chapter

13
The Gamma Camera:
Basic Principles

Radionuclide imaging is the most important little use for imaging because they cannot
application of radioactivity in nuclear medi- penetrate more than a few millimeters of
cine. Radionuclide imaging laboratories are tissue. Therefore they cannot escape from
found in almost every hospital, performing within the body and reach an external radia-
hundreds and even thousands of imaging pro- tion detector, except from very superficial
cedures per month in larger institutions. tissues. Bremsstrahlung (see Fig. 6-1) gener-
In this chapter, we discuss briefly some ated by electron emissions is more penetrat-
general aspects of radionuclide imaging, and ing, but the intensity of this radiation
we describe the basic principles of the most generally is very weak.
widely used imaging device, the gamma Imaging system detectors must therefore
camera, also known as the Anger scintillation have good detection efficiency for rays. It is
camera, named after its inventor, Hal Anger also desirable that they have energy discrimi-
(see Chapter 1, Section C and Fig. 1-3). The nation capability, so that rays that have lost
performance characteristics of this instru- positional information by Compton scattering
ment are discussed in Chapter 14. The use of within the body can be rejected based on
the gamma camera for tomographic imaging their reduced energy (see Chapter 6, Section
is described in Chapter 17. C.3). A sodium iodide [NaI(Tl)] scintillation
detector (see Chapter 7, Section C) provides
both of these features at a reasonable cost;
A. GENERAL CONCEPTS OF for this reason it is currently the detector of
RADIONUCLIDE IMAGING choice for radionuclides with -ray emissions
in the range of 80-300keV.
The purpose of radionuclide imaging is to The first attempts at radionuclide imaging
obtain a picture of the distribution of a radio- occurred in the late 1940s. An array of radia-
actively labeled substance within the body tion detectors was positioned on a matrix of
after it has been administered (e.g., by intra- measuring points around the head. Alterna-
venous injection) to a patient. This is accom- tively, a single detector was positioned manu-
plished by recording the emissions from the ally for separate measurements at each point
radioactivity with external radiation detec- in the matrix. These devices were tedious to
tors placed at different locations outside the use and provided only very crude mappings of
patient. The preferred emissions for this the distribution of radioactivity in the head
application are rays in the approximate (e.g., left-side versus right-side asymmetries).
energy range of 80 to 500keV (or annihilation A significant advance occurred in the early
photons, 511keV). Gamma rays of these 1950s with the introduction of the rectilinear
energies are sufficiently penetrating in body scanner by Benedict Cassen (see Fig. 1-2).
tissues to be detected from deep-lying organs, With this instrument, the detector was
can be stopped efficiently by dense scintilla- scanned mechanically in a raster-like pattern
tors, and are shielded adequately with rea- over the area of interest. The image was a
sonable thicknesses of lead (see Fig. 6-17soft pattern of dots imprinted on a sheet of paper
tissue has attenuation properties similar to by a mechanical printer that followed the
water). Alpha particles and electrons ( par- scanning motion of the detector, printing the
ticles, Auger and conversion electrons) are of dots as the rays were detected.
195
196 Physics in Nuclear Medicine

The principal disadvantage of the rectilin- Computer

ear scanner was its long imaging time (typi- compute X, Y and E
cally many minutes) because the image was linearity & uniformity corrections Image
formed by sequential measurements at many
individual points within the imaged area. The Monitor
first gamma-ray camera capable of record-
ing at all points in the image at one time was

