Application of transmission scan-based attenuation compensation to scatter-corrected thallium-201...

Original article

Application of transmission scan-based attenuationcompensation to scatter-corrected thallium-201 myocardialsingle-photon emission tomographic imagesJun Hashimoto1, Koichi Ogawa2, Atsushi Kubo1, Takashi Ichihara3, Nobutoku Motomura3, Takuzo Takayama3,Shiro Iwanaga4, Hideo Mitamura4, Satoshi Ogawa4

1 Department of Radiology, School of Medicine, Keio University, Tokyo, Japan2 Department of Electronic Informatics, College of Engineering, Hosei University, Tokyo, Japan3 Toshiba Medical Engineering Laboratory, Tochigi, Japan4 Department of Internal Medicine, School of Medicine, Keio University, Tokyo, Japan

&misc:Received 1 September and in revised form 25 October 1997

&p.1:Abstract. A practical method for scatter and attenuationcompensation was employed in thallium-201 myocar-dial single-photon emission tomography (SPET orECT) with the triple-energy-window (TEW) techniqueand an iterative attenuation correction method by usinga measured attenuation map. The map was reconstruct-ed from technetium-99m transmission CT (TCT) data.A dual-headed SPET gamma camera system equippedwith parallel-hole collimators was used for ECT/TCTdata acquisition and a new type of external sourcenamed “sheet line source” was designed for TCT dataacquisition. This sheet line source was composed of anarrow long fluoroplastic tube embedded in a rectangu-lar acrylic board. After injection of 99mTc solution intothe tube by an automatic injector, the board was at-tached in front of the collimator surface of one of thetwo detectors. After acquiring emission and transmis-sion data separately or simultaneously, we eliminatedscattered photons in the transmission and emission datawith the TEW method, and reconstructed both images.Then, the effect of attenuation in the scatter-correctedECT images was compensated with Chang’s iterativemethod by using measured attenuation maps. Our meth-od was validated by several phantom studies and clini-cal cardiac studies. The method offered improved ho-mogeneity in distribution of myocardial activity and ac-curate measurements of myocardial tracer uptake. Weconclude that the above correction method is feasiblebecause a new type of 99mTc external source may notproduce truncation in TCT images and is cost-effectiveand easy to prepare in clinical situations.

&kwd:Key words: Thallium-201 – Myocardial single-photonemission tomography – Sheet line source – Attenuationcorrection – Triple-energy-window scatter correction.

Eur J Nucl Med (1998) 25:120–127

Introduction

For the reconstruction of quantitative single-photonemission tomography (SPET or ECT) images, it is im-portant to apply proper attenuation correction to thescatter-corrected SPET images that consists of only pri-mary photons. With regard to scatter correction, wehave previously reported the triple-energy-window(TEW) method, which corrects scattered photons in theprojection data pixel by pixel [1–3]. For attenuation cor-rection, gamma-ray transmission CT (TCT) has beenused by many investigators to obtain maps of attenua-tion coefficients [4–15]. Very recently, several gammacamera manufacturers have adopted attenuation correc-tion algorithms based on measured attenuation maps ob-tained by gamma-ray transmission scans. However, notso many investigators have published reports in whichscatter and attenuation corrections are considered to-gether [16–21]. We have proposed a scatter and attenua-tion correction method which is practically applicableusing commercially available SPET instrumentation.For brain imaging in a triple-headed gamma camerasystem, we carried out scatter correction with the TEWmethod and attenuation correction using attenuationmaps measured by a line source with a fan-beam colli-mator [19, 20]. This geometry, however, could not beapplied to myocardial imaging because of the insuffi-cient field of view. In this article, we present a practicalmethod for scatter and attenuation compensation inmyocardial SPET by using a new type of flood source

European Journal of Nuclear MedicineVol. 25, No. 2, February 1998 – © Springer-Verlag 1998

Correspondence to:A. Kubo, Department of Radiology, School ofMedicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo,160 Japan&/fn-block:

121

European Journal of Nuclear Medicine Vol. 25, No. 2, February 1998

and a dual-headed SPET system equipped with parallel-hole collimators.

