You are on page 1of 10

Society of Nuclear Medicine Procedure Guideline

for General Imaging
Version 3.0, approved May 30, 2004

Authors: J. Anthony Parker, MD, PhD (Beth Israel Hospital, Boston, MA); Margaret E. Daube-Witherspoon, PhD (Uni-
versity of Pennsylvania, Philadelphia, PA); L. Stephen Graham, PhD, FACR (West Los Angeles VA Medical Cen-
ter/UCLA School of Medicine, Los Angeles, CA); Henry D. Royal, MD (Mallinckrodt Institute of Radiology, St. Louis,
MO); Andrew E. Todd-Pokropek, PhD (University College of London, United Kingdom); and Michael V. Yester, PhD
(University of Alabama at Birmingham, Birmingham, AL)

I. Purpose II. Background Information and Definitions
The purpose of this document is to provide nuclear The strength of nuclear medicine lies in its ability to
medicine practitioners with general guidelines on im- image and quantify regional physiologic biochemical
aging in the practice of nuclear medicine. The guide- processes. This information may complement anat-
line includes recommendations common to most nu- omic information provided by other modalities.
clear medicine imaging procedures. Recommenda- A. Single-Photon Scintillation Cameras
tions specific to individual procedures are included in Single-photon scintillation cameras provide static,
their respective procedure guidelines. Among the dynamic, or gated images of the distribution of
many topics covered are: radiopharmaceuticals within the body. Single-
III. Procedure photon emission computed tomographic (SPECT,
A. Patient Preparation also known as SPET) images may be obtained by
(including considerations for women of child- reconstruction of a number of planar images taken
bearing age [III.A.2]) at different angles. CT scanners have been com-
B. Information Pertinent to Performing the Pro- bined with some single-photon cameras that have
cedure SPECT capability in order to provide attenuation
C. Precautions (including radiation precautions correction capability as well as localizing infor-
[III.C.1–4]) mation. Use of combined SPECT/CT scanners is
D. Quality Control likely to increase in the future.
IV. Single-Photon Scintillation Cameras B. Positron Cameras
A. Planar Imaging Positron cameras provide static, dynamic, or
B. SPECT Imaging gated images of the distribution of positron-
V. Nuclear Medicine Computer Systems emitting radionuclides within the body by detect-
VI. Reporting (including report contents [VI.D]) ing pairs of photons produced in coincidence by

The Society of Nuclear Medicine (SNM) has written and approved these guidelines as an educational tool designed to promote the cost-
effective use of high-quality nuclear medicine procedures or in the conduct of research and to assist practitioners in providing appropriate care
for patients. The guidelines should not be deemed inclusive of all proper procedures nor exclusive of other procedures reasonably directed to
obtaining the same results. They are neither inflexible rules nor requirements of practice and are not intended nor should they be used to es-
tablish a legal standard of care. For these reasons, SNM cautions against the use of these guidelines in litigation in which the clinical deci-
sions of a practitioner are called into question.
The ultimate judgment about the propriety of any specific procedure or course of action must be made by the physician when considering the
circumstances presented. Thus, an approach that differs from the guidelines is not necessarily below the standard of care. A conscientious
practitioner may responsibly adopt a course of action different from that set forth in the guidelines when, in his or her reasonable judgment,
such course of action is indicated by the condition of the patient, limitations on available resources, or advances in knowledge or technology
subsequent to publication of the guidelines.
All that should be expected is that the practitioner will follow a reasonable course of action based on current knowledge, available resources,
and the needs of the patient to deliver effective and safe medical care. The sole purpose of these guidelines is to assist practitioners in achiev-
ing this objective.
Advances in medicine occur at a rapid rate. The date of a guideline should always be considered in determining its current applicability.

