High-Resolution L(Y)SO Detectors Using PMT-Quadrant-Sharing for Human and Animal PET Cameras

862 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 55, NO. 3, JUNE 2008

High-Resolution L(Y)SO Detectors UsingPMT-Quadrant-Sharing for Human

and Animal PET CamerasRocio A. Ramirez, Shitao Liu, Jiguo Liu, Yuxuan Zhang, Soonseok Kim, Hossain Baghaei, Hongdi Li,

Yu Wang, and Wai-Hoi Wong

Abstract—We developed high resolution �� detectorsfor human and animal PET applications using Photomulti-plier-quadrant-sharing (PQS) technology. The crystal sizeswere � � � � � mm� for the animal PQS-blocks and� �� mm� for human ones. Polymer mirror filmpatterns (PMR) were placed between crystals as reflector. Theblocks were assembled together using optical grease and wrappedby Teflon tape. The blocks were coupled to regular round PMTsof 19/51 mm in PQS configuration. List-mode data of Ga-68source (511 keV) were acquired with our high yield pileup-eventrecovery (HYPER) electronics and data acquisition software.The high voltage bias was 1100 V. Crystal decoding maps andindividual crystal energy resolutions were extracted from thedata. To investigate the potential imaging resolution of the PETcameras with these blocks, we used GATE (Geant4 Application forTomographic Emission) simulation package. GATE is a GEANT4based software toolkit for realistic simulation of PET and SPECTsystems. The packing fractions of these blocks were found to be95.6% and 98.2%. From the decoding maps, all 196 and 225 crys-tals were clearly identified. The average energy resolutions were14.1% and 15.6%. For small animal PET systems, the detectorring diameter was 16.5 cm with an axial field of view (AFOV) of11.8 cm. The simulation data suggests that a reconstructed radial(tangential) spatial resolution of 1.24 (1.25) mm near the center ispotentially achievable. For the whole-body human PET systems,the detector ring diameter was 86 cm. The simulation data sug-gests that a reconstructed radial (tangential) spatial resolution of3.09(3.38) mm near the center is potentially achievable. From thisstudy we can conclude that the PQS design could achieve highspatial resolutions and excellent energy resolutions on human andanimal PET systems with substantially lower production costs andinexpensive readout devices.

Index Terms—Animal and human PET detectors,��,��,PET, photomultiplier-quadrant-sharing (PQS) technology.

I. INTRODUCTION

HIGH-RESOLUTION detector blocks have been devel-oped for human and animal PET systems using the

photomultiplier-quadrant-sharing (PQS) decoding technologyand crystals. Our goal is to develop high-resolutionand high-sensitivity PET cameras with lower production costs

Manuscript received May 15, 2007; revised October 10, 2007. This workis supported in part by the National Institutes of Health under Grants NIH-CA58980 PHS, NIH-CA61880 PHS, NIH-CA76246 PHS, NIH-CA58980SPHS, NIH-EB01481 PHS, NIH-EB00217 PHS, and NIH-EB01038 PHS, theU.S. Army Breast Cancer Research Grant, and by Texas ARP/ATP under Grant003657-0058-2001.

The authors are with the University of Texas M.D. Anderson Cancer Center,Houston, TX 77030 USA (e-mail: [email protected]).

Digital Object Identifier 10.1109/TNS.2008.922832

[1]. In previous work, we demonstrated the decoding capabilityof the PQS technology with BGO and GSO scintillation de-tectors [2]–[4]. Due to the fact that have a highoptical light yield ,a short attenuation length (10.4: 14.1: 11.4, in millimeters),and a short decay time (300: 60: 40, in nanoseconds) [5], itshould be possible to develop PQS-PET detectorswith better performance than the previous ones. The aim ofthis study was to test the decoding limits of the lower costPQS analog decoding technique, to develop high-resolutionPQS PET detectors with crystals, to characterizethe detector performance, and to estimate the potential PETimaging resolutions that can be achieved with such a detectordesign, using Monte Carlo simulations.