Materials and methods

Instruments. &p.2:All images in this study were acquired with a dual-headed digital gamma camera (GCA-7200A/DI, Toshiba Corpora-tion, Tokyo) and processed by a medical image processor (GMS-5500A/DI, Toshiba Corporation, Tokyo). For the transmissionmeasurement, a new type of external source named “sheet linesource” was designed (Fig. 1). To prepare the source, 1 GBq oftechnetium-99m was injected to a narrow fluoroplastic tube (inter-nal diameter: 1 mm; length, 50 m) embedded in an acrylic board(rectangular shape: 68 cm×54 cm; thickness, 7 mm). To reducethe absorbed dose for the technologist, injection was performedby an automatic injector. It took about 10 min to complete the in-jection. The board was attached in front of the collimator surfaceof one of the two detectors. The field of view for ECT and TCTwas 55 cm in diameter, which was sufficiently large to acquireprojection data without truncation.

Imaging protocol. &p.2:We examined two different protocols for dataacquisition as described below (Fig. 2). Emission and transmis-sion CT scans were performed simultaneously in the simultaneousmode. First, thallium-201 (111 MBq) was injected to the patientand then a simultaneous TCT/ECT scan was carried out by 360°

continuous rotation with a total scan time of 20 min. In this scan,six energy windows were used for data acquisition because theTEW method requires three windows for the 201Tl photopeak andthe other three for the 99mTc photopeak as indicated in Fig. 3. A64×64 matrix was used for data acquisition.

Emission and transmission CT data were acquired separatelyin the sequential mode. Prior to tracer injection, a TCT scan with360° continuous rotation was carried out over a total scan time of20 min. After the TCT scan, the sheet line source was removedand the patient was injected with 201Tl. The patient was instructedto remain still during TCT and ECT data acquisitions.

Image processing. &p.2:A Butterworth filter (order: 8) was used asthe low-pass filter, and the cut-off frequency was set to 0.28 (cy-cles/pixel) for the projection data acquired by the main windowand to 0.14 for those acquired by the subwindows. The TEWmethod estimates the scattered photons in the main windowbased on the counts in the two subwindows by trapezoidal ap-proximation and rejects them by subtraction [1–3]. With themain window, we acquire the total count (Ctotal), which is com-posed of the count of primary photons (Cp) and that of scatteredphotons (Fig. 3). Cp is estimated by the following equation(Eq. 1):

Cp = Ctotal–(Clow/Ws+ Cup/Ws)×Wm/2, (1)

where Clow and Cup denote the counts in the lower and uppersubwindows, respectively, and Wm and Ws are the widths of the

b

a

Fig. 1. a Geometric configurations of detectors, collimators andtransmission sources. b Sheet line source attached to the detector.A long fluoroplastic tube is embedded in an acrylic board. 99mTcis injected into the tube using an automatic injector&/fig.c:

Fig. 2a, b. Two imaging protocols. a Simultaneous mode. b Se-quential mode. T, transmission raw data; E, emission raw data;N.U., not used; TEW, triple-energy-window scatter correction;ATT, attenuation correction for the acrylic board. Details are de-scribed in the text&/fig.c:

122


main window and the subwindows, respectively. The TEW meth-od was applied to both the TCT and the ECT raw data.

After TEW scatter correction, two sets of corrected projectiondata were obtained because ECT was performed with 360° rota-tion of each detector. In the sequential scan mode, whole projec-tion data were obtained by adding the two sets at each samplingangle. In the simultaneous mode, one of the two sets consisted ofprojections attenuated by the acrylic board. Consequently, to ob-tain whole ECT data, attenuation correction for the acrylic board(7 mm in thickness) was done prior to the data addition. After theaddition, ECT data were reconstructed into a 64×64 matrix imagewith a ramp filter.

Scatter-corrected transmission data were converted to TCTprojection data and they were reconstructed into a 64×64 pixel at-tenuation map using a ramp filter. Then the attenuation map for99mTc was converted to that for 201Tl by multiplying by a factor of1.25 [21]. The conversion factor was obtained by dividing the at-tenuation coefficient of water for 201Tl by that for 99mTc.

