CONTINUING EDUCATION

Acceptance Testing and Quality Controlof Gamma Cameras, Including SPECTPaul H. Murphy

Nuclear Medicine Service, St. Luke's Episcopal—Texas Children ‘sHospitals, and the TexasHeart Institute; and Nuclear Medicine Section, Department ofRadiology, Baylor College ofMedicine, Houston, Texas

Themeasurementof severalperformanceparametersof a scintillationcameraat the timeofinstallationand thereafterat regularintervalsis necessaryto ensurethat the cameraisoperatingwithinspecificationsandto detectchangesovertimethat can initiatea requestforservice.Therearevariedopinionson what constitutesa satisfactoryacceptancetest andquality control program. Few individuals would disagree that either an acceptance test orroutinequalitycontrolmeasurementsarenecessary,yet agreementon the contentsandfrequenciesof thesetests is lacking.This paperdiscussesthe performanceparametersofscintillationcamerasthat are usuallymeasuredfor planarimagingandalsofor SPECTwithrotatingcameras.

J NuclMed 28:1221—1227,1987

n acceptance test is verification following installation that a gamma camera meets the specifications ofperformance that were agreed upon at the time ofpurchase. Following acceptance, periodic checks on theconstancy of selective operating characteristics are necessary to ensure adequate performance and constitutethe quality control program (1—5).When the camera isalso used for SPECT imaging. additional parametersmust be included in both the acceptance test and thequality control program (6—8).There is no universalagreementon a protocol for camera acceptancetestingor a quality control program (9-11). Opinions varywidely on what are the minimum necessary tests andtheir frequency.

Possibly the most widely recognized set of cameraperformance tests is that proposed by the NationalElectrical Manufacturers Association (NEMA) from theNuclear Imaging Section ofthe Diagnostic Imaging andTherapy Systems Division (12,13). They describe howa manufacturer arrives at the specifications of his camera and how these are standardized between manufacturers (14). The NEMA standards do not set minimumperformance requirements. They are not intended foracceptance testing or quality control of scintillationcameras. Special equipment is necessary to make theNEMA measurements and therefore they are not in

ReceivedMar. 20, 1987;revisionaccepted Apr. 14, 1987.For reprints contact: Paul H. Murphy, PhD, Nuclear Medicine

Service,6720 Bertner Ave., Houston, TX 77030.

tended for all camera owners. Although there are criticsof the NEMA standards for “PerformanceMeasurements of Scintillation Cameras― and implementationof these standards by users is difficult and sometimesimpossible (15), familiarity with them does give insightinto the underlying parameters ofcamera performance.

PLANAR IMAGING

For planar imaging, the most important operatingparameters are spatial resolution, flood field uniformity,spatial linearity, energy resolution, count rate response,multi-window spatial registration, and sensitivity.

Spatial ResolutionSpatial resolution is a measure ofan imaging system's

ability to detect two closely spaced objects as two separate entities. The resolving limit of an instrument isthe minimum distance that two objects can be separatedand still be distinguished as two objects. For example,historically the resolving power of a telescope was described by the minimum angular separation oftwo starsthat could be distinguished as two objects. For a scmtillation camera, related parameters that are often usedas surrogate measure of spatial resolution is the fullwidth at half maximum of the point spread function orthe line spread function. This information can be extendedone step further by calculating the modulationtransfer function (MTF). The MTF is a measure of the

Volume28 •Number7 •July1987 1221

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

contrast transfer from object to image as a function ofspatial frequency, i.e., object size.

The NEMA specification for spatial resolution is thefull width at half maximum and full width at tenthmaximum ofline spread functions at multiple positions

within the central and useful (collimated) fields of viewof the camera. The central field of view is defined asthe central region having linear dimensions 75% ofthose of the useful field of view. The resolution shouldbe measured along both the X and the Y axes of thecamera to evaluate any orientational variation in thespread functions. The NEMA protocol requires a special slit phantom and acquisition into a digital matrixfor which there is at least 10 pixels distributed over thefull width at halfmaximum ofthe line spread function.Most nuclear medicine computers do not have a sufficiently large matrix for this unless a magnification(zoom) mode is used. However, zoom mode limitsevaluation to a small region of the useful field of view.

