Positron Emission Tomography: A Review Of Basic Principles, Scanner .

1y ago
1.19 MB
25 Pages
Last View : 3d ago
Last Download : 5m ago
Upload by : Averie Goad

Positron Emission Tomography: A Review of Basic Principles,Scanner Design and Performance, and Current SystemsPat ZanzonicoSINCE THE INCEPTION of positron emissiontomography (PET) several decades ago, PETscanner design and performance have improveddramatically. The number of detector elements hasincreased from 20 to 20,000 and the axial fieldof view from 2 to 20 cm, the spatial resolutionhas improved from 25 to 5 mm, and thesensitivity has increased 1000-fold.1,2 At thesame time, the clinical applications of PET havegrown dramatically as well.3-5 The current articlereviews the technical basis of the remarkableadvance of this modality—the underlying principles of PET and the basic design and performancecharacteristics of PET scanners— emphasizingdedicated “full-ring” devices and including multimodality (ie, PET-CT) and special-purpose (ie,small-animal) devices.PHYSICAL BASIS OF PET AND PET “EVENTS”PET is based on the annihilation coincidencedetection (ACD) of the two colinear 511-keV -rays resulting from the mutual annihilation of apositron and a negatron, its antiparticle (Fig 1).Positron-negatron annihilation occurs at the end ofthe positron range, when the positron has dissipated all of its kinetic energy and both the positronand negatron are essentially at rest. The totalpositron and negatron energy is therefore 1.22MeV, the sum of their equal rest mass energies(Eo 511, keV 0.511 MeV), and their totalmomentum (a vector, or signed, quantity) is zero.Accordingly, to conserve energy and momentum,the total energy of the two annihilation -rays mustequal 1.22 MeV and their total momentum zero.Two equal-energy (511-keV) annihilation -raystraveling in opposite directions, corresponding toequal-magnitude, opposite-sign (positive and negative) momenta, are therefore emitted.In the parlance of ACD, each of the two annihilation photons is referred to as a “single” and thetotal count rate (counts per second (cps)) for theindividual annihilation photons is called the “singles count rate” (Fig 1). Only when signals fromthe two coincidence detectors simultaneously trigger the coincidence circuit is an output, a “truecoincidence event” (“true”), generated by this circuit. The volume between the opposed coincidencedetectors (the shaded area in Fig 1) is referred to asa “line of response (LOR).” LORs are thus definedelectronically, and an important advantage of ACDis that absorptive collimation is not required. As aresult, the sensitivity (measured count rate per unitactivity) of PET is much higher (two to threeorders of magnitude higher) than that of Angercamera imaging.6 Not every annihilation yields acounted event, however, because both annihilationphotons must strike the coincident detectors for anevent to be counted. As a result, the singles countrate in PET is typically much higher than the truescount rate.The 511 keV-in-energy and the simultaneity-ofdetection requirements for counting of a true coincidence event are not absolute. Scintillation detectors typically have a rather coarse energyresolution— up to 30% (expressed as the percentfull-width half-maximum of the 511-keV photopeak)—and therefore photons within a broad energy range (eg, 250 to 650 keV) can be counted asvalid annihilation -rays.7 Compton-scattered annihilation -rays and scattered and unscatterednonannihilation photons may therefore be included, producing spurious or mispositioned coincidence events.Each detected photon (single) is time-stamped( 1 ns 1 10 9 s), and a true coincidenceevent is defined as a pair of annihilation photonscounted by the coincidence detectors within a timeinterval called the “coincidence timing window ,”typically 6 to 12 ns. Such a finite timing window isnecessitated by several considerations. First, depending on the exact position of the positronnegatron annihilation, the annihilation photonsreach the detectors at slightly different times.However, because these photons travel at the speedof light (c 3 1010 cm/s), this effect is verysmall. Second, the transit and processing of thesignal pulses through the detector circuitry is rapidFrom the Memorial Sloan-Kettering Cancer Center, NewYork, NY.Address reprint requests to Pat Zanzonico, PhD, MemorialSloan-Kettering Cancer Center, 1275 York Avenue, New York,NY 10021. 2004 Elsevier Inc. All rights reserved.0001-2998/04/3402-0003 s in Nuclear Medicine, Vol XXXIV, No 2 (April), 2004: pp 87-11187

88PAT ZANZONICOFig 1. The basic principle of ACD. An event is counted only when each of the two 511-keV annihilation -rays are detectedsimultaneously, that is, within a time interval corresponding to the coincidence timing window , by the two detectors. The shadedLOR corresponds to the volume between and defined by the cross-sectional area of the coincidence detectors 1 and 2 and Ċ1 andĊ2 are the singles count rates recorded by detectors 1 and 2, respectively.but not instantaneous. Third, the light signal emitted by the scintillation detectors used in PET isemitted not instantaneously but over a finite timeinterval, called the “scintillation decay time,” ofthe order of 10 to 100 ns.In addition to the true coincidence events (Figs 1and 2A), a number of other types of events occur inPET, events that degrade quantitative accuracy aswell as image quality. Random or accidental coincidence events (“randoms”) occur when annihilation -rays from two separate positron-negatronannihilations are detected in two different detectorswithin the coincidence timing window (Fig 2B).Randoms thus increase the detected coincidencecount rate by contributing spuriously placed coincidence events. Because the total activity-containing volume is typically much greater than the LOR,random coincidences are common and the randomscount rate may actually exceed the trues count rate.Clinically, the ratio of the randoms-to-true countrates range from 0.1 to 0.2 for brain imaging togreater than 1 for whole-body imaging.6The randoms count rate is actually proportionalto the product of the singles count rate and therefore the square of the activity present6:Ċ randoms 2 Ċ 1Ċ 2(1)where Ċrandoms the randoms count rate (cps), the coincidence timing window (sec), and Ċ1and Ċ2 the detector 1 and detector 2 singlescount rates (cps), respectively (Fig 1). Importantly,because the trues count rate is only linearly proportional to the activity, the ratio of the randomsto-trues count rates increases linearly with activity.Therefore, imaging times cannot be reduced simply by using higher and higher administered activities, as the randoms count rate will increase morerapidly than the trues count rate and at some pointprohibitively degrade image quality. By usingabsorptive septa to restrict the activity-containingregion sampled by coincidence detectors to avolume defined by the cross-sectional area of thedetectors—as in two-dimensional (2D) PET (seebelow)—the randoms-to-true count rate ratio canbe reduced substantially. By using “faster” detectors and therefore shorter coincidence timing windows, the randoms-to-true count rate can be reduced further (see below).Annihilation -rays traveling out of an LORmay undergo Compton scatter and be re-directedback into the LOR (Fig 2C). The scattered photonmay, however, retain sufficient energy to fallwithin the energy window set for the 511-keVannihilation -rays and produce a coincidence

PET PRINCIPLES AND TECHNOLOGY89Fig 2. The various events associated with ACD of positron-emitting radionuclides, illustrated for two opposed banks ofcoincidence detectors and assuming only one opposed pair of detectors are in coincidence. (A) A true coincidence (“true”) iscounted only when each of the two 511-keV annihilation -rays for a single positron-negatron annihilation are not scattered andare detected within the timing window of the two coincidence detectors. (B) A random or accidental coincidence (“random”) isan inappropriately detected and positioned coincidence (the dashed line) that arises from two separate annihilations, with one -ray from each of the two annihilations detected within the timing window of the coincidence-detector pair. (C) A scatteredcoincidence (“scatter”) is a mispositioned coincidence (the dashed line) resulting from a single annihilation, with one of the -raysundergoing a small-angle Compton scatter but retaining sufficient energy to fall within the 511-keV energy window. (D) A spuriouscoincidence is an inappropriately detected and positioned coincidence (the dashed line) which arises from an annihilation and acascade -ray, scattered or unscattered but having sufficient energy to fall within the 511-keV energy window. Spuriouscoincidences occur only for radionuclides which emit both positron and prompt cascade -ray(s).event. Such scatter coincidences (“scatter”) resultin mispositioned events. The scatter count rate aswell as the trues count rate are proportional to theactivity present and therefore the scatter-to-truescount rate ratio is independent of activity.6 Because trues and scatter each result from singleannihilation events, the scatter-to-trues count rateratio is likewise independent of the coincidencetiming window. On the other hand, interdetectorsepta used in 2D PET (See below.) reduce thescatter count rate considerably.Many positron-emitting radioisotopes also emitsignificant numbers of high-energy prompt -rays,and such -rays may be in cascade with each otheror with the positrons.8,9 These can result in spurious coincidences which are spatially uncorrelatedbut nonetheless counted as true events (Fig2D).10,11 Although such coincidences degradeoverall quality and quantitative accuracy, isotopessuch as have copper-62, gallium-66, gallium-68,bromine-75, rubidium-82, yttrium-86, and iodine124, nonetheless been used effectively in PET.10,11Table 1 includes, for selected positron emitters, theenergy and abundance of - (and x-) rays withsufficient energy (ie, greater than 250 keV) to fallwithin the 511-keV energy windows typically usedto count annihilation -rays in PET. Besides the -ray energies and abundance, Table 1 includesother pertinent properties of positron emitters suchas the physical half-life (T1/2), the branching ratio(ie, the percentage of total decays resulting inpositron emission instead of electron capture), the

90PAT ZANZONICOTable 1. Physical Properties of Positron-Emitting Radionuclides Used in PET8,9 Range in Water (mm)x- and -rays 0.25 MeVPhysical Halflife T1/2BranchingRatioMaximum Energy (MeV)Re (ref. 75)Rrms (ref. 24)Energy r-62Copper-64Gallium-66Gallium-6820.4 min9.96 min2.05 min1.83 h9.74 min12.7 h9.49 h1.14 e-76Rubidium-8216.1 h1.3 rium-86Iodine-12414.7 h4.18 dionuclide6.07.0maximum positron energies (Emax), the maximumextrapolated range (Rmax), the root-mean-square(rms) positron range (Rrms), and the method ofproduction.PET DETECTORS AND DETECTORCONFIGURATIONSDetector MaterialsTo date, only four detector materials—all inorganic scintillators— have been widely used in PETscanners: thallium-doped sodium iodide (NaI(Tl)),bismuth germanate (BGO), cerium-doped lutetiumoxyorthosilicate (LSO(Ce) or simply LSO), andcerium-doped gadolinium oxyorthosilicate (GSO(Ce)or simply GSO) (Table 2).7,12,13The most important practical features of scintillation detectors include high mass density ( ) andeffective atomic number (Zeff), high light output,and speed (Table 2). A high mass number and higheffective atomic number maximize the crystalstopping power (ie, linear attenuation coefficient ) and therefore the detection of radiations. Inaddition, a higher-atomic number crystal will havea higher proportion of photoelectric than Comptoninteractions,7 facilitating energy discrimination ofscattered photons. High light output reduces statistical uncertainty (noise) in the scintillation andassociated electronic signal and thus improvesenergy resolution and scatter rejection. A fastcrystal (ie, a crystal with a short scintillation decaytime) allows the use of a narrow coincidencetiming window, , reducing the randoms countrate. Other detector considerations include: transparency of the crystal to its own scintillations ronCyclotronCyclotronCyclotronGenerator minimal self-absorption); matching of the index ofrefraction ( ) of the crystal to that of the photodetector (specifically, the entrance window [ 1.5]of a photomultiplier tube [PMT]); matching of thescintillation wavelength to the light response of thephotodetector (the PMT photocathode, with maximum sensitivity in the 390-410 nm, or blue,wavelength range); and minimal hygroscopic behavior.7NaI(Tl) crystals were used in the original PETscanners. Higher-density and -effective atomic materials, such as BGO, LSO, and GSO, haveemerged as the detectors of choice for PET because of their greater stopping power for 511-keVannihilation -rays (Table 2). Note, for example,that the attenuation length for 511-keV -rays is atleast twice as long in NaI(Tl) as in BGO, GSO, orLSO. Among the latter three materials, GSO andLSO have a faster light output—nearly 10-foldfaster—than BGO, with LSO having a much greaterlight output—approximately 3-fold greater—thaneither BGO or GSO. GSO has somewhat betterenergy resolution, and scatter rejection capability,than either BGO or LSO.A notable disadvantage of LSO is the presenceof a naturally-occurring long-lived radioisotope oflutetium, lutetium-177.7 Lutetium-177 has an isotopic abundance of 2.6% and a half-life of 4 1010 years and emits two prompt -rays (88%abundance) of 201 and 306 keV in energy; thesummed energy of 507 keV falls well within the511-keV energy windows commonly used in PETscanners. The presence of lutetium-177 results in ameasured background count rate of 240 cps/cm3 of

1.85Y8410230*The intrinsic efficiency of 2-cm thick coincidence detectors for 511-keV annihilation -rays.41,00017513.7NaI:Tl0.34 2 (2 cm) 0.24*1.82N1042030,000667.4Lu2SiO5:Ce0.88 2 (2 cm) 0.69*3240N94408,00025596.7Gd2SiO5:Ce0.70 2 (2 cm) odium IodideNaI(Tl)Density, (gm/cm3)CompositionMaterial0.95 2 (2 cm) 0.72*4060N12480Scintillationwavelength, (nm)ScintillationDecay Time(nsec)Light Output(photonsper MeV)Linear AttenuationCoefficient, , for511-keV -rays (/cm)RelativeProbability ,ZeffTable 2. Physical Properties of PET Scintillators7Most commonly in dedicated PET scanners,detectors are arranged in rings or polygonal arraysof discrete, small-area detectors completely encircling the patient (Figs 3A–C). In such systems,multi-coincidence fanbeam detection is used, witheach detector element operated in coincidence withmultiple opposed detector elements. For a ringcomprised of N detector elements, a total of N/4 toN/2 fanbeams is acquired. In rings, each element istypically in coincidence with about half of the totaldetectors in the ring and in polygonal arrays withthe opposed detector bank. PET systems with onlypartial detector rings are less expensive but requirerotation of the detector assembly about the longitudinal axis of the patient to complete acquisitionof the projection data (Fig 3B). In addition, continuous, large-area detectors, such as those foundin multi-head Anger camera systems and used forsingle-photon emission computed tomography(SPECT), have now been appropriately modifiedand are used for coincidence imaging of positronemitters (Fig 3D). With two or even three suchdetectors, rotation (180 or 120 , respectively) isrequired for complete angular sampling. Alternatively, large-area detectors may be arranged in apolygon (if flat) or in a circle (if curved) completely encircling the patient (Fig 3E and F); suchsystems have been manufactured using GSO aswell as NaI(Tl).Technical performance improves but cost increases as one progresses from dual-head coincidence Anger cameras at the low end to partial ringsto polygonal arrays to multiple full rings at thehigh end—multiple rings being the prevailing configuration among current dedicated PET scanners.2Clinical performance, specifically, lesion detectability improves as well, as demonstrated in astudy by Kadrmas and Christian15 using a realisticwhole-body anthropomorphic phantom with multiple focal lesions with clinically realistic dimensions and lesion-to-background activity concentration ratios. Lesion detectability was clearly bestEnergyResolutionat 511 keV(% FWHM)Detector Configurations300Hygroscopic(Y/N)?RefractiveIndex, LSO14 and singles and trues count rates of 100,000and 10,000 cps, respectively, in clinical LSO PETscanners. Although the former has a negligibleeffect on typical emission scans, the latter wouldsignificantly increase the statistical uncertainty(noise) in single-photon transmission scans (eg,with cesium-137) used for attenuation correction.71.85912.15PET PRINCIPLES AND TECHNOLOGY

92PAT ZANZONICOFig 3. PET scanner detector configurations. (A) Multiple full rings of detector blocks comprised of discrete, small-area detectorelements. (B) Multiple partial rings of detector blocks comprised of small-area detector elements. (C) Hexagonal array of detectorsbanks comprised of small-area detector elements. (D) Opposed large-area detectors such as Anger cameras. (E) Hexagonal arrayof large-area detectors. (F) Circular arrangement of six large-area, curved detectors. Inset: Multi-coincidence fanbeam detectionused in detector rings and arrays of small-area detectors. Such fanbeam transverse sampling data are generally treated asparallel-beam data. Adapted from Cherry et al6 with permission.among the multiple-ring systems and poorestamong coincidence Anger camera systems, withthe most pronounced differences for the smallestlesions (Fig 4).Early PET detectors consisted of a single scintillation crystal backed by a single PMT, with thecross-sectional dimensions of the crystal definingthe coincidence LOR and thus intrinsic (ie, crystal)spatial resolution. To improve spatial resolution,therefore, greater numbers of smaller crystals arerequired. Thus, the practically achievable miniaturization of PMTs and associated electronics and thecost of large numbers of detectors, PMTs, etc,representing well over half of the costs of PETscanners,16 limit intrinsic resolution. The blockdetector1,2,17 was an ingenious solution to thislimitation.A block detector consists of a large cubic pieceof scintillator (2 2 to 3 3 cm in cross sectionby 2 to 3 cm in depth) partially cut, or scored,depth-wise into a rectangular array of detectorelements (Fig 5A). The cuts are filled with reflective material to optically isolate the detector elements from one another and to maximize lightcollection efficiency by the PMTs backing thescintillator. Crystal elements with a somewhatsmaller cross-section improve spatial resolution—but only to a certain point. As the cross-section ofdetector elements is reduced and the number ofelements increased, the number of cuts and therefore the fraction of the scintillator face occupied bythe filling material increase. As a result, the detector element packing fraction (ie, the fraction of thescintillator face occupied by scintillation material)and therefore the intrinsic sensitivity decrease.The depth of the cuts into the crystal is notuniform but increases from approximately half thethickness at the center to nearly the full thicknessat the edge of each side of the scintillator (Fig 5A);the actual depths of the cuts are determined empirically to yield a spatially linear distribution of lightamong the four PMTs—in a 2 2 array— backingthe scintillator. The position at which the annihilation -ray strikes the scintillator is then determined by Anger arithmetic. The response of theblock detector is not uniform (Fig 5B). Rather,recorded events are clustered at points corresponding to the individual detector elements and thenassigned to a specific element in the two-dimensional array using a look-up table derived byuniform irradiation of the scintillator. The majoradvantage of the block detector is that it allows anarray of many small detector elements (typically8 8 64) to be spatially encoded using only

PET PRINCIPLES AND TECHNOLOGY93Fig 4. Comparative coronal images, obtained with seven commercially available systems, of a wholebody anthropomorphic phantomwith 27 focal lesions filled with thepositron-emitter sodium-22. The lesions were lucite spheres of innerdiameters 7, 8, 12, and 16 mm andvolumes 0.17, 0.27, 0.91, and 2.10mL, respectively, and were filledwith activity concentrations 4, 6, 10,and 16 times that in the backgroundsoft tissues. The total activity in thephantom at the start of each scanwas 3 mCi. The systems evaluatedincluded: three BGO dedicated fullring PET scanners, the Advance(General Electric), the ECAT EXACTHRⴙ (Siemens-CTI, Knoxville, TN),and the ECAT EXACT HR961 (Siemens-CTI); a NaI(Tl) dedicated PETscanner with six large-area curvedcrystals, the C-PET (Philips-ADAC,Milpitas, CA); two NaI(Tl) hybridPET-SPECT scanners with two Anger cameras, the Irix (Marconi Medical Systems, Cleveland, OH) andthe Vertex MCD (Philips-ADAC);and a NaI(Tl) hybrid PET-SPECTscanner with three Anger cameras,the Axis (Marconi Medical Systems).Data were processed using themanufacturer-supplied software withmanufacturer-suggested default processing parameters. Lesion detectability performance was clearly bestamong the BGO systems and poorest among the hybrid PET-SPECTsystems. Reproduced from Kadrmasand Christian15 with permission.four PMTs rather than one PMT per element,yielding high spatial resolution while minimizingcosts.In modern ring-detector PET scanners (Table2),7,16 there are typically three to four rings of 100to 200 block detectors each. There are about 6 to 8cuts per block detector, yielding an array of 6 6 36 to 8 8 64 elements 4 4 to 6 6 mmeach. Overall, therefore, there are a total of 10,000to 20,000 detector elements. Ring diameters rangefrom 80 to 90 cm, the patient ports and transaxialfields of view from 50 to 70 cm, and the axial (orlongitudinal) fields of view from 20 to 30 cm,typically yielding about 50 transaxial image planeseach 2 to 4 mm thick.An important refinement of the block detector is“quadrant (or light) sharing,6” where a two-by-twoarray of four larger PMTs not only backs a singlescintillator block but each PMT in the array alsobacks the corner of an adjacent block (Fig 5C).This reduces the total number of PMTs by a factorof four and thus reduces overall cost. Disadvan-

94PAT ZANZONICOFig 5. (A) Diagram of a typical block detector composed of a partially and variably scored scintillator crystal (photo) coupled toa 2 ⴛ 2 array of PMTs. (B) “Uncorrected” block detector image of a uniform radiation source. The detector response is nonuniform,with the recorded events clustered at points corresponding to the individual detector elements. (C) Arrangement of PMTs in astandard and in a quadrant-sharing block detector. In a standard block detector, a 2 ⴛ 2 array of PMTs backs a single scintillatorcrystal. In a quadrant-sharing block detector, each PMT in the 2 ⴛ 2 array backs the corners (or quadrants) in adjacent crystals. (D)Diagram of a phoswich block detector, comprised of adjacent layers of two materials with different scintillation decay times.Adapted from Cherry et al6 with permission. (Color version of figure is available online.)tages of quadrant sharing include higher deadtimecount losses and more involved detector servicingbecause of the nonmodular design.A notable refinement of PET scintillators hasbeen the use of adjacent layers of two differentmaterials with significantly different scintillationdecay times (such as LSO and GSO, with decaytimes of approximately 40 and 60 nsec, respectively); this is known as phoswich (Fig 5D).6Based on the pulse shape of the scintillation signal,the interaction of the annihilation -ray can therefore be localized to one or the other half of thephoswich detector. The resolution-degradingdepth-of-interaction effect is therefore reduced bya factor of two. However, the fabrication ofphoswich is more complex than that of singlecomponent detectors, and to date it has not beenwidely used in commercial PET scanners.A recently developed alternative to the blockdetector is the pixelated detector matrix,6,16,18wherein individual small-area detectors elements(typically 4 6 mm in cross-section by 20 mm indepth) are fixed onto a continuous light guidebacked by a close-packed array of PMTs (Fig 6).Pixelated detectors (PIXELAR ) are used in thePhilips-ADAC (Milpitas, CA) Allegro PET scanner and Gemini PET-CT scanners.Two-Dimensional (2D) Versus ThreeDimensional (3D) Data AcquisitionPET ring scanners originally employed lead ortungsten walls, or septa, positioned between andextending radially inward from the detector elements (Figs 7A–C). The Advance PET scanner(General Electric Medical Systems, Milwaukee,WI), for example, uses tungsten septa 1 mm thickand 12 cm long. In this approach, known as 2DPET, these interring annular septa define plane-by-

PET PRINCIPLES AND TECHNOLOGY95Fig 6. The Pixelar (Philips-ADAC) pixelated GSO detector, comprised of individual small-area detectors elements (A) fixedonto a continuous light guide backed by a close-packed array of PMTs (B). (Color version of figure is available online.)plane LORs and largely eliminate out-of-planeannihilation -rays. By minimizing the contribution of out-of-plane randoms and scatter, imagequality is optimized, especially for large-volumesources (ie, as in whole-body PET). However, 2DPET also eliminates most trues and thus reducessensitivity considerably. Typically, both “direct”and “cross” image planes are reconstructed fromLORs within the same detector ring (corresponding to a so-called “ring difference ( )” of 0) andbetween two adjacent detector rings (ring difference of 1), respectively. In the EXACT HR 2D (Siemens-CTI) PET scanner, for example, 32detector rings span an axial field of view (FOV) of15.5 cm, yielding a total of 63 contiguous imageplanes comprised of 32 direct and 31 cross planes;in general, a scanner with n rings of detectorelements yields a total of (2n 1) image planes.The cross-planes lie halfway between the directplanes defined by detector elements and, conceptually, can be assigned to a “virtual” ring ofdetectors lying midway between two adjacent detector rings. Because the cross-plane images resultfrom two LORs and the direct-plane images fromonly one, the cross-plane image sensitivity is abouttwice that of the direct-plane images (Figs 5A–C).This results, in an uncorrected PET study of auniform volume source, in alternating lower-countand higher-count transverse section images. In thenewer 2D PET systems, LORs among as many asthree adjacent rings, corresponding to a ring difference of 3, are used to improve sensitivity.Increasing the ring difference does, however, degrade spatial resolution somewhat.Sensitivity can be increased substantially byremoving the septa altogether and including coincidence events from all of the LORs among all thedetectors (Fig 7D)—a system with 10,000 detector elements has approximately 100 million LORs.This is known as three-dimensional (3D) PET,19and is widely used among state-of-the-art PETscanners. (“Collimator-less” Anger camera-basedcoincidence imaging of positron emitters is inherently 3D.) Sensitivity is increased approximatelyfivefold in 3D relative to 2D PET— but with aconsiderable increase in the randoms and scattercount rates. Clinically, the scatter-to-true countrate ratios range from 0.2 (2D) to 0.5 (3D) in brainand from 0.4 (2D) to 2 (3D) in the whole body.6 Tocompensate for the increase in scatter count rates,

96PAT ZANZONICOFig 7. 2D and 3D PET data acquisition schemes (axial cross-sectional views of a multi-ring scanner) and the corresponding axialsensitivity profiles. (A–C) 2D data acquisition with a ring difference of 0 (direct planes only), 1, and 3, respectively. (D) 3D(septa-less) data acquisition. The sensitivity profiles show the nonuniformity of response as a function of position along the axialFOV. Adapted from Cherry et al6 with permission.detectors (such as GSO and LSO) with betterenergy resolution7 and accurate scatter-correctionalgorithms20 are required for 3D PET. And, tominimize the increased randoms count rates anddeadtime count-rate losses, shorter coincidencetiming windows, and therefore faster detectors(such as GSO and LSO), are required. Data processing time, for 3D PET is about an order ofmagnitude longer than for 2D PET.16,20In contrast to the relatively uniform axial sensitivity for 2D PET, the axial sensitivity profile for a3D PET scanner is triangular and peaked at thecenter of the field of view (Fig 7D). Thus, whole-body 3D PET studies require considerable overlapof adjacent bed-position acquisitions— optimally,one-half of the axial FOVs16,20—to yield uniformsensitivity over the resulting whole-body images.In PET in general and 3D PET in particular, it isimportant that the ends of the detector assemblyare adequately shielded to minimize the contribution of counts from activity outside the axial FOV.PET PERFORMANCEAn extensive series of parameters have beendeveloped over the years to characterize PETscanner performance, and detailed data acquisition

PET PRINCIPLES AND TECHNOLOGY97Fig 8. Physical aspects of positron-negatron annihilation and their effects on PET spa

positron and a negatron, its antiparticle (Fig 1). Positron-negatron annihilation occurs at the end of the positron range, when the positron has dissi-pated all of its kinetic energy and both the positron and negatron are essentially at rest. The total positron and negatron energy is therefore 1.22 MeV, the sum of their equal rest mass energies

Related Documents:

Tomography and Hybrid Modalities Positron Emission Tomography (PET) is an imaging technique performed by using positron emitting radiotracers. Positron decay occurs with neutron-poor radionuclides and consists in the conversion of a proton into a neutron with the simultaneous emission of a positron (β ) and a neutrino (ν). The positron has a .

Anti-3-[18F]FACBC positron emission tomography-computerized tomography correctly up-staged 18 of 70 cases (25.7%) in which there was a consensus on the presence or absence of extraprostatic involvement. Conclusions: Better diagnostic performance was noted for anti-3-[18F]FACBC positron emission tomography-computerized tomography than for 111In .

Positron Emission Tomography Single Photon Emission Computed Tomography. PET and SPECT Properties of ideal imaging nuclides, biological, chemical , physical Production of radionuclides . Positron emitters Image removed due to copyright restrictions. Chart of the Nuclides The "organic" elements 13N[13N]NH 3 15O[15O]H 2 O

Context: Positron emission tomography is a nuclear medicine imaging that deals with physiological function using radioisotopes. With the most PET (Positron Emission Tomography) scanners in integration with the CT scan-ners of late, this technology has registered phenomenal growth. The small amount of radioactive material is called Radiotracers.

cay by positron emission and the most commonly used positron emitting nu-clides are fluorine-18 (18F), carbon-11 (11C) and oxygen-15 (15O). When the radioisotope decays, the emitted positron travels a very short distance from the initial decay site before it interacts with an electron and a positron-electron annihilation occurs.

Medical Imaging - Positron Emission Tomography (thanks to Bill Moses, Life Sciences Div. LBNL) What is Positron Emission Tomography (PET)? Patient injected with drug having β emitting isotope. Drug localizes in patient. Isotope decays, emitting β . β annihilates with e- from tissue, forming back-to-back 511 keV photon pair.

the positron emission tomography tracer 18F-florbetapir. Positron emission tomography agents have shown great promise in mapping fibrillar amyloid deposition in the brain. 18F-florbetapir [(E)-4-(2-(6-(2-(2-(2-18F-fluoroethoxy) e] is a novel imaging agent that binds with high affinity (Kd

Aliens' Behaviour Connectives Game This game was originally developed in 2006 for Year 5/6 at Dunkirk Primary School in Nottingham. It has also been used at KS3. We have chosen this topic because we hope it will encourage children to produce their own alien names (a useful use of phonically regular nonsense words!), portraits and sentences .