Amp/ADC

Amp/ADC

Amp/ADC

Amp/ADC

Amp/ADC

Amp/ADC
described by Hal Anger in 1953. He used a
pinhole aperture in a sheet of lead to project
a -ray image of the radionuclide distribution
onto a radiation detector composed of a
NaI(Tl) screen and a sheet of x-ray film. The
film was exposed by the scintillation light
PM tube array
flashes generated by the rays in the NaI(Tl)
screen. Unfortunately, this detection system Light guide
(especially the film component) was so ineffi- NaI(Tl) crystal
cient that hour-long exposures and therapeu-
tic levels of administered radioactivity were Collimator
needed to obtain satisfactory images.
In the late 1950s, Anger replaced the film-
screen combination with a single, large-area,
NaI(Tl) crystal and a photomultiplier (PM)
tube assembly to greatly increase the detec- Patient
tion efficiency of his camera concept. This
instrument, the Anger scintillation camera,1
or gamma camera, has been substantially FIGURE 13-1 Basic principles and components of a
refined and improved since that time. modern gamma camera. The outputs of each photo
Although other ideas for nuclear-imaging multiplier (PM) tube are amplified and digitized using an
instruments have come along since then, analog-to-digital converter (ADC). The X-Y locations for
none, with the exception of modern positron each gamma ray that interacts in the NaI(Tl) crystal are
computed from the digitized signals. The energy depos-
emission tomography systems (see Chapter ited by the gamma ray, E, which is proportional to the
18), has matched the gamma camera for a total measured pulse amplitude, also is computed by
balance of image quality, detection efficiency, summing the individual PM tube signals. If E falls within
and ease of use in a hospital environment. the selected energy window, the event is accepted and
placed at the appropriate X-Y location in the image.
The gamma camera has thus become the
most widely used nuclear-imaging instru-
ment for clinical applications.
image of the -ray distribution on the surface
of the NaI(Tl) crystal (see Section B.3). The
B. BASIC PRINCIPLES OF THE GAMMA second is that the NaI(Tl) crystal is viewed
by an array of PM tubes, rather than a
CAMERA
single PM tube. Signals from the PM tubes
are fed to electronic or digital position logic
1. System Components circuits, which determine the X-Y location of
Figure 13-1 illustrates the basic principles of each scintillation event, as it occurs, by
image formation with the gamma camera. using the weighted average of the PM tube
The major components are a collimator, a signals (see Section B.2).
large-area NaI(Tl) scintillation crystal, a Individual events also are analyzed for
light guide, and an array of PM tubes. Two energy, E, by summing the signals from all
features that differ from the conventional PM tubes. When the pulse amplitude of an
NaI(Tl) counting detectors described in event falls within the selected energy window,
Chapter 12 are crucial to image formation. it is accepted and the X and Y values are
The first is that an imaging collimator is binned into a discrete two-dimensional array
used to define the direction of the detected of image elements, or pixels. An image is
rays. The collimator most commonly consists formed from a histogram of the number of
of a lead plate containing a large number of events at each possible X-Y location. Large
holes. By controlling which rays are numbers of events are required to form an
accepted, the collimator forms a projected interpretable image because each pixel must
 13 The Gamma Camera: Basic Principles 197

have a sufficient number of counts to achieve reflective material such as TiO2 to maximize
an acceptable signal-to-noise level. Because light output and hermetically sealed inside a
images often are formed in 64- 64-pixel or thin aluminum casing to protect it from mois-
128- 128-pixel arrays, the counting require- ture. An optical glass window on the back
ments are some 103 to 104 times higher than surface of the casing permits the scintillation
for a simple counting detector. light to reach the PM tubes. A cross section of
Images are displayed on a computer a typical gamma camera crystal assembly is
monitor, where image brightness and con- shown in Figure 13-2. The choice of thickness
trast may be manipulated and different color of the NaI(Tl) crystal is a trade-off between
tables may be employed. More sophisticated its detection efficiency (which increases with
digital image processing is discussed in increasing thickness) and, as shown in
Chapter 20. Chapter 14, Section A.1, its intrinsic spatial
Most modern gamma cameras are com- resolution (which deteriorates with increas-
pletely digital, in the sense that the output of ing thickness). Most general-purpose gamma
each PM tube is directly digitized by an analog- cameras have crystal thicknesses of approxi-
to-digital converter (ADC). The calculation of mately 9.5mm. For lower-energy emitters,
X-Y position and pulse-height are performed such as 99mTc and 201Tl, however, detection
in software based on the digitized PM tube efficiency is adequate even with 6-mm-thick
signals, and errors in energy and positioning detector crystals.
caused by noise and pulse distortions caused An array of PM tubes is coupled optically
by the analog positioning circuitry are elimi- to the back face of the crystal with a silicone-
nated. This approach also permits improved based adhesive or grease. Round PM tubes
handling of pulse pile-up at high counting are arranged in a hexagonal pattern to maxi-
rates, as described in Section B.2. mize the area of the NaI(Tl) crystal that is
The gamma camera can be used for static covered. Some cameras use hexagonal (or
imaging studies, in which an image of an rarely, square) cross-section PM tubes for
unchanging radionuclide distribution can be better coverage of the NaI(Tl) crystal. Typical
recorded over an extended imaging time (e.g., PM tube sizes are 5cm in diameter. Most
minutes). Single contiguous images of the modern cameras employ between 30 and 100
whole body can be obtained by scanning PM tubes. Figure 13-3 shows a photograph of
the gamma camera across the entire length of a 30-tube model. The PM tubes are encased
the patient. This can be achieved by moving in a thin magnetic shield (Chapter 7, Section
either the bed or the gamma camera while C.2) to prevent changes in the gain caused by
adjusting the event positioning computation changes in the orientation of the gamma
to account for this movement. Clinically camera relative to the earths magnetic field.
important whole-body studies include bone The ultrasensitivity of PM tubes to magnetic
scans of the skeleton, and the localization of fields also makes gamma cameras susceptible
tumors or their metastases in the body. to the stray fields from magnetic resonance
The gamma camera also can be used for imaging systems.
dynamic imaging studies, in which changes
in the radionuclide distribution can be
observed, as rapidly as several images per
second. This allows physiologic information to
be obtained, such as the rate of tracer uptake Scintillation
or clearance from an organ of interest. Images NaI(Tl) crystal light exits to Glass
(typically 6- to PM tubes entrance
also can be synchronized to electrocardiogram 12.5-mm thick) window
signals, permitting images of the heart in dif-
ferent phases of the cardiac cycle to be formed.
These gated images can provide important
information on cardiac function.
2. Detector System and Electronics
Gamma rays
The gamma camera employs a single, large- enter from
area, rectangular NaI(Tl) detector crystal, Thin aluminum casing this side Reflector
usually 6- to 12.5-mm thick with sizes of up to hermetically seal material
to 60 40cm. Round crystals of 25 to 50cm crystal
in diameter were used in many older systems. FIGURE 13-2 Schematic cross-section of a NaI(Tl)
The NaI(Tl) crystal is surrounded by a highly crystal assembly for a gamma camera.
198 Physics in Nuclear Medicine

