Chapter

2 Radiography

Introduction consequently, the corresponding photon energies are

X-rays were discovered by Wilhelm Konrad Röntgen on the order of keV (1 eV = 1.602 × 10−19 J).
in 1895 while he was experimenting with cathode X-rays are generated in an X-ray tube, which con-
tubes. In these experiments, he used fluorescent sists of a vacuum tube with a cathode and an anode
screens, which start glowing when struck by light emit- (Figure 2.2(a)). The cathode current J releases elec-
ted from the tube. To Röntgen’s surprise, this effect trons at the cathode by thermal excitation. These
persisted even when the tube was placed in a carton electrons are accelerated toward the anode by a voltage
box. He soon realized that the tube was emitting not U between the cathode and the anode. The elec-
only light, but also a new kind of radiation, which he trons hit the anode and release their energy, partly
called X-rays because of their mysterious nature. This in the form of X-rays, i.e., as bremsstrahlung and
new kind of radiation could not only travel through characteristic radiation. Bremsstrahlung yields a con-
the box. Röntgen found out that it was attenuated tinuous X-ray spectrum while characteristic radiation
in a different way by various kinds of materials and yields characteristic peaks superimposed onto the
that it could, like light, be captured on a photographic continuous spectrum (Figure 2.2(b)).
plate. This opened up the way for its use in medicine.
Brehmsstrahlung
The first “Röntgen picture” of a hand was made soon
after the discovery of X-rays. No more than a few The energy (expressed in eV) and wavelength of the
months later, radiographs were already used in clinical bremsstrahlung photons are bounded by
practice. The nature of X-rays as short-wave electro-
hc
magnetic radiation was established by Max von Laue E ≤ Emax = qU , λ ≥ λmin = , (2.2)
in 1912. qU

where q is the electric charge of an electron. For

X-rays example, if U = 100 kV, then Emax = 100 keV.
X-rays are electromagnetic waves. Electromagnetic
radiation consists of photons. The energy E of a Characteristic radiation
photon with frequency f and wavelength λ is The energy of the electrons at the cathode can release
an orbital electron from a shell (e.g., the K-shell), leav-
hc ing a hole. This hole can be refilled when an electron
E = hf = , (2.1) of higher energy (e.g., from the L-shell or the M-shell)
λ
drops into the hole while emitting photons of a very
where h is Planck’s constant and c is the speed of light specific energy. The energy of the photon is the differ-
in vacuum; hc = 1.2397 × 10−6 eV m. The electro- ence between the energies of the two electron states;
magnetic spectrum (see Figure 2.1) can be divided for example, when an electron from the L-shell (with
into several bands, starting with very long radio energy EL ) drops into the K-shell (getting energy EK )
waves, used in magnetic resonance imaging (MRI) (see a photon of energy
Chapter 4), extending over microwaves, infrared, vis-
ible and ultraviolet light, X-rays, used in radiography, E = EL − EK (2.3)
up to the ultrashort-wave, high energetic γ-rays, used
in nuclear imaging (see Chapter 5). The wavelength is emitted. Such transitions therefore yield character-
for X-rays is on the order of Angstrøms (10−10 m) and, istic peaks in the X-ray spectrum.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

Visible light

Radio Waves Micro waves Infrared light Ultraviolet light X-rays γ-rays

2 –1 –2 –3 –4 –5 –6 –7 –8 –9 –10 –11 –12 –13

10 10 1 10 10 10 10 10 10 10 10 10 10 10 10 10

400nm
Wavelength (m)
700nm

MRI
endoscopy radiography nuclear
CT imaging

Photon energy (eV)

–8 –7 –6 –5 –4 –3 –2 –1 2 3 4 5 6 7
10 10 10 10 10 10 10 10 1 10 10 10 10 10 10 10

Figure 2.1 The electromagnetic spectrum.

X-ray beam intensity

6

Characteristic
radiation
5 Ka

25 kV
4

20 kV Continuous
3 radiation

lead shield absorbing most X-rays Kb

15 kV
vacuum 2
cooling
cathode e– anode
10 kV
1
focus shield
x-rays
l min 5 kV
0
0 0.5 1 1.5 2 2.5 3
30–100 kV
(a) (b) Wavelength l, Å

Figure 2.2 (a) Scheme of an X-ray tube. (b) Intensity distribution in the Röntgen spectrum of molybdenum for different voltages. The
excitation potential of the K-series is 20.1 kV. This series appears as characteristic peaks in the 25 kV curve. The peaks Kα and Kβ are due to 15
L-shell and M-shell drops respectively.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

The important parameters of an X-ray source are the incoming photon was traveling. This mecha-
the following. nism is called photoelectric absorption.
• A second possibility is that the photon transfers
• The amount of electrons hitting the anode and, only part of its energy to eject an electron with a
consequently, the amount of emitted photons con- certain kinetic energy. In that case, a photon of the
trolled by the cathode current multiplied by the remaining lower energy is emitted and its direction
time the current is on (typically expressed in mA s). deviates from the direction of the incoming pho-
Typical values range from 1 to 100 mA s. ton. The electron then escapes in another direction.
• The energy of the electrons hitting the anodes This process is called Compton scattering.
and, consequently, the energy of the emitted pho- • A third mechanism is pair production. If the energy
tons (typically expressed in keV), controlled by of a photon is at least 1.02 MeV, the photon can
the voltage between cathode and anode (typically be transformed into an electron and a positron
expressed in kV). For most examinations the val- (electron–positron pair). A positron is the antipar-
ues vary from 50 to 125 kV. For mammography ticle of an electron, with equal mass but opposite
the voltage is 22–34 kV. The energy of the atom charge. Soon after its formation, however, the
defines the upper limit of the photon energy. positron will meet another electron, and they will
• The total incident energy (typically expressed in annihilate each other while creating two photons
joules, 1 J = 1 kV mA s) at the anode, defined by of energy 511 keV that fly off in opposite direc-
the product of the voltage, the cathode current and tions. This process finds its application in nuclear
the time the current is on. Note that almost all of medicine.
this energy is degraded to heat within the tube. Less • At still higher energies, photons may cause nuclear
than 1% is transmitted into X-rays. reactions. These interactions are not used for
medical applications.

