Nuclear Medicine Resources Manual
Подождите немного. Документ загружается.
CHAPTER 4. INSTRUMENTATION
160
filter order. This is a further parameter used to define the shape of a filter
function. It is specific to the Butterworth filter and defines the slope of
the filter function (i.e. the rate at which the curve drops to zero).
FOV. This abbreviation stands for field of view.
full width at half maximum (FWHM). This term refers to resolution measure-
ments (e.g. spatial and energy resolutions). FWHM is usually measured
from a profile through an image of a line or point source, or, in the case of
energy, from the energy spectrum of a single gamma emitting radionu
-
clide. The spread is due to resolution effects and is measured by the full
width of the profile at a point which is half the maximum height of the
profile.
gated SPECT. An ECG can be used to synchronize SPECT data acquisition
on each projection with the patient’s heartbeat in the same way as in
conventional planar gated blood pool studies. Each time interval can be
reconstructed to produce a set of 3-D images throughout the cardiac
cycle.
Interfile. This has become a widely used standard file format for exchange of
image data in nuclear medicine. The method was developed by the ANZ
Society of Nuclear Medicine’s Technical Standards Committee on the
basis of an early AAPM recommendation before being adopted by the
EEC’s COST B2 project.
Internet. Historically this referred to any network that utilized the Internet
Protocol (i.e. TCP/IP), a protocol that is now widely used. It tends now to
be used to refer to the general worldwide network that interconnects the
many thousands of networks that exist.
intrinsic. Refers to measurements performed without a collimator.
iterative reconstruction. This general term applies to a number of recon-
struction algorithms that involve a repetitive process of comparison to
find the best estimate of the activity distribution that matches the
measured projections.
JPEG (Joint Photographic Experts Group). A commonly used method for
image compression was developed by JPEG, and so the term JPEG now
refers to an image compressed by this method. Compression can be lossy
4.7. GLOSSARY OF TECHNICAL TERMS
161
or lossless depending on the degree of compression applied. It is one of
the most commonly used compression methods, particularly for trans
-
mission via the Internet.
line source. A thin line (such as a capillary tube) filled with activity, which is
used for measuring resolution. The diameter of the line source should
typically be 1 mm.
local area network (LAN). A set of computers located in a department or
institution that are connected, usually via Ethernet cabling, so that they
can communicate with each other.
lutetium oxyorthosilicate (LSO). This is a new detector material currently
being considered for PET systems.
multidetector SPECT. It is common for SPECT systems to have two or three
separate detectors rather than a single gamma camera. These multi
-
detector systems simply acquire more than one projection at a time; each
detector being a standard gamma camera. A few special instruments have
been designed that use many individual detectors which are more similar
to the most common PET systems.
multiple coincidence. When more than two photons are detected at the same
time it is not possible to identify which two events are related to a single
positron emission. This event is called a multiple coincidence.
NEMA (National Electrical Manufacturers Association). NEMA develops
standard specifications for imaging equipment including gamma cameras
(SPECT) and PET. These form the basis for specification and acceptance
testing of equipment, and some tests, with modification, can also be used
for routine quality control.
Nyquist frequency. The Nyquist frequency F
N
defines the maximum frequency
that can be recorded for a given pixel size, F
N
= 1/(2 × pixel size). It is
common to refer to frequencies relative to the Nyquist frequency as a
form of unit, for example the cut-off frequency for a Butterworth filter
can be defined as 0.5F
N
. Care should be taken as the Nyquist frequency
will be different for images with different pixel sizes.
positron emission tomography (PET). Tomography based on detection of the
dual annihilation photons that originate from positron emission. The
CHAPTER 4. INSTRUMENTATION
162
technique involves detection of the dual photons in coincidence (at the
same time).
primary photons. These are gamma rays that have not undergone any
scattering.
projections (count profiles). This term refers to the counts recorded during
tomographic acquisition. The counts in a single row of the images
recorded in SPECT at a given angle represent a projection of the emitted
counts. These can also be referred to as count profiles. The set of projec
-
tions, recorded at different angles, form the data that are used for
tomographic reconstruction.
pulse height analyser (PHA). These instruments analyse the size of the energy
signal and produce an output only if the size of the energy signal is within
the range specified by the predefined energy windows.
random coincidence. When two gammas originating from quite independent
sources (e.g. two separate positron emissions) are detected at the same
time, the path defined by the points of detection does not correspond to a
positron emission. This incorrectly located coincidence event is referred
to as a random event.
reconstruction. This is the process of obtaining a cross-sectional image from a
set of projections.
resolution. This refers to the ability of imaging systems to distinguish between
two closely spaced small sources. Usually expressed in terms of FWHM,
which describes the spread of the image obtained from a line or point
source.
resolution recovery. This is the opposite of smoothing, and is achieved by
filtering or deconvolution. By use of an appropriate filter the loss of
resolution due to some measurable effect (e.g. due to a detector’s finite
resolution) can be partially recovered. However, any noise in the original
image will normally be amplified.
ring artefacts. These are a common error in reconstructed images which are
caused by a localized non-uniformity in the detector.