II. MATERIALS AND METHODS

A. Materials

crystals from Crystal Photonics, Inc. (Sanford,Florida, U.S.A.) and circular 19-mm PMT (XP1912) from Pho-tonis Imaging Sensors (Beauregard, Bride Cedex, France) wereused for the animal PET detector studies. crystals fromSiemens Molecular Imaging (Malvern, PA, U.S.A.) and circular51-mm PMT (XP52Y2) from Photonis Imaging Sensors wereused for the human PET detector studies. The crystalswere segmented with dimensions of mm .The surface was finished by polishing with lapping motion to4 m (lapped to 4 m) on five crystal surfaces and polishedsmoothly 1 m on one end. The LYSO individual crystalswere assembled in a 14 14 array block detector for animalPET application. The dimensions of the crystals were

mm and surface finishing was lapped inour lab, from 10 m to 4 m on five crystal surfaces and oneend was polished 1 m . For human PET application,individual crystals were used to develop a block detector with15 15 array. For controlling light sharing, we placed reflectorpatterns between adjacent crystals, [3] and [4]. For the reflectorspatterns, we used 3M Vikuiti ESR (enhanced spectral reflector)film from 3M (St. Paul, MN, U.S.A.); which is a multilayeredpolymer film with high reflectance of 98.5% across the visiblelight spectrum and a thickness of 0.065 mm [6].

B. Methods

To achieve the 4- m crystal surface finishing in thescintillators, we polished with lapping motion, five surfacesof individual crystals samples with 10- m finishing. For the

0018-9499/$25.00 © 2008 IEEE

RAMIREZ et al.: HIGH-RESOLUTION L(Y)SO DETECTORS USING PMT-QUADRANT-SHARING FOR HUMAN AND ANIMAL PET CAMERAS 863

Fig. 1. Experimental reflector patterns.

Fig. 2. Physical animal (a) and human (b) �� PQS-blocks.

lapping process, aluminum oxide lapping film placed in agrinding–polishing machine (MultiPrep System) from AlliedHigh Tech Products, Inc. (Rancho Dominguez, CA, U.S.A.)[7], was used.

PQS-block detectors were assembled from the individualcrystals with mirror-film reflector patterns inserted preciselybetween them. The crystals and reflector patterns were assem-bled together to build a PQS-detector blocks, using opticalcoupling grease (Rexon 688). The PQS detector blocks werecovered by mirror film over five of the external surfaces andlater, wrapped with six layers of Teflon tape. The averagethickness of the mirror film and grease between two adjacentcrystals was 0.08 mm. The initial patterns of the mirror-filmreflectors were obtained from a simulator, a theoretical and sta-tistical calculation tool for the light sharing in the PQS decodingmethod, which was developed in our laboratory, [8]. Thereafter,the patterns were literally refined by several trial-and-errorexperimental attempts. Fig. 1 shows a typical set of reflectorpatterns. Fig. 2 shows physical PQS-blocks for twoconfigurations: (a) 14 14 and (b) 15 15, for animal andhuman PQS-PET blocks, respectively.

To assess the scintillator performance, the detector blockswere optically coupled to regular round PMTs of 19/51 mm inPQS configuration using optical coupling grease (Rexon 688).A mirror-film reflector was placed under the central gap insidethe four circular PMTs to enhance the light-collection efficiencyfor the crystals in the central region of a block. Compared to theregular detector design, the PQS method is more efficient in itsuse of PMTs. In this configuration, a quarter of each PMT canreach a maximum decodable ratio of . In other words, eachPMT serves four block detectors [9], instead of one as in the

Fig. 3. Comparison between PQS block and regular block/PMT design.

regular detector design. This design can lower the PMT cost byup to 75% compared to a PET camera using four PMTs to de-code just one block detector, as shown in Fig. 3.