Interaction with matter Interaction of an X-ray beam with tissue

Interaction of photons with matter Consider an X-ray beam and a material of thickness
X-rays and γ-rays are ionizing waves. Such photons are d = xout − xin (see Figure 2.3(a)). Inside the mate-
able to ionize an atom, i.e., to release an electron from rial, the beam is attenuated by the different types of
the atom. Photons with energy less than 13.6 eV are interaction explained above. Although the individual
nonionizing. These photons cannot eject an electron interactions are of statistical nature, the macroscopic
from its atom, but are only able to raise it to a higher intensity of the beam follows a deterministic expo-
energy shell, a process called excitation. Ionizing nential law: The intensity of the outgoing beam Iout is
photons can interact with matter in different ways. related to the intensity of the incoming beam Iin by

• The energy of X-ray photons can be absorbed by an Iout = Iin e−µ d , (2.4)
atom and immediately released again in the form where µ is called the linear attenuation coefficient (typ-
of a new photon with the same energy but traveling ically expressed in cm−1 ). This simple law is only valid
in a different direction. This nonionizing process when the material is homogeneous and the beam con-
is called Rayleigh scattering or coherent scattering sists of photons of a single energy. Actually, µ is a
and occurs mainly at low energies (<30 keV). The function of both the photon energy and the material,
lower the energy the higher is the scattering angle. that is, µ = µ(E, material), for example:
In most radiological examinations it does not play
a major role because the voltage used is typically µ(10 keV, H2 O) = 5 cm−1
in the range from 50 to 125 kV. For mammogra-
µ(100 keV, H2 O) = 0.17 cm−1
phy, however, the voltage is lower (22–34 kV) and (2.5)
Rayleigh scatter cannot be neglected. µ(10 keV, Ca) = 144 cm−1
• A photon can be absorbed by an atom while µ(100 keV, Ca) = 0.40 cm−1 .
16 its energy excites an electron. The electron then
escapes from its nucleus in the same direction as Equation (2.4) can be generalized as follows.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

(a) (b) µ, cm–1

100
K-edge
I in I out

10

Pb I
PbII
1 Pb
Al PbIII
AlII

0.1
x in x out Al I

AlIII
0.01
0.01 0.1 1 10 100
Photon energy, MeV

Figure 2.3 (a) X-ray beam traveling through a slab of material. (b) Linear attenuation coefficient for photons in aluminum and lead. The solid
curves represent the total linear attenuation coefficient. The dashed lines show the partial linear attenuation coefficient for each of the three
effects: I for photoelectric absorption, dominant at low energies; II for Compton scattering, dominant at higher energies; III for pair production,
dominant at very high energies.

• When a beam of single-energy photons trav- atomic number Z . With increasing Z , photoelectric
els through a nonhomogeneous medium, Iout is absorption increases more rapidly than Compton
related to Iin by scattering.
Often the mass attenuation coefficient (µm ) is used

xout
− µ(x) dx instead of the linear attenuation coefficient:
Iout = Iin e xin . (2.6)

µm = µ/ρ, (2.8)
• A real X-ray beam does not contain a single photon
energy but a whole spectrum of energies. Making
the intensity distribution of the incoming where ρ is the mass density of the attenuating medium.

∞ beam a For example, for water ρ = 1 g/cm3 and for calcium
function of the energy, that is, Iin = 0 σ (E) dE,
the intensity of the outgoing beam is equal to ρ = 1.55 g/cm3 . Hence,

 ∞
xout µm (10 keV, H2 O) = 5 cm2 /g

− µ(E,x) dx
Iout = σ (E) e xin dE. (2.7)
0 µm (100 keV, H2 O) = 0.17 cm2 /g
(2.9)
µm (10 keV, Ca) = 93 cm2 /g
Figure 2.3(b) shows that at low energies photoelectric
absorption is most prominent while at intermediate µm (100 keV, Ca) = 0.258 cm2 /g.
energies the Compton scattering dominates. Pair pro-
duction exists only at very high energies. Photoelectric
absorption occurs at photon energies higher than the X-ray detectors
binding energy of K-shell electrons. Hence, the atten- To produce an image from the attenuated X-ray beam,
uation coefficient suddenly increases at this energy, the X-rays need to be captured and converted to
known as the K-edge. Figure 2.3(b) also shows that with image information. Some detectors for digital radio-
increasing energy, photoelectric absorption decreases graphy are relatively recent developments. Older but
more rapidly than Compton scattering. Note also that still in use are the screen–film detector and the image 17
the linear attenuation coefficient increases with the intensifier.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

Screen–film detector stopped. If the delay to reach peak emission is

Screen longer than 10−8 seconds or if the material con-
tinues to emit light after this period, it is said to
Photographic film is very inefficient for capturing
phosphoresce. Phosphorescence in screens is an
X-rays. Only 2% of the incoming X-ray photons con-
undesirable effect, because it causes ghost images
tribute to the output image on a film. This percentage
and occasionally film fogging.
of contributing photons corresponds to the probabil-
ity that an X-ray photon (quantum) is absorbed by
the detector. It is known as the absorption efficiency. Film
The low sensitivity of film for X-rays would yield The film contains an emulsion with silver halide crys-
prohibitively large patient doses. Therefore, an inten- tals (e.g., AgBr). When exposed to light, the silver
sifying screen is used in front of the film. This type of halide grains absorb optical energy and undergo a
screen contains a heavy chemical element that absorbs complex physical change. Each grain that absorbs
most of the X-ray photons. When an X-ray photon a sufficient amount of photons contains dark, tiny
is absorbed, the kinetic energy of the released elec- patches of metallic silver called development centers.
tron raises many other electrons to a higher energy It is important to note that the amount of photons
state. When returning to their initial state they pro- required is independent of the grain size. When the
duce a flash of visible light, called a scintillation. Note film is developed, the development centers precipi-
that these light photons are scattered in all directions. tate the change of the entire grain to metallic silver.
Consequently, two intensifying screens can be used, The more light reaching a given area of the film, the
i.e., one in front and one behind the film, to increase more grains are involved and the darker the area after
the absorption efficiency further. The portion of the development. In this way a negative is formed. After
light that is directed toward the film contributes to the development, the film is fixed by chemically removing
exposure of the film. In this way, the absorption effi- the remaining silver halide crystals.
ciency can be increased to more than 50% instead of In radiography, the negative image is the final
the 2% for film. Because the light is emitted in all direc- output image. In photography, the same procedure
tions, a smooth light spot (the PSF, see p. 3) instead of has to be repeated to produce a positive image. The
a sharp peak hits the film and causes image blurring. negative is then projected onto a sensitive paper car-
X-ray intensifying screens consist of scintillating rying silver halide emulsion similar to that used in the
substances that exhibit luminescence. Luminescence is photographic film.
the ability of a material to emit light after excitation, Typical characteristics of a film are its graininess,
either immediately or delayed. speed, and contrast.