For attenuation correction we used a modified correction matrix(MCM) method, which is a modified version of Chang’s iterativemethod [4, 22]. The method uses a measured attenuation coefficientmap, which is the two-dimensional distribution of linear attenuationcoefficients obtained by gamma ray TCT mentioned above. Attenu-ation correction is performed by multiplying the scatter-correctedECT image by a correction matrix, i.e. a matrix containing the cor-rection factor for each pixel. Next, the attenuated projection datawere calculated from this corrected image. In this process, we didnot consider collimator aperture. Then, the attenuated projection da-ta were compared with the original projection data and their dis-crepancies formed error projection data which were reconstructedto an error image. This error image was multiplied by the correctionmatrix and added to the first corrected image. In the next iteration,the attenuated projection data were calculated and correction wasperformed with the above procedure. Three iterations were carriedout to avoid the amplification of the statistical noise.

Uniformity of the transmission source. &p.2:To evaluate the uniformityof the transmission source, TCT raw data were obtained withoutany object between the two detectors. The distance between thetwo detectors was set to 45 cm to simulate the clinical situation.The first scan was performed by using a uniform flood source andthe second one was done by the sheet line source.

Phantom studies. &p.2:All of the following phantom studies were con-ducted by a chest phantom (Kyoto Kagaku Co.) including chestwall, mediastinum, vertebrae, lung chambers, and a cardiac insert.Lung chambers were filled with a mixture of Styrofoam and wa-ter. Transmission images of the phantom are shown in Fig. 4.

Absorbed dose was measured by thermoluminescent dosimetry(TLD) during a 20-min transmission scan performed with thesheet line source containing 1 GBq of 99mTc. Six dosimeters(BeO, 1 mm×10 mm, cylindrical shape) were laid out on the sur-face of the chest phantom (anterior and posterior median, RAOand LAO 60°, RPO and LPO 60°) and measured twice. We cali-brated the dosimeters in a Lucite phantom using 37 GBq of cesi-um-137 as a calibration source.

The cardiac insert was filled with the 201Tl solution(15.6 MBq/whole heart) and the sheet line source was filled withthe 99mTc solution (1 GBq). A 15-mm-diameter cold defect wasattached to the septum of the cardiac insert to evaluate the correc-tion accuracy. Emission and transmission CT data were acquiredin the sequential or simultaneous mode.

Next, the cardiac insert was removed, and a rod (cylindrical,20 mm ∅ × 50 mm) was placed in the cardiac position. The rodwas filled with 11.8 MBq of 201Tl solution. Counts in the rod wereobtained by setting regions of interest (ROIs) in the reconstructed

Fig. 3. Triple-energy-window scatter correc-tion. Cp, count of primary photons; Cs, countof scattered photons; Cup, count in the uppersubwindow; Clow, count in the lower subwin-dow; sub, subwindow; Wm, width of the mainwindow; Ws, width of the subwindow. Scatterof photons in the main window is estimatedby trapezoidal approximation (shaded areas) &/fig.c:

a b

Fig. 4a, b.Transmission images of the chest phantom&/fig.c:

123


images. A cross-calibration scan was performed to quantify theactual activity as follows [20]. 201Tl (0.3 ml) was placed in a sy-ringe approximately 5.0 mm in diameter and suspended in the airto minimize the effect of photon scattering and attenuation. Afterdata acquisition and image reconstruction, a thick single-sliceSPET image containing the entire activity was obtained. A largeROI was then set to cover the entire source area. The cross-cali-bration factor was calculated using the counts in the ROI and theactivity measured by a scintillation dose counter (Aloka IGC-3,Tokyo, Japan).

We carried out another experiment using the chest phantom toevaluate the noise in corrected and uncorrected images. The cardi-ac insert in the chest phantom was filled with 28 kBq/ml of 201Tlsolution, and the total acquisition time per scan ranged from 2 to30 min. A 3.2-mm-thick short axial SPET image was reconstruct-ed after data acquisition by the simultaneous mode. Square ROIs(9.6 mm×9.6 mm) were set in the myocardial areas. We measuredmean counts and standard deviations of SPET values and calculat-ed the coefficient of variation (SD/mean), which reflects the sta-tistical noise in the image.