List-mode data was taken with Ga-68 source (511 keV)and home-made electronics and data acquisition software,[10]–[12]. For data analysis, a crystal-decoding map was ac-quired with an energy threshold, typically set to the Comptonedge of the block-composite energy spectrum. A crystallookup table was created from crystal identifications through acrystal-decoding map. Using the lookup table and the list data,we extracted spectral information such as pulse height andenergy resolution from the energy spectra of individual crystalsby fitting the spectra with Gaussian curves. Profiles of eachrow and column of crystals were obtained from the decodingmap. The profiles were fitted with Gaussian functions to obtainthe heights and position of peaks and valleys. To minimize theinfluences of PMT nonuniformities, we averaged values forcrystals with the same rotational symmetry (symmetry restora-tion), i.e., the same up–down and left–right crystal locationsrelative to the center of the block. Light-collection efficiencyis obtained by measuring the pulse height of each individualcrystal spectrum.

Due to the large number of middle crystals sitting in spacewithout coupling to any photocathode; the data were alsoanalyzed according to the crystal location in the block andcrystal-PMT coupling. For this analysis the block was dividedinto four types of crystal-to-PMT coupling regions. Regiontype 1 designated crystals that are coupled in the center of thephotocathode. Region type 2, crystals are those coupled in theperipheral area of the photocathode, region type 3 are crystalscoupled in the PMT wall and, region type 4, are crystals withoutPMT coupling. Fig. 4 shows the PQS-block crystal-to-PMTcoupling regions distribution, for two configurations, animaland human PQS-PET, respectively.

We also investigated the potential PET imaging performancefrom the decoded block geometry using GATE (Geant4 Appli-cation for Tomographic Emission) simulation package. GATEis a GEANT4-based software toolkit for realistic simulation ofPET and SPECT systems for small-animal and human applica-tions, [13] and [14]. For the animal PET simulation, we adoptedthe same geometry as our BGO rodent-research PET (RRPET),[3], with 16.5-cm detector-ring diameter and 11.8-cm AFOV,except we substituted 14 14 detectors. Both point


Fig. 4. Crystal-to-PMT coupling regions for animal (a) and human (b) detectorblocks.

Fig. 5. Crystal decoding map of animal (a) and human (b) PQS-detector blocks.

sources and the Micro Deluxe hot rod phantom were simu-lated. The positron range and the gamma acollinearity at theannihilation points were taken into accounts. The micro deluxephantom is a 4.5-cm diameter cylinder with rods of 1.2, 1.6, 2.4,3.2, 4.0, and 4.8 mm, and the center-to-center distance betweenrods was twice the rod diameters. The methodology for thisstudy is described in [15]. For the human PET simulation, weconfigured a PET system with a ring diameter of 86 cm with54 block detectors (15 15 array) or 810 crystals perdetector ring. Both point sources and a resolution Derenzo-likephantom of 128.0 mm in diameter and hot rod diameters of2.0, 2.4, 3.2, 4.0, 4.8, 6.4 mm were simulated using small binsizes (0.44 , 0.5 mm). For spatial resolution simulation, anideal point source ( positron embedded in water spherein the GATE code) was used. For the image reconstructionof the simulated data, filtered-back projection was used. Allsimulation results were based on the assumption where theenergy resolution was fixed for all crystals animal PETand human PET, at 511 keV. The energy resolution was15.0%, and the energy window 300–750 keV.

III. RESULTS

A. Characterization of the PQS Detectors

The decoding maps from all crystals were well identified inboth, 14 14 and 15 15 detector arrays, as shown in Fig. 5(a)

Fig. 6. Performance of crystal energy from 14 � 14 �� animal PETdetectors. (a) Typical crystal energy spectrum. (b) Profile of external crystalcolumn. (c) Profile of the central crystal column in decoding map.

and (b). To determine the crystal spectra, relative light output,energy resolution at several gamma-ray energies, and light col-lection efficiency of PQS-detector blocks, we used astandard nuclear spectroscopy technique.

The energy spectra from PQS-blocks are shown in Fig. 6(a),for animal and 7(a), for human PET; there we show a com-parison between corner crystals, center crystals, and normal-ized block average energy spectrum. Examples of crystal pro-files are shown in Fig. 6(b) and Fig. 7(b) for external crystalcolumn (corner to corner) and Fig. 6(c) and 7(c) for centralcrystal column.