• Fluorescence is the prompt emission of light when • Graininess The image derived from the silver crys-
excited by X-rays and is used in intensifying tals is not continuous but grainy. This effect is
screens. A material is said to fluoresce when light most prominent in fast films. Indeed, because the
emission begins simultaneously with the exciting amount of photons needed to change a grain into
radiation and light emission stops immediately metallic silver upon development is independent
after the exciting radiation has stopped. Initially, of the grain size, the larger the grains, the faster the
calcium tungstate (CaWO4 ) was most commonly film becomes dark.
used for intensifying screens. Advances in tech- • Speed The speed of a film is inversely propor-
nology have now resulted in the use of rare tional to the amount of light needed to produce
earth compounds, such as gadolinium oxysulfide a given amount of metallic silver on develop-
(Gd2 O2 S). A more recent scintillator material is ment. The speed is mainly determined by the
thallium-doped cesium iodide (CsI:Tl), which has silver halide grain size. The larger the grain size
not only an excellent absorption efficiency but also the higher the speed because the number of pho-
a good resolution because of the needle-shaped or tons needed to change the grain into metallic silver
pillarlike crystal structure, which limits lateral light upon development is independent of the grain size.
diffusion. Speed is expressed in ASA (American Standards
• Phosphorescence or afterglow is the continuation Association) or in ISO (International Standards
18
of light emission after the exciting radiation has Organization). These units are the same.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

For X-ray imaging with a screen–film combina- D

4
tion, it makes more sense to speak about the speed
of the screen–film combination: how many X-ray 3.5
photons are needed to produce a certain density
3
on the film. The speed then depends on the prop-
erties of the intensifying screen and the film, but 2.5
also on the quality of film–screen contact, and on a
good match between the emission spectrum of the 2

screen and the spectral sensitivity of the film used. 1.5

Because light is emitted in all directions, a sig-
nificant proportion, about 50%, of that light is 1
not directed toward the film. A reflective layer 0.5
behind the screen–film–screen redirects it toward
the film, ensuring that it contributes to expo- 0
log E
sure. This has the advantage of increasing the
speed of the screen–film–screen combination with Figure 2.4 Typical sensitometric curve for radiographic film. D is
a corresponding reduction in patient dose. the optical density and E the exposure.

• Contrast The most widely used description of the

photosensitive properties of a film is the plot of
the optical density D versus the logarithm of the
Image intensifier
An image intensifier works as follows (see Figure 2.5).
exposure E. This curve is called the sensitometric
A fluorescent screen converts the X-rays into visible
curve. The exposure is the product of incident light
light. The emitted light hits a photocathode, and the
intensity and its duration. The optical density is
energy of the photons releases electrons from this cath-
defined by
ode. A large potential difference between the cathode
and the output accelerates the ejected electrons. The
Iin resulting electron beam is directed onto a small flu-
D = log , (2.10)
Iout orescent screen by electrostatic or magnetic focusing
and converted to light photons again. This focusing
where Iin and Iout are the incoming and outgoing makes the system suitable to be coupled to a camera
light intensity when exposing the developed film without any loss of light. The main advantage of an
with a light source. Note that Iin and Iout are image intensifier system is that it is capable of pro-
different from the incident light intensity in the ducing dynamic image sequences in real time at video
definition of the exposure E, in which it refers to rate, a process known as fluoroscopy. However, when
the light emitted by the intensifying screen during compared with film–screen systems, the images are
X-ray irradiation. degraded in three ways.
Figure 2.4 shows a typical sensitometric curve. • The spatial resolution will generally be less than
It is S-shaped. In low- and high-density areas, con- that of a film–screen system because of the limited
trast is low and there is little information. Only the camera resolution.
linear part is really useful and its slope character-
• Because of the additional conversions (light →
izes the film contrast. The maximal slope of the
electrons → light), the noise increases slightly.
curve is known as the gamma of the film. Note that
• Geometric distortion occurs, called pin-cushion
a larger slope implies a higher contrast at the cost
of a smaller useful exposure range.∗ distortion, particularly toward the borders of the
image.

∗ Double contrast films also exist. Their sensitometric curve con- Detectors for digital radiography
tains two linear parts with a different slope, i.e., a high-contrast part Storage phosphors
as usual, continued by a low-contrast part at high optical densities.
This increases the perceptibility in hyperdense regions, such as in A special case of phosphorescence is when part of 19
mammography at the border of the breast. the absorbed energy is not released immediately in

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

Input Figure 2.5 Scheme of an image

Phosphor intensifier. The camera is placed against
the output screen to minimize light loss.
Input (Reprinted with permission of Kieran
Photocathode Evacuated Tube Maher.)
Screen

X-rays Output
Screen

light

Electron lenses

Output
Phosphor

High Voltage Power Supply

the form of light. The temporarily stored energy

can be released upon stimulation by other forms of
energy such as laser light. This phenomenon is called
photostimulated luminescence and is used in digital
radiography. This type of scintillator is called a storage
phosphor or photostimulable phosphor. The screen–
film combination is then replaced by a screen coated
with such a scintillator. When X-rays are absorbed
by the phosphor, electrons are pumped up from the
valence band to the conduction band. In a classical
scintillator plate such an electron falls back to the
valence band while releasing its energy in the form of
a light photon. In a storage phosphor, however, these
electrons are trapped by electron traps, which are impu-
rities in the scintillator. In this way, the incident X-ray
energy is converted into stored energy. After expo-
sure a latent image is trapped in the scintillator. The
latent image can be stored in the phosphor plate for a
considerable period after exposure. It takes 8 hours to
decrease the stored energy by about 25%. The stored
energy can be extracted by pixelwise scanning with a
laser beam. This way the trapped electrons receive a
new energy shot that allows them to escape from their
trap and fall back into the valence band. The latent Figure 2.6 This system scans the latent image with a laser beam
and erases the residual image on the storage phosphor after which
image information is thereby released as visible light, the screen can be reused for new X-ray exposure.
which is captured by an optic array and transmitted
to a photomultiplier. The photomultiplier converts
the detected light into an analog electrical signal. This exposure. As soon as the radiologic technician puts the
analog signal is then converted in an A/D converter to cassette into the scanner (Figure 2.6), this whole laser
a digital bit stream. The residual information on the scanning and cleaning process is done automatically.
20 scintillator screen is erased by a strong light source, Storage phosphor screens provide a much wider
after which the screen can be reused for new X-ray useful exposure range than conventional film–screen