the annihilation of a positron and an electron. tions are required if the potential radiation
Positron-emission tomographic (PET) images are dose to the infant is likely to exceed 1 mSv
produced by reconstruction from the coincidence (100 mrem). Pathways for exposure in-
pair data. CT scanners have been combined with clude ingestion of tracer concentrated in
some positron cameras to provide attenuation cor- breast milk and external exposure as a re-
rection capability as well as localizing informa- sult of close contact during breastfeeding.
tion. Use of combined PET/CT scanners is rap- There is considerable uncertainty about the
idly increasing. actual dose to the infant since little data
C. Nuclear Medicine Computer Systems are available for excretion into breast milk
Nuclear medicine computer systems collect, for most radiopharmaceuticals.
quantitate, analyze, and display the imaging in- h. For a few diagnostic studies (e.g., 131I
formation. whole-body imaging), the radiation dose
to the lactating breast can be large (10 cSv
[10 rem] or greater)––as much as 10 times
III. Procedure
the dose to the nonlactating breast. When
A. Patient Preparation possible it is best to delay these studies for
1. For many procedures, no patient preparation is at least 4 wk after breastfeeding has
necessary; however, when patient preparation stopped.
is necessary for specific procedures, the pre- 3. Before the study is begun, the procedure
paration is outlined in the relevant guideline. should be explained to the patient and ques-
2. In women of childbearing age, pregnancy and tions about the procedure should be answered.
lactation status should be determined. 4. When sedation is needed, appropriate proce-
a. Elective diagnostic nuclear medicine pro- dures should be followed (e.g., Procedure
cedures should be delayed until a patient is Guideline for Pediatric Sedation in Nuclear
no longer pregnant or breastfeeding. Medicine).
b. Nonelective diagnostic procedures (e.g., 5. Written informed consent is not required for
ventilation perfusion imaging) usually can standard diagnostic imaging procedures.
be modified to make a small fetal dose B. Information Pertinent to Performing the Pro-
even smaller. cedure
c. Care should be taken not to decrease the The nuclear medicine practitioner and staff can
diagnostic accuracy of the study when provide the best results when fully integrated into
modifications are made to reduce the ra- the clinical management team. Pertinent informa-
diation dose. tion includes:
d. A few diagnostic nuclear medicine studies 1. Indication.
(e.g., 131I whole-body imaging) may have 2. Relevant history and physical findings,
severe consequences for the developing fe- including medications and recent imaging
tus. The risk of performing these studies in with contrast-containing material.
a pregnant woman should be weighed 3. History of prior administration of radiophar-
against the benefits. maceuticals.
e. When performing diagnostic studies that 4. History of prior therapy, including surgery,
may have severe consequences for the de- that might affect radiopharmaceutical distribu-
veloping fetus, a pregnancy test should be tion.
obtained in women with childbearing po- 5. Results of pertinent imaging studies and labo-
tential. Under exceptional circumstances, ratory results.
the pregnancy test may be omitted. 6. In large or small patients, consideration
f. Breastfeeding should be interrupted for an should be given to adjustment of the adminis-
amount of time appropriate for the radio- tered activity.
pharmaceutical used. For a few diagnostic C. Precautions
studies (e.g., 131I whole-body imaging), 1. Medical personnel should be instructed in the
breastfeeding must be stopped for 1–2 mo; proper care of patients after radioisotope ad-
therefore, it is impractical to resume ministration. Generally, the hazard to person-
breastfeeding for that child. nel is small and measures taken to protect
g. Current Nuclear Regulatory Commission workers from biological hazards are sufficient
regulations require that written instruction to protect workers from radiological hazards.
be given to breastfeeding women if the po- In general, radioactive human waste is dis-
tential radiation dose to the infant is likely posed of in the same way as nonradioactive
to exceed 5 mSv (500 mrem); oral instruc- human waste.

2. In general, precautions taken to avoid biologi- IV. Single-Photon Scintillation Cameras
cal hazards from patient excreta are more than
sufficient to avoid the often much smaller ra- A. Planar Imaging
diation hazard. 1. Image Acquisition
3. Instructions should be provided on methods of a. Peaking
minimizing radiation exposure to the patient’s Scintillation camera should be peaked cor-
family and to the general public, where ap- rectly for the energy (energies) to be used
propriate. at least daily. A 15%–20% energy window
4. In general, there is no scientific or regulatory is typically used. The window is placed
reason why a pregnant nurse cannot provide symmetrically about the peak, or asym-
routine care to a patient who has had a diag- metrically if an appropriate energy correc-
nostic imaging study. The risk of caring for a tion is available. A physicist can help in
patient receiving therapy is small; however, it determining the limits of asymmetry that
may be prudent not to assign pregnant nurses can be used for a range of energies.
