Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print September 16, 2005, 10.1148/radiol.2372041387
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2372041387v1
237/2/662    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Utsunomiya, D.
Right arrow Articles by Yamashita, Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Utsunomiya, D.
Right arrow Articles by Yamashita, Y.
(Radiology 2005;237:662-669.)
© RSNA, 2005

Technical Developments

Object-specific Attenuation Correction at SPECT/CT in Thorax: Optimization of Respiratory Protocol for Image Registration1

Daisuke Utsunomiya, MD, Takeshi Nakaura, MD, Tsuyoshi Honda, MD, Shinya Shiraishi, MD, Seiji Tomiguchi, MD, Koichi Kawanaka, MD, Shoji Morishita, MD, Kazuo Awai, MD, Hisao Ogawa, MD and Yasuyuki Yamashita, MD

1 From the Departments of Diagnostic Radiology (D.U., T.N., S.S., S.T., K.K., S.M., K.A., Y.Y.) and Cardiovascular Medicine (T.H., H.O.), Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan. From the 2003 RSNA Annual Meeting. Received August 10, 2004; revision requested October 14; revision received January 18, 2005; accepted February 16. Address correspondence to D.U., Diagnostic Imaging Center, Saiseikai Kumamoto Hospital, 5-3-1 Chikami, Kumamoto-shi, Kumamoto 861-4193, Japan (e-mail: d-utsunomiya{at}skh.saiseikai.or.jp).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Appendix
 References
 
Institutional review board approval was obtained for multiple imaging examinations in healthy volunteers and patients and for the analysis of images. The purpose of the study, and the risks associated with radiation exposure with regard to stochastic effects that might result in cancer and/or genetic mutations, were explained to all subjects, and all questions from subjects were answered. Each subject provided written informed consent. The purpose of the study was to prospectively determine the respiratory protocol at computed tomography (CT) that results in the best registration of CT images with images acquired at single photon emission computed tomography (SPECT) in the thorax. Errors of registration between myocardial SPECT images and CT images obtained with different respiratory protocols (postinhalation breath hold, postexhalation breath hold, and free breathing) in 13 healthy subjects were compared. CT scans obtained with free breathing and postexhalation breath hold better matched SPECT images than did those obtained with postinhalation breath hold (one-way analysis of variance, P < .01). Fewer SPECT/CT images showed artifacts with registration performed by using internal landmarks (four, two, and one of 13 images with postinhalation breath-hold, postexhalation breath-hold, and free-breathing protocols, respectively) than with registration performed by using external markers (nine, four, and two of 13 images). CT data acquisition with a free-breathing or postexhalation breath-hold protocol and image registration by using internal landmarks are recommended for attenuation correction.

© RSNA, 2005


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and Methods
 Results
 Discussion
 Appendix
 References
 
Cardiac single photon emission computed tomography (SPECT) is widely used as a noninvasive procedure for the examination of myocardial perfusion. With use of SPECT, however, photon attenuation correction is important because of the nonuniform distribution of attenuation coefficients (1–6). An attenuation correction method that involves the use of an attenuation map derived from images obtained with x-ray–based computed tomography (CT) was proposed to counter this problem (7–9). For reliable attenuation correction, however, anatomically accurate registration between SPECT images and x-ray–based CT images is vital. The use of a hybrid gamma camera–CT scanner system, which enables consecutive imaging sessions with the two modalities in a single examination, has been reported (7,10). However, there are some drawbacks to such hybrid systems. One important disadvantage is the poor performance of the x-ray–based CT scanner. The lower-spatial-resolution CT images (obtained with a matrix of 128 x 128) may be sufficient for attenuation correction but often are insufficient for image interpretation in clinical circumstances. To ensure both the robust registration of SPECT and CT images and the high resolution of CT images (with a matrix of 512 x 512), we combined our SPECT and multi–detector row CT scanners. By using this system, we can acquire images sequentially with SPECT and with CT, without the need to transfer the patient from one table to another. Both attenuation correction and accurate localization of myocardial perfusion are possible with this dual-modality scanner, with results similar to those obtained with a previously described integrated positron emission tomography (PET)/CT scanner (10,11). There is still the possibility that registration errors may arise from physiologic motion due to breathing (12–15), but the influence of respiratory movement on attenuation correction with use of a combined scanner for functional and anatomic imaging has been investigated in only a few studies (12,13). Thus, our aim in this study was to prospectively determine the respiratory protocol at CT that results in the best registration of CT images with SPECT images in the thorax.