4.7. GLOSSARY OF TECHNICAL TERMS
163
scatter coincidence. When one or both photons originating from a positron
event are detected in coincidence, the path defined by the points of
detection does not necessarily correspond to the point of positron
emission. This event is referred to as a scatter coincidence.
scattered photon. A gamma ray which has changed direction at least once due
to Compton interaction and loss of energy in the material through which
it is travelling.
sensitivity. Fraction of the emitted gamma rays which pass through the
collimator (collimator sensitivity) or are detected by the gamma camera
(system sensitivity).
single event. In a PET system, when a photon is detected without a corre-
sponding coincident photon, this is referred to as a single event. Owing to
the probability of detection, there are many more single events detected
than coincidences.
single photon emission computed tomography (SPECT). Cross-sectional slices
of the radionuclide distribution in the patient are generated by taking
images of the patient from various angles and then using these to
construct the slices with a computer. SPECT refers to tomography based
on the use of a single photon emitter.
sinogram. The image formed by placing projection values in sequential rows
(i.e. arranging pixels corresponding to projection position versus
projection angle) is called a sinogram. It is so called since the projections
from a single point describe a perfect sine wave when plotted in this form.
slit phantom. A phantom consisting of a lead mask with thin slits cut into it.
Typically the slits are 1 mm wide and 30 mm apart. They are used for
measuring intrinsic FWHM resolution and also linearity.
smoothing. Smoothing is an operation that involves spreading values across
neighbouring pixels; the averaging effect reduces statistical noise but
degrades resolution. Smoothing is a filtering operation often achieved by
convolution.
spatial frequency. Frequency normally refers to cyclic variations as a function of
time (units: s
–1
). However, if a curve represents variations in values over
CHAPTER 4. INSTRUMENTATION
164
distance (units: 1/distance), the number of oscillations per unit distance is
referred to as a spatial (rather than temporal) frequency.
tomography. Literally this means ‘drawing a body slice’. Tomography involves
measurement from different angles around an object with the intention to
‘reconstruct’ an image of the internal distribution of some parameter (e.g.
activity in SPECT or PET).
true coincidence. When two annihilation photons originating from a single
positron annihilation are detected in coincidence (without being
scattered), this is referred to as a true coincidence.
uniformity. A measure of how uniform the observed counts across the FOV are
when the detector is irradiated by a uniform source. Integral uniformity is
a measure of the maximum count deviation ((max – min)/(max + min))
over a given FOV. Differential uniformity is a measure of the maximum
rate of change over a specified distance. Both shall be measured for the
UFOV and the CFOV.
voxel. If one considers a digitized 3-D volume rather than a digitized 2-D
image, each digital value within the volume can be considered to occupy a
small volume element (e.g. a small cube) or voxel. One therefore refers to
planar projections as having pixels, but to each reconstructed slice as
having voxels, which also have a thickness corresponding to the spacing
between adjacent slices.
BIBLIOGRAPHY TO CHAPTER 4
ANZSNM TECHNICAL STANDARDS COMMITTEE: Minimum quality control
requirements for nuclear medicine practices, ANZ Nucl. Med. 27 (1996) 16–23.
BARBER, R.W., “Design, calibration and use of equipment for the assay of
radioactivity”, Textbook of Radiopharmacy Theory and Practice, 3rd edn (SAMPSON,
C.B., Ed.), Gordon and Breach, Amsterdam (1999) 135–144.
BRITTON, K.E., VAUROMO, E., Cost b2: The quality of nuclear medicine software,
Eur. J. Nucl. Med. 20 (1993) 815–816.
BIBLIOGRAPHY TO CHAPTER 4
165
BUSEMANN-SOKOLE, E., KUGI, A., BERGMANN, H., Influence of high energy
photons from cobalt-57 flood sources on scintillation camera uniformity images, Eur. J.
Nucl. Med. 23 (1996) 437–442.
CRADDUCK, T.D., et al., A standard protocol for the exchange of nuclear medicine
image files, Nucl. Med. Commun. 10 (1989) 703–713.