to care for these patients. b. Multiple Energy Windows
5. Adverse reactions associated with radiophar- The use of multiple energy windows for
maceutical administration should be reported radionuclides with more than 1 energy
to the Society of Nuclear Medicine/U.S. peak is advantageous. It is necessary to
Pharmacopeia Drug Problem Reporting Pro- check the multiple window spatial
gram. registration for the combination of the
6. Weight and size tolerances of equipment windows. A physicist can help determine
should be observed when imaging large pa- if the coregistration for all of the windows
tients. is proper in order to maintain the best
D. Quality Control spatial resolution and contrast. A col-
1. A procedure manual should exist, containing limator that will offer adequate resolution
the following: for the most energetic photons should be
a. A step-by-step procedure for each exam used. Intrinsic uniformity should be
including: radiopharmaceutical and ad- checked for imaging multiple energy
ministered activity, other pharmaceuticals windows for such radionuclides. A
and doses, patient preparation, views, physicist can help determine the need for
type of collimator, timing, instrument set- special uniformity corrections.
up, acquisition parameters, and data c. Dual Radionuclides
processing. When using 2 radionuclides in a sequential
b. A detailed description of the quality con- study, images from the lowest energy ra-
trol procedures for all instruments. This dionuclide should be obtained first.
should include the testing frequency, im- In principle, it is possible to use several
aging or data format, and data analysis and energy windows to image two radionu-
action levels. clides simultaneously. Such a technique
c. Detailed information on all aspects of ra- involves many pitfalls, however, and the
diation safety and emergency procedures. results will depend on the equipment used
2. Quality Control Records and special quality control tests. The dif-
a. Records of all quality control procedures ferent energies will have different spatial
should be maintained for the time speci- resolutions. The procedure must account
fied by regulatory agencies. These should for the detection of scatter from the higher
include all pertinent information relative to energy photons into the energy window
the acquisition of data and should include used for lower energy photons. Such a
the identification of the person who ac- procedure should be designed with care by
quired the data. an individual who has the necessary exper-
b. A log of all instrument problems should be tise.
maintained, and all problems should be re- d. Static Imaging
ported to the chief technologist or supervi- The specific imaging parameters for a
sor. This information will serve to identify given exam will vary depending on the de-
trends in instrument performance and to sired clinical information.
alert the clinical staff to factors that may For computer-acquired images, matrix
affect interpretation of patient data. size will vary depending upon the specific
c. All service records should be maintained requirements of each type of study.
in appropriate files for each instrument. Whole-body scans require large matrices.

When large matrices are used for smaller of at least 5 million counts in the entire
areas, higher resolution images can be ob- study will provide sufficient statistics
tained, but they have more statistical fluc- for quantitative and functional image
tuation (noise). Statistical fluctuation in processing.
large matrix sizes can be reduced by h. Pinhole Imaging
smoothing, but spatial resolution will de- i. Pinhole imaging provides the spatial
crease. The digital appearance of smaller resolution that most closely approaches
matrix sizes can be reduced by interpola- the intrinsic limit of the camera at the
tion to large matrices for display. expense of counting sensitivity. The
e. Whole-Body Imaging distance between the collimator and
Scan time varies depending on the count the patient determines both the degree
rate and count density required. Matrix of magnification and the sensitivity (or
size will vary depending on the radionu- count rate). Smaller pinhole apertures
clide, number of counts, and resolution re- (2–3 mm) provide better resolution but
quired. Because a whole-body image cov- lower sensitivity. The largest pinhole
ers about 200 cm, the matrix dimension in routine clinical use is 5 mm.
along the length of the patient should be at ii. For typical collimators with a 25-cm
least 512 pixels. Acquisition times greater diameter field of view, a matrix size of
than about 30 min are not practical for 256 × 256 or 128 × 128 is generally
routine use in unsedated ill patients. sufficient.