A more common and less quantitative approach isthe visual evaluation ofthe image ofa resolution phantorn. The quadrant bar phantom is most often used.The minimum perceptible bar spacing in a transmissionimage is used as an index of camera spatial resolution.This can be performed either with or without collimation. An estimate of the full width at half maximum ofthe line spread function at the same energy can beobtained by multiplying the smallest resolvable barspacings by 1.75 (3). With collimation on moderncameras, particularly medium or high energy collimators where the intrinsic resolution may approach thecollimator hole size, there is the potential for generatingMoire patterns that can complicate image interpretation. Moire patterns can also be observed on an intrinsicbar phantom image if it is collected in a digital matrixwhere the pixel size is close to the intrinsic spatialresolution. Generally, for routine quality control a resolution test pattern is used to evaluate intrinsic resolution at least weekly to ensure that there has not beendegradation of resolution in comparison to a benchmark image that was acquired at time of acceptance.At time of acceptance the resolution is also measuredwith the collimators in place in order to verify that thecollimators comply with agreed upon specifications.

Intrinsic UniformityThe intrinsic flood field uniformity of a scintillation

camera is the ability ofthe camera to produce a uniformimage when exposed to a homogeneous spatial distribution of gamma rays. Most modern cameras are notdesigned to be intrinsically uniform because gains inspatial resolution can be obtained by sacrificing intrinsic uniformity. Therefore, these systems require somemechanism of uniformity correction. Typically, theuniformity correction involves computer corrected registration of regional photopeak Z signal and X and Ycoordinate position signals. The monitoring of camera

uniformity is probably the most sensitive indicator ofcamera performance and should be performed dailyprior to patient studies. Most nuclear medicine facilitiesperform daily uncorrected (when possible) and corrected flood images that are subjectively evaluated. Fora more quantitative evaluation, the flood image can bedigitized and numerically or graphically analyzed. TheNEMA protocol for intrinsic flood field uniformityanalyzes both differential and integral uniformity over

the useful and central fields of view. The integral urnformity represents the maximum pixel count ratechange over the indicated field of view expressed as apercent. The differential uniformity is the maximumchange over a five pixel distance in either the X or Ydirections thereby representing the maximum rate ofchangeof the regional count rate.

Flood field images should be evaluated under thesame energy conditions used for patient imaging. If offpeak technetium-99m (99mTc) imaging is performedthen the flood should be checked for uniformity at thispulse height setting. Also, the user should verify thatfloods obtained with gamma ray energies other thanthose used for uniformity correction are acceptablyuniform for clinical studies. This is particularly true forthallium-201, gallium-67 (67Ga), and iodine-131. Ifthese floods do not appear adequately uniform then acorrection flood of the same energy is indicated.

LinearitySpatial linearity is one of the parameters that influ

ence flood field uniformity. In the ideal system, astraight line source of gamma rays should yield astraight line in the image. Any deviation from a straightline represents distortion. Because ofthe finite numberof PM tubes in scintillation cameras there is a wavelike distortion in the image ofa line source. Quantitativelinearity correction is accomplished by many manufacturers by storing in a microprocessor a correction algorithm that shiftsthe positionsof scintillation eventsthe appropriatedirection and distanceto yield a straightline.

The NEMA protocol for measuring linearity involvesthe acquisition along the X and Y directions of animage from a multi-slit phantom, the same one used

for the spatial resolution measurement, followed by ananalysis ofthe line spread peak positions. Deviations ofthe peak position from the true location of the centerof the slits is a measure of the deviation from linearity.Typically, most departments do not measure linearityseparate from either spatial resolution or flood fielduniformity. A subjective evaluation of linearity is ohtamed when a bar phantom or an orthogonal holephantom is imaged.