FIGURE 13-3 A rectangular gamma camera detector with the cover removed showing the photomultiplier (PM) tubes
mounted on the NaI(Tl) crystal. In this example, the gamma camera detector measures 50 15cm and is read out by
30 PM tubes 5cm in diameter. This is a digital camera in which each of the PM tube outputs is individually digitized.
(Courtesy Dr. Joel Karp, University of Pennsylvania, Philadelphia, PA.)

Many manufacturers employ plastic light The amount of light detected by a particu-
guides between the detector crystal and PM lar PM tube is inversely related to the lateral
tubes, whereas others couple the PM tubes distance between the interaction site and the
directly to the crystal. The functions of the center of that PM tube. This is illustrated in
light guide are to increase the light collection one dimension in Figure 13-4. Ideally, the
efficiency, by channeling scintillation light relationship between signal amplitude and
away from the gaps between the PM tubes, location with respect to the center of a PM
and to improve the uniformity of light collec- tube would be linear. This would enable the
tion as a function of position. The latter effect position of an event to be determined by
is achieved by painting or etching a carefully taking a weighted average or centroid of the
designed pattern onto the entrance face of the PM tube signals using the simple relation-
light guide. The use of the PM tubes with ships shown in Figure 13-4. In practice,
hexagonal or square cross-sections that can however, the response is more complex, with
be tiled without gaps on the NaI(Tl) crystal a plateau directly beneath the PM tube
may in some cases allow elimination of the (because the PM tube is not a point detector)
light guide, assuming there is sufficient and long, flat tails caused by reflections of
spreading of the scintillation light in the glass light from the back and side surfaces of the
entrance window of the PM tube for accurate NaI(Tl) crystal. Therefore a calibration for
positioning. spatial nonlinearity is required (see Chapter
The detector crystal and PM tube array are 14, Section B).
enclosed in a light-tight, lead-lined protective Figure 13-5A shows a schematic drawing
housing. In most modern cameras, most of for an eight-PM tube version of the gamma
the electronics (such as preamplifiers, pulse- camera and is used to illustrate the principles
height analyzers, automatic gain control, of scintillation event localization in an analog
pulse pile-up rejection circuits and ADCs) are detector. The position is determined by split-
mounted directly on the individual PM tube ting the signal from each PM tube onto four
bases within the detector housing to minimize output lines, whose signals are denoted X +,
signal distortions that can occur in long cable X , Y+, and Y (Fig. 13-5B). The fraction of
runs between the detector head and control the PM tube current that goes to each output
console. line is determined by the value of the resistors
 13 The Gamma Camera: Basic Principles 199

S2
S1

If linear, then:
D1  (S2  D) / (S1  S2)
D2  (S1  D) / (S1  S2)
where D  D1  D2

PM tube array

NaI(Tl) crystal
D1 D2

Collimator

Gamma ray
FIGURE 13-4 Illustration of light sharing between photomultiplier (PM) tubes. The PM-tube signal, S, is inversely
related to the distance of the interaction site, D, from the center of the PM tube. Equations for a linear relationship
are shown.