f. Dynamic Imaging 2. Processing
The time per frame should be selected de- a. Windowing
pending on the temporal resolution needed To optimally evaluate clinical studies, it is
for the process being studied. For quantita- often necessary to adjust the upper and
tive functional studies, shorter times are lower threshold on a computer display in
preferred, consistent with obtaining ade- order to enhance the contrast of the perti-
quate statistics, in order to measure physi- nent parts of the image. A lesion with in-
ologic changes. For purposes of imaging tense activity may make other lesions dif-
alone, longer times are generally used in ficult to visualize. Similarly, small varia-
order to provide sufficient imaging statis- tions in activity of large organs may be
tics for each frame. seen more clearly if contrast is enhanced.
For computer-acquired images, a ma- b. Cine
trix size of 64 × 64 or 128 × 128 may be A cine (movie) can be used to improve the
suitable, depending on the type of study. A detection of abnormalities such as in gas-
particular point to note is that it may be trointestinal bleeding studies or cardiac
necessary to choose between “word” and wall motion studies. Cines are also useful
byte mode acquisitions. If there is any in SPECT studies for visualizing spatial
question, word mode should be used to relationships and checking the projection
avoid pixel saturation that may occur in data for patient motion.
byte mode. 3. Quality Control
For high-count-rate studies, dead time a. Scintillation Camera
effects may be important. In such cases, i. The pulse height analyzer should be set
count rate loss should be ascertained by appropriately for the radioisotope to be
dead time measurements. used.
g. Gated Imaging ii. Field uniformity should be tested each
i. Electrocardiographic gating is used to day a scintillation camera is used.
synchronize image acquisition with the A solid- or liquid-filled uniform flood
patient’s heart rate. source can be used to evaluate system
ii. The number of frames per R-R interval (collimator on) uniformity. An accept-
should be no less than 16 for ejection able alternative is to remove the colli-
fraction measurements and 32 for time- mator and produce an image using a
based measurements (filling rates, etc). distant (5 × the maximum detector di-
iii. For gated cardiac blood pool imaging, mension) point source.
electronic zoom generally should be The point source must be on the cen-
used to magnify the field of view to tral axis (an imaginary line extending
approximately 25 cm. A matrix size of out perpendicularly from the center of
64 × 64 is sufficient. Typically, a total the detector). For planar imaging, the

general rule is that images for small- ages of a resolution test pattern can be
field cameras must contain at least 1.25 placed in all positions of the formatter and
million counts; for large-field circular a film carefully examined for uniformity
cameras, 2.5 million counts; and for and variation in intensity, spatial resolu-
large rectangular field cameras, 4 mil- tion, distortion, and the presence of arti-
lion counts. Alternatively, the manufac- facts. Monitor quality control may be a
turer’s recommendations can be used. particular problem in the telenuclear medi-
All imaging parameters and the tech- cine setting (see SNM Telenuclear Medi-
nologist’s initials should be recorded. cine Guideline).
iii. Spatial resolution and linearity (distor- c. Film Processors
tion) should be tested at least once i. Perform daily temperature and sensi-
each week. This measurement can be tometric checks.
made with the collimator on or off. ii. Perform periodic cleaning and mainte-
However, if the collimator is on, it nance.
should have the highest resolution that iii. The chemicals should be tested on a
is available and be designed for 99mTc. regular basis, as recommended by the
General-purpose collimators may show manufacturer. Maintain a log of all
moiré patterns. measurements and maintenance.
If the imaging system is digital, the d. Printers
finest matrix that is available should be Inexpensive hardcopy printers can be used
used. A coarse matrix will produce im- for distribution of demonstration quality
ages that appear to show a loss of spa- images. Printers with 600 or more dots per
tial resolution and may also show inch (dpi) can provide images of reason-
moiré patterns. Medium- and high- able quality.
energy collimators will produce moiré B. SPECT Imaging
patterns even in analog imaging sys- 1. Image Acquisition
tems because of the interplay between a. The parameters employed will depend
the collimator holes and the resolution greatly on the number of detectors incor-
pattern. The count density should be porated into the camera. For a single-head
the same as for uniformity images; camera, the general matrix size will nor-
thus, typically the bar pattern would be mally be 64 × 64. For multihead cameras,
imaged for one-half the number of the matrix size will be 64 × 64 or 128 ×
counts used for the uniformity images. 128 for higher resolution studies. The
All imaging parameters should be re- manufacturer’s processing protocols
corded. should be consulted for compatibility with
iv. All collimators except the pinhole specific data acquisitions.