The corner crystals have higher light-collection efficiencythan all others because of they are coupled in the center of thephotocathode and the central crystals in the block are sittingwithout any PMT coupling. Photo peaks of the spectra are welldefined with high peak-to-valley ratios.


Fig. 7. Performance of crystal energy from 15 � 15 �� human PET detec-tors. (a) Typical crystal energy spectrum. (b) Profile of external crystal column.(c) Profile of the central crystal column in decoding map.

The measured energy resolutions of 14 14 animalPET detector block, are ranged from 12.9% to 15.8% after sym-metric restoration with an average of 14.1%. For 15 15human PET detectors, the energy resolutions varied from 13.5%to 17.3% after symmetric restoration with an average of 15.6%.

The crystal energy resolutions symmetrically restored fromboth blocks are plotted in Fig. 8 (animal detectors) and Fig. 9(human detectors). Fig. 8(a) and Fig. 9(a) show the results fromthe individual crystals and Fig. 8(b) and Fig. 9(b) show the re-sults of the four crystal-to-PMT coupling regions, where gooduniformities were observed for both the animal and human de-tector blocks.

The distribution of light-collection efficiencies is shown inFig. 10 (animal detectors) and Fig. 11 (human detectors), wherethe individual values were expressed as a percentage of thehighest value obtained for each detector block. Fig. 10(a) and

Fig. 8. Energy resolution distribution of 14� 14 �� animal PET detec-tors. (a) Individual crystals. (b) Four coupling regions.

Fig. 9. Energy resolution distribution of 15 � 15 �� human PET detectors.(a) Individual crystals. (b) Four coupling regions.

Fig. 11(a) show the symmetry-restored crystal signal pulse-height distribution of individual crystals, and Fig. 10(b) and


Fig. 10. The distribution of light-collection efficiencies of animal PQS-detec-tors. (a) Pulse height of individual crystals. (b) Pulse height of the four couplingregions.

Fig. 11(b) show the crystal pulse–height distribution in the fourregions of the PQS-detector blocks. The higher light loss of17.3% for central crystals in the animal PET detectors mightreflect the difficulty for light to go through the tiny crystal nee-dles and many interfaces before reaching the PMT.

B. Results From Monte Carlo Simulations

For the animal PET, the simulation data suggests thatthe new scanner could have a reconstructed radial and(tangential) spatial resolution of 1.24 (1.25) mm, 1.29 (1.33)mm, 1.78 (1.41) mm, 2.28 (1.68) mm, and 3.09 (1.75) mm at0.5 cm, 1 cm, 2.5 cm, 3.5 cm, and 5 cm off center, respectively.The first number is radial resolution, while the number in paren-theses is for tangential resolution. Fig. 12 shows (a) simulatedradial and tangential resolution results and (b) the reconstructedimage of the simulated data for the micro deluxe phantom withhot rod inserts for the new animal camera. The 1.2-mmhot rods are clearly distinguishable. Certainly, additional spa-tial blurring of the hit crystal, due to block decoding effects andmechanical misalignment of the detector blocks, which are dif-ficult to simulate and which change with production techniquesand production errors, are likely, making these simulation re-sults too optimistic.

For the human PET, the simulation data suggests that the newhuman LSO scanner could have a reconstructed radial (tangen-tial) spatial resolution of 3.09 (3.24) mm, 3.08 (3.14) mm, 3.51(3.19) mm, 4.09 (3.48) mm, 4.79 (3.41) mm and, 5.68 (3.85) mm

Fig. 11. (a) Pulse-height distribution of human PET detectors. (b) Pulse-heightdistribution of the four coupling regions for human PET detectors block.

at 1 cm, 5 cm, 10 cm, 15 cm, 20 cm and 25 cm off center, respec-tively. Fig. 13(a) shows the simulated radial and tangential reso-lution results and (b) the reconstructed image of the phantom forthe human system. The resolution sectors down to 2.4 mmwere resolved. The simulation shows that a central spatial res-olution of 3 mm is potentially achievable using the human PETdetectors developed in this study.