   Materials and Methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
 Discussion
 Appendix
 References
 
Subjects
Each of 13 subjects (seven men and six women) underwent cardiac SPECT with 111 MBq (3 mCi) of thallium 201 (201Tl) and CT with three different respiratory protocols. The median age of subjects was 58 years, and the age range was 33–80 years. The study group consisted of six healthy volunteers (four men and two women; median age, 50 years; age range, 33–64 years) and seven consecutive patients (three men and four women; median age, 62 years; age range, 50–80 years) without a history of coronary artery disease. The six volunteers were considered healthy on the basis of their medical history, physical examination, and electrocardiogram. The seven patients were referred for resting myocardial SPECT imaging because of atypical chest pain during the period between November 2002 and January 2003. Each of these seven patients subsequently underwent echocardiography and coronary angiography, with no resultant findings of coronary artery disease, and they too were therefore considered healthy.

Institutional review board approval was obtained for multiple imaging examinations in healthy volunteers and patients and for the analysis of the resultant images. We explained to all subjects the purpose of our study and the risks associated with radiation exposure, with regard to stochastic effects that might result in cancer and genetic mutations, and we answered all their questions. Each subject provided written informed consent for this study.

Respiratory Protocols
All subjects underwent three CT examinations each: One examination was performed during free breathing; the second, during breath holding after normal inhalation; and the third, during breath holding after normal exhalation. Normal inhalation was defined as the respiratory level reached when the subject inhaled without forcing inhalation. Normal exhalation was defined as the respiratory level reached when the subject first inhaled and then exhaled without forcing exhalation.

System Overview
Our combined SPECT/CT system includes a SPECT scanner and a multi–detector row CT scanner that are commercially available (Fig 1). The SPECT scanner is gantry free (Skylight; ADAC Laboratories, Milpitas, Calif), and the CT scanner has eight detector rows (LightSpeed Ultra; GE Medical Systems, Milwaukee, Wis). The SPECT and CT scanners were positioned so that the patient could be moved directly from the CT scanner to the SPECT scanner by means of a simple extension of the CT table.



View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1a. (a) Schematic and (b) photograph of SPECT/CT system. Position of the SPECT scanner adjacent to the multi–detector row CT scanner (MDCT) enables movement of the patient with a simple extension of the table from the CT scanner into the gantry-free SPECT scanner.

 


View larger version (122K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1b. (a) Schematic and (b) photograph of SPECT/CT system. Position of the SPECT scanner adjacent to the multi–detector row CT scanner (MDCT) enables movement of the patient with a simple extension of the table from the CT scanner into the gantry-free SPECT scanner.

 
SPECT Imaging
Dual-head detectors were equipped with low-energy general-purpose parallel-hole collimators, and the two heads were oriented at an angle of 90° to each other. The 35% window was centered at approximately 74 keV for SPECT image data, and digital images were obtained by using a 64 x 64 matrix and a 41-cm field of view. A total of 32 projections were sampled within a 90° arc with each detector, and the time for each acquisition was 60 seconds. SPECT images were reconstructed by using filtered back projection, and the level of high-frequency noise was decreased by using postreconstruction Butterworth filtering (cutoff, one-half cycle per pixel; order, 5) at a postprocessing and display workstation (Pegasys; ADAC Laboratories). Thoracic images were then displayed in transverse, coronal, and sagittal planes for image registration. Horizontal and vertical (long-axis and short-axis) cardiac sections also were displayed.

CT Imaging
On the basis of the results of a preliminary study (Appendix), we adopted a low-current CT acquisition protocol for image attenuation correction. CT with postinhalation and postexhalation breath-hold protocols was performed in the helical mode with a 0.7-second rotation time, 8 x 2.5-mm collimation (eight detector elements and a 2.5-mm individual section thickness), 120 kV, and 10 mA. CT scans were reconstructed at 5.0-mm section thickness. For CT during free breathing, we selected a slower rotation time so as to obtain averaging of image data throughout the respiratory cycle and to avoid respiration-related artifacts caused by displacement of the diaphragm. CT with the free-breathing protocol was performed in the incremental mode with 4.0-second rotation time, 4 x 5.0-mm collimation, 140 kV, and 10 mA. CT images were reconstructed at a 5.0-mm section thickness by using a standard reconstruction algorithm with a 512 x 512 matrix and 50-cm field of view. The volume CT dose index was 1.0 and 3.4 mGy with the breath-hold and free-breathing protocols, respectively.