should be tested annually for damage. b. Because counting statistics are very impor-
This can be done with 5- to 10-million tant in the reconstruction process, long im-
count images. aging times are sometimes needed. How-
v. Several additional tests should be per- ever, total imaging time should be less
formed periodically. These include than 30–45 min to minimize problems
tests of dead time, uniformity for nu- from patient motion.
clides other than 99mTc, and multiple- c. SPECT data can be acquired using step-
window spatial registration. and-shoot, continuous motion, or a hybrid
vi. The imaging system should be in- technique, depending on the camera de-
spected regularly for mechanical and sign and the type of study to be performed.
electrical hazards. Minor problems Continuous rotation will provide the most
should be corrected as soon as possi- efficient image-gathering capability. If the
ble; hazardous situations must be cor- stepping time between views is greater
rected immediately. Interlocks should than 10% of the imaging time per stop,
be tested on a regular basis as specified continuous motion is preferred.
in the procedure manual. d. The number of stops or views should be
b. Monitors and Imaging Formatter equal to or greater than 60 for single-head
A Society of Motion Picture and Televi- cameras when acquiring a 360° acquisi-
sion Engineers (SMPTE) or similar test tion. At least 30 views should be obtained
pattern may be used to ensure the faithful for 180° imaging. For high-resolution im-
reproduction of images. Alternatively, im- ages, 120 views would be used for a 360°

acquisition and 60 views would be used transmission scan that is used to gener-
for a 180° acquisition. ate an image (map) of the attenuation
2. Processing factors for the patient. The attenuation
a. Preprocessing factors are then applied to the recon-
i. Prefiltering of the projection data is structed images.
appropriate in many SPECT studies, c. Reformatting (Oblique Rotation)
because smoothing in the axial direc- Reformatting of SPECT images is used
tion may be included. When necessary, for organizing reconstructed images
guidance on the choice of filter and fil- along the primary axis of specific organs,
ter parameters (cutoff and order) most commonly the heart and brain.
should be sought from experts. Software is provided to guide the opera-
ii. Correction for patient motion may re- tor in the selection of the axis, for exam-
duce reconstruction artifacts. The high- ple, the long and short axis of the heart.
est quality study will be obtained by This technique is also used to generate
minimizing patient motion. slices of the brain that are parallel to the
b. Reconstruction orbital-meatal line.
i. Filtered Backprojection. A ramp filter 3. Quality Control
is always used in filtered backprojec- a. All quality control procedures described
tion. It corrects for the smoothing for planar cameras should be performed on
caused by the backprojection process. SPECT imaging systems.
Filters exist that “restore” some of the b. Center-of-rotation calibrations should be
resolution lost in the reconstruction performed according to the vendor’s rec-
process. The particular filter that is ommendations.
used depends upon the imaging c. High-count floods should be acquired pe-
equipment, the depth of the organ of riodically and used to apply uniformity
interest, and the radius of rotation. correction to SPECT projection data. The
Care should be taken with image en- vendor’s recommendation should be fol-
hancement (“restoration”), because it lowed. There are significant variations be-
may produce artifacts. tween vendors in terms of the number of
ii. Iterative Reconstruction. With the in- counts and the matrix size that should be
creasing power and memory of com- used. Typically, at least 30 million counts
puter hardware and with development are used for 64 × 64 matrices, and 120
of the more efficient ordered subsets million for 128 × 128 matrices. For some
expectation maximization (OSEM) al- cameras, the same isotope should be used
gorithm, iterative reconstruction of for the flood and imaging.