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

systems. Moreover, the storage phosphor is a linear light photons negatively influences the PSF because of
detector. This means there is no contrast reduction the light distribution in different directions. A more
in the low- and high-density areas of the image, as is recent technique eliminates the need for a scintillator
the case with the S-shaped sensitometric curve. Con- by using a photoconductor, such as amorphous sele-
sequently, the system is much more tolerant to over- nium (a-Se) or cadmium telluride (CdTe), instead of
exposure and underexposure, and retakes caused by a phosphor. When exposed to radiation, the photo-
suboptimal exposure settings (mA s, kV) are reduced. conductor converts the energy of the X-ray photons
In theory, a reduction of the radiation dose per image directly into an electrical conductivity proportional to
is also possible because of the available contrast at low the intensity of the radiation. To scan this latent image,
exposure. However, dose reduction adversely affects the photoconductor layer is placed upon an active
the SNR of the resulting image. Therefore, reduc- matrix array that consists of a 2D array of capacitors
ing the dose per examination must be considered in (instead of photodiodes) deposited onto the amor-
relationship to the diagnostic information required. phous substrate. These capacitors store the electric
Often, the greed for diagnostic detail slightly increases charge produced by detected X-ray photons until it
the dose rather than reducing it. is read out by the electronic circuit of the active matrix
A second advantage of digital radiography is that array. This technology is known as direct radiography
the image is available for computer postprocessing as against the indirect approach where light is pro-
such as image enhancement and quantification. More- duced by a scintillator as an intermediate step in the
over, the image can easily be stored and transported transformation of X-rays to a measurable signal.
in digital form, making the images more accessible Active matrix flat panel detectors have become an
and making large film archives unnecessary. Today, accepted technology for mammography because of
digital picture archiving and communication systems their overall performance (see p. 24 below on DQE).∗
(PACS) are part of hospital information systems, mak-
ing the medical images immediately available through
the digital network in the same way as the other patient Dual-energy imaging
information. By taking two radiographic images, each capturing
a different energy spectrum, the image of substances
with a high atomic number (e.g., bone, calcifications,
Active matrix flat panel detectors stents) can be separated from that of the soft tissue
Newer detector technologies for digital radiography by proper image processing. This way two different
are flat panel detectors with fast-imaging capability. selective images are obtained, for example, a soft-
These systems produce nearly real time images, as tissue image and a bone image. Several methods have
opposed to storage phosphor systems which require a been proposed to calculate tissue selective images.
readout scan on the order of a minute and a workflow The method explained here is also used in computed
similar to that for screen–film systems. tomography (p. 48).
Traditional electronic capturing devices, including Two system configurations have been used. The
CCDs (charge-coupled devices), are almost exclusively first captures two radiographic images in a short time
based upon Si-crystal technology, and for manufac- interval (e.g., 200 ms) and at different X-ray tube volt-
turing reasons this restricts the devices to small areas. ages (e.g., peaks at 110–150 kV and at 60–80 kV). The
This is because it is difficult and expensive to create a second configuration contains two layers of scintillator
large defect-free semiconductor crystal. A flat, large- detectors and acquires the images in a single expo-
area integrated circuit, called an active matrix array, sure. The top layer detects and filters most low-energy
can easily be made by depositing a 2D array of identical photons, while the bottom layer detects primarily
semiconductor elements onto an amorphous material, high-energy photons. A third configuration is promis-
such as hydrogenated amorphous silicon (a-Si:H). ing but immature. It uses photon counting detectors
A light-sensitive active matrix array can be pro-
duced by depositing an array of photodiodes onto the
a-Si:H substrate. By coupling it to a fluorescent plate it ∗ Commercial mammography systems exist that are able to count

functions as a large and fast flat panel X-ray detector. the individual X-ray photons with a very high absorption effi-
ciency. To obtain their unsurpassed DQE they make use of
In spite of the technological progress in scintilla- crystalline silicon strip detectors in combination with a slit-scanning 21
tor materials, the conversion of X-ray radiation into technology.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

3.00
cm–1 be written as a linear combination of the attenuation
Calcium
coefficient of two selected materials, provided that
2.00
the attenuation properties of both basis materials are
1.50
sufficiently different (e.g., bone and soft tissue)
1.00

0.70 µS (E) = a1 · µ1 (E) + a2 · µ2 (E). (2.13)

Iodine
0.50
Substituting Eq. (2.13) in Eq. (2.7) for a spectrum with
0.30 energy range [Emin , Emax ] yields
0.20 Water
 Emax

− µ(E,s) ds
0.15 I (x, y) = σ (E) e Lx,y dE
20 40 60 80 100 120 140 keV Emin
Figure 2.7 Linear attenuation coefficient as a function of the
 Emax

− Lx,y (a1 (s)·µ1 (E)+a2 (s)·µ2 (E)) ds
energy for calcium (Ca), water and the contrast agent iodine (I) in = σ (E) e dE,
water (10 mg/ml). Note the K-edge discontinuity of I at 33.2 keV. Emin
(2.14)

that count and measure the energy of the photons. where Lx,y is the projection line arriving at pixel (x, y)
This multi-energy technique has an improved spectral of the radiographic image. When taking images, the
sensitivity, needs only one radiographic image and is tissue-dependent coefficients a1 and a2 are unknown.
insensitive to patient motion. Defining
Dual-energy imaging relies on the dependence 
of the attenuation coefficient µ on the energy E A1 (x, y) = a1 (s) ds
(Figure 2.7). In the absence of K-edge discontinu- Lx,y
 (2.15)
ities in the used energy range [Emin , Emax ] the linear
A2 (x, y) = a2 (s) ds,
attenuation coefficient can be approximated as Lx,y

