
Follow-up of patients with Only if signs or symptoms have changed, when
Not indicated
heart disease or CXR (I) M M M M comparison with the CXR obtained at presentation
routinely (B)
hypertension may be helpful.
F. Thoracic system
Conditions such as Tietze’s disease show no
Not indicated
Non-specific chest pain CXR (I) M M M M abnormality on CXR. Main purpose is
routinely (C)
reassurance.
Not indicated Showing a rib fracture after minor trauma does not
Chest trauma CXR (I) M M M-H M
routinely (C) alter management (see Trauma Section K).
Not justified except in a few high-risk categories

Pre-employment or (e.g. at risk immigrants with no recent CXR).
CXR (I) M M M M Not indicated
screening medicals Some have to be done for occupational (e.g.
divers) or emigration purposes (UK category 2).
Exception before cardio-pilmonary surgery, likely

admission to ITU, suspected malignancy or
possible TB. Anaesthetists may also request
Not indicated CXRs for smokers, dyspnoeic patients, those with
Pre-operative CXR (I) M M M M
routinely (B) known cardiac disease and the very elderly. Many
patients with cardio respiratory disease have
recent CXR available; a repeat CXR is then not
usually needed.
Upper respiratory-tract Not indicated

CXR (I) M M M M
infection routinely (C)
Chronic obstructive
Not indicated
airways disease or CXR (I) M M M M Only if signs or symptoms have changed.
routinely (B)
asthma; follow-up
To confirm clearing, etc. Pointless to re-examine

Pneumonia adults: follow-
CXR (I) L L M L Indicated (A) at less than 10-day intervals as clearing can be
up
slow (especially in the elderly).
Small effusion can be missed, especially on a

Pleural effusion CXR (I) M M M M Indicated (B)
fronta CXR.l
Haemoptysis CXR (I) M H M M Indicated (B) PA plus lateral view.
A CXR is most helpful when there has been a

change in symptoms or insertion or removal of a
ITU/HDU patient CXR (I) M M M M Indicated (B)
device. The value of the routine daily CXR is
being increasingly questioned.
G. Gastrointestinal system
Gastrointestinal tract
CXR may be sufficient, unless localisation for
Oesophageal perforation CXR (I) M M M M Indicated (B)
surgical repair is planned.
Acute GI bleeding: Not indicated

AXR (II) L L L L Of no value.
haematemesis routinely (B)
Decubitus AXR to show free air if CXR supine.

Acute abdominal pain- CXR (I) (erect) Supine AXR usually sufficient to establish
L M M L Indicated (B)
perforation- obstruction and AXR (II) diagnosis and point to an anatomical level of
obstruction.
Inflammatory bowel
AXR (II) L M M-H L Indicated (B) Often sufficient for evaluation.
disease of colon
General abdominal problems
58
Local policy will determine strategy. Supine AXR
Acute abdomen pain;
(for gas pattern, etc.) is usually sufficient. Erect
(warranting hospital AXR (II) plus
L M M L Indicated (B) AXR not indicated routinely. Increasing use of CT
admission and surgical erect CXR (I)
as a ‘catch- all’ investigation here. US widely used
consideration)
as a preliminary survey.
Not indicated
Palpable mass AXR (II) L L M L
routinely (C)
Many normal adults show extensive faecal

material; although this may be related to
Not indicated prolonged transit time it is impossible to assess
Constipation AXR (II) L L L L
routinely (C) significance on AXR alone. But AXR can help
certain specialists (e.g. geriatricians) in refractory
cases.
Biliary disease, (e.g. Not indicated

AXR (II) M M M M Plain XRs only show about 10 % of gallstones.
gallstones) routinely (C)
Unless diagnosis in doubt; then AXR needed to

exclude other causes of acute abdomen pain (see
Not indicated
Pancreatitis: acute AXR (II) M M M M G19). Some patients presenting with acute
routinely (C)
pancreatitis have underlying chronic pancreatitis
which may cause calcification evident on AXR.
Pancreatitis: chronic AXR (II) M M M M Indicated (B) To show calcification.
H. Urological, adrenal and genito-urinary systems

Hypertension (without
Not indicated IVU is insensitive for renal artery stenosis. See
evidence of renal IVU (II) M M M M
routinely (A) H3.
disease)
Imaging should be performed as an emergency

xamination whilst the pain is present, as
radiological signs disappear rapidly after passage
IVU (II) or US (0)
of a stone. Delayed films (up to 24 hrs) may be
Renal colic, loin pain and AXR (II) or M M M M Indicated (B)
needed to show the site of obstruction. A plain
CT (III)
AXR on its own is of little value. Both CT and US
are increasingly being used, especially in those
with contraindications to contrast medium.
Not indicated CT or MRI preferable for further evaluation. NM

Renal mass AXR (II) + IVU (II) M M M M
routinely (C) may be needed to determine relative function.
US can also assess upper tract and bladder

Not indicated
Prostatism IVU (II) M M M M volumes before and after voiding, preferably with
routinely (B)
flow rates. It can also show bladder calculi.
US for diagnostics of upper urinary tract (after

Not indicated
Urinary retention IVU (II) L L L L catheterisation and relief of bladder distension),
routinely (C)
particularly if urea levels remain raised.
I. Obstetrics and gynaecology

NB: Transvaginal (TV) US equipment should be available in all departments
performing pelvic US
Not indicated
Lost IUCD AXR (II) M M ? M Unless IUCD is not seen in uterus on US.
routinely (C)
The need for pelvimetry is increasingly being

questioned. Local policy should be determined in
Suspected cephalopelvic Not indicated agreement with obstetricians. Furthermore MRI or
XR (II) Pelvimetry L L L L
disproportion routinely (B) CT should be used wherever possible. MRI is best
as it avoids x-irradiation. CT generally offers a
lower dose than standard XR pelvimetry.
59
K. Trauma
Head: general
Head: low risk of intracranial injury
• Fully orientated
• No amnesia These patients are usually sent home with head
• No neurological defects Not indicated injury instructions into the care of a responsible
SXR (I) M M M M
• No serious scalp routinely (C) adult. They may be admitted to hospital if no such
laceration adult is available.
• No haematoma
Head: medium-risk of intracranial injury

• Loss of consciousness
or amnesia
• Violent mechanisms of
injury
• Scalp bruise, swelling or
CT is increasingly being used first and ONLY to
laceration down to bone
exclude cranial injury. If no fracture is seen,
or > 5 cm
patients will usually be sent home with head injury
• Neurological symptoms
instructions into the care of a responsible adult. If
or signs (including
no responsible adult is available or if a fracture is
headache, vomiting twice
CT (II) or M M M M Indicated (B) present, the patient will usually be admitted. See
or more, return visit)
Section M (M13) for non-accidental injury in
• Inadequate history or
children. MRI of the brain is the preferred
examination
investigation for intracranial injuries in NAI, but
(epilepsy/alcohol/child/etc
SXR may still be needed to exclude fractures
.)
missed on CT.
• Child below 5 yrs:
suspected NAI, tense
fontanelle, fall of more
than 60 cm or on to hard
surface
Head: very high risk of intracranial injury

Unless requested by a specialist. Poor correlation
between radiological findings and presence of
SXR (I) XR facial
Not indicated external deformity. Management of the bruised
Nasal trauma bones (I), XR H H H H
routinely (B) nose will depend on local policy: usually follow-up
nasal bones (I)
at an ENT or maxillo-facial clinic will determine the
need for XR.
Especially in those where ‘blow-out’ injury

Orbital trauma: XR facial bones possible MRI or low dose CT may eventually be
H H H H Indicated (B)
blunt injury (I) required by specialists, especially when XRs or
clinical signs equivocal.
When: (1) Radio-opaque intra-ocular FB is a

Orbital trauma: possibility (see A16). (2) Investigation requested
XR orbits (I) M M H M Indicated (C)
penetrating injury by ophthalmologist. (3) Suspicion of damage to
orbital walls.
But patient cooperation essential. Advisable to

XR facial bones
Middle third facial injury M M M M Indicated (B) delay XR in uncooperative patients. In children,
(I)
XR often unhelpful.
XR Mandible (I)
Mandibular trauma or orthopantomo- M M M M Indicated (C) For non-traumatic TMJ problems see B11.
gram (OPG) (I)
Cervical spine
In those who meet all of the following criteria:
Conscious patient with
Not indicated (1) Fully conscious (2) Not intoxicated (3) No
head and/or face injury XR C spine (I) M M M M
routinely (B) abnormal neurological findings.
only
(4) No neck pain or tenderness.
Must be of good quality to allow accurate

Unconscious head injury evaluation. But radiography may be very difficult in
XR C spine (I) H H H H Indicated (B)
(see K3/4) severely traumatised patient and must avoid
manipulation (see K11 also K12).
60
Cervical spine XRs can be very difficult to
evaluate. Radiography also difficult
1. Must show C7/T1.
Neck injury: with pain XR C spine (I) H H H H Indicated (B) 2. Should show odontoid peg (not always possible
at time of initial study)
3. May need special views, CT or MRI especially
when XR equivocal or complex lesions.
Neck injury: with

XR (I) H H H H Indicated (B) For orthopaedic assessment.
neurological deficit
Neck injury: with pain but Views taken in flexion and extension (consider
XR C spine; Specialised
XR initially normal; fluoroscopy) as achieved by the patient with no
flexion and M M M M investigation
suspected ligamentous assistance and under medical supervision. MRI
extension (I) (B)
injury may be helpful here.
Thoracic and lumbar spine

Physical examination is reliable in this region. If
Trauma: no pain, no Not indicated
XR (II) M M M M the patient is awake, alert and asymptomatic, the
neurological deficit routinely (B)
probability of injury is low.
A low threshold to XR when there is

Trauma: with pain, no
pain/tenderness, after a significant fall, if a high
neurological deficit or XR painful area
M M M M Indicated (B) impact RTA or other spinal fracture present or it is
patient not able to be (II)
not possible to clinically evaluate the patient.
evaluated
Increasing use of CT and MRI here.
Trauma: with neurological

XR (II) M M M-H M Indicated (B)
deficit - pain
Pelvis and sacrum

Physical examination may be unreliable. Check
for femoral neck fractures, which may not show on
Fall with inability to bear XR pelvis (I) plus
M M M M Indicated (C) initial XR, even with good lateral views. In
weight lateral XR hip (I)
selected cases NM or MRI or CT can be useful
when XR normal or equivocal.
To show urethral integrity, leak, rupture. Consider

Urethral bleeding and Retrograde
M M M M Indicated (C) cystogram if urethra normal and suspicion of
pelvic injury urethrogram (II)
bladder leak.
Trauma to coccyx or Not indicated Normal appearances often misleading and

XR coccyx (I) M M M M
coccydynia routinely (C) findings do not alter management.
Upper limb
Some dislocations present subtle findings. As a
minimum, orthogonal views are required. US, MRI
Shoulder injury XR shoulder (I) M M M M Indicated (B)
and CT arthrography all have a role in soft tissue
injury.
Indicated in cases of effusion with no obvious

Elbow injury XR elbow (I) M M M M Indicated (B) fracture’ (see also Section M). Increasing use of
CT and MRI here.
Scaphoid fractures can be invisible at

Indicated (B) presentation. Most centres repeat XR at 10–14
XR wrist (I) NM Specialised days if there are strong clinical signs and initial XR
Wrist injury M M M M
(II) or MRI (0) investigation negative. Some departments use CT, NM or MRI
(B) to exclude fracture earlier than this. Increasing
use of MRI as the only examination.
Lower limb
Especially where physical signs of injury are
minimal. Inability to bear weight or pronounced
Knee injury Not indicated bony tenderness, particularly at patella and head
XR knee (I) M M M M
(fall/blunt trauma) routinely (B) of fibula, merit radiography. CT/MRI may be
needed where further information is required (see
D23).
61
Features which justify XR include age ( elderly
Not indicated
Ankle injury XR ankle (I) M M M M patients), malleolar tenderness, marked soft tissue
routinely (B)
swelling and inability to bear weight.
Not indicated unless there is true bony

tenderness. Even then the demonstration of a
fracture rarely influences management. Only
Not indicated
Foot injury XR foot (I) M M M M rarely are XRs of foot and ankle indicated
routinely (B)
together; both will not be done togetherwithout
good reason. Clinical abnormalities are usually
confined to either foot or ankle.
Stress fracture XR (I) M M M M Indicated (B) Although often unrewarding.
Provides a means of early detection as well as

NM (II) or MRI (0) M M M M Indicated (B) visual account of the biomechanical properties of
the bone. Some centres use US here.
Foreign Body (FB)

All glass is radio-opaque; some paint is radio-
Soft tissue injury: opaque. Radiography and interpretation may be
FB (metal, glass, painted XR (I) H H H H Indicated (B) difficult; remove blood-stained dressings first.
wood) Consider US, especially in areas where
radiography difficult.
Soft tissue injury: Not indicated Plastic is not radio-opaque, wood is rarely radio-
XR (I) M M M-H M
FB (plastic, wood) routinely (B) opaque.
After direct examination of oropharynx (where

most FBs lodge), and if FB likely to be opaque.
Differentiation from calcified cartilage can be
Swallowed FB suspected
XR soft tissues of difficult. Most fish bones invisible on XR. Maintain
in pharyngeal or upper M M M M Indicated (C)
neck (I) a low threshold for laryngoscopy or endoscopy,
oesophageal region
especially if pain persists after 24 hours (see
K33). NB: for possible inhaled FB in children see
Section M (M23).
The minority of swallowed FBs will be radio-

opaque. In children a single, slightly over-
exposed, frontal CXR including neck should
Swallowed FB: smooth suffice. In adults, a lateral CXR may be needed in
and small (e.g. coin) addition if frontal CXR negative. Majority of FBs
that impact, do so at crico pharyngeus. If the FB
has not passed (say within 6 days), AXR may be
useful for localisation.
Most swallowed foreign bodies that pass the