SPECT studies is now common in the d. Periodically, a high-count SPECT study of
clinical environment. This method a uniformly filled cylindrical phantom
makes it possible to incorporate correc- should be performed to assess tomo-
tion for many physical effects, such as graphic uniformity. Tomographic uniform-
nonuniform attenuation correction, ity should also be evaluated after major
scatter reduction or removal, variation repair of the camera or installation of new
of spatial resolution with distance, etc. or updated software. Some phantoms also
iii. Attenuation Correction. Attenuation contain structures that enable the user to
correction is used to correct for the fact evaluate contrast and spatial resolution.
that many photons that originate in the e. For multiple head cameras, the manufac-
body are absorbed within the body be- turer’s or physicist’s recommendation
fore they are detected. Most vendors should be followed for assessing the geo-
supply software for this operation, but metrical and electronic alignment of all
the algorithms are fairly simple and detectors with respect to each other.
only work correctly if the part of the
body that is imaged is homogeneous.
V. Nuclear Medicine Computer System
The operator may need to define the
boundaries of the body before applica- A. Components
tion of the algorithm. 1. Camera head
Some vendors now offer software A substantial (and increasing) amount of
for nonuniform attenuation correction. processing normally takes place not in the nu-
They are based on the acquisition of a clear medicine computer system itself but in

the camera head. However, such functions can display levels, number of overlays, num-
sometimes be provided in the associated im- ber of lookup, and display tables.
age processing system. These functions in- b. Hardcopy
clude: (a) image size, position, and zoom; (b) A hardcopy device may be provided for
energy correction; (c) spatial distortion correc- output of images onto paper or film. The
tion; (d) other corrections (scatter correction, method of generating such displays can be
dead time correction, depth of interaction cor- analog (video) or purely digital. Important
rection, sensitivity correction); and (e) digital parameters include print size, print resolu-
position computation. tion (number of points per unit of surface,
2. Interface number of gray levels or colors per point),
a. The 2 general types of interfaces are: (1) and print time.
analog: the signals from the gamma cam- c. Labeling
era are analog pulses and are digitized by Monitor or hardcopy display should in-
analog-to-digital converters (ADCs) in the clude patient name, patient identifier such
interface; and (2) digital: the signals as a medical record number, date, and type
passed from the camera head have already of study so that the displayed data can be
been digitized, and only digital signals are uniquely identified. Other information
handled in the interface. such as the phase of the study, view and
b. The interface handles the data in 2 basic institution may facilitate interpretation,
modes: (1) frame mode: complete images particularly when viewed out of the nu-
or matrices are available to the attached clear medicine department or at another
computer; and (2) list mode: data are institution.
passed on to the attached computer as a 5. Archive
list of event x, y coordinates, to which time A storage device with a database for lookup
information and energy information may and backup capability may be used for saving
be also attached. studies.
c. The interface also handles some special 6. Networks
functions. Dynamic study acquisitions It is common for the computers involved in
are commonly multiple images acquired processing nuclear medicine data to be linked
at different times and frame rates. The together in the form of a network. Such a net-
timing of such studies needs to be con- work is called a local area network (LAN) and
trolled by the acquisition system and in- may, in turn, be connected to other networks,
terface. for example, wide area networks (WAN) in-
For cardiac studies in particular, time cluding the Internet. Formats for sending nu-
lapse averaging is required, such that each clear medicine data over such networks in-
image acquired at some specific time clude: (a) manufacturer-specific formats, (b)
within the cardiac cycle is added to others INTERFILE, (c) DICOM, and (d) general
acquired at similar times. This operation is purpose graphic formats such as GIF, JPEG,
handled by the acquisition system and in- or TIFF.
terface, which must also handle the correc- B. Acquisition
tions needed when ectopic beats occur. 1. The types of acquisitions are described in sec-
3. Processing system tion IV.
The processing system is a computer com- 2. Matrix size and number of bytes
prised of: (a) 1 or more central processing a. The matrix size used for nuclear medicine
units (CPUs); (b) computer memory (random studies is almost always a power of 2,
access memory [RAM]); (c) backup storage, typical values being 64 × 64, 128 × 128,
such as hard disk drives; and (d) peripherals, 256 × 256, and 512 × 512. Nonsquare ma-
such as removable disk drives. trix sizes also exist, for example, for
4. Display whole-body studies.
The display is an important component of a b. Each pixel can be represented with a sin-
medical imaging system and may be specifi- gle byte (pixel values ranging from 0 to
cally designed for this purpose. 255 counts) or 16-bit words (pixel values
a. Monitors ranging up to a maximum of 32,000 or
Monitors typically display color and are 64,000).
large enough to display several images si- c. Overflow occurs when the number of
multaneously. Important parameters in- counts recorded at some given position (in
clude the display matrix size, number of some given pixel) exceeds the maximum

number of counts. Overflow is more likely pared against some truth. This assumes
to occur when using a byte matrix. that the simulated software is also correct.