µ(E) ≈ µp (E) + µC (E) Eq. (2.14) can be written as

1
≈ ap m + aC fKN (E). (2.11) I (x, y)
E  Emax
The two components express the attenuation due to = σ (E) e−(A1 (x,y)·µ1 (E)+A2 (x,y)·µ2 (E)) dE.
photoelectric interaction and Compton scatter respec- Emin

tively. The exponent m is an empirically defined (2.16)

parameter (e.g., m = 3 [2]). fKN (E) is the so-called A1 (x, y) and A2 (x, y) represent the equivalent thick-
Klein–Nishina function. The tissue-dependent coeffi- ness of the basis materials along ray Lx,y . In this
cients ap and aC are related to the physical material equation A1 (x, y) and A2 (x, y) are unknown, but they
properties: can be retrieved. Indeed, if two radiographic images
ρ are acquired, each at a different energy with corre-
ap ≈ Kp Z n , n≈4 sponding spectra σLE and σHE , the following system
A
ρ (2.12) of two nonlinear equations must be solved to calculate
aC ≈ KC Z A1 (x, y) and A2 (x, y) in pixel (x, y)
A
where Kp and KC are constants, ρ is the mass density, IHE (x, y)
A the mass number and Z the atomic number of the  Emax
attenuating medium [2]. = σLE (E) e−(A1 (x,y)·µ1 (E)+A2 (x,y)·µ2 (E)) ds dE
Using Eq. (2.11) it can easily be shown that the Emin

attenuation coefficient of an arbitrary substance S can ILE (x, y)

 Emax
[2] R. E. Alvarez and A. Macovski. Energy-selective reconstructions = σHE (E) e−(A1 (x,y)·µ1 (E)+A2 (x,y)·µ2 (E)) ds dE.
22 in x-ray computerized tomography. Physics in Medicine and Biology, Emin
21(5): 733–744, 1976. (2.17)

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

In case of a single exposure with spectrum σLE and two • The size of the focal spot. The anode tip should
detector layers, the second spectrum σHE is defined as make a large angle with the electron beam to
produce a nicely focused X-ray beam.
σHE (E) = σLE (E) e−µf (E) tf , (2.18) • The patient. Thicker patients cause more X-
ray scattering, deteriorating the image resolution.
where µf (E) and tf are the known attenuation coef- Patient scatter can be reduced by placing a collima-
ficient and thickness of the filtering top detector tor grid in front of the screen (see p. 24). The grid
layer. allows only the photons with low incidence angle
Various approaches exist to solve Eqs. (2.17). For to reach the screen.
• The light scattering properties of the fluorescent
example, the inverse relationship can be modeled by a
second- or third-order polynomial. If more than two screen.
measurements and corresponding equations are avail- • The film resolution, which is mainly determined
able, an optimization strategy is required to solve the by its grain size.
overdetermined system. This is for example the case • For image intensifier systems and digital radiogra-
when photon-counting detectors can be used. More phy, the sampling step at the end of the imaging
information on numerical optimization can be found chain is an important factor.
in [3].
Using the obtained values of A1 (x, y) and A2 (x, y) The resolving power (i.e., the frequency where
the original radiographic image can be separated the MTF is 10%) of clinical screen–film combinations
into two material equivalent images (e.g., bone and varies from 5 up to 15 lp/mm. In most cases, spatial
soft tissue). Note that the above theory needs some resolution is not a limiting factor in reader perfor-
modification in the presence of a substance with an mance with film. For images with storage phosphors,
observable K-edge in the energy range [Emin , Emax ]. In a resolving power of 2.5 up to 5 lp/mm (at 10% con-
that case Eqs. (2.11) and (2.13) have to be extended trast) is obtained. This corresponds to a pixel size of
with a third component and corresponding coeffi- 200 to 100 µm, which is mostly sufficient except for
cients aK and a3 respectively. This yields a third mammography, for which more recent detector tech-
unknown A3 in Eq. (2.16) and, hence, requires a multi- nology (see active matrix flat panel detectors, p. 21) is
energy approach with at least three different measure- needed. Depending on the size of the object, it is clear
ments [4]. The original image is then separated into that images with 2000 by 2000 pixels and even more
three instead of two basis images, the third being an are needed to obtain an acceptable resolution.
image of the substance with K-edge. The strength of
K-edge imaging is that the energy dependence of a
material with K-edge is very different around its K- Contrast
edge, resulting in a high sensitivity for multi-energy The contrast is the intensity difference in adjacent
imaging. K-edge imaging is immature but offers regions of the image. According to Eq. (2.7) the
opportunities for target-specific contrast agents and image intensity depends on the attenuation coeffi-
drugs, particularly in multi-energy CT (see p. 48). cients µ(E, x) and thicknesses of the different tissue
layers encountered along the projection line. Because
the attenuation coefficient depends on the energy of
Image quality the X-rays, the spectrum of the beam has an important
Resolution influence on the contrast. Soft radiation, as used in
mammography, yields a higher contrast than hard
The image resolution of a radiographic system radiation.
depends on several factors. Another important factor that influences the con-
trast is the absorption efficiency of the detector, which
[3] J. Nocedal and S. Wright. Numerical Optimization, Volume XXII is the fraction of the total radiation hitting the detec-
of Springer Series in Operations Research and Financial Engineering. tor that is actually absorbed by it. A higher absorption
Springer, second edition, 2006. efficiency yields a higher contrast.
[4] E. Roessl and R. Proksa. K-edge imaging in x-ray computed
tomography using multi-bin photon counting detectors. Physics in In systems with film, the contrast is strongly deter- 23
Medicine and Biology, 52: 4679–4696, 2007. mined by the contrast of the photographic film. The

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

higher the contrast, the lower the useful exposure influence the DQE, particularly the absorption effi-
range. In digital radiography, contrast can be adapted ciency of the detector, the point spread function of
after the image formation by using a suitable gray value the detector and the noise introduced by the detector.
transformation (see p. 4). Note however that such a Figure 2.8 shows an example of the DQE as a function
transformation also influences the noise, thus keeping of frequency for three different detector technologies.
the CNR unchanged.