Sharp or potentially oesophagus eventually pass through the
poisonous swallowed FB: AXR (II) M M M M Indicated (B) remainder of the gastrointestinal tract without
(e.g. battery) complication. But location of batteries is important
as leakage can be dangerous.
Dentures vary in radio-opacity; most plastic

Swallowed FB: large dentures are radiolucent. AXR may be needed if
object (e.g. dentures) CXR negative, as may barium swallow or
endoscopy. Lat CXR may be helpful.
Chest
Not indicated The demonstration of a rib fracture does not alter
Chest trauma: minor CXR (I) M M M M
routinely (B) management.
Frontal CXR for pneumothorax, fluid or lung

contusion. A normal CXR does not exclude aortic
Chest trauma: moderate CXR (I) M M M M Indicated (B)
injury and arteriography/CT/MRI should be
considered.
PA and/or other views to show pneumothorax,

Stab injury CXR (I) M M M M Indicated (C) lung damage or fluid. US useful for pleural and
pericardial fluid.
62
Indicated (C) In addition to CXR. Think of thoracic
Sternal fracture XR lateral M M M M Indicated (C)
spinal and sternum (I) aortic injuries too.
Abdomen (including US valuable for detecting haematoma and

AXR (II) + erect
Supine kidney). Blunt or M M M M Indicated (B) possible injury to some organs, e.g. spleen, liver.
CXR (I)
stab injury CT may be needed (see K40–K42).
Major trauma
Stabilise patient’s condition as a priority. Perform
only the minimum XRs necessary at initial
Major trauma — general C-spine XR (I), assessment. C-spine XR can wait as long as
screen on unconscious or CXR (I), pelvis M M M M Indicated (B) spine and cord suitably protected, but CT C-spine
confused patient XR (I), may be combined with CT head. Pelvic fractures
often associated with major blood loss. See Head
Injury K1–K4.
Pneumothorax must be excluded. Pelvic fractures

Major trauma — CXR (I), Pelvis
M M M M Indicated (B) which increase pelvic volume often associated
abdomen/pelvis XR (I)
with major blood loss.
Allows immediate management (e.g.

Major trauma – chest CXR (I) M M M M Indicated (B)
pneumothorax).
Especially useful to exclude mediastinal

CT Chest (III) M M M M Indicated (B) haemorrhage. Low threshold for proceeding to
arteriography.
L. Cancer
Lung
CXR PA and Lat But can be normal, particularly with central
Diagnosis M M M M Indicated (B)
(I) tumours.
Bladder
To assess kidneys and ureters for further
Staging IVU (II) M M M M Indicated (B)
urothelial tumours.
Musculoskeletal tumours
Imaging and histology complementary. Best
Diagnosis XR (I) + M M M M Indicated (B) before biopsy: See Musculoskeletal Section D.
NM needed to ensure that lesion is solitary.
M. Paediatrics
Minimise x-irradiation in children, especially those with long term problems
(for head injury in children see Trauma Section K)
CNS
Abnormal head US indicated where anterior fontanelle is open.
Specialised
appearance — Where sutures are closed/closing. MRI indicated
SXR (I) M M M M investigation
hydrocephalus — odd for older children. (CT may be appropriate if MRI
(C)
sutures not available).
Not indicated
Epilepsy SXR (I) M M M M Poor yield.
routinely (B)
Hydrocephalus —shunt
XR (I) M M M M Indicated (B) XR should include whole valve system.
malfunction (see A10)
Not indicated If persistent or associated with clinical signs refer

Headaches SXR (I) M M M M
routinely (B) for specialised investigations.
Not indicated before 5 years as the sinuses are

poorly developed; mucosal thickening can be a
Sinusitis see also A13 Not indicated
XR (I) M M M M normal finding in children. A single under-tilted
Sinus routinely (B)
OM view may be more appropriate than the
standard OM view depending on the child’s age.
63
Neck and spine — For trauma see Section K
Deformity is usually due to spasm with no
significant bone changes. If persistent, further
Torticollis without trauma XR (I) L L L L Not indicated
imaging (e.g. CT) may be indicated following
consultation.
Back pain is uncommon in children without an

Back or neck pain XR (I) L L L L Indicated (B) apparent cause. Follow-up is needed if infection is
suspected.
A common variation and not in itself significant

Not indicated
Spina bifida occulta XR (I) L L L L (even in enuresis). However, neurological signs
routinely (B)
would require investigation.
Not indicated
Hairy patch, sacral dimple XR (I) L L L L May be helpful in older children.
routinely (B)
Musculoskeletal
Local policies will apply; close clinical/radiological
liaison essential. Skeletal survey for those under
Non accidental injury —
XR (I) of two injury see Section K) years after clinical
child abuse (for head M M M M Indicated (B)
affected parts consultation. May occasionally be required in the
injury see Section K)
older child. CT/MRI of brain may be needed, even
in the absence of cranial apparent injury.
Limb injury: opposite side Not indicated

XR (I) M M L-M M Seek radiological advice.
for comparison routinely (B)
2–18 yrs: left (or non-dominant) hand/wrist only.

Premature infants and neonates: knee
Indicated at
Short stature, growth XR (I) for bone (specialised investigation). May need to be
M M M M appropriate
failure age supplemented with a skeletal survey and MRI for
intervals (B)
hypothalamus and pituitary fossa (specialised
investigations).
Gonad protection is used routinely unless shields

will obscure area of clinical suspicion. If slipped
Limp XR pelvis (I) M M M M Indicated (C)
epiphyses is likely, lateral XRs of both hips are
needed.l
XR may be normal initially. US can be helpful

Focal bone pain XR (I) M M M M Indicated (B)
particularly in osteomyelitis.
Although bony radiological changes are visible in

Osgood–Schlatter’s disease these overlap with
Osgood–Schlatter’s Not indicated
XR knee (I) M M M M normal appearances. Associated soft tissue
disease routinely (C)
swelling should be assessed clinically rather than
radiographically.
Cardiothoracic
Initial and follow-up films are indicated in the
presence of persisting clinical signs or symptoms
Not indicated
Acute chest infection CXR (I) M M M M or in the severely ill child. Consider the need for
routinely (B)
CXR in fever of unknown origin. Children may
have pneumonia without clinical signs.
Children with recurrent chest infection tend to

have normal CXRs (apart from bronchial wall
Recurrent productive Not indicated
CXR (I) M M M M thickening). Routine follow-up CXR not indicated
cough routinely (C)
unless collapse present on initial CXR. Suspected
cystic fibrosis requires specialist referral.
History of inhalation often not clear. Bronchoscopy

is indicated, even in the presence of a normal
Inhaled FB (suspected) CXR. NM/CT may be helpful to show subtle air
(see Section K) trapping. Wide variation in local policy about
expiratory films, fluoroscopy, CT and NM
(ventilation scintigraphy).
64
Children with asthma usually have normal CXR
Not indicated apart from bronchial wall thickening. Sudden
Wheeze CXR (I) M M M M
routinely (B) unexplained wheeze CXR indicated, may be due
to inhaled FB (above).
Not indicated Epiglottitis is a clinical diagnosis, but consider FB

Acute stridor XR neck (I) M M M M
routinely (B) (above).
Not indicated Specialist referral may be needed; cardiac US

Heart murmur CXR (I) M M M M
routinely (C) often may be indicated.
Gastrointestinal — see also Section G for more general abdominal problems

Local policies require close paediatric, radiological
and surgical liaison. Where expertise is available,
Intussusception AXR (II) M M M M Indicated (C)
both US and contrast enema (air or barium) can
confirm diagnosis and guide reduction.
Except for sharp or potentially poisonous FBs, e.g.

Swallowed FBs (see Not indicated batteries. See Section K. If there is doubt whether
AXR (II) M M M M
Section K) routinely (C) the FB has passed, an AXR after 6 days may be
indicated.
US may be used as initial investigation but CT is

more specific, particularly in visceral trauma. XRs
Not indicated may show bone injury in severe trauma. The
Minor trauma to abdomen AXR (II) M M M M
routinely (C) principles for the investigation of major trauma in
children similar to those in adults (see Major
Trauma, K40–K42).
Many normal children show extensive faecal

Not indicated material; impossible to assess significance of
Constipation AXR (II) M M M M
routinely (C) radiological signs. But AXR can help specialists in
refractory cases.
Uroradiology
Both examinations may be needed to evaluate
Continuous wetting IVU (II) L L L L Indicated
duplex system with ectopic ureter.
GLOSSARY
Classification of the typical effective doses of ionising radiation from common imaging
TABLE
procedures
Typical effective Dose

Class Examples
(mSv)
0 0 US, MRI
I <1 CXR, limb XR, pelvis XR
IVU, lumbar spine XR, NM (e.g. skeletal scintigram), CT
II* 1–5
head & neck
III 5–10 CT chest and abdomen, NM (e.g. cardiac)
IV >10 Some NM studies (e.g. PET)
* The average annual background dose in most parts of Europe falls in Band II.
Quality Dose Level

L Low
M Middle
H high
65
Chapter V
Optimisation and Quality Assurance

for Digital Radiography
Diagnostic quality is determined by the justification for the examination itself

(evidence based medicine), the optimal choice of imaging parameters, the
quality of the imaging chain, the results being displayed optimally on monitors,
and the quality of the diagnostic evaluation by experienced radiologists.
Quality assurance has to guarantee a constant high quality of the overall
imaging and diagnostic procedure. An important part of this procedure is the
quality of the imaging equipment.
There are a number of published discussions on strategies and methods for

optimisation and standardization of image quality (see references). Many of
these studies describe interesting results, but lack a methodical framework.
The new concept can be considered in terms of four steps:
1. Optimisation (use clinical criteria)

2. Objectivation (“thumbprint” — description with phantom exposures)
3. Standardisation (defined bandwidth of image quality)
4. Quality assurance (Constancy testing)
Past experience shows that optimal post-processing is not independent of the

dose range. Therefore, optimisation strategies depends on the organ field, the
clinical problem and the demanded image quality (dose) class. Any
optimisation has to be done by clinical criteria. Several years of experience in
this field has shown that optimisation by “physical test phantoms” only is not
possible (although test phantom images such as CDRAD 2.0 are well suited to
objectivation (“thumbprints”) and standardisation of image quality). A QA
protocol has been developed for different digital detectors that uses the new
potential of numerical and quantitative evaluation of image quality that digital
technology provides (such as signal/noise relation). The protocol is based on
the German standards for constancy testing, DIN 6868-58 and DIN 6868-13,
but also includes digital evaluation of additional parameters.
66
Optimisation Strategies
Any optimisation strategy should be designed with consideration to the
diagnostic requirements of a given clinical situation. The referral guidelines
mentioned in chapter IV are an excellent starting point for developing such
optimisation strategies. The ideal clinical practice strategy should, in the first
instance, lead to a reduction in the number of patients referred for investigation
and, therefore, to a reduction in radiation exposure to patients. Additionally,
the choice of imaging method must be made on the basis of evidence-based
medicine. When developing clinical pathways for digital projection radiography
as a diagnostic tool, the goal is to meet the requirements of the referring
physician – which means choosing an imaging method and adjusting
technique parameters to fit the given clinical task with the lowest risk to both
patient and staff.
After deciding which imaging method to use, the next issue is deciding the
correct image quality. This, again, depends simply on the clinical application;
on the “question” that has to be answered. Evaluating a fracture without
dislocation, for example, requires high image quality, establishing the
position of a fracture requires medium image quality, and locating a metal
object requires a low image quality. Depending on the imaging method, three
dose levels (or speed classes, if referring to conventional films/screens) can
be assigned (e.g. speed class 400, 800, 1600). This represents the
amendment of conventional radiography referral guidelines for use in digital
radiography by adding an additional parameter: the image quality class (low,
medium or high) when using digital radiography.
According to this definition, imaging parameters, including post-processing,

have to be optimised to reach the necessary quality with lowest dose. This can
be achieved only by clinical studies, not test phantom exposures. In this
sense, new optimisation strategies have to be defined for digital radiography. It
is important that optimisation includes post-processing, the results of which
depends on the different detectors in use and their exposure/dose parameters.
One example of an optimising strategy is shown in fig. 1.
67
Selection of the Organprogram (e.g. Spine lat)
Selection of the
Exposure parameters
Film/Screen exposures of
20 Patients (Reference point)
Digital Exposures of 20 patients

without technical problems
yes Image quality ok no

for at least 16 Patients?
Discussion of the
Parameter setting
Documentation
Phantom exposures
End Film/Screen +digital
Optimisation
no yes
Image Quality ok?
Exposures of 3 patients
Optimisation (small, normal, thick)
no yes
Image Quality ok?
Fig 1: Optimisation of image quality (post-processing default values)
Imaging chains that include post-processing and documentation require a tool