Registration Methods
SPECT and CT image registration was performed by using two different methods. First, semiautomated registration (external registration) was performed by using two external markers (three-port cocks with locks made of polycarbonate, which contained an aqueous solution of 201Tl and contrast medium) that were attached to the cushion of the imaging table. Second, manual registration (internal registration) was performed by using the heart border as an internal landmark.

Use of CT Images for Attenuation Correction
Attenuation correction was performed by using software (Hyogo CM Attenuation Correction, Hyogo College of Medicine, Nishinomiya, Hyogo, Japan) that was developed specifically for this purpose and installed on the workstation mentioned earlier (7). A flowchart of the attenuation correction process is shown in Figure 2. Registration of SPECT and CT images was performed at this same workstation by using the following steps: CT image data were retrieved from the CT workstation (Advantage Windows 4.0P; GE Medical Systems) and were converted to a SPECT-compatible format (5.9 x 5.9 x 5.9 mm) for image registration. After SPECT/CT image registration, the CT numbers were converted with linear scaling to the attenuation coefficients (µ) corresponding to the x-ray energy of 74 keV. The following equation defines the relationship between the CT number (CTn) and the attenuation coefficient (1/cm): µ = 0.181 · (CTn + 1000)/1000.



View larger version (38K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Flowchart shows process in which CT images are used to correct attenuation on SPECT images. MLEM = maximum likelihood expectation maximization.

 
The attenuation-corrected SPECT images were reconstructed by using the resultant attenuation coefficient map and a maximum likelihood expectation maximization algorithm. Three SPECT images on which attenuation was corrected by using corresponding CT images obtained with the different respiratory protocols were produced with each registration method. The transverse source image data were automatically reformatted in the horizontal and vertical (long-axis and short-axis) directions, and the reformatted and reconstructed images then were displayed by using the bull's-eye viewing option for evaluation of cardiac sections.

Measurement of Image Registration Error
The validity of the image registration was evaluated at a computer workstation with consensus by two diagnostic radiologists (T.N. and S.T., with 8 and 20 years of experience, respectively, in CT and cardiac SPECT). Errors of registration between the SPECT image data and the CT image data that were obtained with the three different respiratory conditions were measured separately for each registration method. The reference points measured to determine the extent of registration error were the following: (a) the dome of the diaphragm, (b) the left heart border, and (c) the chest wall border (Fig 3). The coordinates of each reference point on the x-, y-, and z-axes on CT and SPECT images were determined, and three-dimensional distances were calculated between the reference points on CT images and those on SPECT images. All measurements were performed by a radiologist (D.U., with 9 years of experience in both CT and cardiac SPECT imaging), who had obtained preliminary measurements in five additional subjects to optimize the techniques prior to this study.



View larger version (74K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Left: Coronal SPECT image. Right: Coronal SPECT/CT fusion image with grid overlay that shows the three reference points used to determine image registration error. Error was measured as the distance between the reference point on the SPECT image and that on the CT image, with regard to the following parameters: vertical displacement of the diaphragmatic dome (1); lateral displacement of the heart border (2); and lateral displacement of the chest wall border (3). Double white lines with opposing arrows indicate the limits of displacement.

 
Visual Evaluation of Attenuation-corrected SPECT Images
Attenuation-corrected SPECT images were reviewed at a workstation by two observers (T.H. and S.S., with 7 and 19 years of experience, respectively, with cardiac SPECT) using a linear rainbow color scale with seven colors that ranged from purple to red. Observers were aware that all subjects were healthy. Observers visually evaluated the attenuation-corrected SPECT images, which were grouped according to the CT respiratory protocol used, for the presence and location of artifactual evidence of reduction in radiotracer uptake. Images were classified according to findings of no artifact, mild artifact (wall thinning of ≤50%), or severe artifact (wall thinning of >50%). The location of designated artifacts was classified as the anterior, inferior, septal, or lateral wall of the myocardium. Visual evaluations were performed independently to enable assessment of the level of interobserver agreement between the physician (T.H.) and the radiologist (S.S.). When the two observers disagreed, the classification that indicated the more severe reduction was accepted as the final determination.