Noise Artifacts
Quantum noise, which is due to the statistical nature Although other modalities suffer more from severe
of X-rays, is typically the dominant noise factor. A artifacts than radiography, X-ray images are generally
photon-detecting process is essentially a Poisson pro- not artifact free. Scratches in the detector, dead pixels,
cess (the variance is equal to the mean). Therefore, unread scan lines, inhomogeneous X-ray beam inten-
the noise amplitude (standard deviation) is propor- sity (heel effect), afterglow, etc., are not uncommon
tional to the square root of the signal amplitude, and and deteriorate the image quality.
the SNR also behaves as the square root of the sig-
nal amplitude. This explains why the dose cannot be Equipment
decreased unpunished. Doing so would reduce the
Let us now take a look at the complete radiographic
SNR to an unacceptable level. Further conversions
imaging chain, which is illustrated schematically in
during the imaging process, such as photon–electron
Figure 2.9. It consists of the following elements.
conversions, will add noise and further decrease
the SNR. • The X-ray source.
To quantify the quality of an image detector the • An aluminum filter, often complemented by a cop-
measure detective quantum efficiency (DQE) is often per filter. This filter removes low-energy photons,
used. The image detector is one element in the imag- thus increasing the mean energy of the photon
ing chain and to quantify its contribution to the SNR, beam. Low-energy photons deliver doses to the
the DQE is used, which expresses the signal-to-noise patient but are useless for the imaging process
transfer through the detector. The DQE can be cal- because they do not have enough energy to travel
culated by taking the ratio of the squared SNR at the through the patient and never reach the detector.
detector output to the squared SNR of the input signal Because low-energy photons are called soft radia-
as a function of spatial frequency: tion and high-energy photons hard radiation, this
removal of low-energy photons from the beam is
SNR 2out (f ) called beam hardening.
DQE = . (2.19)
SNR 2in (f ) • A collimator to limit the patient area to be
irradiated.
It is a measure of how the available signal-to-noise
ratio is degraded by the detector. Several factors
low-energy
collimating
absorbing
scatter grid
filter

Figure 2.8 DQE of four digital X-ray detection systems, obtained X-ray source collimator detector
under standardized measurement conditions. The DQE curve is cut
24 off at a frequency close to the Nyquist frequency, i.e., half of the Figure 2.9 Schematic representation of the radiographic imaging
sampling frequency. (Courtesy of Agfa HealthCare.) chain.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

• The patient, who attenuates the X-ray beam and storage phosphor cassette, as well as an image inten-
produces scatter. sifier beneath the table. More recent X-ray systems
• A collimating scatter grid. This is a collimator contain an active matrix flat panel detector with fast-
that absorbs scatter photons. It stops photons with imaging capability, which replaces the cassette and
large incidence angle, whereas photons with small image intensifier (Figure 2.11).
incidence angle can pass right through the grid. Figure 2.12 shows a 3D rotational angiography sys-
The grid can be made of lead, for example. Note tem (3DRA). Images of the blood vessels can be made
that a scatter grid is not always used in paediatrics from any orientation by rotating the C-arm on which
because in small children the scatter is limited. the X-ray tube and image detector are mounted at
•
both ends. By continuously rotating the C-arm over
The detector. This can be a screen–film combina-
a large angle (180◦ and more), sufficient projection
tion in which a film is sandwiched between two
images are obtained to reconstruct the blood vessels
screens (see p. 18), an image intensifier coupled to
in three dimensions (3D) (Figure 2.13). The mathe-
a camera (see p. 19), a cassette containing a storage-
matical procedure used to calculate a 3D image from
phosphor plate (see p. 19), or an active matrix flat
its projections is also used in computed tomography
panel detector (see p. 21) or dual-layer detector
(CT) and is explained in Chapter 3.
(see p. 21).

Figure 2.10 shows a general purpose radiographic Clinical use

room. The table can be tilted in any orientation, from Today, the majority of the radiographic examinations
the horizontal to the vertical position. The X-ray sys- in a modern hospital are performed digitally. X-ray
tem contains a tray for a conventional film-based or a images can be static or dynamic. Static or still images

(a) (b)

Figure 2.10 Multipurpose radiographic room. The table can be tilted in any orientation. Both an image intensifier and a storage phosphor 25
are available.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

Radiographic images
These are made of all parts of the human body. They
are still responsible for the majority of radiologi-
cal examinations. The most common investigations
include the following.
• skeletal X-rays (see Figure 2.14),
• chest images (radiographs of the thoracic cavity
and heart, see Figure 2.15),
• mammography (images of the breasts, see
Figure 2.16),
• dental X-rays (images of the teeth and jaw).

Fluoroscopic images
These are image sequences produced in real time.
Consequently, their application field focuses on inves-
tigations in which motion or the instant availability of
the images, or both, are crucial. This application field
is obviously narrower than that of radiographic exam-
inations, which explains why the number of fluoro-
scopic guided examinations is an order of magnitude
lower. The most typical applications, in decreasing
order of occurrence, include the following.
• Interventional fluoroscopy (see Figure 2.17). This
application is responsible for the majority of fluo-
roscopic sequences. Typically, the images are used
Figure 2.11 In more recent X-ray systems the cassette and image to guide and quickly verify surgical actions, partic-
intensifier are replaced by an active matrix flat panel detector. This ularly in bone surgery, such as for osteosynthesis
picture shows a Siemens system with large-area amorphous silicon
detector coupled to a CsI scintillator plate.
(traumatology, orthopedics).
• Angiography (see Figure 2.18), which takes images
of blood vessels through the injection of an iodine-
containing fluid into the arteries or veins. Usually,
are made with a film–screen combination or with digi- subtraction images are made by mathematically
tal radiography, whereas dynamic images are obtained subtracting postcontrast and precontrast images
with an image intensifier or an active matrix flat panel followed by a simple gray level transformation to
detector and viewed in real time on a TV monitor increase the image contrast of the vessels. The
or computer screen. Dynamic image sequences are result is an image in which the blood vessels appear
commonly known as fluoroscopic images as against as contrasting line patterns on a homogeneous
radiographic images, which refer to static images. background. Obviously, it is essential that the
In X-ray images, the attenuation differences of patient does not move during the imaging pro-
various nonbony matter are usually too small to distin- cedure, to avoid motion blurring and subtraction
guish them. A contrast agent or dye (i.e., a substance artifacts in the images. Traditionally, angiogra-
with a high attenuation coefficient) may overcome phy has been used for diagnosis of conditions
this problem. It is especially useful for intravascu- such as heart ischemiae caused by plaque buildup.
lar (blood vessels, heart cavities) and intracavitary However, today radiologists, cardiologists, and
(kidney, bladder, etc.) purposes. vascular surgeons also use the X-ray angiography
Following are a number of typical examples of fre- procedure to guide minimally invasive interven-
26 quently used examinations. They are subdivided into tions of the blood vessels, such as for vascular
radiographic images and fluoroscopic images. repermeabilization (Figure 2.12).