to both describe and then standardise the chain. The post-processing methods
used depend on the manufacturer of the equipment. User-defined default
post-processing and individual post-processing are only partly known.
Consequently, we have to handle the issue of post-processing using the “black
box” model.
Characterisation of the imaging procedure
“Black box”
Input Output
Image
post-processing
Monitor
Test image
(CDRAD)
Fig.2: The “Black box” model
This "black box” can be described in terms of the relation between an input
function and an output function. A useful input function would be a digital test
68
pattern, placed like a trojan horse within a digital image. The output function
can then be analysed in a numerical way. A more practical approach would be
to perform an exposure of a test phantom with typical organ parameters in
order to study the image quality on dedicated display monitors or on laser films
via a viewing box. A good example of this phantom is the CRRAD 2.0 phantom
(Artinis Medical Systems B.V.). Several of the squares (see below) contain a
number of cylindrical objects each, each of different diameters and depths.
These have to be detected by human observers. Exposures with different dose
values and scattering material can indicate that the system being used has low
contrast detectability. Using this phantom enables one to study the whole
imaging chain. Phantom measurements cannot optimise the imaging chain,
but can assess the imaging chain for quality control, standardization and
comparison purposes.
Fig. 3: The CDRAD 2.0 phantom
The CDRAD Phantom is a good tool for evaluating the imaging characteristics
of digital radiographic systems, including computed radiographic imaging
taken using storage phosphor systems or flat detector systems. One of the
principle concerns in using digital radiography is the potential reduction in
detail, which can have an adverse affect on diagnosis. Resolution (bar
phantom) test objects, which are used to evaluate conventional x-ray imaging
systems, are generally not appropriate for evaluating digital systems. The
CDRAD Phantom allows reliable evaluation of the detail information, with the
main disadvantage being that evaluation is time-consuming and observer-
dependant. Therefore, we developed a program that automatically evaluates
detail information (see annex 2). We first checked the sensitivity (correctly
detected hole patterns) of the evaluation, looked for for alterations caused by
post-processing parameters and the reproducibility of the results.
69
Fig. 4 shows the sensitivity (detected holes) and the standard deviation of 10
exposures in a group depending on the dose (mAs). Detected by the
automatic detection program.
140
77kV / 63mAs
120
77kV / 28mAs
100
80
77kV / 6,3mAs
60
40
20
0
0 1 2 3 4
Fig. 4: Sensitivity(correct detected holes) of CDRAD images depending on the dose.
Standard deviation of 10 exposures in a group.
Fig. 5 demonstrates the influence of different post-processing on the

evaluation of the CDRAD images.
Multi-Frequency B alance (77kV /12,5m As)
100
90
80
70
Sesitivity
60
50
40
30
20
10
0
A B C D E F G H I
M u lti-F re q u e n c y B a la n c e C u rv e s
Fig. 5: Influence of different post-processing parameters on CDRAD evaluation
70
To characterise the exposure conditions and the post-processing, a
“thumbprint” of the unit for a special organ program is necessary. Without
automatic post-processing optimisation, different local parts of an organ
exposure can be simulated by CDRAD Phantom exposures, each using
different dose values (fig. 6). Fig. 7 demonstrates examples of thumbprints of
different organ programs. Results from different reference centres are also
useful in developing guidelines by which images and post-processing can be
standardised - but this is a job for a future program.
Signal Thumbprint of a unit

Fixed mode
Dose
Without automatic optimisation by post-processing
Fig. 6: Thumbprint of a unit
71
Image quality thumbprint
Sensitivity
Hand ap. BWS ap.
100 100
90 90
80 80
70 70
60 60
% 50 % 50
40 40
30 30
20 20
10 10
0 0
0 2 4 6 mAs 8 10 12 14 0 10 20 30 mAs 40 50 60 70
Dose value
BWS lat.
Abdomen 100
100
90
90
80
80
70
70
60
60
% 50
% 50
40
40
30
30
20
20
10
10
0
0
0 10 20 30 mAs 40 50 60 70
0 5 10 15 mAs 20 25 30 35 40
Fig. 7: Image quality thumbprint of different organ programs (hand a.p., abdomen,
spine a.p., spine lat)
Individual post-processing can either increase or decrease image quality.

To demonstrate the influence of post-processing on individual images, it is
helpful to integrate a fixed digital test pattern, a form of trojan horse, into the
raw image. This digital test object could, for example, simulate a lead bar
pattern and contrast detail phantom and have similar post-processing applied
to this pattern and the whole image. Evaluation of this pattern on image
display monitors or hard copy films will give in impression of the influence of
post-processing on image quality.
Quality control
Quality control can be performed by phantom exposures that assess the whole
imaging chain. Qualitative evaluation can be obtained at the level of digital
stored images, monitors or film documentation. This can be achieved by
subjective assessment of image quality displayed on dedicated monitors or
films (for example, spatial resolution or contrast detectability) or by direct
evaluation of the digital images using computer QA programs (for example,
signal/noise).
72
Quality Assurance
Constancy testing
CR CR CR CR
Laserimager
Flat detector
Test images
SMPTE
Testpattern
Fig. 8: Central system for quality assurance and constancy testing
Digital radiography imaging chains can be divided into two parts: imaging
acquisition and documentation on monitors and laser films. The image quality
of test phantom exposures(DIN 6868-58) can be evaluated by digital
parameters such as signal/noise ratio, dynamic range or homogeneity. These
parameters can be assessed by directly analysing the digital image. Image
documentation testing can be performed through the use of digital test
patterns, such as the SMPTE test. These can be evaluated either on a monitor
or on laser film. The data should be stored and displayed as curves to
demonstrate the results over a given period of time.
Fig. 9: Test phantom DIN 6868-58 SMPTE test
In conclusion, digital projection radiography does offer new possibilities for

dose and quality management and quality control. The goal is optimal
adaptation of image quality to the diagnostic task, with a reduction of risk to
73
patients and staff through lowered dose levels. Constancy testing should keep
the image quality constantly high by giving relative measurement values with
reference to the acceptance test at the beginning. Follow ups and trends, for
example, are of great interest to constancy testing.
QA measurements:
Phantom: Test phantom DIN 6868-58

Exposure parameters:
70 kV, small focus, with grid, focus detector distance 100 cm
Field size equal to phantom size
Filtration: 25 mm Al
With automatic exposure control/
Without automatic exposure control (fixed mode)
Evaluation:
Parameters:
homogeneity
signal/noise
high contrast (lead bar pattern)
low contrast
dynamic range
The exposures has been evaluated by an automatic program (see annex 3).
An example of the results is shown in fig. 9. The quality of monitor and laser
film images has been described by the SMPTE phantom images.
74
Automatic
evaluation of
test images
Constancy testing 1.660

MA 3: Std.abwg.
1.640
• Dynamic range
1.620
Gra uwert
1.600
• Signal/Noise 1.580
• Contrast
1.560
1.540
15 1
0 2. 1
06 1
17 1
23 1
1. 01
01
01
2. 01
02 1
01
09 1
01
01
.01
26 1
01
09 1
01
01
01
01
30 1
01
1 .0
.0
.0
3 .0
3 .0
.0
.0
4 .0
.0
.0
4 .0
0
2.
2.
.03.
3.
3.
3.
3.
.04.
4.
4.
.04.
4.
02
03
03
04
04
04
.02
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
1 2.
1 6.
2 9.
1 4.
1 9.
• Homogeneity
26
29
04
07
18
05
12
19
22
12
26
Da tum
MA 3 : G r a u w e r t m it t e l/ k le in s t e r / g r ö ß t e r G r a u w e r t
• Resolution
3 5. 0 00
3 0. 0 00
2 5. 0 00
G r auw e rt mit t e l
2 0. 0 00
k el ni s t e r
g r ö ß t e r G r auw er t
1 5. 0 00
1 0. 0 00
5. 0 00
Parameters S MPTE-Te stbilder: De nsitometrie
3,5
• Density
MP 0
2,5 MP 1
MP 2
MP 3
op tische Dichte
2 MP 4
• Homogeneity
MP 5
MP 6
1,5
MP 7
MP 8
MP 9
• Contrast
1
M P 10
0,5
0
07 .01
09 .01
26 1
30 1
29 1
04 1
12 1
15 1
18 1
02 1
05 1
12 1
16 1
19 1
22 .0 1
29 1
02 1
06 1
09 1
12 1
14 .01
17 1
19 1
23 1
26 .0 1
03 1
07 1
11 1
1
2 .0
2.0
2.0
3 .0
3.0
3.0
.0
3 .0
3.0
4.0
4 .0
4.0
4.0
.0
4 .0
5 .0
5.0
5.0
1.0
.0
2 .0
.0
.0
4 .0
2
.03
.04
.01
.03
.03
.04
.04
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
26
Datum
Fig. 10: Constancy testing
When seen in context of the recommended reference values, documentation

and evaluation of the dose values used in patient examinations is very
important. For digital units such as the flat detector (Philips: DigitalDiagnost)
the dose value is stored automatically, giving a very good dose monitoring tool
(fig. 11).
75
Dose watching
generatorareadose Thorax pa Okt. 2002 - Jan 2003

12,0
10,0
8,0
uGy
6,0
4,0
2,0
0,0
Exposure
EI Thorax pa. Okt. 2002 - Jan 2003

3 000
2 500
2 000
EI
1500
1000
500
Exposure
Fig. 11: Monitoring dose values for an organ programme (“dose watching”)
76
Chapter VI
Image Quality and Dose for Digital Projection Radiography
S. Busch, Chr. Decker, C. Schilz, H.P. Busch
Abstract :
Purpose: Comparison of the imaging capabilities of storage phosphor

(computed) radiography and flat detector systems with conventional
film/screen radiography to find new strategies for image quality and dose
management, i.e. optimising image quality and dose depending on the imaging
method and clinical situation.
Material and methods: Images of a CDRAD-phantom, hand-phantom,

abdomen-phantom and lung-phantom were processed using different digital
systems (Flat detector: DigitalDiagnost (Philips); Storage phosphor: ADC-70
(Agfa), ADC-Solo (Agfa), FCRXG1 (Fuji)) and a conventional film/screen
system (Ortho Regular/HT1000G (Agfa)) with different exposure voltages
(50kV, 73 kV, 109 kV) and different speed classes (200, 400, 800, 1600). The
processed images where then evaluated.
Results: The evaluation of CDRAD images demonstrated that the flat detector
system delivered the highest contrast detectability, followed by the FCRXG1,
the ADC-Solo, and the ADC70 systems. Comparison of organ-phantom
images showed that the flat detector system in particular demonstrated
excellent image quality at low speed classes in comparison to film/screen and
storage phosphor systems.
Conclusions: Flat detector systems demonstrated the highest potential to

produce high image quality with low dose. Storage phosphor systems showed
an increase of image quality depending on the system generation, but limited
possibilities to lower dose significantly.
77
Introduction
In recent years, the range of digital image information available to radiologists
has been growing rapidly. This has been fuelled to a large extent by the
improvements to workflow brought by fast image transfer, safe storage and
images being much more easily accessible in digital form. Despite the
dramatic progress in the fields of CT and MRI, projection radiography of the
lung, skeletal and gastrointestinal tracts still form the largest areas of routine
clinical radiology use. The development of storage phosphor plates and
introduction of flat detectors has, however, significantly enlarged and improved
the spectrum of digital methods for projection radiography (1). The milestones
of reached by these new imaging methods include the significant increase of
image quality and the new potential for dose reduction (2, 3, 4).
Over a several years, there have been a number of national and international
efforts to increase the diagnostic value and to minimise side effects of imaging
with x-rays (Council directive 97/45 Euratom). The ALARA principle (dose As
Low As Resonably Achievable) means that imaging has to be done with
sufficient image quality, yet a low-as-possible dose value. This prerequires
optimised technical equipment and sufficient knowledge of its imaging
capabilities for certain dose values. This publication evaluates the relation
between image quality and dose for different generations of storage phosphor
and flat detector units in comparison to film/screen units. The results can serve
as a basis on which to develop new strategies for optimal use of these
methods (5).
Material and method:

Image quality was evaluated by images from four digital projection radiography
units and one film/screen unit.
Film/screen: Ortho Regular/HT1000G (AGFA) (speed class 400)

Storage phosphor: ADC-70 (Agfa)
ADC-Solo (Agfa)
FCRXG1 (Fuji)
Fat detector: DigitalDiagnost(Philips)
78
Images of 4 phantoms were evaluated:
1. CDRAD 2.0 phantom

2. Hand phantom
3. Abdomen phantom
4. Thorax phantom
Imaging capabilities can be demonstrated by a CDRAD 2.0 phantom. The

phantom consists of a 1cm perspex plate, which is divided into 225 squares
with cylindrical holes. The diameter decreases in the direction of the vertical
axis logarithmically from 0.3 - 8.0 mm. The columns represent the depth in a
logarithmic row from 0.3 - 8.0 mm. Each square has a specific drill pattern.
Fig. 1: CDRAD 2.0 phantom
The hand, abdomen and thorax phantoms consist of organ parts embedded in
perspex.
79
Exposure conditions
Storage phosphor plates and film/screens were exposed at a Bucky table
(Philips: Bucky Diagnost), flat detector images at a dedicated unit (Philips:
DigitalDiagnost).
The process started with a film/screen image at speed class 400. Orientated to
this fixed mAs value for the speed class 400 of film/screen, exposures of the
other phantoms were taken for speed classes 200, 400, 800, 1600. For thorax
exposures, the dose values were limited to speed 200, 400 and 800 because
of the limitation of switch time of the generator. Sequences of exposures had
been taken with the described phantoms for 50 kV, 73 kV and 109 kV.
Standard post-processing was run for 50 kV using the parameters of the hand
program, for 73 kV using the organ program “abdomen”, and for 109 kV with
the organ program “thorax”. Exposures of the CDRAD phantom with 50 kV
(hand parameters) had 3 cm of perspex as additional scatter material;
exposures with 73 kV (abdomen parameters) had 14 cm perspex, and
exposures with 109 kV (thorax parameters) had 10 cm perspex. Holes of small
diameter should be imaged in a direction nearly parallel to the central beam by
asymmetric (decentralised) positioning (19). The grey values of the
documentation on laser image films were normalised to a reference density of
1.2 +/- 0.2.
Evaluation of CDRAD exposures