Clinical evaluation. &p.2:A normal volunteer (44-year-old man) and apatient (55-year-old man) with posterolateral myocardial infarc-tion took part in the clinical study. Informed consent was obtainedfrom each subject after detailed explanation of the study. The sub-jects were injected with 111 MBq of 201Tl and data were acquiredby a simultaneous TCT/ECT scan and an ECT scan without TCTmeasurement to obtain corrected ECT data and pure ECT datawithout correction.

Results

Count profile curves obtained from the TCT raw datawithout any object are shown in Fig. 5. The line-shapedpattern (5 mm/line) due to the sheet line source did notappear in the image.

From TLD at 12 measurement points, the mean ab-sorbed dose was 0.12 mSv (SD: 0.03). The maximumand minimum values were 0.17 (anterior median) and0.08 (posterior median) mSv, respectively.

As shown in Fig. 6, scatter and attenuation correctionimproved homogeneity in the distribution of myocardial

activity. The mean ratios of posterior-to-anterior myo-cardial counts without and with corrections were 0.86and 0.98, respectively. Septal-to-lateral count ratios were0.84 and 1.03, respectively.

Cross-calibrated SPET values became stable whenmore than two iterations were carried out (Table 1). Inquantification, there were no appreciable differences be-tween the reconstructed values obtained in the sequentialmode and those obtained in the simultaneous mode. Al-though corrections increased the statistical noise, the CVvalues were stable when data were acquired for morethan 10 min (Fig. 7).

SPET images and count profiles obtained from thenormal volunteer are shown in Figs. 8 and 9. Measuredattenuation coefficients for the mediastinum and lungsfor 201Tl were 0.177 (cm–1) (SD: 0.009) and 0.048 (SD:0.007), respectively. The septal-to-lateral count ratio im-proved from 0.80 for the original image without any cor-rection to 0.97 for the image with scatter and attenuationcorrection. The value was 0.88 in the attenuation-cor-rected image without scatter correction. The posterior-to-anterior count ratios were 0.93, 1.05 and 1.27 for theimages without correction, with scatter and attenuationcorrection, and with attenuation correction only, respec-tively.

Figure 10 shows SPET images of the patient withmyocardial infarction. The infarcted area in the postero-

a b

Fig. 5a, b.Count profilecurves obtained fromTCT data without anyobject. a Obtained by auniform flood source. b Obtained by the sheetline source. Verticalsection (upper right)and horizontal section(lower left). Homogene-ous count distribution isobserved in spite of theline-shaped pattern ofthe sheet line source&/fig.c:

Table 1 Cross-calibrated SPET values and the number of itera-tions&/tbl.c:&tbl.b:

Iterations Sequential mode Simultaneous mode(MBq/rod) (MBq/rod)

0 9.5 9.61 11.3 11.42 11.7 12.03 11.7 12.0

True value 11.8 11.8

&/tbl.b:

124


lateral wall was visualized more clearly after scatter andattenuation correction.

Discussion

Both attenuation correction and scatter correction areneeded to obtain a quantitative SPET image, and attenu-ation-corrected SPET images without scatter correctionmay result in misreading of myocardial images [23, 24].

The main advantage of the TEW scatter correction meth-od is that it considers the position dependency of photonscattering using a simple algorithm that is applicable inroutine clinical studies [25, 26]. However, counts in nar-row subwindows are sometimes too low in number andmay increase the noise in scatter-corrected images. Webroadened the width of the subwindows to reduce thestatistical instability [3]. The noise in the attenuationmap also affects the corrected ECT images. We carriedout some basic experiments in previous studies to evalu-

Fig. 6. Bull’s eye maps in the chest phantom study. A Withoutcorrection; B with scatter and attenuation correction&/fig.c:

Fig. 7. Acquisition time and image noise. The coefficient of varia-tion (SD/mean) reflects the noise in the image. SC, Scatter correc-tion; AC, attenuation correction&/fig.c:

Fig. 8. ECT and TCT images of a normalvolunteer. ECT images without correc-tion (upper left), with attenuation correc-tion only (lower left), and with scatterand attenuation correction (lower right),and TCT images (upper right). Spurious-ly increased uptake in the inferoposteriorwall was observed when attenuation cor-rection was performed without scattercorrection&/fig.c:

125


ate the noise in reconstructed brain SPET images [27].In the study, the scan time ranged from 1 to 30 min andreconstructed SPET values were obtained. The resultsshowed that the standard deviations of SPET values inthe ROIs, reflecting the noise, were acceptably stablewhen data were acquired for more than 9 min, which isconsistent with the results in the present study (Fig. 7).