GOLDSTONE, K.E., “Monitoring”, Radiation Protection in Nuclear Medicine and
Pathology Report 63 (GOLDSTONE, K.E., JACKSON, P.C., MYERS, M.J., SIMPSON,
A.E., Eds), Institute of Physics and Engineering in Medicine, York (1991) 104–128.
GRAHAM, L.S., FAHEY, F.H., MADSEN, M.T., VAN ASWEGEN, A., YESTER,
M.V., Quantitation of SPECT performance: Report of Task Group 4, Nuclear Medicine
Committee, Med. Phys. 22 (1995) 401–409.
HART, G., JARRITT, P., COSGRIFF, P., MACKIE, A., McGILL, G., Gamma Camera/
Data Processor Tender Specification, British Nuclear Medicine Society, London (1996).
HINES, H., et al., Recommendations for implementing SPECT instrumentation quality
control., Eur. J. Nucl. Med. 26 (1999) 527–532.
HUNT, C., TCP/IP Network Administration, O’Reilly and Associates, Sebastopol, CA
(1994).
INTERNATIONAL ATOMIC ENERGY AGENCY, IAEA Quality Control Atlas for
Scintillation Camera Systems, IAEA, Vienna (2003).
INTERNATIONAL ELECTROTECHNICAL COMMISSION, Radionuclide Imaging
Devices — Characteristics and Test Conditions, Part 1: Positron Emission Tomographs,
IEC, Geneva (1998).
—
Radionuclide Imaging Devices — Characteristics and Test Conditions, Part 2: Single
Photon Emission Tomography, IEC, Geneva (1998).
KESHTGAR, M.R.S., WASSINGTON, W.A., LAKHANI, S.R., ELL, P.J., The Sentinel
Node in Surgical Oncology, Springer-Verlag, Berlin (1999) Ch. 3 and Ch. 4.
KING, M.A., TSUI, B.M.W., PAN, T.S., Attenuation compensation for cardiac single-
photon emission computed tomographic imaging, Part 1: Impact of attenuation and
methods of estimating attenuation maps, J. Nucl. Cardiol. 2 (1995) 513–524.
LINKS, J.M., Advances in nuclear medicine instrumentation: Considerations in the
design and selection of an imaging system, Eur. J. Nucl. Med. 25 (1998) 1453–1466.
CHAPTER 4. INSTRUMENTATION
166
MEIKLE, S.R., DAHLDOM, M., “Positron emission tomography”, Nuclear Medicine
in Clinical Diagnosis and Treatment, Churchill Livingstone, Edinburgh (1998) 1603–
1616.
NATIONAL ELECTRICAL MANUFACTURERS ASSOCIATION (Rosslyn, VA)
Performance Measurements of Scintillation Cameras (1994).
Performance Measurements of Positron Emission Tomographs (1994).
Digital Image Communication in Medicine (DICOM) (1994).
Performance Measurements of Scintillation Cameras (2001).
Performance Measurements of Positron Emission Tomographs (2001).
PARKIN, A., “Protocol for establishing and maintaining the calibration of medical
radionuclide calibrators, and their quality control”, Quality Standards in Nuclear
Medicine Report 65, Institute of Physics and Engineering in Medicine, York (1992) 60–
77.
PATTON, J.A., TURKINGTON, T.G., Coincidence imaging with a dual-head
scintillation camera, J. Nucl. Med. 40 (1999) 432–441.
SIEBERT, J.A., FILIPOW, L.J., ANDRIOLE, K.P., Practical digital imaging and PACS,
Medical Physics Publishing, Madison, WI (1999).
TODD-POKROPEK, A., CRADDUCK, T.D., DECONNINCK, F., A file format for
the exchange of nuclear medicine image data: A specification on Interfile Version 3.3,
Nucl. Med. Commun. 13 (1992) 673–699.
167
Chapter 5
GUIDELINES FOR GENERAL IMAGING
5.1. INTRODUCTION
5.1.1. Purpose
The purpose of this chapter is to provide nuclear medicine practitioners
with general guidelines on imaging using a single photon scintillation camera.
Recommendations specific to individual procedures are included in their
respective procedure guidelines.
5.1.2. Planar image acquisition using a gamma camera
5.1.3. Peaking
The scintillation camera must be peaked correctly for the energy
(energies) of the emitted photon for the radionuclide used. This should be
checked at least once daily and when different radionuclides are used. A 15 or
20% energy window is typically used. The window is placed symmetrically
about the photopeak, or asymmetrically if an appropriate energy correction is
available, in order to minimize scattered radiation.
A physicist can help in determining the limits of asymmetry that are
desirable for a range of energies. Caution must be taken not to affect
uniformity.
5.1.3.1. Multiple energy windows
The use of multiple energy windows for radionuclides that have more
than one energy peak is advantageous. It is necessary to check the spatial regis
-
tration for the combination of windows.
A physicist can help determine if the co-registration is adequately
adjusted for all of the windows in order to maintain the best spatial resolution
and contrast. A collimator offering adequate resolution for the most energetic
photons must be used. Intrinsic uniformity should be checked for imaging
multiple energy windows for such radionuclides. Again, a physicist can help
determine the need for special uniformity corrections.
CHAPTER 5. GUIDELINES FOR GENERAL IMAGING
168
5.1.3.2. Dual radionuclide studies
When using two radionuclides in a sequential study, images from the
lower energy radionuclide should be obtained first.
In principle, it is possible to use multiple energy windows to image two
radionuclides simultaneously. Such a technique involves many pitfalls however,
and the results will depend on the equipment used and special quality control
tests. The different energies will render different spatial resolutions. The
procedure must account for the detection of scatter from the higher energy
photons into the energy window used for lower energy photons (normally
referred to as downscatter). This procedure should be designed carefully by an
individual with the necessary expertise. If, on the other hand, the two radio
-
nuclides are imaged separately, it will be necessary to consider the effects
of
motion, especially if subsequent processing of the two images assumes
co-registration.
5.1.3.3. Matrix size for planar imaging
The matrix size is primarily dependent on resolution and is independent
of counting statistics. It is a requirement of sampling theory that the pixel size
be less than the FWHM divided by 2 for planar imaging, otherwise resolution is
lost and aliasing artefacts may occur. The matrix size is determined by the
field size divided by the pixel size. Increasing the matrix size beyond this limit
does not improve resolution. Magnification (zoom) can be applied for imaging
of small areas. If desired, a larger matrix size can be used for display using
interpolation.
5.1.3.4. Static imaging
The specific imaging parameters for a given static acquisition will vary in
accordance with the above sampling considerations, depending on the desired
clinical information. For computer acquired images, matrix size will depend on
the specific requirements of each type of study. For example, whole body scans
require large matrices to ensure that sampling is maintained over the large area
scanned. Typical matrix sizes for spot images are 128 × 128 and 256 × 256.
When large matrices are used for smaller areas, statistical fluctuation (noise)
may be excessive unless reduced by smoothing: this will result in decreased
spatial resolution. The digital appearance of smaller matrix sizes can be
improved by interpolation to large matrices for display, although this will not
improve resolution.
5.1. INTRODUCTION
169
5.1.3.5. Whole body imaging
Scan time varies depending on the count rate and count density required.
Matrix size must match the expected resolution for a given radionuclide.
Because a whole body image covers about 200 cm, the matrix dimension along
the length of the patient should be at least 512 pixels. Acquisition times greater
than about 30 min are not practical for routine use in unsedated patients.
5.1.3.6. Dynamic imaging
The time per frame selected depends on the temporal resolution needed
for the processing of the study and the organ function under investigation.
Shorter times are preferred for quantitative functional studies, provided
adequate statistics are obtained, in order to measure physiological changes. For
purposes of qualitative imaging alone, somewhat longer times are generally
used or multiple frames summed together in order to provide sufficient
imaging statistics for each frame.
For computer acquired images, the matrix size chosen for dynamic studies
may be smaller than that required for static imaging provided that the resultant
loss of resolution is acceptable for image interpretation. It is worth noting that
sometimes a choice has to be made between word and byte mode acquisitions.
If there is any doubt, word mode should be used to avoid pixel saturation that
may occur in byte mode. For high count rate studies, dead time effects may be
important. Count rate loss should be ascertained by dead time measurements,
about which a physicist can provide advice.
5.1.3.7. Gated imaging
ECG gating is used in functional cardiac studies to synchronize image
acquisition with the patient’s heart rate.
The number of frames per RR interval (the time duration between two
consecutive R waves of an ECG) should be no less than 16 for planar cardiac
blood pool ejection fraction measurements, 32 for time based measurements
(filling rates, etc.) and 8 for gated myocardial SPECT studies (although 16
frames should preferably be used if accurate volumetric parameters are
required).
For planar gated cardiac blood pool imaging, an electronic zoom should
normally be used to magnify the FOV to approximately 25 cm during acqui
-
sition. A matrix size of 64 × 64 is sufficient provided that magnification is
applied.