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

(a) (b)

Figure 2.12 3D rotational angiography (3DRA). (a) C-arm with X-ray tube and image intensifier at both ends. (b) More recent system in which
the image intensifier has been replaced by an active matrix flat panel detector with an acquisition frame rate of up to six 2048 × 2048 images
(12 bbp) per second. By rotating the C-arm on a circular arc (e.g., 240◦ in 4 s) around the patient, a series of projection images are acquired
that can be used to compute a 3D image of the blood vessels. (Courtesy of Professor G. Wilms, Department of Radiology.)

• Barium fluoroscopy of the gastrointestinal tract

after the patient swallows barium contrast solu-
tion and/or where the contrast is instilled via the
rectum (see Figure 2.19).
• Urography (image of the kidneys and bladder)
using an iodine-containing contrast fluid.

Biologic effects and safety

Even at very low X-ray doses the energy deposited by
ionizing radiation, such as X-rays, may be sufficient
to damage or destroy cells. Although this generally
has no negative consequence, the probability always
exists that modifications in single cells could lead to
malignancy (cancer) or genetic changes. There is no
evidence of a threshold dose below which the proba-
bility would be zero. If the X-ray dose increases, the
frequency of cell damage and the occurrence of cancer
Figure 2.13 3D image of the cerebral blood vessels reconstructed increases, but not the severity of the cancer.
from a series of 2D projection images around the patient, obtained Malignant disease and heritable effects, for which
with the 3DRA system shown in Figure 2.12(b). Selective injection of the probability but not the severity is proportional to
the right internal carotid artery in a patient with a subarachnoid
hemorrhage showing an aneurysm of the anterior communicating the dose, without any threshold, are stochastic effects of 27
artery. (Courtesy of Professor G. Wilms, Department of Radiology.) radiation. Deterministic effects of radiation also exist.

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

(a) (b)

Figure 2.14 (a) Double mandibular fracture with strong displacement to the left. (b) Solitary humeral bone cyst known as ‘‘fallen leaf sign’’.
(Courtesy of Dr. L. Lateur, Department of Radiology.)

(a) (b)

Figure 2.15 Radiographic chest image showing multiple lung metastases. (Courtesy of Professor J. Verschakelen, Department of Radiology.)

28

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

(a) (b)

Figure 2.16 (a) Dense opacity with spicular border in the cranial part of the right breast; histological proven invasive ductal carcinoma.
(b) Cluster of irregular microcalcifications suggesting a low differentiated carcinoma. (Courtesy of Dr. Van Ongeval, Department. of Radiology.)

They are injuries to a large population of cells where

repair mechanisms fail and the complete tissue is
damaged. Deterministic effects are characterized by
a threshold dose and an increase in the severity of the
tissue reaction with increasing dose.
The SI unit of absorbed dose, D, is the gray (Gy).
One Gy is one joule per kilogram of irradiated mate-
rial. If the average absorbed dose, DT , in organ or
tissue T is known, it is, for example, possible to predict
the onset of deterministic effects. Tables of threshold
values can be found in the literature. For example, a
dose of 5 Gy in a single exposure at the level of the eye
lens can cause visual impairment due to cataract. In
clinical practice deterministic effects are rare.
The probability of stochastic effects from radiation
depends heavily on the type of radiation. Some types
Figure 2.17 Postoperative fluoroscopic control of bone fixation of radiation are more detrimental per unit of absorbed
with plate and screws after a complete fracture of the humerus. dose than others. To assess the risk of the stochastic
(Courtesy of Dr. L. Lateur, Department of Radiology.)
effects of radiation in a particular organ or tissue T ,
29

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

To assess the overall radiation detriment † from

stochastic effects, the effective dose,†† also expressed
in sieverts (Sv), is used. The effective dose is the tis-
sue weighted sum of equivalent doses in all irradiated
tissues or organs of the body, that is,

E = T (wT · HT ), (2.21)

where HT is the equivalent dose in tissue or organ

T and wT the tissue weighting factor. The weights wT
represent the (rounded) relative radiation detriments
of the individual organs and tissues. The sum of all
weights equals 1:

T wT = 1. (2.22)

For example, the nominal risk coefficient, expressed in

cases per 10 000 persons per sievert, for the liver is 30.
The detriment, i.e., the radiation detriment adjusted
nominal risk coefficient, is 26.6. Given a total detri-
Figure 2.18 Cerebral angiogram showing an aneurysm or saccular
dilation of a cerebral artery. (Courtesy of Professor G. Wilms,
ment for all organs and tissues of 574, the relative
Department of Radiology.) detriment is 0.046 (rounded: wliver = 0.04).
Tissue weighting factors can be found in the “2007
Recommendations of the International Commission
on Radiological Protection” (Annals of the ICRP, pub-
the equivalent dose, HT , is used:
lication 103). They are averaged over all ages and both
sexes and therefore do not apply to particular indi-
HT = R (wR · DT ,R ), (2.20) viduals. Red bone marrow, colon, lung, stomach and
breast have a tissue weighting factor of 0.12. Gonads
where DT ,R is the average absorbed dose from radi- have 0.08; bladder, oesophagus, liver and thyroid have
ation of type R in tissue or organ T and wR is the 0.04; bone surface, brain, salivary glands and skin have
radiation weighted factor, which is a measure of its 0.01. Thirteen remainder tissues have been defined,
relative biological impact per unit of absorbed dose. i.e., adrenals, extrathoracic region, gall bladder, heart,
The SI unit of equivalent dose is the sievert (Sv). The kidneys, lymphatic nodes, muscle, oral mucosa, pan-
radiation weighting factor for X-rays, electrons and creas, prostate (male), small intestine, spleen, thymus
muons is 1, for protons and charged pions it is 2, and and uterus/cervix (female). Because the sum of all
for heavier particles 20. For neutrons there is no sin-
gle value but the weighting factor is a function of the † “Radiation detriment is a concept used to quantify the harmful