(R) that are used. By Ohms law, this current on the pulse height. Note that the possible
is proportional to 1/R. A separate circuit sums range of X and Y values is from 1 to +1. The
the outputs of all the PM tubes to form the resistor values shown in Figure 13-5C were
Z-signal. The Z-signal is proportional to the chosen such that the calculated X- and
total amount of light produced by a scintilla- Y-position signals vary linearly with distance
tion event in the crystal, and therefore the in the X and Y directions. In a perfect gamma
total energy deposited by the gamma ray, and camera, measured (X, Y) values would change
is used for pulse-height analysis. linearly from (1, 1) in the bottom left-hand
The X +, X , Y +, and Y signals are com- corner to (+1, +1) at the top right-hand corner
bined to obtain X-position and Y-position of the camera face. The X and Y values can
signals. The X-position of the scintillation be scaled by the detector size to determine the
event is given by the difference in the X + and absolute position of an event on the gamma
X signals, divided by the total X signal (X + camera face.
+ X ) However, Equations 13-1 and 13-2 do not
give a perfect mapping of source position
X = ( X + X )/( X + + X ) (13-1) because, as was discussed previously, the PM
tubes signal does not actually vary linearly
Similarly, for the Y-position with interaction position. This gives rise to a
pincushion artifact, which is illustrated in
Y = (Y + Y ) / (Y + + Y ) (13-2) Figure 14-9. There are also effects caused
by nonuniformities in the crystal, light
The X- and Y-position signals are normalized reflections at the edge of the crystal, and non-
to the total X and Y signals, so that the cal- uniform response across the face of the PM
culated position of interaction does not depend tubes that can cause further nonlinearities in
200 Physics in Nuclear Medicine

NaI(Tl) crystal RX RX RY RY

1  12.5 12.5 
Y
2 25 25 12.5 

1 2 3 3 12.5  12.5 
X
4 5 4 50 16.7 25 25

PM tube 6 7 8 5 16.7 50 25 25

6  12.5  12.5
A
7 25 25  12.5

Y 8 12.5   12.5
C
RY

RX RX
X X

RY

Y
B
FIGURE 13-5 Illustration of analog positioning in a gamma camera. A, Schematic representation of an eight-
photomultiplier (PM) tube camera. B, Signals from individual PM tubes are split using resistors onto four output
lines, designated X +, X , Y +, and Y . C, Representative resistor values (in k) for the eight PM tubes. Resistor
values are chosen such that the X and Y positions computed from Equations 13-1 and 13-2 vary linearly with inter-
action position in the detector, ranging from a value of 1 in the bottom left hand corner to +1 in the top right corner.

position determination. These effects and cor- significant pulse amplitude, the noise from
rection techniques for them are discussed in the PM tubes that produce negligible signal
Chapter 14, Section B. amplitude (and that therefore contribute
In digital cameras, the output signal from little to position information) is not included
each PM tube is digitized and the event posi- in the position calculation. Second, with
tion is calculated in software. Often, this is signal thresholding, only a small number of
simply analogous to the resistor readout PM tubes surrounding the interaction loca-
described earlier; the inverse of the resistor tion are used for position determination. This
values are used as weighting factors for the allows a gamma camera to detect multiple
individual PM tube signals, and Equations events simultaneously when they occur in dif-
13-1 and 13-2 are used to determine the X ferent portions of the gamma camera and
and Y values. However, digital cameras also their light cones (the projection of the scintil-
can use more sophisticated algorithms that lation light on the PM tube array) do not
incorporate information regarding the non- significantly overlap. This improves the
linearity of PM tube response with position counting rate performance of the gamma
into the weighting factors to provide better camera, reducing dead time losses.
positioning accuracy. Energy selection is important for imaging
A commonly used tactic that is employed in because it provides a means to discriminate
both digital and analog cameras to improve against rays that been scattered within the
the positioning accuracy is to include in the body and therefore lost their positional infor-
position calculation only PM tubes with mation. By choosing a relatively narrow pulse-
signals above a certain threshold. This has height analyzer window that is centered on
two important benefits. By using the signal the photopeak, only rays that undergo no
only from those PM tubes that produce a scatter or small-angle scatter will be accepted.
 13 The Gamma Camera: Basic Principles 201

Two different methods can be used to select Energy window Z

the photopeak events. The first approach uses
simple energy discrimination on the Z-signal.
However, because of nonuniformities in the
NaI(Tl) crystal (small variations in light pro-
duction with position), in light collection