Four observers evaluated the CDRAD images without knowledge of the
exposure parameters and the exposure method. After the results were input
into a computer, a processing program tested the surrounding squares and
graded a square as correctly detected if the correct drill pattern was found and
the correct pattern was identified in at least two of any four squares
immediately surrounding that square.
A mean value of these results was then taken. It was demonstrated in

diagrams for each phantom and exposure sequence (50 kV, 73 kV, 109 kV). A
standardisation of their individual evaluation of a common film/screen
exposure (speed 400) was then performed, before the mean value was
calculated. For the calculation of the relative detectability, the number of
correctly detected squares was evaluated in relation to the absolute number of
squares. Contrast detectability was evaluated by looking at the drill depth up to
which a pattern with constant diameter was found correctly.
The shape of the diagrams were difficult to compare. Therefore, the areas
under the curves (sum of drill depths) were evaluated. To take in account the
different parts of the curve, the results were shown for two ranges of diameters
(diameter 0.5 – 2.0 and diameter 2.5 – 8 mm). The lower the sum of the drill
80
depths is, the more the contrast detectability of the different imaging methods
improves.
Evaluation of the hand, abdomen and thorax phantom exposures

Hand, abdomen and thorax phantom exposures were evaluated by eight
radiologists. The attention of the observers was focused only on a selection of
special areas by blackening out the remaining regions.
Fig. 2a: Hand phantom with Fig. 2b: Abdomen phantom with
windowing for evaluation windowing for evaluation
Fig. 2c: Thorax phantom with windowing for evaluation
Fig. 2: Image of the organ phantoms

81
Image quality was graded into the quality classes (high, medium, low). The
film/screen exposure was classified as medium quality. The exposure
conditions and imaging method were unknown to the observers. Averaging
had been done by classifying the image quality to the class to which it was
graded most frequently.
Results:
Fig. 3 shows a boxplot diagram of the relative detectability for 50 kV with
speed classes 200, 400, 800 and 1600. Lines above and below show the
maximum and minimum values. The line within the box demonstrates the
median and divides it into a second and third quartile. The straight line gives
the value of film/screen images with speed class 400.
70 60
% %
60 50
50 40
40 30
DiDi FCRXG 1 Solo ADC70 DiDi FCRXG1 Solo ADC70

1I
Fig. 3a: Dose class 200(8mAs) Fig. 3b: Dose class 400(4mAs)
50 %%
42 DiDi
% %
DiDi
40
DiDi
38
40 36
34
32
30 30
28
26
20 24
DiDi DiDi FCRXG1 Solo ADC70

FCRXG1 Solo ADC70
Fig. 3c: Dose class 800 (2mAs) Fig. 3d: Dose class 1600 (1mAs)
82
Fig. 3: Boxplot diagram of relative detectability
(400 speed class of film/screen (50%))
DiDi: DigitalDiagnost - Philips (flat detector)
FCR XG1: FUJI (storage phosphor)
Solo: AGFA (storage phosphor)
ADC-70: AGFA (storage phosphor)
The following figures show the mean values of the CDRAD evaluation for 50
kV (Fig. 4), 73 kV (Fig. 5) and 109 kV (Fig. 6).
70
60
Digital Diagnost
relative Detectability (%)
50
FCR XG 1
40 ADC-Solo
30 ADC-70
Film/Folien
20
10
0
8mAs 4mAs 2mAs 1mAs
(200) (400) (800) (1600)
mAs
(speed classes)
Fig. 4: CDRAD evaluation 50 kV- relative detectability
83
60
50
relativedetectability (%)
40
Digital Diagnost
FCR XG 1
30
ADC-Solo
ADC-70
20
Film/Folie
10
0
16mAs (200) 8mAs (400) 4mAs (800) 2mAs (1600)
mAs (speed classes)
Fig. 5: CDRAD evaluation 73kV- relative detectability
60
50
relative detectability (%)
40
Digital Diagnost
FCR XG 1
30
ADC-70
20 F/F
10
0
2mAs (200) 1mAs (400) 0,5mAs (800)
mAs (speed classes)
Fig. 6: CDRAD evaluation 109 kV- relative detectability
The contrast detectability shows the depth up to which the drill pattern with
fixed diameters was detected correctly. The following diagrams show the
mean value of the results noted by all observers for the sum of depth values.
Fig. 7a shows the results for drill depth of 0.5-2.0 mm, figure 7b for drill depth
2.5 -8mm.
84
Sum of depth(mm)
35
30
25 Digital Diagnost:
FCR XG 1
20
Film/Folie
15
ADC-70
10
ADC- Solo
5
0
8mAs 4mAs 2mAs 1mAs
(200) (400) (800) (1600)
mAs
Dose (Speed Class)
Fig. 7a: Sum of drill depths 0.5-2.0mm
10
9
Sum of depth (mm)
8
7 Digital Diagnost
6 FCR XG 1
5 Film/Folie
4 ADC-70
3
ADC- Solo
2
1
0
8mAs 4mAs 2mAs 1mAs
(200) (400) (800) (1600)
mAs
Dose(Speed Class)
Fig. 7b: Sum of drill depths 2.5-8mm

Fig. 7: Contrast detectability
85
Hand, abdomen and thorax phantom exposures
The following tables show the results of the evaluation of hand, abdomen and
thorax exposures (Table 1, 2, 3).
Image quality Unit

High Digital Diagnost 200 400 800
FCR XG1 200 400
ADC-Solo 200
Medium Digital Diagnost 1600
FCR XG1 800 1600
ADC-Solo 400 800
ADC-70 200
Film/Slide 400
Low ADC-Solo 1600
ADC-70 400 800 1600
Tab. 1: Evaluation of hand images (50 kV) depending on the

speed classes 200 (6mAs) / 400 (3mAs) / 800 (1,5mAs) /
1600 (0,8mAs)
86
Image quality Unit
High Digital Diagnost 200 400 800
Medium Digital Diagnost 1600
FCR XG1 200 400
ADC-Solo 200 400
ADC-70 200 400
Film/Slide 400
Low FCR XG1 800 1600
ADC-Solo 800 1600
ADC-70 800 1600
Tab. 2 : Evaluation of abdomen images (73kV) dependant on the

speed class 200 (25mAs) / 400 (12mAs) / 800 (6mAs) / 1600 (3mAs)
Image quality Unit

High Digital Diagnost 200
FCR XG1 200
Medium Digital Diagnost 400 800
FCR XG1 400 800
Film/Slide 400
Low ADC-70 200 400 800
Tab. 3: Evaluation of thorax images (109 kV) depending on the

speed class 200 (1.8mAs) / 400 (0.9mAs) / 800 (0.5mAs)
87
Discussion:
The goal of this study is to optimise digital imaging by assessing the imaging
capabilities of different methods and systems, depending on the dose level.
Therefore, a comparison between different digital systems and a film/screen
system was performed. Exposures were taken using four different phantoms
(CDRAD phantom, hand phantom, abdomen phantom and thorax phantom).
Comparisons of the imaging capabilities of digital systems have already been

conducted using CDRAD phantoms (6, 7, 8, 9). This study involved exposures
being taken with the CDRAD phantom at 50 kV, 73 kV and 109 kV. CDRAD
images were then evaluated by four observers. To take into account the
individual detection and grading levels, the individual results were
standardised in relation to the grading of a defined film/screen exposure
(speed class 400). The boxplots clearly show that this standardisation leads to
a clear separation of the results of the specific systems.
The evaluation of relative detectability for exposures with 50 kV shows that

image quality of the flat detector was graded highest by all observers. Next in
line were the units FCRXG 1, ADC-Solo and ADC- 70. There was no overlap
in the results of the specific systems.
The relative detectability of the specific systems in relation to dose shows that
flat detectors with speed class 800 demonstrate the same detectability as the
storage phosphor system ADC 70 with speed class 200. This result means
that, at the same image quality, the flat detector system allows a dose
reduction of 75%. It should be noted, however, that the ADC-70 system is six
years old and therefore no longer the latest generation of storage phosphor
systems. The results of state of the art systems shows that development of
technology over this period has led to a significant increase in image quality.
In terms of contrast detectability of the exposure series 50 kV, the first part
of the curve (diameter 0.6-2.0mm) demonstrates the superiority of the flat
detector for all speed classes. The storage phosphor systems FCRXG1, ADC-
Solo and ADC-70 show a lower quality. The second part of the curve (diameter
2.5 – 8.0) demonstrates the the same sort of superiority of the flat detector
with the exception of the 400 speed value.
For CDRAD exposures with 73kV and 109 kV, the flat detector system is at the
top of the row for all speed classes, followed by the storage phosphor systems
FCRXG1, ADC-Solo and ADC-70.
Neitzel (6), Geijer (7) and Peer (9) used the same CDRAD phantom for their
studies of imaging capabilities of different digital imaging systems. Our results
prove the statements of these publications – that flat detector systems
demonstrates a higher image quality in comparison to film/screen systems with
a significant lower dose value. This study additionally shows that the image
88
quality of storage phosphor systems is dependent on the age (generation) of
the system. The ADC-70, the oldest system, had the worst results. The
FCRXG1, as a representative of the new generation systems, is even
equivalent to the flat detector in some cases. A comparison of flat detector
system and film/screen showed that a similar quality had been reached with
significantly lower dose.
Anatomical phantoms (hand, abdomen, thorax) were also evaluated to give

the relation to clinical image aspects. Imaging method and exposure
parameters (such as dose) were not known by the observers. Only parts of the
whole image were evaluated by the observers, in order for them to concentrate
on the same regions. The overall result of all observers was classified into the
quality classes “high”, “medium” and “low”.
Exposures of the hand were taken at the voltage of 50 kV. Observers graded
the total quality of flat detector images highest, followed by FCRXG1, ADC-
Solo and ADC 70. DigitalDiagnost, FCRXG1 and ADC solo were graded as
producing “high” image quality. DigitalDiagnost and FCRXG1 provided the
highest dose reduction (up to 75%). Film/screen images were graded as
“medium”. The ADC 70 needed double the dose to produce similar image
quality. DigitalDiagnost and FCRXG1 required only 25% of the dose compared
to film/screen. This potential for dose reduction has also been described by
other authors mentioned in the references (2, 3, 4, 10, 11, 12).
For exposures of the abdomen with a voltage of 73 kV the flat detector (speed
class 200, 400, and 800) had been graded to the quality class “high”.
Film/screen images with speed 400 were in the “medium” class. The flat
detector image with speed 1600 was also graded to this class. In this case, a
dose reduction of 75% was possible – without loss of quality. But in
comparison to film/screen, there is storage phosphor systems at speed 200
are of no improvement. Doubling the dose also gives no additional image
information. In the quality class “low”, storage phosphor images (FCRXG1,
ADC-Solo, ADC-70) with speed classes 800 and 1600 were allocated. Overall,
the results of the abdomen exposures demonstrated the high potential for
dose reduction for the flat detector. This would also confirm what has been
stated by various publications (11, 13).
For the comparison of lung exposures; the DigitalDiagnost, the FCRXG1, the
ADC 70 and the film/screen system were available. The results show that flat
detector images were graded similar to FCRXG1 images. Imaging methods in
both the quality classes “high” and “medium” had the same speed class. The
ADC 70 images were graded lower in all speed classes. These advantageous
imaging capabilities of digital systems for the thorax had been described
already by some authors (14, 15, 16, 17, 18, 19).
The results show that there is a significant gap in image quality between the
flat detector system, the storage phosphor system, and the film/screen system.
89
The flat detector demonstrated the highest potential for dose reduction for
each of the evaluated phantom images. It was also shown that the gap in
image quality between state of the art storage phosphor systems and flat
detector systems has narrowed. The ADC 70, the oldest unit, produced the
lowest quality.
In comparison to film/screen radiography, digital systems produce high image

quality with low dose values. Increased doses lead directly to significant
increases in quality. Whether this is always connected to an increase of
diagnostic information has to be decided for the specific case. The broad
range of possible dose values gives radiologists the ability to adapt image
quality and dose to the clinical situation – a strategy that has the potential to
lower the radiation dose significantly.
90
References:
1. Busch HP. Digitale Projektionsradiographie: Technische Grundlagen,
Abbildungseigenschaften und Anwendungsmöglichkeiten.
Radiologe 1999; 39: 710-724
2. Reissberg S, Hoeschen C, Kästner A, Theus U, Fiedler R, Krause U,
Döhring W. Erste klinische Erfahrung mit einem großformatigen
Flächendetektorsystem bei Aufnahmen des peripheren Skelettsystems
Fortschr. Röntgenstr 2001; 173: 1048-1052
3. Heyne JP, Merbold H, Sehner J, Neumann R, Adler R, Freesmeyer M,
Kaiser WA. Reduktion der Strahlendosis mittels
Speicherfolienradiographie am Handskelett.
Fortschr. Röntgenstr 2000; 172: 386-390
4. Hamers S, Freyschmidt J, Neitzel U. Digital Radiography with a large-
scale electronic flat-panel detector vs. screen-film radiography: observer
preference in clinical skeletal diagnostics. Eur Radiol 2001; 11: 1753-
1759
5. Busch HP, Bosmans H, Faulkner K, Peer R, Vano E, Busch S. Dose
management with new digital imaging technique. European Radiology
2002; B-0561/Scientific paper
6. Neitzel U, Böhm A, Maack I. Comparison of low- contrast detail
detectability with five different conventional and digital radiographic
imaging systems. SPIE Medical Imaging 2000; 3981: 31
7. Geijer H, Beckman KW, Andersson T, Persliden J. Image quality vs.
radiation dose for a flat- panel amorphous silicon detector: a phantom
study. Eur Radiol 2001; 11: 1704-1709
8. Harrell G, Chotas MS, Carl E, Ravin MD. Digital Chest Radiography with
a Solid-state Flat-Panel X-ray Detector: Contrast –Detail Evaluation with
Processed Images Printed on Film Hard Copy. Radiology 2001; 218:
679-682
9. Peer S, Neitzel U, Giacomuzzi S M, Peer R, Gassner E, Steingruber I,
Jaschke W. Comparison of low- contrast detail perception on storage
phosphor radiographs and digital flat- panel detector images. IEEE
2001; 20-3: 239-242
10. Ludwig K, Lenzen H, Kamm KF, Link M, Diederich S, Wormanns D,
Heindel W. Performance of a flat-panel detector in detecting artificial
bone lesions: Comparison with conventional screen-film and storage-
phosphor radiography. Radiology 2002; 222:453-459
11. Okamura T, Tanaka S, Koyama K, Norihumi N, Daikokuya H, Matsuoka
T, Kishimoto K, Hatagawa M, Kudoh H, Yamada R. Clinical evaluation of
digital radiography based on an large-area cesium iodide-amorphous
silicon flat- panel detector compared with screen-film radiography for
skeletal system and abdomen. Eur Radiol 2002; 12: 1741-1747
12. Strotzer M, Völk M, Wild T, Landenberg P, Feuerbach S. Simulated
bone erosions in a hand phantom: detection with conventional screen-
film technology vs. caesium iodide- amorphous silicon flat- panel
detector Radiology 2000; 215: 512-515
91
13. Persliden J, Beckman KW, Geijer H, Andersson T. Dose-image
optimisation in digital radiography with a direct digital detector: an
example applied to pelvic examinations. Eur Radiol 2002; 12: 1584-1588
14. Floyd CE, Warp RJ, Dobbins JT, Chotas HG, Baydush AH, Vargas-
Voracek R, Ravin CE. Imaging characteristics of an amorphous silicon
flat- panel detector for digital chest radiography. Radiology 2001; 218:
683-688
15. Hennigs S P, Garmer M, Jaeger HJ, Classen R, Jacobs A, Gissler HM,
Christmann A, Mathias K. Digital chest radiography with a large-area flat
panel silicon X-ray detector: clinical comparison with conventional
radiography. Eur Radiol 2001; 11: 1688-1696
16. Kim TS, Im JG, Lee HJ, Lee YJ, Kim SH, Kim S. Detection of Pulmonary
Edema in Pigs: storage Phosphor versus Amorphous Selenium-based
Flat –Panel-Detector Radiography. Radiology 2002; 223: 695-701
17. Schaefer-Prokop C, Eisenhuber E, Fuchsjäger M, Puig S, Prokop M.
Aktuelle Entwicklung auf dem Gebiet der digitalen
ThoraxradiographieRadiologe 2001; 41:230-239
18. Aufrichtig R. Comparison of low contrast detectability between a digital
amorphous silicon and a screen-film based imaging system for thoracic
radiography. Medical Physics 1999; 26-7: 1349-1358
19. Hermann KA, Bonel H, Stäbler A, Kulinna C, Glaser C, Holzknecht N,
Geiger B, Schätzel M, Reiser MF. Chest imaging with flat-panel detector
at low and standard doses: comparison with storage phosphor
technology in normal patients. Eur. Radiol 2002; 12: 385-390
This article was translated from a publication in German language:

S. Busch, Chr. Decker,C. Schilz, H.P. Busch: Bildqualität und Dosis in der
Digitalen Projektionsradiographie. Fortschr.Röntgenstr. 175, 32 – 37
92
Chapter VII
Digital Radiography – Quality Criteria

References
1999
Busch, H.P., Lehmann, K.J., Freund, M.C., Georgi (1991), “Digitale

Projektionsradiographie” (Digital Projection Radiography), Röntgenpraxis 44:
329-335
Busch, H.P., Lehmann, K.J., Freund, M.C., Georgi, M. (1992), “Digitale

Projektionsradiographie: Klinische Anwendungsmöglichkeiten” (Digital
Projection Radiography: Clinical application), Röntgenpraxis 45: 35-43
Lehmann, K.J., Busch, H.P., Georgi, M. (1992), “Digitale Bildverstärker

Radiographie – welche Aufnahmedosis für welche Fragestellung (Digital
Image Intensifier Radiography – What dose for what diagnostic question?)”,
Akt Radiol 2: 11-15
Neitzel, U. (1993), “Selenium: a new image detector for digital chest

radiography”, Medica Mundi 38: 89-93
Vuysteke, P., Schoeters, E. (1994), “Multiscale image contrast amplification

(MUSICA)”, SPIE 2167: 551-560
Lee, D.L., Cheung, L.K., Jeromin, L.S. (1995), “A new digital detector for
projection radiography”, SPIE 2432: 237-249
“Leitlinien der Bundesärztekammer zur Qualitätssicherung in der

Röntgendiagnostik”, (1995), Deutsche Ärzteblatt 92, 49: 1691-1703
Lee, D.L., Cheung, L.K., Jeromin, L.S., Palecki, E. (1996), “Imaging

performance of a direct digital radiographic detector using selenium and a thin
film transistor array”, In: Lemke HU (ed), “Computer assisted Radiology”,
Elesevier: 41-46
Schaefer-Prokop, C.M., Prokop, M. (1996), “Digitale Radiographie des

Thorax – der Selendetektor im Vergleich zu anderen Abbildungssystemen”,
Kontraste: 14-22
93
Busch, H.P., Hoffmann, H.G., Kruppert, H., Mörsdorf, M. (1997), “Digitale BV-
Radiography – Eine Methode hat sich durchgesetzt” (Digital II-Radiography –
A method has prevailed),.Electromedica 65: 62-64
Busch, H.P. (1997), “Digital radiography for clinical application”, Eur Radiol 7
(Suppl 3): 66-72
Prokop, M., Schaefer-Prokop, C.M. (1997), “Digital image processing”, Eur

Radiol 7 (Suppl 3): 73-82
Shaber, G.S., Maidment, A.D.A., Bell, J., Jeromin, L.S., Lee, D.L., Powell,
G.F. (1997), “Full field digital projection radiography system: Principles and
image evaluation”, In: Lemke HU (ed), “Computer assisted radiology and
surgery”, Elsevier: 1-7
Tapiovaara, M.J., “Efficiency of low-contrast detail detectability in fluoroscopic

imaging”, MEDICAL PHYSICS (1997) 24, 5, 655-64
Yaffe, M.J., Rowlands, J.A., “X-ray detectors for digital radiography”,

PHYSICS IN MEDICINE AND BIOLOGY (1997) 42, 1, 1-39
Bruijns, T.J.C., Alving, P.L., Bury, R., Cowen, A.R., Jung, N., Luijendijk, H.A.,
Meulenbrugge, H.J., Stouten, H.J. (1998), “Technical and clinical results of an
experimental Flat Dynamic (digital) X-ray image Detector (FDXD) system with
real-time corrections”, SPIE 3336: 33-44
Busch, H.P., Jaschke, W., “Adaptation of the Quality Criteria Concept to

digital radiography”, Rad. Prot Dosimetry, 80 (1998) 1-3, 61-63
Chaussat C., Chabbal, J., Ducourant, T., Spinnler, V., Vieux, G., Neyret, R.,
“New CsI / a-Si x17 X-ray flat panel detector provides superior detectivity and
immediate direct digital output for General Radiography systems”, SPIE
Medical Imaging (1998) 3336
Hamers, S., Freyschmidt, J. (1998), “Digital radiography with an electronic flat-

panel detector: First clinical experience in skeletal diagnostics”, Medica Mundi
42,3: 2-6
Kamm, K.F. (1998), “Grundlagen der Röntgenabbildung” (Principles of x-ray

imaging), In: Ewen E (Hrsg), “Moderne Bildgebung”, Georg Thieme: 45-60
Neitzel, U. (1998), “Grundlagen der digitalen Bildgebung (Principles of x-ray

imaging)”, In Ewen, K. (Hrsg), “Moderne Bildgebung”, Georg Thieme: 71-76
Neitzel, U. (1998), “Systeme für die digitale Bildgebung (Systems for digital
imaging)”, In: Ewen K (Hrsg), “Moderne Bildgebung”, Georg Thieme: 127-135
94
Strotzer, M., Gmeinwieser, J., Völk, M., Fründ, R., Seitz, J.; Manke, C.,
Albrich, H., Feuerbach, S., “Clinical application of a flat-panel X-ray detector
based on amorphous silicon technology: image quality and potential for
radiation dose reduction in skeletal radiography”, AJR. AMERICAN JOURNAL
OF ROENTGENOLOGY (1998), 171; 23-27
Vetter, S., Heckmann, H., Strecker, E.P., Busch, H.P., Kamm, K.F.,
Allmendinger, H. (1998), “Klinische Aspekte zu Bildqualität und Dosis bei
gittergesteuerter gepulster Durchleuchtung”, Akt.Radiol. 8: 191-195
Vetter, S., Faulkner, K., Strecker, E.P., Busch, H.P. (1998), “Dose reduction
and image quality in pulsed fluoroscopy”, Radiation Protection Dosimetry, 80:
299-301
Aufrichtig, R., “Comparison of low contrast detectability between a digital

amorphous silicon and a screen-film based imaging system for thoracic
radiography”, MEDICAL PHYSICS (1999), 26, 7, 1349-58
Busch, H.P., Klose, K.J., Braunschweig, R., Neugebauer, E. (1999), “Digitale

Radiographie – Ergebnisse einer Anwendungsumfrage und einer
Konsensuskonferenz” (Digital radiography – results of a user survey and a
consensus conference), Akt.Radiol. 7: 56-63
Busch, H.P., “Digitale Projektionsradiographie: Technische Grundlagen,

Abbildungseigenschaften und Anwendungsmöglichkeiten” (Digital
Radiography: Basics, imaging and application), Radiologe (1999), 39: 710-724
Chotas, H.G., Dobbins, J.T., Ravin, C.E. (1999), “Principles of digital

radiography with large-area, electronically readout detectors: A review of the
basics”, Radiology 210: 595-599
Frija, G. (1999), “Flat Panel Sensors: Questions and Answers”, Medical

Imaging Technology, Vol 17, 2: 99-104
Neitzel, U. (1999), “Integrated digital radiography with a flat-panel sensor”,

Medical Imaging Technology, 17, 2: 123-129
Strotzer, M., Gmeinwieser, J., Völk, M., Fründ, R., Feuerbach, S. “Digitale
Flachbilddetektortechnik basierend auf Cäsiumjodid und amorphem Silizium:
Experimentelle Untersuchungen und erste klinische Ergebnisse”, Fortschr
Röntgenstr 170 (1999): 66-72
95
Reiff, K.J. “Flat panel detectors – closing the (digital) gap in chest and skeletal
radiology”, European Journal of Radiology (1999)
Vano, E., Fernandez, J.M., Gracia, A., Guibelalde, E., Gonzalez, L., “Routine
quality control in digital versus analog radiography”, Physica Medica Vol XV
(1999), 319-321
96
2000
Aufrichtig, R., Xue, P., “Dose efficiency and low-contrast detectability of an

amorphous silicon x-ray detector for digital radiography”, Physics in Medicine
and Biology; 45 (2000) 9, 2653-69
Heyne, J.P., H. Merbold, J. Sehner, R. Neumann, R. Adler, M. Freesmeyer,

W. A. Kaiser: “Reduktion der Strahlendosis mittels Speicherfolienradiographie
am Handphantom/ Reduction of radiation dosage by using digital
luminescence radiography on a hand phantom”, Fortschr Röntgenstr (2000)
172, 4, 386-90
Ludwig, K., Link, T. M., Fiebich, M., Renger, B., Diederich, S., Oelerich, M.,
Lenzen, H., Heindel, W., “Selenium-based digital radiography in the detection
of bone lesions: preliminary experience with experimentally created defects”,
RADIOLOGY (2000) 216, 1, 220-224
Neitzel, U., Böhm, A., Maack, I., “Comparison of low-contrast detail

detectability with five different conventional and digital radiographic imaging
systems”, SPIE Medical Imaging (2000) 3981-31
Strotzer, M., Völk, M., Wild, T., Landenberg, P., Feuerbach, S., “Simulated
bone erosions in a hand phantom: detection with conventional screen-film
technology versus cesium iodide-amorphous silicon flat-panel detector”,
RADIOLOGY (2000) 215, 2, 512-515
Vlk, M., Strotzer, M., Holzknecht, N., Manke, C., Lenhart, M., Gmeinwieser, J.,
Link, J., Reiser, M., Feuerbach, S., “Digital radiography of the skeleton using a
large-area detector based on amorphous silicon technology: image quality and
potential for dose reduction in comparison with screen-film radiography”,
Clinical radiology, 55 (2000) 8, 615-21
97
2001
Nuclear Associates Diagnostic Radiology and Radiation Therapy Catalog,

“CDRAD contrast detail digital/conventional radiography phantom”, Scientific
program RSNA 2001
Chotas, H.G., Ravin, C.E., “Digital Chest Radiography with a Solid-state Flat-
Panel X-ray Detector : Contrast-Detail Evaluation with Processed Images
Printed on Film Hard Copy”, Radiology 201; 218: 679-682
Floyd, C.E., Warp, R.J., Dobbins, J. T. Dobbins, H. G. Chotas, A. H. Baydush,

R. Vargas-Voracek, C. E. Ravin: “Imaging characteristics of an amorphous
silicon flat-panel detector for digital chest radiography”, RADIOLOGY (2001)
218, 3, 683-688
Geijer, H., Beckman, K.-W., Andersson, T., Persliden, J., “Image quality vs.
radiation dose for a flat-panel amorphous silicon detector: a phantom study”,
Eur. Radiol. (2001) 11, 1704-1709
Hammers, S., Freyschmidt, J., Neizel, U., “Digital radiography with a large-
scale electronic flat-panel detector vs. screen-film radiography: observer
preference in clinical skeletal diagnostics”, Eur. Radiol. (2001) 11, 1753-1759
Harrell, G., Chotas, M.S., Carl, E., Ravin, M.D, “Digital Chest Radiography
with a Solid-state Flat-Panel X-ray Detector: Contrast-Detail Evaluation with
Processed Images Printed on Film Hard Copy”, Radiology 2001; 218: 679-682
Hennis, S.P., Garmer, M., Jaeger, H.J., Classen, R., Jacobs, A., Gissler,
H.M., Christmann, A., Mathias, K., “Digital chest radiography with a large-area
flat panel silicon X-ray detector: clinical comparison with conventional
radiography”, Eur. Radiol. (2001) 11: 1688-1696
Jessen, K.A., “The quality criteria concept: an introduction and overview”.