3. Patient information This type of analysis is very helpful in cer-
a. When an acquisition is performed, patient tain cases (e.g., testing uniformity, compu-
information must be associated with the in- tation of ejection fraction) but still does
formation acquired. The minimal informa- not reflect true clinical reality.
tion would be a pointer to another database d. Use of reference patient data
containing the full patient information. Clinical studies are acquired from a vari-
b. Normally, the patient name and unique ety of centers under careful control, and
identification are stored with the study, the results of clinical parameters computed
plus other information such as date of by different software packages are com-
birth, sex, referring clinic, type of study, pared. No independent reference standard
provisional diagnosis, etc. exists, and normally only limited clinical
C. Processing Regions of Interest groups (normal/abnormal) exist. Nonethe-
The accuracy of regions of interest should be con- less, analysis of a standard data set is im-
firmed by viewing cine with the regions of inter- portant in evaluating the precision of
est overlaid to make certain that the patient did analysis software.
not move. e. Clinical audit
Periodic reanalysis of studies should be done The final judge of any analysis method is a
to measure intraobserver variability. When multi- clinical audit: the correctness and impact
ple individuals perform quantitative image analy- of the decisions made with respect to any
sis, periodic reanalysis of studies should be done method and process. Such methods are very
to measure interobserver variability. slow, requiring many years to establish
D. Quality Control conclusions and are not entirely appropriate
1. Trend analysis for software quality assurance, which nor-
A nuclear medicine computer system should mally requires much faster feedback.
be able to store and analyze the results of Any useful quality assurance system
gamma camera quality control tests so that must study both individual components and
trend analysis (the change in performance as a overall performance. This is common in the
function of time) can be analyzed and fault quality assurance of detectors but less
conditions predicted before they happen. common with respect to software. Thus a
When faults occur, an analysis can also be chain of tests is appropriate for studying in
made to assist in the determination of the sequence: (1) correct timing of a gated
cause. study acquisition; (2) correct handling of
Some systems are assisted by a remote bad beats; (3) accuracy of the definition of a
connection with the manufacturer. Many qual- region of interest; (4) correct values deter-
ity assurance functions can be performed re- mined for a time activity curve; (5) correct
motely and the resulting data then analyzed. computation of ejection fraction from a
2. Software quality control time activity curve; and (6) correct func-
a. Analysis of software code, line by line tioning of the complete clinical protocol.
The software code, more commonly the Each of these, except for the last, is a
process of producing the code, is con- single test with a well-defined input and
trolled by some process (e.g., as defined in output where correctness can be tested on
the ISO9000 series of specifications). For known data with known values. The final
complex software, analysis of the code it- test needs to be performed on many stud-
self cannot guarantee correct performance. ies—normal, abnormal, and intermediate
b. Phantom studies (if possible)—to establish the limits of
Phantoms with given physical parameters performance of the complete system.
are analyzed such that the computed re-
sults are compared with precisely known
VI. Reporting
distributions of activity. Such studies are
helpful in certain cases (e.g., the estima- A. Goals of a Nuclear Medicine Report
tion of size) but, in most cases, lack clini- 1. Provide the referring physician with a timely
cal realism. answer to the clinical question within the lim-
c. Simulated studies its of the test
Simulated data can be created with certain 2. Document the appropriateness, necessity, and
properties and the software assessed com- performance of the procedure

3. Expedite and assure correct billing ing or, if no diagnosis is known, the
B. Direct Communication pertinent symptom or sign that led to
1. Findings likely to have a significant, immedi- the procedure.
ate influence on patient care should be com- 3. Procedure Description
municated to the requesting physician or an a. Radiopharmaceutical.
appropriate representative in a timely manner. b. Administered dose.