Number of counts
efficiency and in PM tube gains, the position
of the photopeak varies somewhat from posi-
tion to position in the detector. If a single
discriminator level is applied across the whole
detector, the window must be widened to
accommodate the fluctuations in photopeak
position, thus accepting more scatter (Fig.
13-6, top).
In the second method, suitable only for
digital cameras, the photopeak positions and
appropriate discriminator level settings are Pulse height, Z
computed and stored for many different loca-
tions across the detector face (Fig. 13-6,
bottom). When an event is detected, the X, Y Z1
values are calculated based on Equations Z2
13-1 and 13-2, and a look-up table is used to Z3
find the appropriate discriminator levels for
that location. If the event amplitude Z falls
Number of counts

within the pulse-height analyzer settings, the

event is accepted.
A modern gamma camera has an energy
resolution of 9% to 10% at 140 keV (99mTc).
Typically, the energy window (the difference
between upper-level and lower-level discrimi-
nators) is set to 14%, or 20 keV, centered
around 140 keV. The gamma camera soft-
ware adjusts the discriminator levels for
radionuclides other than 99mTc based on the
relationship (approximately linear over a Pulse height, Z
small energy range) between the -ray energy FIGURE 13-6 The pulse-height response is not uniform
deposited and the light output of NaI(Tl) (see across the whole face of the gamma camera, leading to
Fig. 10-11). variation in the pulse height of photopeak events. Using
a global energy window requires a wide window that
3. Collimators leads to the inclusion of larger amounts of scatter and
also results in nonuniform efficiency (top). Using local
To obtain an image with a gamma camera, it energy thresholds, in which the energy window is tai-
is necessary to project rays from the source lored to the event location, improves scatter rejection and
uniformity (bottom).
distribution onto the camera detector. Gamma
rays cannot be focused; thus a lens principle
similar to that used in photography cannot Gamma rays not traveling in the proper direc-
be applied. Therefore most practical -ray tion are absorbed by the collimator before
imaging systems employ the principle of they reach the detector. This projection by
absorptive collimation for image formation.* absorption technique is an inherently ineffi-
An absorptive collimator projects an image of cient method for using radiation because most
the source distribution onto the detector by of the potentially useful radiation traveling
allowing only those rays traveling along toward the detector actually is stopped by the
certain directions to reach the detector. absorptive walls between the collimator holes.
This is one of the underlying reasons for the
relatively poor quality of radionuclide images
*An important exception is imaging of the two 511-keV (e.g., as compared to radiographic images), as
annihilation photons from positron-emitting radionu-
clides, in which electronic coincidence detection can be
discussed in Chapter 15.
used to replace the collimator as described in Chapter 18, Four basic collimator types are used with
Section A.1. the gamma camera: pinhole, parallel-hole,
202 Physics in Nuclear Medicine

diverging, and converging. The different smaller than the collimator cone length f; it
types of collimator are introduced subse- is minified when the source distribution is
quently. Their effects on the spatial resolu- farther away. The image size I and object
tion and sensitivity of the gamma camera are (source) size O are related according to
discussed in Chapter 14 , Sections C and D.
A pinhole collimator (Fig. 13-7A) consists I/O = f / b (13-3)
of a small pinhole aperture in a piece of lead,
tungsten, platinum, or other heavy metal The size of the imaged area also changes with
absorber. The pinhole aperture is located at distance from the pinhole collimator. If the
the end of a lead cone, typically 20 to 25cm detector diameter is D and the magnification
from the detector. The size of the pinhole can (or minification) factor is I/O (Equation 13-3),
be varied by using removable inserts and is the diameter of the image area projected onto
typically a few millimeters in diameter. the detector, D, is
The imaging principle of a pinhole collima-
tor is the same as that employed with inex- D
D = (13-4)
pensive box cameras. Gamma rays passing I/O
through the pinhole project an inverted image
of the source distribution onto the detector Thus a large magnification factor, obtained at
crystal. The image is magnified when the dis- close source-to-collimator distances, results in
tance b from the source to the pinhole is a small imaged area.

I I

f
O

b
O
Pinhole Parallel hole
A B

f
b

O f
I

Diverging Converging
C D
FIGURE 13-7 A-D, Four types of collimators used to project -ray images onto the detector of a gamma camera.
O, Radioactive object; I, its projected image.
 13 The Gamma Camera: Basic Principles 203

Image size changes with object-to-pinhole directly head-on, it can be positioned closer
distance b. Therefore the pinhole collimator to the patient for better image detail in some
provides a somewhat distorted image of three- imaging studies (e.g., left anterior oblique
dimensional objects because source planes at cardiac views).
different distances from the collimator are A diverging collimator (Fig. 13-7C) has
magnified by different amounts. Pinhole col- holes that diverge from the detector face. The
limators are used primarily for magnification holes diverge from a point typically 40-50cm
imaging of small organs (e.g., thyroid and behind the collimator, projecting a minified,
heart) and for small-animal imaging. noninverted image of the source distribution
Another type of pinhole collimator, the onto the detector. The degree of minification
multi-pinhole collimator, has an array of mul- depends on the distance f from the front of the
tiple pinholes, typically seven, arranged in collimator to the convergence point, the dis-
a hexagonal pattern. This collimator was tance b from the front of the collimator to the
employed in the past for tomographic imaging. object (source), and the collimator thickness t
This type of tomography now is seldom used
clinically; however, multi-pinhole approaches I/O = ( f t) / ( f + b) (13-5)
are being widely employed for some small-
animal imaging applications. where I and O are image and object size,
The parallel-hole collimator (Fig. 13-7B) is respectively. The useful image area becomes
the workhorse collimator in most imaging larger as the image becomes more minified
laboratories. Parallel holes are drilled or cast (Equation 13-4).
in lead or are shaped from lead foils. The lead
EXAMPLE 13-1
walls between the holes are called collimator
septa. Septal thickness is chosen to prevent What is the minification factor for a diverging
rays from crossing from one hole to the next collimator 5-cm thick, with f = 45cm, and a
(see Chapter 14, Section C.2). A magnified source distribution 15cm from the collima-
view of a parallel-hole collimator is shown in tor? If the detector diameter is 30cm, what is
Figure 13-8. The parallel-hole collimator proj- the imaged area at this distance?
ects a -ray image of the same size as the
source distribution onto the detector. A varia- Answer
tion of the parallel-hole collimator is the From Equation 13-5,
slant-hole collimator, in which all of the holes I/O(minification factor)
are parallel to each other but angled, typically
= (45 5) / (45 + 15) = 0.67
by approximately 25 degrees, from the per-
pendicular direction. This type of collimator From Equation 13-4,
has characteristics that are similar to those Diameter of imaged area
of the parallel-hole type. Because it views the
= 30 cm/0.67 = 44.8 cm
source distribution from an angle rather than

FIGURE 13-8 Close-up view of a

typical parallel-hole collimator
showing the hole structure. Indi-
vidual holes are approximately
1.9mm across in this example.
(Courtesy Nuclear Fields, Des
Plaines, IL.)
204 Physics in Nuclear Medicine

As shown by Example 13-1, a typical diverg-

ing collimator decreases the size of the image
on the detector and increases the diameter of
the imaged area, by approximately one-third
as compared with a parallel-hole collimator.
As with the pinhole collimator, image size
changes with distance; thus there is a certain
amount of image distortion. Diverging colli-
mators are used primarily on cameras with
smaller detectors to permit imaging of large
organs such as the liver or lungs on a single
view. Parallel-hole collimator Converging collimator
A converging collimator (Fig. 13-7D) has
holes that converge to a point 40-50cm in FIGURE 13-9 Examples of geometric image distortions
created by the converging collimator.
front of the collimator. For objects between
the collimator face and the convergence point,
the converging collimator projects a magni-
fied, noninverted image of the source distribu-
tion. Image size I and object size O are related cameras having large-area detectors to permit
according to full use of the available detector area for
imaging of small organs.
I/O = ( f + t) / ( f b) (13-6) Converging collimators project an inverted
magnified image when the object is located
between the convergence point and twice the
where f is the distance from the collimator convergence length of the collimator, and an
face to the convergence point, b is the distance inverted minified image beyond that distance;
from the collimator face to the object, and t is however, they are used rarely at distances
collimator thickness. beyond the convergence point.
Some manufacturers provide a single, One consequence of the magnification or
invertible collimator insert that can be used minification effects of these collimators is
in either converging or diverging mode. that the contribution of the intrinsic detector
resolution (see Chapter 14, Section A) to the
EXAMPLE 13-2 resulting image resolution may be reduced
Suppose the collimator described in Example (magnification > 1) or increased (magnifica-
13-1 is inverted and used as a converging col- tion < 1). Thus magnifying collimators can be
limator to image a source distribution 15cm useful in situations in which high spatial
in front of the collimator, also with a 30-cm resolution is required, for instance in imaging
diameter detector. What are the image mag- of small organs such as the thyroid and in
nification factor and the size of the imaged small-animal imaging applications.
area?
4. Event Detection in a Gamma
Answer Camera
When the collimator is inverted, the back face There are four types of events that may be
becomes the front face, and the convergence detected by a gamma camera, as illustrated
distance f becomes (45 5cm) = 40cm. Thus in Figure 13-10. Of these, only one provides
from Equation 13-6 correct positional information. The four events
I/O (magnification factor) types (labeled to correspond with Fig. 13-10)
are the following:
= (40 + 5) / (40 15) = 1.8
A: valid eventa ray is emitted parallel
From Equation 13-4, to the collimator holes, passes through
Diameter of imaged area a hole and interacts photoelectrically in
the NaI(Tl) crystal, depositing all of its
= 30 cm/1.8 = 16.7 cm
energy at a single location.
B: detector scatter eventa ray is emitted
Again, because magnification depends on parallel to the collimator holes, passes
distance, there is some image distortion with through a hole and interacts by Compton
the converging collimator (Fig. 13-9). Con- scattering in the NaI(Tl) crystal. The
verging collimators are used primarily with scattered ray can either interact a
 13 The Gamma Camera: Basic Principles 205