In transmission CT, there are several geometries forthe transmission source and the collimator. The first typeis a combination of a cone-beam collimator and a pointsource [6, 7]. The second is a moving line source and a

parallel-hole collimator [12]. The third is a line sourceand a fan-beam collimator [8, 10, 11, 14, 15]. The fourthis a combination of a flood source and a parallel-holecollimator [4, 9, 13]. We employed the third method forthe correction in brain imaging [19–21]. This combina-tion provides some advantages including small fractionsof scattered photons in TCT and ECT data, and easyhandling of the transmission source. In addition, themethod uses lower activities for the transmission sourcethan those in the second method. However, it was notpreferable to use the method in myocardial imaging be-cause of the truncation artefacts. On the other hand, ad-vantages of the fourth combination are absence of trun-cation artefacts and lower cost of the instruments. Largeramounts of scattered photons, however, adversely affectthe images, and preparation of the transmission source isinconvenient in this method. To overcome these prob-lems, we designed a new type of external source and ap-plied the TEW method to both the TCT and the ECT rawdata. Although the source was in a narrow tube in aboard, it worked as a flood source at a given distance be-tween the detector and the source. Thus, the sheet linesource is almost the same as a uniform flood source, butoffers easy handling of the external source during prepa-ration and reduces absorbed dose to technologists andphysicians. Celler et. al. reported attenuation correctionusing a combination of parallel-hole collimation and ten

Fig. 9. Count profile curves obtained from ECT images withoutcorrection, with attenuation correction, and with scatter and atten-uation correction&/fig.c:

Fig. 10.ECT and TCT images of a patientwith myocardial infarction (MI). ECT im-ages without correction (upper left) andwith scatter and attenuation correction (low-er left), and TCT images (upper right). Theinfarcted area (arrow) is clearly visualizedafter scatter and attenuation correction&/fig.c:

126


line sources spaced by 3.5 cm [28]. Their method usedcollimators at the source side to reject scattered photonsin TCT data, while our method employed the TEWmethod to eliminate scattered photons in both the TCTand the ECT data. Scatter correction is indispensable inour method because the sheet line source engenders asignificant amount of scattered photons compared to theline sources. The TEW method combined with TCT-based attenuation correction provided quantitative SPETimages. Scan time is the same as that in the conventionaldata acquisition when the simultaneous mode is selected.Taking all of the above into consideration, this compen-sation method is a practical and cost-effective approachto obtain quantitative myocardial SPET images in rou-tine clinical examinations.

As shown in Figs. 8 and 9, spuriously increased up-take in the inferoposterior wall was observed when scat-ter correction was not performed but attenuation correc-tion was, suggesting the importance of scatter correc-tion. The results showed that the spillover of scatteredphotons into the inferoposterior myocardium from near-by abdominal activity reduces attenuation effects in con-ventional images, and that attenuation correction withoutscatter correction amplifies the spillover activity, result-ing in spuriously increased uptake in the inferoposteriorwall and relatively reduced accumulation in the anteriorwall. Both of the scatter and attenuation correctionsshould be implemented to achieve diagnostic perfor-mance with high specificity.

In the patient study (Fig. 10), an infarcted area waseasily recognized after scatter and attenuation correc-tion. Coronary angiography revealed 90% stenosis in asmall branch of the left circumflex artery (LCx) of thispatient. Attenuation correction visualized infarcted myo-cardium more clearly by increasing counts in the posteri-or septum which is perfused by intact right coronary ar-tery (RCA). It is reported that improved sensitivity indetecting perfusion abnormalities in RCA and LCx terri-tories was achieved by performing attenuation correction[29], which is compatible with our result. Scatter correc-tion increased the contrast between intact and infarctedmyocardium by eliminating the spillover activity in theinfarcted area from adjacent normal sites. It is possiblethat scatter correction also contributes to improvement indetecting coronary artery stenosis, especially when thestenotic vessel perfuses small areas.