2. Actual or attempted communication should be c. Route of administration.
documented as appropriate. d. Timing of imaging relative to radiophar-
3. Significant discrepancies between an initial maceutical administration.
and final report should be promptly reconciled e. Other drugs used, including name, dose,
by direct communication. route, rate of administration, and timing
C. Written Communication relative to images.
1. Include the items in section D, Contents of a f. Catheters or devices used.
Nuclear Medicine Report, that are appropriate g. Imaging technique, including alteration in
for a specific study. normal procedure.
2. Many items, such as patient identification or h. Complications or patient reactions.
radiopharmaceutical information, can be 4. Description of Findings
transferred to the report automatically or en- a. Significant positive findings as well as
tered by a technologist or secretary. pertinent negative findings should be men-
3. The final report should be proofread. tioned.
4. Electronic signature instead of a written signa- b. Image quality or other causes of study
ture is acceptable if access to the signature limitations (e.g., patient motion).
mechanism is secure. c. A reference range may be useful for quan-
5. Copies of the report should be sent to the re- titative values.
questing physician, made available to other d. See individual guidelines for specifics.
identified health care workers, and archived 5. Impression
for an appropriate period of time. a. A separate impression should be included
D. Contents of a Nuclear Medicine Report for all but the shortest reports.
1. Study Identification b. The impression should address the clinical
a. Patient name. indication for the scan.
b. Other information to uniquely identify the c. Diagnoses, differential diagnoses, and
patient, such as sex, date of birth, social judgments about the significance of the
security number, medical record number, pertinent findings may be included in the
or universal patient code. impression.
c. Requesting physician and other appropri- 6. Comments
ate health care providers such as the pri- a. Study limitations.
mary care physician. b. Recommendations for further procedures,
d. Type of study. if appropriate.
e. Date of study. c. Documentation of direct communication
f. Time of study, if relevant. of results including the name of the physi-
g. Study accession number (in a well- cian or physician designate and time/date
integrated information system, the study of contact.
accession number may not need to be visi- d. Comments may be included in the Impres-
ble). sion section, especially when brief.
h. Completion dates and times.
2. Clinical Information
VII. Issues Requiring Further Clarification
a. Indication.
b. Other relevant history—see specific pro- A guideline for General PET Imaging including
cedure guideline for details. PET/CT should be developed.
c. Information needed for billing such as re-
ferral number, patient status (e.g., inpa-
VIII. Concise Bibliography
tient/outpatient), or diagnostic codes (e.g.,
ICD-9-CM code). A. American Association of Physicists in Medicine.
i. Insurance carriers may not accept Rotating Scintillation Camera SPECT Acceptance
phrases such as “rule out” or “possible.” Testing and Quality Control. AAPM Report 42.
ii. List the diagnosis to the highest level Woodbury, NJ: American Institute of Physics;
of specificity known at the time of bill- 1987.

B. Basic sciences. In: Harbert JC, Eckelman WC,
Neumann RD, eds. Nuclear Medicine: Diagnosis
and Therapy. New York, NY: Thieme Medical
Publishers; 1996:1–358.
C. Hendel RC, Corbett JR, Cullom SJ, DePuey EG,
Garcia EV, Bateman TM. The value and practice
of attenuation correction for myocardial perfusion
SPECT imaging: a joint position statement from
the American Society of Nuclear Cardiology and
the Society of Nuclear Medicine. J Nucl Med.
D. Hutton BF. Basic sciences. In: Murray IPC, Ell
PJ, eds. Nuclear Medicine in Clinical Diagnosis
and Treatment. London, UK: Churchill Living-
stone; 1994:1277–1388.
E. Performance Measurements of Positron Emission
Tomographs. NEMA Standards Publication NU
2-2001. Rosslyn, VA: National Electrical Manu-
facturers Association; 2001.
F. Rosenthal MS, Cullom J, Hawkins W, et al.
Quantitative SPECT imaging; a review and rec-
ommendations by the focus committee of the so-
ciety of nuclear medicine computer and instru-
mentation council. J Nucl Med. 1995;36:1489–