IV. DISCUSSION

Since our PQS block detector also plays the role of aninternal light guide for distributing scintillation light to the fourPMTs, this design increases light output, thereby improvingdecoding resolution, as the scintillation light from each crystalis dispersed (sideways through the cylinder surfaces) intomany other crystals to conduct the light to the decoding PMT,instead of transporting all the light along its own narrow anddeep geometry first and then going though another externallight guide system before reaching the PMT. Both light col-lection and crystal decoding depend on the PMT photocathodecharacteristics and the internal light-guide properties suchas refractive index of crystals and optical grease, number ofinterfaces, surface finish of the interfaces, and the reflectorpatterns. For this study and with the application of the PQSanalog decoding technology for lutetium-based crystals, wedeveloped and tested high resolution, position-sensitive blockdetectors. The developed lutetium PQS-detector blocks foranimal/human PET applications could decode 196/225 crystalsper PMT using regular round 19/51 mm PMTs, respectively.


Fig. 12. (a) Projected spatial radial and tangential resolution curves from Monte Carlo simulation of the animal detector block. (b) Reconstructed image of thesimulated data for the micro deluxe phantom for �� animal PET cameras.

Fig. 13. (a) Projected radial and tangential resolution curves from Monte Carlo simulation of the human detector block. (b) Reconstructed image of the simulateddata for the resolution phantom for �� human PET cameras.

For animal PET detectors, we decoded a 14 14 array with acrystal size of mm . This animal detector block hada packing fraction of 95.6% and an energy resolution rangingbetween 12.9% 15.8% for individual crystals (average energyresolution of 14.1%); the pulse height for the least favorablecrystal was 81.7% of the most favorable crystal.

For human PET detectors, we decoded a 15 15 array witha crystal size of . The human detector-blockhad a packing fraction of 98.2%, and an energy resolution rangeof 13.5% 17.3% (average energy resolution 15.6%). The pulseheight of the least favorable crystal was 83% of the most fa-vorable crystal, thereby achieving a nearly uniform light outputthroughout the block. The results of this study are summarizedin Table I.

From Monte Carlo simulations for small animal PET,a central (near the center of field of view) spatial resolution of1.1–1.2 mm may potentially be achieved using low cost 19-mmcircular PMT. For human PET systems, 3-mm central spatialresolution may potentially be achieved using a large 51-mm cir-cular PMT.

With the PQS- detector blocks developed in thisstudy, we may be able to realize a lower cost PET systems thathas higher resolution than existing cameras without requiringcostly position sensitive photomultiplier tubes (PSPMTs) or

avalanche photodiodes (APDs), [16]–[22]. The comparison ofdetector specifications between commercial PET systems andthe PQS- system development are shown in Table II. Thecost savings may be used so as to extend the AFOV sufficiently,that the whole mouse/rat body can be imaged in one fixed bedposition. This would in turn allow for dynamic pharmaceu-tical studies while tapping the dynamic arterial input functiondirectly from the cardiac images being acquired concurrently.An extended AFOV also increases the detection sensitivity,especially for the two ends of a mouse/rat (brain and hindquarters) where tumors were often implanted.

There may be some potential drawbacks such as: 1) Increasein costs of manufacturing process, due to surface-finish require-ment. However, all animal PET systems already use crystalsthat are chemically polished (etched) or mechanically polishedto maximize light output, which is even more smooth than the4–5- m finish used here. To achieve the high mechanical pre-cision required in animal PET systems with 1–2-mm detectorsizes, the crystals are not cut to size but lapped to the exact sizeanyway; therefore, the lapping process used in this work is al-ready an integral process in the manufacturing of other animalsystems, although the grit sizes of the palling particles are dif-ferent. Using the slab–sandwich–slice production method [23],that we developed, only large slabs of crystals, 10–20 detectors