Statistical Analysis
Before the study, to determine the appropriate sample size for statistical analysis, a power analysis was performed by one of the authors (D.U.) by using preliminary measurements of the registration error obtained in five additional subjects who were not included in the final study population. The difference in mean registration error among the three respiratory protocols was approximately two times the standard deviation (SD). Therefore, a minimum of 10 subjects was considered appropriate to achieve an intended power of 80% or greater at subsequent statistical analysis.

One-way analysis of variance (ANOVA) was used to test for a statistically significant difference in mean registration error between the three respiratory protocols. If a significant difference was found, the Bonferroni-Dunn test was used for post hoc analysis. A P value of less than .01 was considered to indicate a statistically significant difference. All data are presented as the mean ± SD.

The degree of agreement between the two observers at visual evaluation of the attenuation-corrected SPECT images was measured with the {kappa} statistic. {kappa} values were reported as follows: 0, no agreement; more than 0 but less than 0.20, poor agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–0.99, almost perfect agreement (16). Software was used for the statistical analyses (SAS 8.01 for Windows; SAS Institute, Cary, NC) and for the power analysis (G-Power, version 2.1.2; A. Buchner, F. Faul, and E. Erdfelder, University of Duesseldorf, Duesseldorf, Germany; available at http://www.psycho.uni-duesseldorf.de/aap/projects/gpower/index.html).


   Results
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
Discussion
 Appendix
 References
 
Measurement of Image Registration Error
The mean distance ± SD between the reference points on the SPECT and CT images obtained with each of the three respiratory protocols and with external registration is shown in Table 1. Statistical analysis with ANOVA revealed that the results of external registration performed by matching of the diaphragm and the heart border on SPECT and CT images were significantly different with the three different respiratory protocols. The results of post hoc analysis also indicated that matching of the diaphragmatic dome and the heart border on SPECT and CT images was statistically far better with free breathing and postexhalation breath holding than with postinhalation breath holding (Table 2). The mean difference between the level of the diaphragmatic dome on CT scans obtained during free breathing, postexhalation breath holding, and postinhalation breath holding and the level of the dome on the corresponding SPECT scans was 5.56 mm ± 7.55, –3.15 mm ± 14.14 (median, –9.80 mm; range, –22.30 to 19.00 mm), and 29.88 mm ± 5.37, respectively. The corresponding difference for the location of the heart border was 8.57 mm ± 2.43, 13.30 mm ± 2.69, and 18.87 mm ± 6.29, respectively.


View this table:
[in this window]
[in a new window]

  		TABLE 1. SPECT/CT Image Registration Error Generated by External Registration, according to Respiratory Protocol Used at CT

 

View this table:
[in this window]
[in a new window]

  		TABLE 2. Post Hoc Analysis of Distance between Reference Points with External Registration of CT and SPECT Images

 
The mean distance ± SD between reference points on the SPECT and CT images obtained with each of the three respiratory protocols and with internal registration are shown in Table 3. The results of ANOVA revealed that matching of the diaphragmatic dome and the chest wall border between SPECT and CT images during internal registration was significantly different between respiratory protocols, and the results of the post hoc analysis revealed that matching was significantly better with the free-breathing and postexhalation breath-hold protocols than with the postinhalation breath-hold protocol (Table 4). The mean differences in movement of the diaphragmatic dome on the CT scans obtained during free breathing, postexhalation breath holding, and postinhalation breath holding, and the level of the dome on the corresponding SPECT scan, were 5.30 mm ± 4.32, –1.14 mm ± 9.92 (median, –7.33; range, –10.06 to 14.80 mm), and 24.16 mm ± 5.45, respectively. The corresponding differences for the chest wall border were 3.72 mm ± 3.08, 8.51 mm ± 5.03, and 20.60 mm ± 5.85, respectively.