neutron energy. In most medical imaging applications effects of radiation exposure in different parts of the body. It is
determined from nominal risk coefficients, taking into account the
only X-rays are involved and wR is simply 1. In the lit- severity of the disease in terms of lethality and years of life lost.
erature, factors can be found that relate the equivalent Total detriment is the sum of the detriment for each part of the
organ or tissue dose to the risk of stochastic effects. For body (tissues and/or organs)” (Annals of the ICRP, publication 103,
2007).
example, lung cancer occurs on average in 114 cases †† “The concept of “effective dose” associated with a given expo-
per 10 000 persons per sievert, yielding a so-called sure involves weighting individual organs and tissues of interest
nominal risk coefficient ∗ of lung cancer induction of by the relative detriments for these parts of the body. In such a
system, the weighted sum of the tissue-specific dose equivalents,
1.14%/Sv. called the effective dose, should be proportional to the total esti-
mated detriment from the exposure, whatever the distribution of
equivalent dose within the body. The components of detriment are
∗ “Nominal risk coefficients are derived by averaging sex and age- essentially the same for cancer and heritable disease and, if desired,
30 at-exposure lifetime risk estimates in representative populations” these detriments may be combined” (Annals of the ICRP, publication
(Annals of the ICRP, publication 103, 2007). 103, 2007).

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

(a) (b)

Figure 2.19 (a) Double contrast (barium + gas insufflation) enema with multiple diverticula in the sigmoid colon (arrows). (b) Polypoid mass
proliferating intraluminally (arrowhead on the spotview). (Courtesy of Professor E. Ponette, Department of Radiology.)

weights equals 1, the weight for the remainder tissues According to the International Commission on
is 0.12. It must be applied to the arithmetic mean dose Radiological Protection (ICRP) the relative radiation
of the 13 organs and tissues. detriment adjusted nominal risk coefficient for cancer
Effective dose is a valuable measure to compare dif- is 5.5%/Sv and for heritable effects up to the second
ferent examinations. Examples of effective doses for generation is 0.2%/Sv. For adults (18 to 64 years), these
some typical radiographic examinations are: dental risk factors are a little lower, i.e., 4.1%/Sv and 0.1%/Sv
0.005–0.02 mSv; chest 0.01–0.05 mSv; skull 0.1– respectively.
0.2 mSv; pelvis 0.7–1.4 mSv; lumbar spine 0.5–1.5 mSv. Because of the potential risk of medical irradia-
Note that many examinations require more than one tion, the ICRP recommends keeping the magnitude
or a continuous X-ray exposure, which increases of individual examination doses as low as reason-
the dose. The use of fluoroscopy for diagnostic and ably achievable (ALARA principle). There are no dose
therapeutic reasons may yield doses around 5 mSv. limits for patients, but every exposure should be jus-
Examples are intravenous urography (3 mSv), barium tified. This is, to a large extent, a medical decision.
enema (8 mSv) and endoscopic retrograde cholan- The physician should have as much knowledge as
giopancreatography (4 mSv). Interventional proce- possible about previous examinations of the patient
dures, such as performed in the angiography room and about the patient’s condition. Pregnancy, for
or in the catheterization lab, may have much higher example, is a state where risks are increased. Most
doses, and occasionally even skin doses that reach countries have now introduced diagnostic reference
the thresholds for deterministic effects. Relatively low levels and can verify in this way whether the X-ray
doses are seen with cerebral angiography (5 mSv) doses for typical examinations in medical centers are
and much higher doses for transjugular intrahep- too high or too low. Particular attention is given to
atic portosystemic shunt procedures (TIPS) (70 mSv). screening examinations because they are performed
Compare this to the dose equivalent due to natural on asymptomatic people. In this regard, there is a lot of
31
sources, which is 2–3 mSv per year. experience in breast cancer screening programs, where

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009
 Chapter 2: Radiography

European Guidelines are widely applied. Special atten- Flat panel detectors for a large field of view and
tion should also be given to children and to high-dose with a fast readout capability will become available
imaging, such as interventional radiology. for 3D imaging. Hence, 2D projective imaging will
Furthermore, the ICRP recommends limiting all further be augmented by 3D volumetric imaging (see
exposed workers from regulated radiation practices to also Chapter 3).
20 mSv per year when averaged over five years and It can also be expected that the DQE of the
the public to 1 mSv per year. In particular, physicians detectors will continue to improve, yielding reduced
may receive a significant exposure when doing proce- radiation doses or images with enhanced contrast-to-
dures under fluoroscopy, but they too must not exceed noise ratio. Furthermore, photon counting detectors,
20 mSv per year. There are strict protection protocols which count the number of photons and measure
they have to follow, among which is the protection of their energy, will become commercially available by
the body and the thyroid gland with a lead apron and employing direct radiography with very fast readout
collar. A dosimeter, which is a small device clipped capability.
to the personnel’s clothing, measures the cumulative Currently all medical images can be fully integrated
absorbed dose. into the hospital information system. The images can
be interpreted on screen by the radiologist and elec-
tronically transmitted to the referring physician. It
Future expectations can be expected that manual interventions during the
Today, other imaging modalities, such as ultra- image acquisition process, such as cassette handling
sound, CT, and MRI, have largely replaced a number and parameter setting, will be further reduced. This
of X-ray examinations. Examples are arthrography will have a strong impact on the work flow in a med-
(joints), myelography (spinal cord), cholangiogra- ical imaging department. Furthermore, the computer
phy (bile ducts), cholecystography (gall bladder), and will behave as an intelligent assistant for the radiolo-
pyelography (urinary tract). Although radiography gist and will improve his/her performance. Computer
will remain an important imaging modality, this aided diagnosis (CAD) is discussed in more detail in
evolution can be expected to continue. Chapter 7.

32

Cambridge
[Link] Published online byBooks Online
Cambridge © Cambridge
University Press University Press, 2009