Energy ResolutionThe energy resolution of a scintillation camera is a

measure of its ability to separately distinguish the ener

1222 Murphy TheJournalof NuclearMedicine

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

gies of two gamma rays that differ only slightly inenergy.Traditionally the parameter that is measuredisthe full width at half maximum of the photopeak expressed as a percentage of the photopeak value. SinceCompton scatter in soft tissue is so dominant at theenergies used for gamma-ray imaging, the ability todiscriminate against these, which reflects the energyresolution of the system, has a substantial impact onimage quality. Energy resolution is a function of gammaray energy and therefore the energy at which it ismeasured must be specified. The NEMA protocol suggests measurement at 140 keV with uniform irradiationof the useful field of view, without a collimator. Mostmodern cameras have energy resolution for @mTcinthe range of 10—12%.Energy resolution is not routinelymeasured as part of a quality control program. If possible, it should be measured as part of an acceptancetest. For an accurate measurement a multichannel analyzer is usually required so that, as suggested by NEMA,at least 50 channels are distributed over the full widthat half maximum of the photopeak. In some camerasthe energy pulse is not readily accessible so the measurement may be impracticable.

Count Rate PerformanceFollowing the interaction of a gamma ray in a scm

tillation crystal, a finite period of time is necessary forthe scintillation event to evolve to the output “count―.During this time interval the camera cannot processother scintillations as distinctly separate events. Theminimum time required between two events for theevents to be processed as distinct events is the deadtime.This limits the maximum processing rate and results inincreasing data loss as the input rate increases. Forquantitative image analysis it is important that thepercentage data loss due to camera deadtime be knownand corrected. This is usually a problem only duringvery high count rate studies such as first-pass cardiacscintiangiography. It is only during this type of studythat a large amount of radioactivity (possibly 10—30mCi of 99mTc)is simultaneously within the field viewof the camera. In this situation the input rate to thecollimated detector can exceed several hundred thousand gamma rays per second. Under most other imaging situations the radioactivity is distributed throughoutthe body and that in the field of view results in muchlower input rates to the detector. The NEMA protocolsuggests specifying the count rate performance as theinput rate that results in a 10% data loss and also as themaximum observablecount rate.NEMA definesatechnique for measuring the deadtime ofthe camera underthe assumption that the system behaves as fully paralyzable in the range where a 20% data loss is expected.Few laboratories measure either camera deadtime orthe count rate response ofa scintillation camera as partof a routine quality control protocol. It is, however, anappropriate parameter to measure during acceptance

testing and it is one of the specifications quoted bymost camera manufacturers.

Multi-Window Spatial RegistrationThe number of light photons incident onto a photo

multiplier tube in a scintillation camera is determinedby how close the photomultiplier tube is to the interaction site in the crystal and how much energy isdeposited in the crystal and, therefore, the total numberof light photons emitted. Thus, the signal from a photomultiplier tube reflects some combination of spatiallocation and energy deposition. The signals from all ofthe photomultiplier tubes are analyzed in order to derive positioning voltages for the scintillation event. Ifthese positioning voltages are not normalized for theenergy of the gamma ray, or equivalently the windowsetting of the pulse height analyzer, gamma rays of

different energies originating from the same locationwould be recorded at different locations on the image.

Therefore, correction circuits are necessary to ensurethat images ofdifferent energies spatially coincide whenthegammaraysoriginate from the samelocations.Thisis most important when summed images from multienergy emitters are used. For example, if two or threeofthe 67Gapeaks are summed to form a single image itis important that the images derived from each windoware spatially coincident. The NEMA approach for measuring this multi-window spatial registration consists ofimaging a point source of67Ga at two locations on eachofthe X and Y axes and analyzing the displacement inthe location ofthe image ofthese point sources for eachof the windows. This procedure is usually performedduring an acceptance test but rarely as part ofa routinequality control program.

SensitivityThe sensitivity of a scintillation camera is measured

as the number ofdetected counts per unit time per unitsource activity for a specified energy window and geometry of measurement. The intrinsic sensitivity istypically tens of thousands times higher than thatthrough low-energy, parallel hole collimators. Since thecamera is never used for imaging without a collimatorit is the sensitivity of the total system that is of interestas well as the relative sensitivity between the availablecollimators for that particular camera. Manufacturersofcollimators label sets ofcollimators quite differently.What is called a high resolution collimator by onemanufacturer might be a general purpose collimator toanother. Therefore, it is important to know the relativesensitivity of the available collimators. The NEMAapproach to measuring sensitivity is to measure thecount rate per microcurie for a small source of knownactivity in a 10 cm diameter flat dish. This should bedone through each collimator so that the relative sensitivities can be calculated. It is usually done onceduring acceptance testing but not routinely as part of

Volume28 •Number7 •July1987 1223

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

the quality control program. The activity should be lowenough so that data loss because ofcamera deadtime isnegligible. It is possible to monitor the sensitivity of thecamera over time as part ofthe quality control programif the amount of activity used each day for the floodimage or for the bar phantom image is known. Bystandardizing the geometry for the acquisition of eitherof these images and noting the time for acquisition ofa preset number of counts, a measure of relative sensitivity can be obtained by dividing by the activity.

Other Important ParametersThere are several very important performance char

acteristics ofscintillation cameras that are not addressedby the NEMA standard which should be included ineither acceptance tests or routine quality control tests.These include:

1. High count density flood images obtained througheach collimator for evaluation of any defects in thecollimator construction. This should be done duringacceptance testing and at any time thereafter when thereis a suspicion that the collimator has been damaged.

2. On cameras capable of acquiring whole-body images by scanning either by moving the detector or thepatient bed, spatial resolution and alignment of thezipper between multiple passes should be checked. Thiscan be done by scanning a quadrant bar phantom setat a diagonal to the direction of motion. Comparisonof the image with and without scanning will indicateresolution loss or misalignment ofthe interface betweenmultiple passes. Also the speed should be calibratedperiodically.

3. Ifa formatter is used, the fidelity ofimage qualityover all areas ofthe film should be checked. This shouldbe studied for all of the standard formats that are usedclinically by recording a flood image at each film leation and evaluating the images for constancy of intensity and shape.

4. When the camera is used with a computer thecamera/computer interface must be tested to ensurethat information is not being lost or distorted duringthis transfer. The X-Y gains of the interface should bechecked by imaging an object of known dimensionsand calculating the gains along the two axes. Also thecount rate recorded at the camera console should coincide with that recorded by the computer.

5. Most cameras have built-in safety switches todeactivate the system, particularly with respect to motion of the detector. Some of these are manually operated (“panicbuttons―)and others are supposed to beautomatic, such as stopping the detector motion inwhole-body scans. These safety switches should betested at installation and periodically thereafter.

There are other special accessories on some camerasthat should be included in a quality control program,but will not be discussed here. Table 1 summarizes theparameters previously discussed with common tech

1224 Murphy

niques and suggested frequencies for monitoring. It isemphasized that the selection of parameters measuredroutinely as part of a quality control program are verysubjective. If there is lack of universal agreement onwhat should be measured and there is even less agreement on how often these should be measured. In themajority of nuclear medicine departments a minimumquality control protocol consists offlood field uniformity measured daily and spatial resolution measured atleast weekly.

SPECF

A great deal of attention has been given in the pastfew years to quality control protocols for SPECT. Thesubstantial increase in the number of publications onartifact formation and quality control procedures forSPECT are not without justification. It is now wellknown that there are significant problems with respectto accurate presentation of the spatial distribution ofradioactivity in transverse sections and slight degrees of“sloppiness―in data acquisition or processing can havesubstantial adverse effects on the results. In fact, poorlyperformed SPECT can be more detrimental to patientcarethan no SPECTat all, particularly ifSPECT imagesare used in lieu of planar images. The potential advantages of tomography over planar imaging can beachieved only by emphasizing the strict adherence toproven acquisition protocols and by understandingwhat the processing protocols are doing to the data.This means that the physician and technologist mustbe familiar with options of data acquisition, availablefilters, and attenuation correction techniques. Theymust know where each is most appropriate, i.e., different ifiters may be indicated for different clinical studies.The appropriate processing is a function of the spatialfrequencies in the object and the count density in theplanar images. Filters should be evaluated under clinically meaningful conditions, not from high count phantom images. With this increased demand for attentionto detail in SPECT, the need for more sophisticatedtraining of technologists and physicians and frequentreviews is obvious. In addition to the camera performance parameters discussed previously, the followingmust be considered in a SPECT system acceptance testand quality control program.

Camera/Collimator UniformityBecause the spatial orientation of the camera and

object changes during a SPECT acquisition, isolatedareas of nonuniformity are spread throughout the dataset and therefore into the reconstructed image duringback projection. For 360°data acquisition the mostcommon presentation are rings of increased or decreased intensity suggesting a bull's-eye appearance.

Over and above the camera's built-in uniformity

TheJournalof NuclearMedicine

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

TABLE IAcceptanceTest andQualityControlParameters

Performanceparameters Acceptancetest protocol

Full width half maximum and fullwidthtenthma)dmumof linespreadfunctionfor @“Tcalongx andVaxes.Alsobarphantornimagesasbenchmark.

Integralanddifferentialoveruseful fle@of-vlewandcentralfield-of-view.Floodimagesforbenchmarks.Offpeakandother energies.

Linesourcesor orthogonalholephantomevaluatedsubjectivaly.

Fullwidthathalfmaximumof@“Tcphotopeakexpressedas

percent.Twentypercentdataloss,resolv

ingtime,maximumcountratefor 20%wkidow.

Two @GapointsourcesimagedalongX andV axeswithmdivkiualwmdowsandsummedwmdows. Meas@ed @ecementson indMdualwindowimages.

CountrateperMCithroughallcollimatorswfth20% @ndow,calculateabsolutesensitivityandrelativesensitivityfor availablecoknators.

TenmillioncountfloodsthrougheachcollimatorforevaluationofcoUk@atordefects.

Floodimagesat all locationsandfor all imagewas.

Scanbarphantomalongdiagonalandcompareto StatiOneryimage,calibratespeed.

Qualitycontrolprotocol

Resolutionphantom(quadrantbarphantom,rotate90@betweenacquisItIons)

Intrinsic or extrinsic flood evaluatedqualitatively

Noseparatetest,subjectiveevaluationfrombarphantomresolutiontestandIntrinsicto theuniformItymeasurement

Same

Same

Frequency

Spatialresolution

lkiiformity

Uneahty

Energy resolution

Countrateresponse

Mufti-windowregistration

Weekly

Daily(beforefirstpatient)

Annualor when suspiciousofdamage

Annual

Annual

Annual

Annual

Annual

Annual

Same

Same

Same

Same

Same

Sensmvfty

Highcountdensitycollimatorfloods

Formatter

Whole-bodyaccessory

correction scheme (linearity and energy correction), ahigh count correction flood is also necessary for SPECTimaging. The temporal and spatial stability of the uniformity is very important. The high count density correction flood is collected in one camera orientation andat one point in time. Data acquisition for SPECT resultsfrom multiple camera orientations which could changethe uniformity, for example, due to effects of environmental magnetic fields changing orientation with theaxes of the photomultiplier (PM) tubes, slippage of PMtubes, shift of the collimator, etc. The stability of theuniformity over time should be checked for each camera to determine the necessary frequency of collectingthe high count density correction flood. Most cameramanufacturers have a recommended protocol forSPECT uniformity correction but this should be verifled with respect to frequency of application.

Center of RotationIn the reconstruction of images from projections it is

assumed that the image matrix, representing the activitydistribution in a section, has a constant relationship tothe data acquisition matrix. If one matrix shifts withrespect to the other for different angles of data acquisition, then the image reconstructed from back projectingthe data will be blurred because of the relative motionof the two matrixes. The simplest example is that of apoint source. For 360-degree acquisition the count rateversuspixel number of the point should be a straightline along the axial direction and should have the shapeof a sine wave along the other camera axis. The amplitude ofthe sine wave is the distance ofthe point sourcefrom the center ofrotation. Deviation from this suggeststhat the center of rotation is moving during the acquisition and the reconstructed image of the point source

1225Volume28 •Number7 •July1987

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

will be blurred, resulting in loss of resolution. Sinceactivity distributions other than points can be considered as an assemblage of many point sources of variousintensities, the reconstructed image of an extendedsource will be blurred (over and above the blurring dueto the resolution ofthe camera) ifthe center of rotationis not centered in the image matrix.

There are several ways to measure the center ofrotation and apply corrections to the data after acquisition or apply off sets to the ADCs before data acquisition. Each camera manufacturer supplies a recommendedprotocol for their systems,including software.

As with the uniformity correction flood image, thelocation of the center of rotation must be checkedfrequently and the data retained if future reconstructions from the original projections are anticipated. It isthe value ofthese parameters at the time ofdata acquisition that is relevant, not the condition of the cameraat a future time when the original data set is reconstructed, i.e., you should not correct projections acquired last month with this month's uniformity orcenter of rotation values.

Alignment of Camera with Axis of RotationWith parallel hole collimators the plane ofthe detec

tor must be parallel to the axis ofrotation. Ifthe camerahead is angled with respect to the axis of rotation, theprojections will not be orthogonal to the object, andthe slice volume will vary from this effect over andabove that due to changes in camera resolution withdepth. For this reason it is extremely important thatthe camera stand be level and perpendicular to thefloor, and that the camera head be parallel to the floorin the detector up or down orientation. During camerainstallation the leveling of the detector stand and thepatient pallet should receive close attention from themanufacturer's service representatives. Prior to eachdata acquisition, the operator must level the detectorhead. There are various ways to do this including useofa spirit level on the detector and this can be checkedby acquisitions ofimages ofpoint sources of 180 degreeopposed views. Misalignment results in resolution lossin the reconstructed images similar to misalignment ofthe centerofrotation. By aligning the cameraheadwiththe axis of rotation, it is assumedthat the collimatorholes are perpendicular to the detector and thus the axisof rotation.

Adjustment of X-Y GainsThe gains in the X and Y directions, i.e., over the

two axes of the detector, should be matched so thatpixel cross sections are square and distance factors inboth axes are equal. The gains can be measured byimaging point or line sources separated by known distances along each axis. The separation of the peaks ofthe count profiles should be the same number of pixelsalong the X and the Y axes. Note, this is important for

both the camera X-Y gains and at the camera/computerinterface. Pixel dimensions in the third direction of thereconstructed image, i.e., the Z axis, can be determinedfrom the measurements in the X-Y plane but this pixeldimension should not be confused with the slice thickness. The slice thickness can be defined by a line spreadfunction obtained along the Z axis and expressed as afull width ofhalfmaximum ofthis line spread function.

In summary, the thorough characterizationof a scmtillation camera's performance immediately after installation is important for at least two reasons: (a) as amechanism to verify that the system performs as designated by the manufacturer's specifications and agreedupon in the purchase order; and (b) to serve as a benchmark against which future measurements can be compared. The camera quality control protocol should bedesigned to verify acceptable performance ofthe cameraprior to each patient study or at least daily. Exactlywhat constitutes the specific test within a quality controlprogram is still very much a matter of personal preference. The recommendations presented here are thoseofthe author and certainly do not represent a consensusby any organized group within nuclear medicine.

REFERENCES

1. Quality Control for Scintillation Cameras. Rockville:U.S. Department of Health, Education, and Welfare,1976.

2. Paras P, Van Tuinen RJ, Hamilton DR. Quality control for scintillation cameras. In: Rhodes BA,ed. Quality control in nuclear medicine. Radiopharmaceuticals, instrumentation, and in vitro assays. St. Louis:The C.V. Mosby Company, 1977, 336—348.

3. Scintillation camera acceptance testing and performance evaluation. AAPM Report No. 6, Chicago:American Institute of Physics, 1980.

4. Graham IS. Quality assurance of Anger cameras. In:Rao DV, Chandra R, Graham MC, eds. Physics ofnuclear medicine. Recent advances. Medical PhysicsMonograph No. 10, Chicago: American Institute ofPhysics, 1984: 68—82.

5. Computer-aided scintillation camera acceptancetesting. AAPM Report No. 9, Chicago: American InstituteofPhysics,1981.

6. Croft B. Quality assurance: acceptance testing andquality control. In: Croft B. Single photon emissioncomputed tomography. Chicago: Year Book MedicalPublishers, Inc., 1986: 177—234.

7. Greer K, Jaszczak R, Harris C, et al. Quality controlin SPECT. JNuclMed Tech 1985; 13:76—85.

8. Harkness BA, Rogers WL, Clinthorne NH, et al.SPECT: quality control procedures and artifact identification.JNuclMedTech1983;11:55—60.

9. LewellenTK, Graham MM. A low-contrast phantomfor daily quality control. J NucI Med 1981; 22:729—282.

10. Hasegawa BH, Kirch DL, Lefree MT. et al. Qualitycontrol ofscintillation cameras using a minicomputer.JNuclMed 1981;22:1075—1080.

11. Lancaster JL, Kopp DT, Lasher JS, et al. Practicalgamma camera quality control with a four-point phan

The Journal of Nuclear Medicine1226 Murphy

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

torn. JNuclMed 1985;26:300—307.12. Performance measurements of scintillation cameras.

Standards Publication No. NU1, Washington DC:National Electrical Manufacturers Association, 1986.

1 3. Muehllehner 0, Wake RH, Sano R. Standards for

performance measurements in scintillation cameras. JNuciMed 1981; 22:72—77.

14. Hasegawa BH, Kirch DL. The measure ofa camerapathways for future understanding [Teaching EditorialJ.JNuclMed 1981;22:78—81.

15. Raff U, Spitzer VM, Hendee WR. Practicality ofNEMA performance specification measurements foruser-based acceptance testing and routine quality assurance. JNuclMed 1984; 25:679—687.

1227Volume28 •Number7 •July1987

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from

1987;28:1221-1227.J Nucl Med.   Paul H. Murphy  Acceptance Testing and Quality Control of Gamma Cameras, Including SPECT

http://jnm.snmjournals.org/content/28/7/1221This article and updated information are available at:

  http://jnm.snmjournals.org/site/subscriptions/online.xhtml

Information about subscriptions to JNM can be found at:  

http://jnm.snmjournals.org/site/misc/permission.xhtmlInformation about reproducing figures, tables, or other portions of this article can be found online at:

(Print ISSN: 0161-5505, Online ISSN: 2159-662X)1850 Samuel Morse Drive, Reston, VA 20190.SNMMI | Society of Nuclear Medicine and Molecular Imaging

is published monthly.The Journal of Nuclear Medicine

© Copyright 1987 SNMMI; all rights reserved.

by on January 1, 2019. For personal use only. jnm.snmjournals.org Downloaded from