Radiat Prot Dosimetry, 94 (201) 1-2, 29-32
Marshall, N.W., “Optimisation of dose per image in digital imaging.” Radiat

Prot Dosimetry, 94 (201) 1-2, 83-7
Peer, S., Neitzel, U., Giacomuzzi, S.M., Peer, R., Gassner, E., Steingruber, I.,
Jaschke, W. “Comparison of low-contrast detail perception on storage
phosphor radiographs and digital flat panel detector images”, IEEE
TRANSACTIONS ON MEDICAL IMAGING (2001) 20; 3; 239-42
Peer. S., Peer, R., Giacomuzzi, S.M., Jaschke, W., “Comparative reject
analysis in conventional film-screen and digital storage phosphor radiography”.
Radiat Prot Dosimetry, 94 (201) 1-2, 69-71
98
Reissberg, S., Hoschen, C., Kästner, A., Theus, U., Fiedler, R., Krause, U.,
Dühring, W., “Erste klinische Erfahrung mit einem großformatigen
Flächendetektorsystem bei Aufnahmen des peripheren Skelettsystems (First
clinical experience with a flat panel detector for imaging the peripheral
skeleton)”, Fortschr Röntgenstr (2001) 173, 1048-1052
Schaefer-Prokop, C., Eisenhuber, E., Fuchsjäger, M., Puig, S., Prokop, M.,
“Aktuelle Entwicklung auf dem Gebiet der digitalen Thoraxradiographie
(Current developments in the field of thoracic imaging)”, Radiologe 2001; 41:
230-239
99
2002
Busch, H.P., Bosmans, H., Faulkner, K., Peer, R., Vano, E., Busch, S., “Dose
management with new digital imaging techniques”, European Radiology
Association § Congress; 2002 ,SS 1013: Digital radiography/Digital
mammography
Hermann, K.A., Bonel, H., Stäbler, A., Kulinna, C., Glaser, C., Holzknecht, N.,
Geiger, B., Schätzel, M., Reiser, M.F., “Chest imaging with flat-panel detector
at low and standard doses: comparison with storage phosphor technology in
normal patients”, Eur. Radiol 2002; 12: 385-390
Ludwig, K., Lenzen, H., Kamm, K.F., Link, M., Diederich, S., Wormanns, D.,
Heindel, W., “Performance of a flat-panel detector in detecting artificial bone
lesions: Comparison with conventional screen-film and storage-phosphor
radiography”, Radiology 2002; 222: 453-459
Kim, T.S., Im, J.G., Lee, H.J., Lee, Y.J., Kim, S.H., Kim, S., “Detection of
Pulmonary Edema in Pigs: Storage Phosphor Versus Amorphous Selenium-
based Flat-Panel-Detector Radiography”, Radiology 2002; 223: 695-701
Okamura, T., Tanaka, S., Koyama, K., Norihumi, N., Daikokuya, H.,
Matsuoka, T., Kishimoto, K., Hatagawa, M., Kudoh, H., Yamada, R., “Clinical
evaluation of digital radiography based on an large-area cesium iodide-
amorphous silicon flat- panel detector compared with screen-film radiography
for skeletal system and abdomen”, Eur Radiol 2002; 12: 1741-1747
Persliden, J., Beckman K.W., Geijer H., Andersson T., “Dose-image

optimisation in digital radiography with a direct digital detector: an example
applied to pelvic examinations”, Eur Radiol 2002; 12: 1584-1588
Strotzer, M., “Digital radiography with flat-panel detectors: the missing link”,
Eur Radiol (202) 12: 1603-1604
100
2003
Busch, H.P., Busch, S., Decker, C., Schilz, C., “Bildqualität und Dosis in der
Digitalen Projektionsradiographie” (Image quality and exposure dose in Digital
Projection Radiography), Fortschr. Röntgenstr. 2003; 175, 32 – 37
Busch, H.P., “Qualitäts- und Dosismanagement in der Digitalen

Projektionsradiographie” (Dose and Quality Management in Digital
Radiography), Fortschr. Röntgenstr 203; 175; 17 - 19
Dobbins, J.T., Samei, E., Chotas, H.G., Warp, R.J., Baydush, A.H., Floyd,
C.E., Ravin, C.E., “Chest Radiography: Optimization of X-ray Spectrum for
Cesium Iodine-Amorphous Silicon Flat-Panel Detector”, Radiology 203; 226:
221-230
Eisenhuber, E., Stadler, A., Prokop, M., Fuchsjäger, M., Weber, M., Schaefer-
Prokop, C., “Detection of Monitoring Materials on Bedside Chest Radiographs
with the Most Recent Generation of Storage Phosphor Plates: Dose Increase
Does Not Improve Detection Performance”, Radiology 203; 227: 216-221
Ludwig, K., Schülke, Ch., Diedrich, S., Wormanns, D., Lenzen, H., Bernhardt,
Th.M., Brinckmann, P., Heindel, W., “Detection of Subtle undisplaced Rib
Fractures in a Porcine Model: Radiation Dose Requirement – Digital Flat-
Panel versus Screen Film and Storage Phosphor Systems”, Radiology 203;
227: 163-168
Rapp-Bernhardt, U., Roehl, F.W., Gibbs, R.C., Schmidl, H., Krause, U.W.,
Bernhardt, T.M., “Flat-Panel X-ray Detector Based on Amorphous Silicon
versus Asymmetric Screen Film System: Phantom Study of Dose Reduction
and Depiction of Simulated Findings”, Radiology 203; 227: 484-492
101
102
Annex1:
DIMOND III
(Contract: FIGM-CT-2000-00061)
Final Report
1st October 2003
Lead contractor:
Workpackage 1
Clinical Quality Criteria
Leading partner:
WP 1.2: Identification of the three image quality bands
WP 1.4: Report on clinical quality criteria for new digital
technology
WP 2.2: QA protocol for digital detectors and new imaging
devices
WP 6.7: Circulation of mammography images between
centres
103
DIMOND III
Research Issues in Diagnostic Radiology

1st October 2003
Clinical Quality Criteria

(Lead contractor H.P. Busch)
WP 1.1 – 1.10
104
Workpackage 1: Clinical Quality Criteria
Lead contractor: H.P. Busch, Trier
WP 1.1: Publication of equipment requirements for

specific interventional procedures
Lead partner: H. Zoetelief, Delft
WP 1.2: Identification of three image quality bands

Lead partner: H.P. Busch, Trier
WP 1.3: Protocols for optimisation of post-processing

and display conditions
Lead partner: H. Bosmans, Leuven
WP 1.4: Report on clinical quality criteria for new

digital technology
Lead partner: H.P. Busch, Trier
WP 1.5: Publication, compendium document

Lead partner: K. Faulkner, Newcastle
WP 1.6: Quality criteria for fluoroscopic guided

interventional procedures
Lead partner: W. Jaschke, Innsbruck
WP 1.7: Experimental verification of the theoretical

model linking physical measurements to
clinical indices
Lead partner: J. Malone, Dublin
105
WP 1.8. Pilot study on equipment requirements for
fluoroscopic guided interventional radiology
Lead partner: H. Zoetelief, Delft
WP 1.9: Pilot study into the optimisation of viewing

conditions and display settings
Lead partner: J. Kotre, Newcastle
WP 1.10: Pilot study of quality criteria for new detectors

Lead partner: W. Jaschke, Innsbruck
106
DIMOND III
WP 1.2:
Identification of three image quality bands
Lead partner: Brüderkrankenhaus Trier, Germany, partner 3
The new techniques of digital radiography, such as computed radiography and

flat panel radiography, offer new potential for dose reduction. In order to adapt
dose and image quality to diagnostic requirements, we developed the model of
three image quality bands and discussed this extensively at the DIMOND III
meeting in Trier. In WP 1.2, imaging capabilities were determined by different
phantoms to define these three image quality bands using the results of both
the CDRAD phantom and different organ phantoms, for example the
abdomen, lung, hand and pelvis. Images of different dose (speed) classes
(200, 400, 800, 1600) were evaluated or graded by different observers and
assigned to either the “high”, “medium” or “low” image class. This forms the
base of our description of the three image quality bands.
One of the main focuses of WP 1.2 was last year’s development of a computer
program to automatically evaluate CDRAD images by pattern recognition. This
was a breakthrough because of the significant decrease in workload it has
brought and the objectivity it gives to CDRAD image evaluation. We tested the
program intensively and determined its sensitivity and performance. This
program is now available to all DIMOND III partners.
Starting with the “Referral guidelines for imaging” (Radiation Protection 118),
we asked DIMOND partners to define the necessary image quality for all
diagnostic questions. We received responses from four centres.
We then modified our proposal based on these results and distributed them to
all DIMOND partners prior to the Trier meeting in October 2002. As a result of
the discussion that took place at this meeting, we cancelled the trial that had
been planned for February 2003. These results have now been outlined. This
should be the starting point of a broad discussion within the European
community of radiology and radiation protection.
1. Review of existing literature and clinical experience

A summary of the existing literature for digital radiography was sent to all
DIMOND partners in December 2001. This list has now been updated to
coincide with the release of the final report.
107
2. Definition of three image quality bands for digital radiography
Based on the results of phantom measurements taken with the CDRAD
phantom, we defined the three bands for different storage phosphor systems
and a flat detector system. The results were presented as part of a refresher
course at the German Röntgenkongress (radiography conference) 2003.
These results were then discussed extensively during the DIMOND meeting in
Trier with specific reference to digital images. A full set of clinical images of the
image quality band is now available to all DIMOND partners for the purpose of
training and education.
3. Definition of three quality bands by clinical criteria

The three quality bands were defined by clinical criteria based on the
European Commission’s Referral Guidelines for Imaging (European
Communities, 2001 ISBN 92-828-9454-1). A subgroup was created by taking
the applications for which projection radiography examinations are
recommended and graded the image quality into the levels high, medium, low.
We sent this “new” version of the referral guidelines to the DIMOND partners
for discussion.
4. Testing the model (partners)
5. Workshop with physicists and radiologists

At the Trier meeting, there was intensive discussion of the three quality bands.
The discussion resulted in a specific modification of the description for the final
report. The trial that had been proposed for February 2003 was therefore
cancelled.
6. Recommendations for the three image quality bands (with examples)

We worked out the recommendations for the final DIMOND III report. A full set
of clinical images is now available to all DIMOND III partners for training and
education proposes.
Summary:
WP 1.2 achieved what it set out to do at the beginning. We convinced our
partners to take the CDRAD 2.0 as a standard phantom for evaluating image
quality in the project and intensively discussed the results of criteria for the
three image quality bands and their application to clinical problems.
108
List of deliverables
WP 1.2: Identification of three image quality bands

1. Review existing literature and clinical experience (Yes)
2. Define three image quality bands for digital radiography (Yes)
3. Define the three quality bands by clinical criteria (Yes)
4. Test this model (discussion, trial use by DIMOND partners) (Yes)
5. Meet with physicists and radiologists (Trier meeting)(Yes)
6. Make recommendations for the three image quality bands (with examples)
(Yes)
WP 1.2 was finished on time and met its goals.
Partners:
Haughton Institute, Ireland (Prof. Dr. Jim Malone)
Department of Radiology, Leuven (Dr. Hilde Bosmans)
Department of Radiology, Innsbruck (Prof. Dr. W. Jaschke)
Medical Physics Department, Newcastle(Dr. CJ Kotre)
Department of Radiology, Karlsruhe (Dr. S. Vetter)
Linked projects:
1.4 Lead partner: BKT Trier
1.6 Lead partner: Radiology Innsbruck
4.1 Lead partner: Radiology Karlsruhe
5.1 Lead partner: Azienda Ospedaliera Udine
109
WP 1.4:
Report on clinical quality criteria for new digital technology
The main goal of this workpackage was to adapt quality criteria to further
digital imaging procedures involved in the DIMOND II project. The
recommended diagnostic requirements for digital projection radiography are
based on the European Guidelines on Quality Criteria for Diagnostic images
(EC), the diagnostic requirements for digital radiographic procedures
(DIMOND II) and the guidelines published by the Bundesärztekammer
(German Federal Medical Association) for quality assurance in radiological
examinations. We sent these recommendations to all partners for further
examination. Additionally, we asked the partners to send us technical
parameters of their exposures. Using this material, we prepared a contribution
to a book on European guidelines on quality criteria for diagnostic digital
radiographic images, similar to what already existed for conventional
film/screen radiography.
1. Adaptation of quality criteria for new digital technology

Compared to what we had originally planned, the list of examinations covered
by this workpackage was enlarged to include all indications for projection
radiography (all of which were also included in the report on clinical quality
criteria). We limited the report on new digital technologies to storage phosphor
and flat panel technology. As discussed with the partners and the coordinator
at the Vienna meeting in 2002, we did not extend this report to fluoroscopic
procedures because of the possible overlap with WP 4. The quality criteria of
fluoroscopic procedures will be dealt with there.
Post-processing of digital images has a great impact on quality criteria.