PM tubes

Light guide
NaI(Tl) crystal

Collimator

C B A D

Patient with
-emitting
radionuclide
Liver concentrated in
liver

FIGURE 13-10 Illustration of different types of events that may be detected by a gamma camera. Red circles indi-
cate locations of ray interactions. A, Valid event. B, Detector scatter event. C, Object scatter event. D, Septal
penetration.

second time in the detector (as illus- the event is mispositioned, often many
trated in Fig. 13-10), in which case the centimeters from the original site of
full energy of the ray is deposited, or emission. These events lead to a low-
it may escape the detector, in which case spatial-frequency background in the
only part of the -ray energy is depos- images that results in a loss of contrast.
ited. In the former case, energy discrim- (See Chapter 15, Section C.). In clinical
ination cannot be used to reject the imaging situations, a large fraction of
event, and the event will be misposi- the detected events can be due to object
tioned between the two interaction loca- scatter, and good energy resolution in
tions. In the latter case, it is likely that the gamma camera is extremely impor-
the event will be rejected because it does tant (see Chapter 14, Section A.3). The
not satisfy the event energy criteria collimator itself can also be a cause of
established by the upper- and lower- scatter leading to similar effects.
level discriminators. As discussed in D: septal penetrationin this case a ray
Chapter 14, Section A.1, these events is emitted toward the collimator, but not
are relatively rare. parallel to it. Because of incomplete
C: object scatter eventthe ray is not attenuation by the thin collimator walls
emitted toward the collimator holes but (septal penetration), there is a finite
is scattered within the body, then passes chance that the ray will reach the
through a collimator hole and subse- NaI(Tl) crystal and interact with it. This
quently is detected. The ray loses again leads to blurring of the image,
energy during scattering and will there- because all events are considered to
fore produce a smaller signal in the have come from a direction perpendicu-
detector. Some of these events will be lar to the collimator face (for parallel-
rejected by energy discrimination, but if hole collimators). This effect becomes
the angle of scatter is small (45 increasingly important when using
degrees), the energy loss is small and high-energy emitters or high-resolution
the event may be accepted. In this case collimators with thin septa.
206 Physics in Nuclear Medicine

Considerable effort is expended in the positioned in a flexible way over different

design of gamma cameras to reduce or elimi- regions of the patients body. Often, a moving
nate the detection of the events B, C, and D bed is incorporated to permit imaging studies
just described, each of which is a cause of of the whole body. The gamma camera head
blurring and a loss of contrast in the image. often is mounted on a rotating gantry, allow-
Collimators also are carefully designed for ing it to take multiple views around the
specific energies to minimize septal penetra- patient. This feature also is necessary for pro-
tion while maximizing sensitivity for a given ducing tomographic images, or cross-sectional
-ray energy (Chapter 14, Section C). images through the body, as discussed in
In addition to the simple cases illustrated Chapters 16 and 17.
in Figure 13-10, a combination of these event Dual-headed gamma cameras are becom-
types can occur (e.g., scatter in the body and ing increasingly popular. In these systems,
septal penetration, or septal penetration fol- two gamma camera heads are mounted onto
lowed by Compton interaction in the detec- the gantry as shown in Figure 13-12. Usually,
tor). Finally, further complications arise when the two heads can be positioned at a variety
pulse pile-up occursthat is, two or more of locations on the circular gantry. An obvious
events occur almost simultaneously in the advantage of a dual-headed camera is that
gamma camera. This can also lead to event two different views of the patient can be
mispositioning and is discussed in detail in acquired at the same time. For example, in
Chapter 14, Section A.4. Pile-up events can whole-body imaging, the two detector heads
arise from of any combination of the event can be placed at 180 degrees to each other to
types described earlier. provide anterior and posterior views simulta-
neously. Triple-headed systems also exist, pri-
marily for tomographic studies, as described
in Chapters 16 and 17.
C. TYPES OF GAMMA CAMERAS AND An example of a planar image acquired
THEIR CLINICAL USES with a gamma camera system is presented in
Figure 13-13. Dynamic processes can also be
The most common type of gamma camera is measured by taking multiple planar images
the single-headed system (Fig. 13-11). It con-
sists of a gamma camera detector mounted on
a gantry that allows the camera head to be