The accuracy of quantification was improved by ap-plying the scatter and attenuation corrections. Quantita-tive myocardial SPET images reduce the number offalse-positive cases and improve the accuracy of viabili-ty assessment using lesion-to-normal count ratios [29,30]. Quantitative images will also contribute to accuratemeasurements of absolute values of myocardial bloodflow and metabolism.

In summary, we proposed a method for scatter and at-tenuation correction to obtain quantitative myocardialSPET images. The method uses a dual-headed SPETsystem equipped with conventional parallel-hole colli-

mators. The TEW method rejects scattered photons inemission and transmission data, and transmission scan-based attenuation correction is implemented by using anewly designed external source: a sheet line source. Thesource offers easy handling during preparation and re-duces absorbed dose to the technical support personnel.The method yields improved homogeneity in distributionof myocardial activity and accurate measurements ofmyocardial tracer uptake. This correction method ispractical and cost-effective in clinical situations.

&p.2:Acknowledgements.This work was supported in part by Grant-in-Aid for Scientific Research from the Ministry of Education, Sci-ence and Culture, Tokyo, Japan. We would like to express ourthanks to Mr. T. Sammiya, Mr. R. Iwasaki and Mr. K. Okabe, whoassisted us in carrying out our basic experiments.

References

1. Ogawa K, Harata Y, Ichihara T, et al. A practical method forposition-dependent Compton-scatter correction in single-pho-ton emission tomography. IEEE Trans Med Imag1991; 10:408–412.

2. Ichihara T, Ogawa K, Motomura N, et al. Compton scattercorrection using triple-energy window method for single- anddual-isotope SPECT. J Nucl Med1993; 34: 2216–2221.

3. Ogawa K, Ichihara T, Kubo A. Accurate scatter correction insingle-photon emission tomography. Ann Nucl Med Sci1994;7: 145–150.

4. Ogawa K, Takagi Y, Kubo A, et al. An attenuation correctionmethod of single-photon emission computed tomography us-ing gamma ray transmission CT. Kaku Igaku 1985; 22:477–490.

5. Bailey DL, Hutton BF, Walker PJ. Improved SPECT using si-multaneous emission and transmission tomography. J NuclMed1987; 28: 844–851.

6. Manglos SH, Bassano DA, Duxbury CE, Capone RB. Attenu-ation maps for SPECT determined using cone beam transmis-sion computed tomography. IEEE Trans Nucl Sci1990; 31:600–608.

7. Manglos SH, Capone RB, Bassano DA. Detection of nonuni-formity measurements and corrections for cone beam trans-mission CT on a gamma camera. Med Phys 1992; 19:491–500.

8. Frey EC, Tsui BMW, Perry JR. Simultaneous acquisition ofemission and transmission data for improved thallium-201cardiac SPECT imaging using a technetium-99m transmissionsource. J Nucl Med1992; 33: 2238–2245.

9. Cao Z, Tsui BMW. Performance characteristics of transmis-sion imaging using a uniform sheet source with parallel-holecollimation. Med Phys1992; 19: 1205–1212.

10.Tung CH, Gullberg GT, Zeng GL, Christian PE, Datz FL,Morgan HT. Nonuniform attenuation correction using simulta-neous transmission and emission converging tomography.IEEE Trans Nucl Sci1992; 39: 1134–1143.

11. Jaszczak RJ, Gilland DR, Hanson HW, et al. Fast transmissionCT for determining attenuation maps using a collimated linesource, rotatable air-copper-lead attenuators and fan-beam col-limation. J Nucl Med1993; 34: 1577–1586.

12. Tan P, Bailey DL, Meikle SR, et al. A scanning line source forsimultaneous emission and transmission measurements inSPECT. J Nucl Med1993; 34: 1752–1760.

22. Chang LT. A method for attenuation correction in radionuclidecomputed tomography. IEEE Trans Nucl Sci1978; 25:638–642.