Therefore, before beginning with the optimization of post-processing, a
description of the “black box” post-processing model was needed. In addition
to the work planned for WP 1.4, we worked out possible ways of describing the
quality of post-processing and developed strategies for optimization. This was
discussed extensively at the Trier workshop in October 2002. We will present
the results of this at the final DIMOND meeting and in the final DIMOND
report.
110
2. Trial (WP 1.4 partners)
3. Final workshop with clinical partners

Because we extensively discussed the results at the DIMOND Trier meeting in
various working groups and modified the description, we were able to cancel
the proposed trial and final workshop.
4. Development of draft for “European Guidelines on quality criteria for

Digital Radiographic images”
This draft was finished and will be presented at the final meeting and in the
final report in 2004.
Summary:
WP 1.4 is now finished. After intensive discussion with the DIMOND partners,
we are now preparing the draft for the final meeting. A full set of clinical
examples is now available for all DIMOND partners.
List of deliverables:
1. Adapt quality criteria for new digital technology (Yes)
2. Trial at Trier meeting (partners of DIMOND III) (Yes)
3. Final workshop with clinical partners (at Trier meeting) (Yes)
4. Contribution to the draft version of “European Guidelines on Quality
Criteria for Digital Projection Radiography” (Yes)
111
WP 2.2:
QA protocol for digital detectors and new imaging devices
We developed a QA protocol for different digital detectors. This protocol uses

the new capability of numerical and quantitative evaluation of image quality
(e.g. signal/noise). The protocol is based on the German standards for
constancy testing DIN 6868-58 and DIN 6868-13, but includes digital
evaluation of additional parameters. We tested this protocol with a flat detector
system (Philips DigitalDiagnost) and a storage phosphor system (Philips
Compano). We have also sent the protocol to the DIMOND partners for
evaluation before summarising our experience.
1. Development of a QA protocol for digital detectors and new imaging

devices
In WP 2.2, we developed a new QA protocol for flat detectors. We sent this
protocol to all DIMOND partners for evaluation. This protocol intentionally
featured a larger number of parameters than was necessary so that it would
choose only the important and useful parameters. After discussion with our
partners and a year’s experience working with the protocol, we suggested a
final version. This was discussed, among other places, at the Trier DIMOND
meeting in October 2002. Different partners’ suggestions have now been
integrated. We also expanded the application to work with storage phosphor
radiography. The protocol for fluoroscopic procedures will be dealt with in
WP 4.
2. Trial (partners)
3. Meeting to summarise trial results

Trial results were extensively discussed in different working groups at the
DIMOND Trier meeting and comments integrated. We were therefore able to
cancel the proposed trial and final workshop.
We also developed a program to automatically recognise patterns in test
phantom exposures in order to evaluate the images automatically using PCs.
112
Summary:
WP 2.2 is now finished. Over the course of the workpackage, we developed a
new procedure for constancy testing based on the new capabilities of digital
imaging. We then tested this procedure over a long period and discussed the
results intensively with our colleagues and DIMOND partners. In addition, we
developed a new computer program to automatically evaluate the test images.
This program is now available to all DIMOND partners via the internet.
Deliverables:
1. Develop a QA protocol for digital detectors and new image devices (Yes).
2. Discussion/feedback from the partners (Yes)
3. Meet to discuss and summarise results (Trier meeting) (Yes)
4. Contribution to the draft of European guidelines (Yes)
Partners:
Haughton Institute, Dublin( Prof. Jim Malone)
Department of Radiology, Madrid ( Prof. E. Vano)
Department of Radiology, Leuven (Dr. H. Bosmans)
Department of Radiology, Innsbruck (Prof. W. Jaschke)
Department of Radiology, Karlsruhe (Dr. S. Vetter)
Linked projects:
1.2. Lead partner: BKT Trier
1.5. Lead partner: Newcastle City Health Trust
1.6. Lead partner: Radiology Innsbruck
1.10. Lead partner: Radiology Innsbruck
4.1. Lead partner: Radiology Karlsruhe
6.1. Lead Partner: Radiology Innsbruck
113
WP 6.7:
Circulation of mammographic images between centres
1. Preparation
Circulation of mammography images will coincide with the final report in 2004.
This set of images contains both clinical and test images and was put together
from images received from our DIMOND partners in Innsbruck and Madrid.
2. Circulation of mammography images

The mammography images were discussed at the Trier meeting in October
2002. We will send the image folder for further discussion and evaluation
along with the final report.
Summary:
WP 6.7 was not totally successful. If we can obtain more images over the next
few months, we will integrate them into the image file for the final workshop.
However, we were not totally satisfied because of the limited input.
114
Annex 2:
JAnalyser-CDRAD
Quality Testing of Digital Projection Radiography
C. SCHILZ
115
“Contrast-Detail CDRAD phantom”
(Designer’s description: Nuclear Associates / USA)
Introduction
Most definitions of image quality in radiology are based on characterizing the
physical properties of the image chain. However, medical diagnosis is not
made on the basis of the image alone. The observer’s perception of the image
is crucial to the result. This perception can be tested using so-called Contrast-
Detail (CD) phantoms. These CD phantoms make it possible to quantify both
detail and contrast, as observed by the radiologist. The CDRAD phantom can
be used within the entire range of diagnostic imaging systems, such as
fluoroscopy and digital subtraction angiography.
Construction
The CDRAD phantom consists of a Plexiglas tablet with cylindrical holes of
exact diameter and depth (tolerances: 0.02 mm). Together with additional
Plexiglas tablets, which are used to simulate the dimensions of the patient, the
radiographic image of the phantom gives information about the imaging
performance of the whole system.
This image shows 225 squares, 15 rows

and 15 columns. In each square, either one
or two spots are present, these being the
images of the holes. The first three rows
show only one spot, while the other rows
have two identical spots – one in the
middle and one in a randomly chosen
corner.
The optical densities of the spots are
higher compared to the uniform
background. Due to the (exponentially)
increasing depth of the holes in a
horizontal direction, the image shows 15
columns of spots with increasing contrast. In the vertical direction, the
diameter of the holes increases in stages, exponentially from 0.3 to 8.0 mm.
For the image, this means 15 rows of spots with increasing spatial resolution.
Evaluation
To evaluate the phantom image, the observer indicates the location of the
second spot in each square. Correct indication proves that a the observer
really does see contrasts. At the transition from visible to invisible, it is difficult
116
to decide in which corner the second spot is located, and the response equals
pure chance. The line connecting the central spots with smallest visible
diameter and contrast is called the contrast-detail (CD) curve. In order to
compare the imaging performance of different systems, phantom images are
made under identical conditions and evaluated by the same observer and at
the same time. The better system will produce an image in which smaller
contrasts and details are visible. This results in a shift of the CD curve to the
lower left part of the image
Why is computer-based evaluation

of CDRAD phantoms useful?
The aim of the program JAnalyser-CDRAD to provide computer-based

evaluation of CDRAD phantoms.
There are some good reasons for evaluating the examination by computer.
Manual evaluation:
- Depends on the observer

- Is very time-consuming
- Will train the observer (the results change), which means that
- The observer learns to “read” the examination
- The observer learns the sequence of the holes
Computer-based evaluation:
- Is independent of the observer (depends only on the algorithm used)

- Is faster
- Produces results that are reproducible
It is, thus, useful to have a computer-based evaluation of the CDRAD

PHANTOM.
117
How does the program work?
The 15x15 grid of spot-containing squares is surrounded by a thin metallic
grid, which is easily visible during the examination. Using various methods of
digital image processing (edge enhancement, skeletation, outline pursuit the
grid is selected and the adjustment of the phantom is determined. If the grid is
completely shown, any turn of the phantom can be recognized (although due
to the divergence of the x-rays, dertermining the position of the grid is not
simple).
The position and angle of the 225 squares are

now established. We now try to determine the
position of the spots based on the original picture
information.
There are a number of tested methods for
determining the position of the spots:
Hough transformation (for cycles)
Mask contrast procedure
Standardized cross correlation
The result is compared with the CDRAD phantom. If the position of the spots is
determined correctly, then the square is considered as recognized (“1”). If the
position is not correctly determined, then the square is considered as not
recognized (“0”). The results for all 225 fields are then submitted for a
plausibility check, because the evaluation is prone to error.
The plausibility check minimizes the influence of:
Fields that are correctly recognised by chance
Fields that are not recognized fields (but should have been)
This check involves comparing

the results for each square with
the results for each surrounding
square. The figures displayed by
the programme represent
square recognition results after
the plausibility check.
118
Procedures for evaluating the position of the spots
Hough transformation
Original image After edge extraction After hough transformation
A filter is applied to extract the outline of the spots (edge extraction) from the
original picture. Hough transformation is then applied to these outlines,
dependent on the diameters of the spots. This hough transformation transmits
the outer edge of the circle to the centre of the circle. The positions of the
centres are determined by a search for the two maximum values .
In squares with one spot:

- If the maximum value is in the centre of the square, the square is
recognized (“1”)
In squares with two spots:
- If one of the two maximum values is in the centre and the other in a corner
of the square, the square is provisionally recognized. Tests are then carried
out to establish whether the spot is in the correct corner. If it is, then the
square is considered as recognized.
119
Annex 3:
JAnalyser
Constancy Testing of Digital Projection Radiography
C. SCHILZ
120
General information
The computer program JAnalyser allows automatic evaluation of constancy

tests using stored digital radiography images. For this, phantom exposures
need to be made in accordance with DIN6868-12 (Fig1) and constancy tests
must have constant exposure parameter values. Once taken, the digital image
has to be transferred without additional image processing, or otherwise image
processing parameters should be kept constant.
Constancy test evaluation is computer based, with high contrast, low contrast,
homogeneity, the square wave response function of the lead bar pattern and
the light field/xray-field are all measured.
The evaluation conforms to the German DIN standards:

DIN 6868-12
DIN 6868-58
Company: Pehamed Company: Pehamed Company Wellhöfer

Fig.1: Different images of supported phantoms.
D
The program searches the representative marks in the A B
image (A to D)
After these marks have been found, the program calculates
the characteristic values for high contrast, low contrast,
homogeneity, the square Wave Response Function” of the
lead bar pattern and the light field/xrayfield. C D
The entire program structure is a Client/Server solution. On the server there is

a data base (mySQL) for sampling and storage of determined data.
Communication between server and client is made by an JDBC interface. The
client program is written in Java, therefore it can be used in each operating
system environment (Windows, Linux, Sun etc.). For the execution the Java
Runtime Environment in version 1.4.0 or newer is needed.
For mySQL the usual data base interfaces (JDBC; ODBC...) exist, so the
collected data can be evaluated by other statistic tools (e.g. Microsoft
Excel/pivot tables).
121
Constancy Testing for digital x-ray units
The characterisation of imaging systems is based on communication theory.