FIGURE 13-11 Single-headed gamma camera mounted

on a rotating gantry. The camera is operated from the FIGURE 13-12 A dual-headed gamma camera system
computer (left). The flexible positioning of the camera (top). Note that the camera heads can be placed in differ-
head and the bed (right) allows the system to obtain ent orientations to provide two simultaneous views of an
images of many different parts of the body. (Courtesy GE organ or the body (bottom). (Courtesy Siemens Medical
Medical Systems, Milwaukee, WI.) Systems, Inc., Hoffman Estates, IL.)
 13 The Gamma Camera: Basic Principles 207

over time. An example of a dynamic study is

shown in Figure 13-14.
Single- and dual-headed gamma cameras
are the workhorses of clinical nuclear medi-
cine laboratories. However, a range of spe-
cialty gamma cameras have been or are being
developed for specific imaging tasks. Exam-
ples are systems designed specifically for
small-organ imaging (e.g., heart, breast, and
thyroid) and mobile systems for use on
patients who are too sick to be moved to the
nuclear medicine department (e.g., from
intensive care). These systems typically have
smaller detector heads and may not have a
built-in bed. An example of a compact gamma
camera for breast imaging and representative
images from it are shown in Figure 13-15. The
detector typically ranges from 10 10cm2 to
20 20cm2. A number of different detector
technologies are being exploited for these
small-detector cameras, including traditional
NaI(Tl)/PM tube systems, cameras based on
pixellated NaI(Tl) or cesium iodide [CsI(Tl)]
scintillator arrays (see Fig. 13-15B), and
CsI(Tl) scintillator arrays with read-out by
silicon photodiode arrays. There are also
systems being developed that employ arrays
of cadmium zinc telluride elements (see
Chapter 7, Section B) for direct detection of
rays, eliminating the need for a scintillator-
photodetector combination.
FIGURE 13-13 Whole-body bone scan obtained using High-resolution gamma cameras also have
99m
Tc-MDP. These planar images were obtained with a been developed for small-animal imaging.
dual-headed gamma camera on which both anterior (left) The goal is to provide a tool that biologists can
and posterior (right) views can be acquired simultane- use to monitor radiotracers in vivo, particu-
ously. The entire body was imaged by translating the
patient bed through the gamma camera system. (Cour- larly in rats and mice. Most approaches
tesy Siemens Medical Systems, Inc., Hoffman Estates, IL.) involve the use of very small pinhole

Gallbladder

1st minute 2nd minute 3rd minute 4th minute 5th minute

6th minute 7th minute 8th minute 9th minute 10th minute
FIGURE 13-14 Planar gamma camera images over the region of the gallbladder following injection of 99mTc-HIDA. At
approximately 7 minutes, cholecystokinin was given to the patient to stimulate emptying of the gallbladder. The rate
and extent of emptying can be measured from this dynamic sequence of planar images. (Courtesy GE Medical Systems,
Milwaukee, WI.)
208 Physics in Nuclear Medicine

FIGURE 13-15 Example of a compact, mobile

gamma camera system. A, The gamma camera
A head is attached to a cantilevered arm for easy
and flexible positioning. B, Pixellated NaI(Tl)
scintillator array that is coupled to small
position-sensitive photomultiplier tubes to
form the detector head. Each element in the
array is approximately 2 2mm and is sepa-
rated from its neighbor by a reflective mate-
rial. The inset shows a magnified view of the
pixel elements. C, Clinical images of a breast
cancer patient acquired with this camera
following the injection of 99mTc-sestamibi. A
tumor is seen as a hot spot against the low
background uptake of the normal breast.
B (Photographs and images courtesy Dilon Tech-
Superior nologies, Inc., Annapolis, MD.)

Lateral Medial Inferior

Left craniocaudal Left lateral

C

collimators to provide high spatial resolution. BIBLIOGRAPHY

Tomographic small animal imaging systems The principles of the gamma camera are discussed
based on this approach are discussed in in greater detail in the following:
Chapter 17, Section A.3. Simmons GH: The Scintillation Camera, New York, 1988,
Society of Nuclear Medicine.
REFERENCE
1. Anger HO: Scintillation camera. Rev Sci Instr 29:27-
33, 1958.