View this table:
[in this window]
[in a new window]

  		TABLE 3. SPECT/CT Image Registration Error Generated by Internal Registration, according to Respiratory Protocol Used at CT

 

View this table:
[in this window]
[in a new window]

  		TABLE 4. Post Hoc Analysis of Distance between Reference Points with Internal Registration of CT and SPECT Images

 
Visual Assessment of Attenuation-corrected SPECT Images
Results of the visual assessment of attenuation-corrected SPECT images for each registration method are summarized in Tables 5 and 6. With external registration, the numbers of cases with either mild or severe artifacts indicative of decreased radiotracer uptake were nine (69%), four (31%), and two (15%) of 13 with postinhalation breath holding, postexhalation breath holding, and free breathing, respectively. With internal registration, the corresponding numbers were four (31%), two (15%), and one (8%) of 13 with postinhalation breath holding, postexhalation breath holding, and free breathing, respectively (Fig 4 ). No severe artifacts were observed for either the postexhalation breath-hold protocol or the free-breathing protocol. Image registration based on internal landmarks (internal registration) therefore offered improved homogeneity in the display of apparent distribution of myocardial radiotracer uptake (Fig 5). In addition, no artifacts were observed in the septal and lateral walls in any of the subjects, with either registration method.


View this table:
[in this window]
[in a new window]

  		TABLE 5. Visual Evaluation of Artifacts on Final Attenuation-corrected SPECT Images Obtained with External Registration

 

View this table:
[in this window]
[in a new window]

  		TABLE 6. Visual Evaluation of Artifacts on Final Attenuation-corrected SPECT Images Obtained with Internal Registration

 


View larger version (107K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4a. Final attenuation-corrected short-axis cardiac sections obtained with SPECT/CT in a 32-year-old man by using three respiratory protocols and internal landmarks for registration. (a) Image acquired during postinhalation breath hold shows severe artifact that suggests decreased radiotracer uptake in the inferior wall (arrows). (b, c) Images acquired during postexhalation breath hold (b) and during free breathing (c) show no artifact.

 


View larger version (106K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4b. Final attenuation-corrected short-axis cardiac sections obtained with SPECT/CT in a 32-year-old man by using three respiratory protocols and internal landmarks for registration. (a) Image acquired during postinhalation breath hold shows severe artifact that suggests decreased radiotracer uptake in the inferior wall (arrows). (b, c) Images acquired during postexhalation breath hold (b) and during free breathing (c) show no artifact.

 


View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4c. Final attenuation-corrected short-axis cardiac sections obtained with SPECT/CT in a 32-year-old man by using three respiratory protocols and internal landmarks for registration. (a) Image acquired during postinhalation breath hold shows severe artifact that suggests decreased radiotracer uptake in the inferior wall (arrows). (b, c) Images acquired during postexhalation breath hold (b) and during free breathing (c) show no artifact.

 


View larger version (101K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5a. Final attenuation-corrected short-axis cardiac sections obtained with SPECT/CT in a 60-year-old man during postinhalation breath hold. (a) Image obtained with registration based on external markers shows severe artifact (arrows) suggestive of decreased radiotracer uptake in the inferior wall. (b) Image obtained with registration based on internal landmarks shows no artifact.

 


View larger version (102K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5b. Final attenuation-corrected short-axis cardiac sections obtained with SPECT/CT in a 60-year-old man during postinhalation breath hold. (a) Image obtained with registration based on external markers shows severe artifact (arrows) suggestive of decreased radiotracer uptake in the inferior wall. (b) Image obtained with registration based on internal landmarks shows no artifact.

 
For each method of image registration, interobserver agreement at visual evaluation, as indicated by the {kappa} statistic, was either substantial or nearly perfect. With external registration, the level of interobserver agreement in analysis of the anterior myocardial wall was 0.88 for the postinhalation breath-hold protocol, 0.75 for the postexhalation breath-hold protocol, and 0.99 for the free-breathing protocol; in analysis of the inferior myocardial wall, it was 0.88 for the postinhalation breath-hold protocol, 0.83 for the postexhalation breath-hold protocol, and 0.99 for the free-breathing protocol. With internal registration, the level of interobserver agreement in analysis of the anterior myocardial wall was 0.84 for the postinhalation breath-hold protocol and 0.99 for both the postexhalation breath-hold and free-breathing protocols; for analysis of the inferior myocardial wall, it was 0.82 for the postinhalation breath-hold protocol and 0.99 for both the postexhalation breath-holding and free-breathing protocols.


   Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
Appendix
 References
 
Myocardial SPECT with 201Tl is a noninvasive, cost-effective, and routinely available method for obtaining valuable prognostic information about the risk of future cardiac events in a wide spectrum of patients in whom coronary artery disease has been diagnosed or is suspected (17–19). The diagnostic accuracy of myocardial SPECT is limited, however, by specific physical effects, including attenuation artifacts, scatter, and blurring (1–6). Among these effects, attenuation artifacts are arguably the most serious impediment to the accurate assessment of myocardial perfusion (6). Attenuation maps are commonly generated from monoenergetic transmission CT images obtained by using an external gamma radiation source (7,9). This process, however, is not yet widely used in clinical practice because of the low image quality that may result from low photon flux at transmission CT, the expense of both software and hardware, and the long scan times (7). Attenuation maps also have been generated from x-ray–based CT images (9). The quality of the final emission image that has been corrected for attenuation effects largely depends on the quality of registration between the transmission and emission images (20). Because a SPECT image is based on the averaging of data generated over multiple respiratory cycles, the fusion of SPECT images with CT images requires respiratory gating of SPECT scans. We contend that this is not a good approach, however, owing to the resultant substantial increase in imaging time. We believe that efforts to optimize image fusion therefore must be focused on the CT acquisition protocol.

The results of this study show that with the combined SPECT/CT system, the accuracy of image registration was higher with use of the free-breathing and postexhalation breath-hold protocols during CT scanning than with the postinhalation breath-hold protocol. Attenuation-corrected SPECT images generated with the free-breathing and postexhalation breath-hold protocols showed few artifacts, and they had a similarly homogeneous appearance with regard to uptake of 201Tl. Attenuation-corrected images obtained with the postinhalation breath-hold protocol showed artifacts that suggested decreased radiotracer uptake and mimicked perfusion abnormalities in the anterior and inferior myocardial wall. With the postinhalation breath-hold protocol, the attenuation of the heart and the diaphragm might well be underestimated because of registration errors between CT and SPECT images. We conclude, therefore, that CT image data acquired during a postinhalation breath hold are not suitable for SPECT/CT image fusion.

Goerres et al (13) reported that the normal postexhalation breath-hold protocol for CT acquisition gave the best results in comparison with a free-breathing protocol for PET/CT image registration. In their study, data acquisition during free breathing was performed in the helical mode, with a pitch of 1.7–2.5 and a 5-mm section thickness. Free breathing at helical CT may introduce significant geometric uncertainties into the CT image data (13,21). For example, if the diaphragm moves up or down between individual section acquisitions, the sequence of images may be disordered, with inferiorly indexed sections reflecting data that were sampled at superior levels in the structure (and vice versa) (13,21). We considered that our free-breathing protocol, with CT performed in the incremental mode and with 4.0-second rotation time, might help to normalize respiratory movement and reduce CT and SPECT image registration errors. Indeed, by performing CT with a 4.0-second rotation time setting during free breathing, we were able to completely avoid the mushroom-shaped respiratory motion–induced artifacts that result from imaging of the diaphragm at different positions. An additional and important benefit of this method is that free breathing could be performed by each of our patients, whereas breath holding can be quite difficult for some subjects with coronary artery disease.

The quality of the final attenuation-corrected SPECT images obtained with the internal registration method was higher than that obtained with the external registration method, and, hence, we believe that internal registration is more suitable for attenuation correction of myocardial SPECT images. The difference in displacement of the heart between SPECT and CT images was quite high with image registration based on external markers, and the resultant error caused artifacts that mimicked perfusion abnormalities. We concluded, however, that displacement of the chest wall had little influence on the quality of the final attenuation-corrected images.

This study had several limitations. First, the conversion from the CT value to the attenuation coefficient of gamma rays is complicated because the energy spectrum of gamma rays is finite, whereas x-rays are continuous. Second, the SPECT images were not scatter-corrected, although such correction would have improved the image quality and the accuracy of quantification. Third, SPECT images in this study were generated from image data collected within a 180° arc, which was chosen because the use of anterior 180° acquisitions has become standardized in the clinical setting (22). Acquisitions of 180° and 360° were performed in both an anthropomorphic torso phantom and one subject prior to this study, and, when the resultant images were compared at visual evaluation, no substantial differences were observed. Further studies are imperative to confirm these findings.

In conclusion, for proper object-specific attenuation correction of cardiac SPECT images, we recommend SPECT/CT image registration based on internal landmarks and on CT images acquired either with free breathing or with breath holding after normal exhalation.