According to this theory, x-ray devices can be considered in terms of “black
boxes". The black box has a defined transmission characteristic and noise
characteristic. The output signal can be determined from the input signal, if the
characteristic of the “Black box" is known.
Even if the manufacturers keep the characteristics of their systems secret,

tests can be developed using this communications theory.
The aim of constancy tests is not to establish system characteristics of the

black box with absolute values, but to establish the follow up of routine
performance of an x-ray unit .
Demands on an imaging system

An imaging system should represent differences in absorption without
distortion as a two-dimensional picture. It should have the following conditions:
1. Linearity
The input signal changes by the factor α, then the output signal
changes with the same factor α. The intensity value from the sum of
the single intensities is devoted to two objects which are
superpositioned in the object level (absolute detected values, without
calibration or adjustment by exposure index)
Logarithmically linearity
The input signal changes by a factor α, then the output signal
changes with the factor “log α”
2. Local or shift invariance

The position of the object, whether central or peripheral, has no
influence on the image. If the object is shifted, than the position of the
object within the image is shifted in the same way.
3. Time invariance
Images made under comparable conditions but at different times
should be identical.
X-ray systems fulfil only the first two of these criteria, while constancy tests are
based on time invariance.
122
The first criteria is restricted through
- A partial lack of logarithmic linear/linear characteristics
- Change of measured values by post-processing (which is not
influenced by the user)
The second criteria is restricted through

- Inhomogeneous radiation (heel effect, parameter of examination)
- Different local sensitivity of the detector
These first two demands are system-dependent and cannot not be ignored.
For constancy tests, the restriction is not importanted if
1. The conditions of examination are constant

2. The measured variables are comparable to those measured during the
acceptance test
Constancy tests for digital systems can be execuded under the same system
characteristics as as conventional (film/screen) constancy tests.
Measurements of these characteristics must be adapted to the requirements
and conditions of digital radiography.
Measured system characteristics

The image below shows a test phantom for digital radiography (according to
DIN 6868-12).
Fig.2: Image of a test phantom for digital radiography
123
This phantom allows all relevant data to be assessed:
- Dynamic range,
- Detail detectability,
- Spatial resolution
- Homogeneity
- Light field / x-ray field
The measurement of the parameters must be adapted to the requirements and

conditions of digital radiography. The measurement of the variables can be
done by hand or by computer.
124
Dynamic range
In conventional radiography, optical density depends on dose. The dynamic
range of an examination depends on the dose and the film-screen combination
selected (as well as whether a film is over-exposed, correctly-exposed, or
under-exposed). The constancy of an analogue x-ray device can be
established if constant dose values produce the same optical density. The
relationship between dose and optical density is logarithmically linear.
For digital x-ray devices, dynamic range can be adapted using post-processing
techniques. Images will show the same brightness independent of the dose
(see fig. 3). There is no relation between printed grey value / optical density
and dose.
81kV S200 81kV S400
81kV S800 81kV S1600

Fig.3: Digital images with different dose values (speed class 200, 40, 80, 1600)
Nearly all digital x-ray systems have a logarithmically linear relation between
dose and detected value. For constancy tests of digital x-ray systems, a
measured value should be used, which is unchanged or linearly changed
during post-processing. The high contrast is measured in the seven fields
shown above, with the different mean values describing the dynamic range. In
logarithmic linear systems, the noise is defined by the standard deviation. If
post-processing is linear, the standard deviation within homogeneous ranges
is linearly dependent on the dose and the gradient of the post-processing
curve. The post-processing parameters must be kept constant for the
constancy test so the standard deviation depends only on the dose.
The "average value" (signal) and the "standard deviation" (noise) are
calculated (automatically) for different levels of the grey values. If the quality of
the x-ray system changes, the measured “average values” and “standard
deviation” change too.
For the interpretation of the change further measured variables are necessary
(e.g. the area dose product or surface dose)
What JAnalyser does

The high contrast is measured in the seven fields shown above. The following
values are determined in the shown ROIs:
125
Mean value of the ROIs
- Standard deviation of the mean value
- Minimum grey value
- Maximum grey value
- Contrast of the ROIs to the surrounding area
- Contrast of the ROIs to the surrounding area divided with the noise
(standard deviation)
- Area of the ROIs
Detail detectability
Objects can be detected in relation to the contrast of the surrounding area.
However, the next four pictures show that contrast is not the only criterion for
detectability.
The four images have the same tube voltage (81 kV) but different dose
classes (S200, S400, S800, S1600). All images have the same contrast.
You can see that, in addition to the contrast, noise is also important to the
detectability of objects. The higher the noise gets, the more difficult it is to
recognize the objects.
81kV S200 81kV S400
81kV S800 81kV S1600
This regularity can be used during constancy testing.
In the low contrast area, the "contrast" and the relationship between the
object's contrast and noise of the objects can be determined (automatically). If
the quality of an x-ray system (and thus the detail detectability) decreases,
then the contrast/noise ratio decreases too.
126
What JAnalyser does
Contrast measurement can be taken putting by the mean value (MeanObj) of a

ROI (HC) of the same form and size above (Meanup) and below (Meandown). The
contrast has been calculated using the formula:
Meanup − Meandown
Contrast ( HC x ) = MeanObj −
2
Calculation of "contrast / noise" divided the contrast to the standard deviation

( ϑObj ) of the regarded ROI (HC).
Meanup − Meandown
MeanObj HC x −
Contrast ( HC x ) 2
=
N ϑObj HC
x
127
Homogeneity
The demand “local or shift invariance” on x-ray systems is restricted by:
- Inhomogeneous radiation (heel effect, parameters of examination)

- Different local sensitivity of the detector
The influence of these effect should be able to be kept as small as possible by:
- Using the maximum film-focal distance, thereby minimising the heel

effect
- Calibrating the detector (depending on the x-ray system)
With increasing work time of an x-ray unit:
- The emitted dose distribution of the x-ray tube can change

- The local sensitivity of the detector can change
To determine this change, homogeneity should be measured during the

constancy test.
For the description of homogeneity, the absolute or relative difference between

the minimum an maximum grey value can be accepted.
What JAnalyser does

The homogeneity is measured in the fields marked with "hori" and "vert" (see
image below).
Horizontal homogeneity:
For noise reduction, vertical grey values are averaged.
After that, the minimum (gmin) and maximum (gmax) grey
value is determined and the homogeneity can be
calculated.
The results of the calculation are presented in the table

“measurement”. In addition, the minimum and the
maximum grey values are registered and saved in the data base.
The following formula is used to calculate homogeneity (hom):
hom = g max − g inx
Vertical homogeneity is measured in the same way.
128
Spatial resolution / square wave response function
If a sinusoidal distribution of intensity in the object level is represented by an
imaging system, then the distribution of intensity of the image has the same
spatial frequency as that of the object.
However, the amplitudes of intensity distribution in the image with increasing

spatial frequency are reduced by the optical transfer errors of the system. The
reduced amplitudes of the image express the loss of resolution.
The influence of the amplitude of a sinusoidal distribution is called modulation.

The modulation transfer function (MTF) describes the relationship between the
modulation (Mod(u))at the frequency u and the modulation (Mod(0)) at the
frequency 0.
Mod (u )
MTF (u ) =
Mod (0 Lp / mm)
The production of a sinus wave raster is very complex and expensive. For this
reason, constancy tests are made using a lead bar pattern. With lead bar
patterns, the square wave response function (SWRF) is measurable. This
function stands in close relationship to the MTF.
In order to state the constancy, the SWRF measurement and the relative
comparison to the accepting test are sufficient.
129
For an exact determination of the MTF/SWRF, single images of lead bar
pattern are insufficient. Particularly at high spatial frequency, the number of
based measuring points becomes too small.
The SWRF measurement is more precise than the determination of the last
visible linepair in analogue radiography. In the SWRF, the last visible linepair
corresponds with the range, in which the curve leads to zero.
Changes to the characteristic of systems should be able to be recognised

earlier using the measurement of the SWRF than by determining the last
visible linepair.
What JAnalyser does
The square wave response function is

determined within the lead bar (0.6-5.0 lp/mm).
The program itself automatically finds the object
borders of the line pairs of lines.
direction, in which the
mean value of the
intensitivty values is ROI
build
grid group in
y-direction Direction, in which the
profile of intensity is build
Fig. 11: Lead-bar pattern After that, the program calculates the mean
value for each row and searches the three
maxima and two minima (or two maxima and
three minima).
130
Fig. 12: mean values of the rows
The linepair is recognized and the program finds exactly three maxima (max1
until max3) and two minima (min1 until min2). If this is impossible, the
SWRF(lp(n)) is set to 0. (= > line pair cannot be recognized).
The modulation and the SWRF are calculated using the following formula.
max1 (lp (n)) + max 2 (lp (n))

Mod1 (lp (n)) = − min1 (lp (n))
2
Mod 2 (lp (n)) = ...
Mod 3 (lp (n)) = ...
Mod1 (lp (n)) + Mod 2 (lp (n)) + Mod 3 (lp (n))

Mod (lp (n)) =
3
Mod (lp (n))
SWRF (lp (n)) =
Mod (lp (0.6))
131
Light field / X-ray field
a Measurement of the difference

between the light field and the X-
ray-field. (This is only
possible/useful if the exposure of
c the phantom has been made
conforming to DIN6868-12)
The relative difference is

calculated using the following
formula:
a+b
hori _ ffa =
ffd
d
c+d
vert _ ffa =
ffd
ffd = focus _ film _ dista.
b
132
Annex 4:
CD1: Quality and Dose Management for Digital Radiography
WP 6.7 – Circulation of Mammography Images Between Centres
CD2 : Image Quality Evaluation “Digital Mammography CD2”

Prof. Dr. E. Vano (Madrid / Spain)

Prof. Dr. E. Vano (Madrid / Spain)

Prof. Dr. W. Jaschke (Innsbruck / Austria)
133
CD Einlegeblatt 1
134
CD Einlegeblatt 2
135
Annex 5:
Acknowledgements
Those involved in the DIMOND projects (DIMOND I, II, and III) can look back
on 13 successful years of cooperation. This European study has made a
significant contribution to the development of digital radiography as a new
examination technique. Our goal was to achieve optimal image quality (and
with it, better diagnostic information) under minimal radiation doses and to
standardize digital radiography techniques all over Europe.
Now that DIMOND III has come to and end, I would like to express my thanks
to all of the European research groups involved for both the collaboration and
the interesting dialogue that came out of it. In particular, I would like to thank
the projects' founding members, Dr. Keith Faulkner and Prof. Dr. Jim Malone,
who directed and coordinated the projects DIMOND I-III with dedication and
enthusiasm. Our collaboration over the last 13 years has turned into great
friendship – a wonderful example of how Europeans can grow closer on a
scientific and a social level. Besides the scientific research, DIMOND will
surely be remembered by all by their experiences at the various events that
took place over the course of the project.
The DIMOND project would not have been possible without the support of the
European Union, and I would also like to take this opportunity to thank them
for their support and patronage, particularly Dr. Schibilla for her extraordinary
efforts in the early years.
Thanks are also due to the Krankenhaus der Bamherzigen Brüder hospital in
Trier for their support during the project – in particular for their understanding
of the numerous compromises and allowances made for DIMOND.
136
Special thanks go to the DIMOND team at the radiology department at the
Krankenhaus der Barmherzigen Brüder hospital here in Trier for their work and
dedication:
Christian Decker
Clemens Schilz
Anke Jockenhöfer
Stephanie Busch
Marion Anschütz
as well as the rest of the staff at the radiology department, whose extra
workload has now paid off.
Lastly, I would like to thank my wife Hildegard and my children Stephanie and
Thomas for their support and understanding throughout DIMOND, not least for
the large amount of travel I have done over the course of the projects.
I, for one, would be happy to see such successful collaboration between

European research groups continue in the future in the area of digital
radiography.
Trier, 11 January 2004
Prof. Dr. H.P. Busch
137

Dimond Iii: Image Quality and Dose Management For Digital Radiography

Uploaded by

Dimond Iii: Image Quality and Dose Management For Digital Radiography

Uploaded by

DIMOND III

Image Quality and

Prof. Dr. H.P. Busch M.D.

Chapter I Image Quality and 5

Chapter II Digital Projection Radiography: 11

Chapter III Diagnostic Requirements 24

Chapter IV Referral Guidelines 52

Chapter V Optimisation and Quality Assurance 66

Chapter VI Image Quality and Dose for 77

Chapter VII References 93

Annex 4: CD1: Quality- and Dose- Management

Image Quality and Dose Management for

In the field of projection radiography, conventional film/screen radiography is

Digital imaging differs significantly from film/screen radiography in that the

Choosing suitable technique parameters will often provide additional

In recent years, film/screen radiography has been characterised by the search

A clear definition of indications for Digital Projection Radiography should be

The choice of specific imaging method and exposure parameters must be

Part of the DIMOND III project involved developing a three-level concept of

High: Diagnosis of a fracture/fissure

Image quality class

High Medium Low

This new concept consists of three steps:

1. Optimisation (use clinical criteria)

Standardisation of the various manufacturers’ post-processing algorithms and

An important part of quality assurance is constancy testing, something that

1. There will be an extensive replacement of conventional film/screen

3. The capabilities of digital imaging methods differ significantly from

4. Image quality and dose management strategies must be based on

7. No significant differences exist between digital radiography and film/screen

8. Image optimisation must take place according to clinical criteria. However,

9. Economically efficient application of new techniques (e.g. flat detectors)

10. The new technological capability that digital projection radiography

Digital Projection Radiography

The application of the digital image intensifier technique for projection

In the mid-1980s, the storage phosphor technique came into clinical

In recent years, the progress of technology has allowed direct acquisition of

From a radiologist’s point of view, the aim of further developing imaging

1. High image quality (spatial resolution, contrast detectability, dynamic range,

Principles of imaging systems for digital projection radiography

Film/screen imaging is a well-proven system for projection radiography. Its

The digital imaging chain consists of three independent steps: image

1. Image intensifier (II) technique

1. Digital image intensifier technique

2. Storage phosphor radiography

4. Digital CCD technique

5.1. Flat detectors with direct conversion (selenium)

5.2. Flat detector with indirect conversion (scintillator)

The imaging capabilities of digital radiography are determined by the

The parameters characterising digital systems are:

Modulation transfer function (MTF):

Digital Image Intensifier Technique

The diameter of the entrance surface is between 15 cm and 40 cm. With a

Digital storage phosphor radiography

In storage phosphor radiography, pixel size and spatial resolution depend on

Digital radiography with the selenium drum technique

Digital CCD technique

The number of pixels of a CCD camera is 1024x1280. The sum of 4 CCD

Flat detector with direct conversion (selenium)

Flat detectors manufactured by Trixel (a joint venture by Thomson, Siemens,

Clinical applications of digital projection radiography

The advantages of digital imaging techniques are:

1. Large dynamic range (no retakes, imaging of great differences of

The digital image intensifier technique

The digital imaging technique has completely replaced film/screen radiography

Storage phosphor technique

Flat detector technique

Flat detectors with amorphous selenium have the advantage of a direct

Digital radiography - a new challenge

This article was translated from a publication in German language:

Diagnostic Requirements for Digital Projection

1. European Guidelines on Quality Criteria

Suggested radiographic technique

Imaging technique: Flat detector, storage phosphor

Suggested radiographic technique

Imaging technique: Flat detector, storage phosphor