23. Nakajima K, Matsudaira M, Yamada M, Taki J, Tonami N,Hisada K. Effect of scatter correction on quantification ofmyocardial SPECT and application to dual-energy acquisitionusing triple-energy window method. Kaku Igaku 1995; 32:959–967.

24. O’Connor MK, Caiati C, Christian TF, Gibbons RJ. Effects ofscatter correction on the measurement of infarct size fromSPECT cardiac phantom studies. J Nucl Med 1995; 36:2080–2086.

25. Ljungberg M, King MA, Hademenos GJ, Strand SE. Compari-son of four scatter correction methods using Monte Carlo sim-ulated source distributions. J Nucl Med1994; 35: 143–151.

26. Buvat I, Rodriguez VM, Todd PA, Benali H, Di Paola R.Comparative assessment of nine scatter correction methodsbased on spectral analysis using Monte Carlo simulations. JNucl Med1995; 36: 1476–1488.

27. Hasegawa H, Motomura N, Ichihara T, Ogawa K, HashimotoJ, Kubo A. Transmission CT image quality for quantitativebrain SPECT imaging. [Abstract]. J Nucl Med1996; 37: 212P.

28. Celler A, Sitek A, Harrop R. Reconstruction of multiple linesource attenuation maps. IEEE Trans Nucl Sci1997; 44:1503–1508.

29. Ficaro EP, Fessler JA, Shreve PD Kritzman JN, Rose PA, Cor-bett JR. Simultaneous transmission/emission myocardial per-fusion tomography. Diagnostic accuracy of attenuation-cor-rected Tc-99m-sestamibi single-photon emission computed to-mography. Circulation 1996; 93: 463–473.

30. Udelson JE, Coleman PS, Metherall J, et al. Predicting recov-ery of severe regional ventricular dysfunction. Comparison ofresting scintigraphy with Tl-201 and Tc-99m sestamibi. Cir-culation1994; 89: 2552–2561.

127


13. Murase K, Tanada S, Inoue T, Sugawara Y, Hamamoto K. Im-provement of brain single-photon emission tomography(SPET) using transmission data acquisition in a four-headSPET scanner. Eur J Nucl Med1993; 20: 32–38.

14. Kemp BJ, Prato FS, Nicholson RL, Reese L. Transmissioncomputed tomography imaging of the head with a SPECTsystem and a collimated line source. J Nucl Med1995; 36:328–335.

15. Chang W, Loncaric S, Huang G, Sanpitak P. Asymmetric fantransmission CT on SPECT systems. Phys Med Biol1995; 40:913–928.

16. Meikle SR, Hutton BF, Bailey DL. A transmission-dependentmethod for scatter correction in SPECT. J Nucl Med1994; 35:360–367.

17. Welch A, Gullberg GT, Christian PE, Datz FL, Morgan HT. Atransmission-map-based scatter correction technique for SPECTin inhomogeneous media. Med Phys1995; 22: 1627–1635.

18. Kaplan MS, Miyaoka RS, Kohlmyer SK, Haynor DR, Harri-son RL, Lewellen TK. Scatter and attenuation correction forIn-111 based on energy spectrum fitting. Med Phys1996; 23:1277–1285.

19. Ichihara T, Motomura N, Hasegawa H, et al. Simultaneousemission transmission scan for the measurement of 3-dimen-sional absolute activity distribution using dual isotope, TEWscatter compensation method and fan beam collimation. ConfRec IEEE Nucl Sci Symp1995; 1275–1279.

20. Hashimoto J, Kubo A, Ogawa K, Amano T, Fukuuchi Y, Mot-omura N, Ichihara T. Scatter and attenuation correction in tech-netium-99m brain SPECT. J Nucl Med1997; 38: 157–162.

21. Ichihara T, Motomura N, Ogawa K, et al. Evaluation of SPETquantification of simultaneous emission and transmission im-aging of the brain using a multidetector SPET system with theTEW scatter compensation method and fan-beam collimation.Eur J Nucl Med1996; 23: 1292–1299.

Application of transmission scan-based attenuation compensation to scatter-corrected thallium-201...

Documents

Transcript of Application of transmission scan-based attenuation compensation to scatter-corrected thallium-201...