   Appendix
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Appendix
References
 
We preliminarily investigated the optimal x-ray tube current for use in attenuation correction. High currents improve image quality but also increase the radiation dose to the patient. If the CT scan is to be used only for attenuation correction, a low-current CT scan may be sufficient. We compared attenuation-corrected SPECT/CT images obtained by using a low current (10 mA) with those obtained by using a standard current (140 mA) at CT. A SPECT phantom (ECT QA Phantom, type JSP; Kyoto-kagaku, Kyoto, Japan) was used (Fig A1). Cylinder A of the phantom contained water and 201Tl, and cylinders B–D contained diluted contrast material at iodine concentrations of 10.0, 4.5, and 2.5 mg of iodine per milliliter, respectively, as well as 201Tl. The concentration of 201Tl in each of the four cylinders was 0.2 MBq/mL. Two unenhanced CT scans were acquired with 10 and 140 mA (hereafter, these scans are designated as CT 10 and CT 140). Two attenuation-corrected SPECT images were obtained by using CT 10 and CT 140 (SPECTCT10 and SPECTCT140, respectively) with iterative reconstruction (Fig A2). The adequacy of the SPECT/CT attenuation coefficient maps was assessed by comparing the total counts of gamma ray emissions from 201Tl in four regions of interest (one region of interest placed in each of the four cylinders) on the two attenuation-corrected SPECT images. The size of each region of interest was 5 pixels. To verify the reproducibility of these measurements in the phantom, CT at 10 and 140 mA was repeated six times. Two-way ANOVA was used to evaluate the effect of the CT current on attenuation-corrected SPECT images. The mean counts ± SDs for gamma ray emissions in each cylinder on SPECTCT10 and SPECTCT140 images are shown in Table A1. There was no significant difference between the counts on SPECTCT10 and SPECTCT140 (P = .95). In summary, a low-current CT scan is sufficient for attenuation correction.



View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure A1. Transverse CT image of SPECT phantom. Cylinder A contains water and 201Tl; cylinders B, C, and D, diluted contrast material with different iodine concentrations and 201Tl.

 


View larger version (57K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure A2. Attenuation-corrected SPECT/CT image, obtained by using low-current CT in the same phantom as in Figure A1, shows uniform attenuation in cylinders A–D.

 

View this table:
[in this window]
[in a new window]

  		TABLE A1. Mean Attenuation Coefficients for SPECT/CT Images Obtained with Different CT Currents in a SPECT Phantom with Four Cylinders

 


   FOOTNOTES
 

Abbreviations: ANOVA = analysis of variance • SD = standard deviation

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, K.A., H.O., Y.Y.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, D.U., S.M.; clinical studies, D.U., T.N., T.H., S.S., K.K.; experimental studies, D.U., S.T.; statistical analysis, D.U., K.A.; and manuscript editing, D.U., S.T., K.K., S.M., K.A., H.O., Y.Y.


   References
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
 Results
 Discussion
 Appendix
 References
 