TABLE IPERFORMANCE OF �� PQS BLOCK DETECTORS

TABLE IICOMPARISON OF DETECTOR SPECIFICATIONS BETWEEN COMMERCIAL

PET SYSTEMS AND THIS PQS-�� DETECTOR DEVELOPMENT

: Parameter not available or not applicable for the system.

in width, and not individual pixels are handled or processed inmaking the detector array, which lowers the manufacturing cost.2) The increase in dead time and reduction in the dynamic rangeof PET scanner due to the sharing of PMT channels in PQS. This

drawback can be partly mitigated by the fast crystal usedhere and partly by the pileup–event–recovery electronics thathas been shown to increase detection rates by 10 times com-pared to traditional pulse-processing electronics [10]–[12].

For a hypothetical animal PET geometry with a long 12 cm(20 cm) AFOV and a 16–17 cm detector-ring diameter, only 168(280) PMTs will be used with an estimated total PMT cost ofonly $20,000 ($33,000) for such large systems. Existing humanclinical PET systems achieved a lower 4–6-mm detector-pitchresolution using much smaller 19 mm PMT than the 51 mmPMT used in this study (see Table II).

Hence, the human PET detector developed in this study mayprovide higher image resolution than existing clinical PETwhile reducing the total cost of PMT by or seventimes. A typical human PET system with 15–16-cm AFOVwould need only 220 PMT instead of the typical 1200 PMT,thereby, reducing the total PMT cost per camera from $160,000to $28,600, plus the additional savings in the correspondingsupporting electronics for each PMT channel as amplifiers,analog-to-digital converters (ADCs), high voltage dividers (HVdividers), assembly labor, etc.

V. CONCLUSION

With PMT-quadrant-sharing design along and using the latestimproved block, we have achieved the decoding ofvery large arrays, a 14 14 for animal PET system and a 15

15 for human PET system.Despite the disadvantages that the detector sizes are very

narrow and that a large number of the middle crystals are sittingin space without coupling to any photocathode, the crystal lightoutputs, the energy resolution, and the pulse-height distributionsare uniform. For small animal PET application, we achieved ahigh crystal-decoding resolution of mm usingthe low-cost generic 19-mm circular PMT with an excellentaverage energy resolution of 14.1% and a high crystal-to-PMTratio of 196 crystals per PMT. For human PET application, weachieved a high decoding resolution of mmcrystals using a large 51-mm circular PMT with an averageenergy resolution of 15.6% and a crystal-to-PMT ratio of 225crystals per PMT.

With the proposed application of PMT-quadrant-sharing tocrystals, PET systems with higher spatial resolution

together with excellent energy resolution can be achieved ata lower production cost and without using expensive readoutdevices.

REFERENCES

[1] W.-H. Wong et al., “Design and preliminary performance of a low-costhigh-sensitivity rodent research PET and PET/CT,” in Proc. AnimalMolecular Imaging International Symp., Taipei, Taiwan, pp. 75–78, El-sevier Health Sciences Excerpta Medica, Chapter, (11/2005).

[2] J. Uribe et al., “Signal characteristics of individual crystals in highresolution BGO detector designs using PMT-quadrant sharing,” IEEETrans. Nucl. Sci., vol. 50, no. 3, pt. 2, pp. 355–361, Jun. 2003.

[3] S. Xie et al., “A pentagon photomultiplier-quadrant-sharing BGO de-tector for a rodent research PET (RRPET),” IEEE Trans. Nucl. Sci.,vol. 52, no. 1, pp. 210–216, Feb. 2005.

[4] S. Kim et al., “High resolution GSO block detectors using PMT-quad-rant-sharing design for small animal PET,” IEEE Trans. Nucl. Sci., vol.53, no. 1, pp. 40–43, Feb. 2006.


[5] C. L. Melcher et al., “Scintillation properties of LSO: Ce boules,” IEEETrans. Nucl. Sci., vol. 47, no. 3, pt. 2, pp. 965–968, Jun. 2000.

[6] Innovation, ESR Film Vikuiti, Enhanced Specular Reflector (ESR),ESR. PDF.

[7] Allied High Tech Products, Inc., The Multiprep System, UnequalledResults in All of These Sample Preparation Application [Online].Available: www.alliedhightech.com

[8] H. Li et al., “A simulation study on optically decoding reflecting win-dows for PMT quadrant sharing scintillation detector block,” IEEETrans. Nucl. Sci., vol. 53, no. 5, pp. 2557–2562, Oct. 2006.

[9] Wong et al., “The design of a high resolution transformable wholebodyPET camera,” IEEE Trans. Nucl. Sci., vol. 49, no. 5, pp. 2079–2084,Oct. 2002.

[10] W.-H. Wong, H. Li, J. Uribe, H. Baghaei, Y. Wang, and S. Yokoyama,“Feasibility of a high speed gamma camera design using the high-yield-pileup-event-recovery (HYPER) method,” J. Nucl. Med., vol. 42, no. 4,pp. 624–632, Apr. 2001.

[11] H. Li et al., “A high speed position-decoding electronics for BGO blockdetectors in PET,” IEEE Trans. Nucl. Sci., vol. 47, no. 5, pt. 3, pp.1006–1010, Oct. 2002.

[12] H. Li et al., “A new pileup-prevention front-end electronic design forhigh resolution PET and gamma cameras,” IEEE Trans. Nucl. Sci., vol.49, no. 5, pp. 2051–2056, Oct. 2002.

[13] S. Jan et al., “GATE: A simulation toolkit for PET and SPECT,” Phys.Med. Biol., vol. 49, pp. 4543–4561, 2004, Institute of Physics Pub-lishing.

[14] National Electronical Manufactures Association, NEMA StandardsPublication NU 2-2001, Performance Measurements of PositronEmission Tomographs 2001.

[15] H. Baghaei et al., “GATE Monte Carlo simulation of the performanceof a high-sensitivity and high-resolution LSO based small animal PETcamera,” in 2006 IEEE Nuclear Science Symp. Conf. Rec., Oct. 2006,vol. 4, pp. 2476–2479, accepted for publication in IEEE Trans. Nucl.Sci..

[16] T. G. Turkington et al., “Performance of a BGO PET/CT with higherresolution PET detectors,” in IEEE, Nuclear Science Symp. Conf. Rec.,2005, paper M04-2.

[17] W. Worstell et al., “Design and performance of a prototype whole-bodyPET/CT scanner with fiber optic readout,” in IEEE, Nuclear ScienceSymp. Conf. Rec., 2004, vol. 5, pp. 3280–3284.

[18] Imaging Technology News, PET/CT Comparison Charts [Online].Available: http://www.itnonline.net/

[19] R. Gregory, M. Partridge, and M. A. Flower, “Performance evaluationof the Philips ‘Gemini’ PET/CT system,” in 2004 IEEE Nuclear Sci-ence Symp. Conf. Rec., vol. 4, pp. 2605–2609.

[20] B. J. Kemp et al., “Performance measurements of the siemens inveonsmall animal PET scanner,” presented at the 2006 IEEE MedicalImaging Conf., paper M14-69.

[21] R. Fontaine et al., “The architecture of LabPET/spl trade/, a small an-imal APD-based digital PET scanner,” in 2005 IEEE, Nuclear ScienceSymp. Conf. Rec., vol. 5, pp. 2785–2789, paper J02-3.

[22] Yoshida et al., “The jPET-D4: Detector calibration and acquisitionsystem of the 4-layer DOI-PET scanner,” in 2005 IEEE Nuclear Sci-ence Symp. Conf. Rec., paper M11-270.

[23] J. Fontaine et al., “An efficient detector production method for posi-tion-sensitive scintillation detector arrays with 98% detector packingfraction,” IEEE Trans. Nucl. Sci. , vol. 50, no. 5, pp. 1469–1476, Oct.2003.

High-Resolution L(Y)SO Detectors Using PMT-Quadrant-Sharing for Human and Animal PET Cameras

Documents

Transcript of High-Resolution L(Y)SO Detectors Using PMT-Quadrant-Sharing for Human and Animal PET Cameras