  1. Tsui BM, Gullberg GT, Edgerton ER, et al. Correction of nonuniform attenuation in cardiac SPECT imaging. J Nucl Med 1989;30:497–507.[Abstract/Free Full Text]
  2. Ficaro EP, Fessler JA, Ackermann RJ, Rogers WL, Corbett JR, Schwaiger M. Simultaneous transmission-emission thallium-201 cardiac SPECT: effect of attenuation correction on myocardial tracer distribution. J Nucl Med 1995;36:921–931.[Abstract/Free Full Text]
  3. DePuey EG 3rd. How to detect and avoid myocardial perfusion SPECT artifacts. J Nucl Med 1994;35:699–702.[Abstract/Free Full Text]
  4. Hashimoto J, Ogawa K, Kubo A, et al. Application of transmission scan-based attenuation compensation to scatter-corrected thallium-201 myocardial single-photon emission tomographic images. Eur J Nucl Med 1998;25:120–127.[CrossRef][Medline]
  5. Prvulovich EM, Lonn AH, Bomanji JB, Jarritt PH, Ell PJ. Effect of attenuation correction on myocardial thallium-201 distribution in patients with a low likelihood of coronary artery disease. Eur J Nucl Med 1997;24:266–275.[Medline]
  6. Hasegawa BH, Lang TF, Brown JK, et al. Object-specific attenuation correction of SPECT with correlated dual-energy x-ray CT. IEEE Trans Nucl Sci 1993;40:1242–1252.
  7. Kashiwagi T, Yutani K, Fukuchi M, et al. Correction of nonuniform attenuation and image fusion in SPECT imaging by means of separate x-ray CT. Ann Nucl Med 2002;16:255–261.[Medline]
  8. Takahashi Y, Murase K, Higashino H, Mochizuki T, Motomura N. Attenuation correction of myocardial SPECT images with x-ray CT: effects of registration errors between x-ray CT and SPECT. Ann Nucl Med 2002;16:431–435.[Medline]
  9. LaCroix KJ, Tsui BM, Hasegawa BH, Brown JK. Investigation of the use of x-ray CT images for attenuation compensation in SPECT. IEEE Trans Nucl Sci 1994;41:2793–2799.[CrossRef]
  10. Bocher M, Balan A, Krausz Y, et al. Gamma camera-mounted anatomical x-ray tomography: technology, system characteristics and first images. Eur J Nucl Med 2000;27:619–627.[CrossRef][Medline]
  11. Delbeke D, Martin WH, Patton JA, Sandler MP. Value of iterative reconstruction, attenuation correction, and image fusion in the interpretation of FDG PET images with an integrated dual-head coincidence camera and x-ray–based attenuation maps. Radiology 2001;218:163–171.[Abstract/Free Full Text]
  12. Goerres GW, Burger C, Kamel E, et al. Respiration-induced attenuation artifact at PET/CT: technical considerations. Radiology 2003;226:906–910.[Abstract/Free Full Text]
  13. Goerres GW, Kamel E, Heidelberg TN, Schwitter MR, Burger C, von Schulthess GK. PET-CT image co-registration in the thorax: influence of respiration. Eur J Nucl Med Mol Imaging 2002;29:351–360.[CrossRef][Medline]
  14. Beyer T, Antoch G, Blodgett T, Freudenberg LF, Akhurst T, Mueller S. Dual-modality PET/CT imaging: the effect of respiratory motion on combined image quality in clinical oncology. Eur J Nucl Med Mol Imaging 2003;30:588–596.[Medline]
  15. Goerres GW, Kamel E, Seifert B, et al. Accuracy of image coregistration of pulmonary lesions in patients with non-small cell lung cancer using an integrated PET/CT system. J Nucl Med 2002;43:1469–1475.[Abstract/Free Full Text]
  16. Svanholm H, Starklint H, Gundersen HJ, Fabricius J, Barlebo H, Olsen S. Reproducibility of histomorphologic diagnoses with special reference to the kappa statistic. APMIS 1989;97:689–698.[Medline]
  17. Brown KA. The role of stress redistribution thallium-201 myocardial perfusion imaging in evaluating coronary artery disease and perioperative risk. J Nucl Med 1994;35:703–706.[Abstract/Free Full Text]
  18. Singer ME, Applegate KE. Cost-effectiveness analysis in radiology. Radiology 2001;219:611–620.[Abstract/Free Full Text]
  19. Weissman IA, Dickinson CZ, Dworkin HJ, O'Neill WW, Juni JE. Cost-effectiveness of myocardial perfusion imaging with SPECT in the emergency department evaluation of patients with unexplained chest pain. Radiology 1996;199:353–357.[Abstract/Free Full Text]
  20. Patton JA, Delbeke D, Sandler MP. Image fusion using an integrated, dual-head coincidence camera with x-ray tube-based attenuation maps. J Nucl Med 2000;41:1364–1368.[Abstract/Free Full Text]
  21. Balter JM, Ten Haken RK, Lawrence TS, Lam KL, Robertson JM. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int J Radiat Oncol Biol Phys 1996;36:167–174.[Medline]
  22. American Society of Nuclear Cardiology. Updated imaging guidelines for nuclear cardiology procedures, part 1. J Nucl Cardiol 2001;8:G5–G58.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 RadioGraphicsHome page
B. Bybel, R. C. Brunken, F. P. DiFilippo, D. R. Neumann, G. Wu, and M. D. Cerqueira
SPECT/CT Imaging: Clinical Utility of an Emerging Technology
RadioGraphics, July 1, 2008; 28(4): 1097 - 1113.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2372041387v1
237/2/662    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Utsunomiya, D.
Right arrow Articles by Yamashita, Y.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Utsunomiya, D.
Right arrow Articles by Yamashita, Y.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE