Research ArticleBasic Science Investigation

Multicenter Evaluation of a Standardized Protocol for Rest and Acetazolamide Cerebral Blood Flow Assessment Using a Quantitative SPECT Reconstruction Program and Split-Dose 123I-Iodoamphetamine

Hidehiro Iida, Jyoji Nakagawara, Kohei Hayashida, Kazuhito Fukushima, Hiroshi Watabe, Kazuhiro Koshino, Tsutomu Zeniya and Stefan Eberl
Journal of Nuclear Medicine October 2010, 51 (10) 1624-1631; DOI: https://doi.org/10.2967/jnumed.110.078352

Abstract

SPECT can provide valuable diagnostic and treatment response information in large-scale multicenter clinical trials. However, SPECT has been limited in providing consistent quantitative functional parametric values across the centers, largely because of a lack of standardized procedures to correct for attenuation and scatter. Recently, a novel software package has been developed to reconstruct quantitative SPECT images and assess cerebral blood flow (CBF) at rest and after acetazolamide challenge from a single SPECT session. This study was aimed at validating this technique at different institutions with a variety of SPECT devices and imaging protocols. Methods: Twelve participating institutions obtained a series of SPECT scans on physical phantoms and clinical patients. The phantom experiments included the assessment of septal penetration for each collimator used and of the accuracy of the reconstructed images. Clinical studies were divided into 3 protocols, including intrainstitutional reproducibility, a comparison with PET, and rest–rest study consistency. The results from 46 successful studies were analyzed. Results: Activity concentration estimation (Bq/mL) in the reconstructed SPECT images of a uniform cylindric phantom showed an interinstitution variation of ±5.1%, with a systematic underestimation of concentration by 12.5%. CBF values were reproducible both at rest and after acetazolamide on the basis of repeated studies in the same patient (mean ± SD difference, −0.4 ± 5.2 mL/min/100 g, n = 44). CBF values were also consistent with those determined using PET (−6.1 ± 5.1 mL/min/100 g, n = 6). Conclusion: This study demonstrates that SPECT can quantitatively provide physiologic functional images of rest and acetazolamide challenge CBF, using a quantitative reconstruction software package.

Current clinical practice using SPECT relies largely on interpretation of qualitative images reflecting physiologic function. Quantitative functional parametric images may be obtained by applying mathematic modeling to SPECT data corrected for attenuation and scatter. Quantitative regional cerebral blood flow (CBF) (1–3) and cerebral vascular reactivity (CVR) in response to acetazolamide challenge (4–6) have been obtained with these techniques. One major application of such quantitative SPECT (QSPECT) approaches is the evaluation of ischemic status in patients with occlusion or stenosis in their middle cerebral arteries, to provide prognostic information of the outcome of revascularization therapies (7). Quantitative analysis in SPECT has also been demonstrated in the assessment of binding potential for several neuroreceptor ligands (8,9), for the quantitative assessment of regional myocardial perfusion (10,11), and for the assessment of radioaerosol deposition and clearance in healthy and diseased lungs (12). However, providing the standardized quantitative approach required for multicenter clinical trials has so far received only limited attention. Challenges remain in providing consistent quantitative data across institutions using a variety of SPECT equipment and vendor-specific reconstruction strategies (13). This limitation is attributed to a lack of standardized procedures in the reconstruction software offered by vendors, particularly in terms of correcting attenuation and scatter. Kinetic modeling for physiologic parameter estimation is also not part of the vendors’ standard SPECT software. Although separate packages can be purchased for this purpose, they are not integrated and are flexible general-purpose packages, requiring considerable skill and knowledge to effectively use. Thus, they are not ideal for routine clinical use.

Scatter and attenuation occur in the object and are thus object-dependent but are not dependent on the geometry of the imaging equipment (14). Therefore, once a software program is developed to provide accurate image reconstruction with compensation for both attenuation and scatter, the program should be able to provide quantitative images that are intrinsically independent of the geometric design of SPECT cameras. This is an attractive feature of SPECT for multicenter clinical studies.

From the various techniques available to correct for attenuation (15) and scatter (16), one feasible approach for clinical studies is based on a combination of attenuation correction, incorporated into the ordered-subset expectation maximization (OSEM) reconstruction (17), and scatter correction by the transmission-dependent convolution subtraction (TDCS) originally proposed by Meikle et al. (18). This approach has been extensively investigated by our group (11,19) for 99mTc for studies of the brain and heart (18,20) and also in cardiac 201Tl studies (11,21). A recent study also demonstrated the accuracy of this approach in a combined SPECT/CT system (22). By incorporating a correction for collimator septal penetration by high-energy emissions, one can also make the technique applicable to 123I (19).

The QSPECT reconstruction approach has estimated CBF images at rest in a clinical setting (11) and quantified CVR by measuring CBF at rest and after vasodilation in a single SPECT session. This was accomplished by using the dual-table autoradiographic (DTARG) method and a dual administration of 123I-iodoamphetamine (23). In those studies, corrections for attenuation and scatter appeared to be essential for generating quantitative CBF maps that were consistent with those generated by 15O-water PET (11,23).

These studies were, however, validated in a single institution using a limited range of SPECT systems, and the general applicability of this technique for different SPECT systems had yet to be fully established. Thus, the aim of this study was to verify that analysis of data with a standardized reconstruction package incorporating attenuation and scatter correction can provide reproducible results across multiple institutions for quantitative rest and acetazolamide challenge CBF estimation from a single SPECT session.

MATERIALS AND METHODS

Institutions and Subjects

The 12 participating institutions were clinical centers and generally did not have scientific staff dedicated to nuclear medicine software or hardware development. Standard, vendor-supplied software was used for the collection of the studies, with unmodified scanners and collimators clinically used for brain studies. The acquired data were reconstructed with the program package developed for this project. Manufacturers and models of camera systems and the number of detectors and collimators (including fanbeam or parallel hole) used by the institution are listed in Supplemental Table 1 (supplemental materials are available online only at http://jnm.snmjournals.org). All institutions performed experiments on physical phantoms according to the protocol described in the “Phantom Experiment” section. Of the 12 institutions, 9 obtained patient scans, whereas the remaining 3 provided only phantom data. Clinical studies were approved by institutions’ ethics committees or followed guidelines for clinical research protocols authorized by the institution. All subjects at each institution gave written informed consent.

The clinical studies were divided into 3 protocols: intrainstitutional, intrasubject reproducibility (reproducibility); comparison with PET (vs. PET); and intrascan consistency of the dual-time-point split-dose (rest–rest). Studies were excluded from the analysis if there was severe patient motion during one of the studies or if there were changes in the condition of the patients between the first and second studies likely to lead to changes in CBF.

Eight institutions (institutions 1, 3, 4, 6, 8, 9, 11, and 12) participated in the reproducibility arm, in which quantitative CBF values measured on separate days were compared. In this arm, all patients experienced unilateral or bilateral stenosis or occlusion in the extracranial internal carotid artery. The patients’ ages ranged from 43 to 81 y (mean ± SD, 65 ± 9 y). A total of 31 studies in this protocol were analyzed. Four patients had to be excluded from the analysis—2 because of significant changes in their pathophysiologic status between the studies and 2 because of severe motion and mispositioning in the scanner.

One institution (institution 4) performed the versus-PET studies. CBF values obtained by the DTARG method were compared with those by 15O-water and PET. Studies were performed on 6 patients (5 men, 1 woman; age range, 71–74 y; mean age ± SD, 72 ± 1 y) with stenosis or occlusion of the extracranial internal carotid artery unilaterally (n = 3) or bilaterally (n = 3).

Two institutions (institutions 2 and 12) provided data for the rest–rest comparison. Five patients from institution 2 had chronic cerebral infarction, whereas 4 subjects from institution 12 had no sign of cerebral disease. Patients’ ages ranged from 32 to 72 y (mean ± SD, 52 ± 15 y); 5 patients were men and 4 women.

Phantom Experiment

Three experiments were performed by each institution using the SPECT camera fitted with the collimators normally used in clinical brain studies. The first scan determined the absolute sensitivity or the becquerel calibration factor (BCF) of the reconstructed images. For 10 min, a 360° projection set was acquired of a syringe filled with a 123I-iodoamphetamine solution of known radioactivity and placed at the center of the field of view. The syringe was supplied by Nihon-Medi Physics, and its radioactivity was calibrated to 111 MBq at noon on the day before the experiment, with an accuracy better than 3%, decaying to approximately 30 MBq at the time of the experiment, avoiding the dead time of the camera. The BCF was determined by dividing the absolute radioactivity by the total counts for the syringe region in the reconstructed image.

The second experiment determined the collimator septal penetration contribution (24) from high-energy photons into the primary 159-keV energy window for 123I. A line-spread function was obtained from the projection data of a line source filled with 123I-iodoamphetamine. The septal penetration was determined from the background level as described previously (19). A projection line-spread function was also generated from this line source placed in a water-filled cylindric phantom (diameter, 16 cm).

The third experiment used a 16-cm-diameter, 15-cm-long uniform cylindric phantom. The whole radioactivity used for the BCF determination was diluted into the phantom, and projection data were acquired for 30 min, using the clinical scan protocols described in the “Clinical Studies” section. The radioactivity concentration (counting rate per unit mass) of an approximately 0.3-mL sample from the phantom was measured using the well counters available at the various institutions. Both NaI- and plastic scintillator–based well counters were used (Supplemental Table 1). Average pixel counts derived from regions of interest on the reconstructed emission images were referred to the well counter radioactivity counting rate, to determine the cross-calibration factor between the SPECT images and well counter system. This cross-calibration factor was subsequently used for the blood sample counts of the clinical studies. Uniformity of the reconstructed emission images was evaluated.

Clinical Studies

All clinical SPECT studies followed the DTARG protocol, with dual administration of iodoamphetamine (23), depicted in Figure 1. Briefly, 2 dynamic scans were acquired in quick succession, with a 2-min interval between the scans. The first scan covered the initial 0- to 28-min period, and the second was acquired from 30 to 58 min. At 4 min per frame, 7 frames covered each of the 2 dynamic scan periods. 123I-iodoamphetamine (111 MBq at institutions 2–12 or 167 MBq at institution 1) was infused twice over 1 min into the antecubital vein at 0 and 30 min. Acetazolamide (17 mg/kg, 1,000 mg maximum) was administered intravenously at 20 min after the first iodoamphetamine injection, corresponding to 10 min before the second iodoamphetamine injection. Projection data were summed for the acquisition duration of the first and second scans and reconstructed as described in the “QSPECT Reconstruction” section. In contrast to the study of Kim et al. (23), which used full arterial blood sampling, the individual arterial input functions were derived from a population-based standardized input function scaled with the whole-blood counts from a single arterial blood sample taken at approximately 10 min (1,25–28). This sample was also used for arterial blood gas analysis.

FIGURE 1.
FIGURE 1.

Scanning protocol flow for DTARG procedure. 123I-iodoamphetamine was injected at 0 min, and 28-min resting dynamic SPECT scan was commenced. Blood sample for calibration of population input function was drawn at 10 min. Acetazolamide (DMX-diamox) was administered at 20 min. CBF values are scaled by last frame (time, 24–28 min). Second dynamic SPECT scan followed second injection of 123I-iodoamphetamine at 30 min. IMP = iodoamphetamine.

In the reproducibility arm, an additional, non-DTARG CBF study was performed on a separate day. Instead of DTARG, the previously reported123I-IMP autoradiographic (IMPARG) method (1,19,25) was performed within a month of the DTARG study. The IMPARG method is essentially equivalent to the present DTARG method, except that the IMPARG method uses a single iodoamphetamine administration to assess CBF either at rest or after acetazolamide challenge. The same image reconstruction process as for the DTARG protocol was used. In 12 studies, the DTARG protocol was used instead of IMPARG—namely, the DTARG study was performed twice to assess the CBF reproducibility at rest and after acetazolamide.

In the versus-PET protocol, the PET study was performed within 2 d of the DTARG SPECT study. PET scans used intravenous 15O-water both at rest and after the acetazolamide challenge. CBF images were calculated by the 15O-water autoradiography technique (29), with careful corrections for delay and dispersion (30–32). Patients were stable between the SPECT and PET studies.

In the rest–rest protocol, the DTARG scan was obtained without the pharmacologic challenge during the study to evaluate the consistency of CBF values estimated from the 2 scans.

QSPECT Reconstruction

The program package for QSPECT uses a wrapper written in JAVA to run several programs written in C for Microsoft Windows systems. The package includes programs for reconstructing SPECT images, calculating functional images, coregistering images, and reslicing and printing summary logs.

The QSPECT package reconstructs images from the original projection data from commercial SPECT equipment, based on previous work by Iida and his colleagues (19–21,23,33,34). Reconstructed SPECT images are calibrated in Bq/mL, which provides independence from scanning parameters such as the acquisition time, number of views, matrix size, and zoom factor. Uniformity and center-of-rotation corrections and fanbeam–to–parallel beam conversion (for fanbeam collimators) were performed using the clinical routine software before reconstruction by this package.

An overall flow diagram of the correction and reconstruction process is shown in Supplemental Figure 1. The OSEM reconstruction technique includes attenuation correction (17). A threshold-based edge-detection algorithm generated the attenuation μ-map, assuming a uniform attenuation coefficient of 0.166 cm−1 for 99mTc (0.160 cm−1 for 123I) as an average over the brain and skull (19). The threshold was optimized via the user interface to correctly define the brain outline. The attenuation μ-map was generated from the summed 0- to 28-min rest frame and was coregistered to the other images (35) reconstructed with filtered backprojection without attenuation or scatter correction. The attenuation μ-maps were forward projected to provide the transmission projection data for TDCS. The emission projections were scatter-corrected by the TDCS method, as originally proposed by Meikle et al. (18), and further optimized for realistic 99mTc, 201Tl, and 123I data in the brain and thorax regions (20,21,23,33,34). An offset compensated for the septal penetration of high-energy photons for 123I studies, which adds fairly uniform background counts, or direct current (DC) components, to the projections.

Scatter- and attenuation-corrected images were reconstructed with OSEM (5 iterations, 5 subsets using geometric-mean projections, postreconstruction gaussian filter of 7 mm in full width at half maximum) and then realigned to the image set obtained from the first scan. The acquisition parameters and BCF were used to convert the reconstructed raw counts to Bq/mL.

The global CBF over the entire gray matter was estimated from the SPECT frame covering 24–28 min, because this timing minimizes the individual shape variations in individual input function. The look-up table generated for estimating CBF images from the complete dynamic study (0–28 min) was then scaled to provide global cortical gray matter CBF values consistent with the 24- to 28-min frame estimates. A careful detection algorithm was used to reliably exclude extracranial accumulation of 123I-iodoamphetamine (e.g., in the parotid region), which could adversely affect this scaling procedure. The regional CBF was then estimated at each pixel by means of the table look-up procedure (25,28). The background image at the time of the second 123I-iodoamphetamine injection was estimated from the first-phase CBF images, according to the compartment model assumed in this study (23). An additional table look-up procedure was applied to the second dynamic dataset (30–58 min) for calculating the vasodilated (acetazolamide challenge) CBF images as described previously (23). The data were successfully reconstructed, and CBF was estimated at each institution. To facilitate and provide consistent analysis, the data presented are from the reanalysis conducted at the core lab (National Cerebral and Cardiovascular Center).

Data Analysis

The uniform phantom SPECT activity estimates were compared with the known activity in the phantom. Images for the baseline study were displayed with subsequent images using an absolute flow value scale to visually ascertain regional and global differences in flow. Regions of interest were placed on the middle cerebral artery territories of both hemispheres, and the average flow values between the different methods were compared and plotted. Bland–Altman plots and the SD of the differences evaluated the consistency of CBF values obtained from the reproducibility and versus-PET protocols.

All data were presented as mean ± SD. Pearson correlation analysis and linear regression analysis were used to evaluate relationships between the 2 CBF values. A P value less than 0.05 was considered statistically significant.

RESULTS

Phantom Studies

In the 16-cm scattering cylinder line source experiment, the scatter-uncorrected images show background counts extending beyond the phantom, from septal penetration of the high-energy photons. The scatter correction is largely effective in correcting for scatter and septal penetration counts. As shown in Supplemental Figure 2, the Toshiba-ECAM low- to medium-energy general-purpose (LMEGP) collimator, designed for reduced 123I septal penetration, compared with the standard low-energy high-resolution collimator (GE Healthcare), demonstrates reduced scatter and septal penetration counts. The lower septal penetration of the Toshiba-ECAM LMEGP collimators is also supported by a lowered scatter correction offset value (DC = 0.05, compared with DC = 0.20 for the GE low-energy high-resolution collimator). The reduced scatter and septal penetration result in more complete removal of scatter for the LMEGP collimator.

Figure 2 displays reconstructed slices of the uniform phantom for all 12 institutions, scaled to the same maximum activity concentration. The estimated activity concentrations from these studies, compared with the known activity concentration, represented an accuracy of 87.5% ± 5.1% (Supplemental Table 1). The well counter–to–SPECT cross-calibration factor, which represents the sensitivity of the well counter system for 123I, was 0.5–1.0 for NaI systems and 0.1–0.2 for plastic scintillation detector systems. The BCF values were consistent for the same SPECT camera–collimator configurations.

FIGURE 2.
FIGURE 2.

Reconstructed slices through uniform phantom from the participating 12 institutions. Experiment was designed to have same phantom activity concentration for each center's study. Nonuniformities and also differences in absolute activity concentration estimates can be observed, highlighting need for rigorous calibration, flood correction, and quality control. Legend above each image gives institution number (given in Supplemental Table 1), γ-camera model, and collimator used.

Clinical Studies

Figure 3A shows typical CBF images obtained at 4 institutions with 4 different γ-camera vendors, performed as part of the reproducibility arm of the study. Each case shows different CBF distributions both at rest and after acetazolamide challenge. The acetazolamide images obtained using the DTARG method agree well with the images subsequently obtained with the IMPARG method after acetazolamide infusion.

FIGURE 3.
FIGURE 3.

(A) Images from reproducibility study. CBF images obtained at rest and after acetazolamide with DTARG method. Repeated scan (third row) within 1 mo using IMPARG method and acetazolamide stress (columns 1–3) and at rest (last column). Images demonstrate that CVR can be estimated with this technique and demonstrate good reproducibility of measuring both at rest and after acetazolamide challenge CBF. (B) Bland–Altman plot showing difference vs. IMPARG CBF values estimated from DTARG method and repeated IMPARG studies to assess reproducibility. Little systematic bias is detected (mean difference, −0.4 mL/100 g/min), and SD of differences is moderate (5.2 mL/100 g/min). Correlation coefficient of r = 0.93 (P < 0.001) was found.

CBF images of a subject with left middle cerebral artery occlusion are shown in Supplemental Figure 3 for slices covering the whole brain. The images demonstrate reduced CBF after acetazolamide challenge in the left middle cerebral artery territory. The good reproducibility is confirmed by the Bland–Altman plot comparison of DTARG CBF values, with the CBF values obtained at a different imaging session with IMPARG or DTARG (Fig. 3B). The SD of the differences is 5.2 mL/100 g/min, with low bias supported by the mean difference of 0.4 mL/100 g/min. Regression analysis between DTARG and IMPARG values yielded a significant correlation (P < 0.001), with a correlation coefficient of r = 0.93.

Figure 4A shows MR and CBF images at rest and after acetazolamine obtained with DTARG SPECT and 15O-water PET in a 73-y-old male patient (63 kg) with right internal carotid artery occlusion and left internal carotid stenosis. The MR images do not show any evidence of cerebral infarction in either hemisphere. Rest CBF was reduced bilaterally in the frontal-to-parietal regions, and acetazolamide increased CBF in left parietal regions but not in the right parietal area. DTARG CBF indicated the loss of vasoreactivity in the right internal carotid artery stenotic area. These findings were consistent with those from the PET evaluation. An additional example is shown in Supplemental Figure 4 for a 74-y-old female patient (48 kg) with left internal carotid artery stenosis, for whom MR images did not show cerebral infarction. DTARG CBF demonstrated preserved CBF in both hemispheres but reduced CBF reactivity in the left middle cerebral artery territory. The findings were again consistent with those from PET. Figure 4B compares the flow values obtained at rest and after acetazolamide with DTARG with the corresponding values obtained by 15O-water PET. The SD of the differences is 5.1 mL/100 g/min, with the significant underestimation by 15O-water PET, compared with PET by the DTARG method, highlighted by a mean difference of −6.1 mL/100 g/min. The Pearson analysis showed a significant correlation (P < 0.001), with a correlation coefficient of r = 0.88.

FIGURE 4.
FIGURE 4.

(A) MR and CBF images at rest and after acetazolamide stress assessed with corresponding measurements with 15O-water PET (vs. PET evaluation) in patient with right internal carotid artery occlusion and left internal carotid stenosis. Gaussian filter was not applied to SPECT CBF in this display. (B) Bland–Altman plot. Moderate underestimation of CBF determined by DTARG method, compared with PET, is observed (mean difference, −6.1 mL/100 g/min). Correlation coefficient of r = 0.88 (P < 0.001) was found.

The results from the rest–rest protocol are summarized in Figure 5. The differences between the measurements performed with the 2 injections were small, with good agreement between the 2 flow values. The mean ± SD of the differences was 0.6 ± 2.9 mL/100 g/min.

FIGURE 5.
FIGURE 5.

Results from rest–rest evaluation carried out at 2 institutions (2 and 12). In this study, DTARG method was performed as per normal protocol but without pharmacologic stress. CBFs estimated with first injection (left on graph) are in good agreement with those estimated after second injection (right on graph).

DISCUSSION

The QSPECT package provided quantitative images consistent between the participating centers, using dual- or triple-detector SPECT scanners and collimators routinely used for nonquantitative brain studies. All centers successfully acquired the dynamic SPECT images, and the data from the variety of cameras encountered were successfully processed by the software package. Rest CBF and CVR could be readily obtained by the participating institutions in a single, clinically practical, 1-h scanning session. Good reproducibility of CBF estimates was observed in 31 pairs of studies at 8 institutions (Fig. 3), and the CBF estimated with the 123I-iodoamphetamine SPECT agreed well with 15O-water PET CBF at 1 institution (Fig. 4). The CBF values after the second injection of the DTARG were consistent with the values obtained after the first injection when no vasodilating stress was given in 9 studies at 2 institutions (Fig. 5).

Quantitative CBF and CVR in response to acetazolamide challenge can be of significant prognostic value for patients considered for revascularization of cerebral arteries (5–7). The previously validated IMPARG method requires 2 independent scans on different days to assess the CVR (5–7), limiting it for routine clinical studies. The DTARG protocol to quantitatively assess CBF both at rest and after acetazolamide from a single dynamic SPECT session with the dual administration of 123I-iodoamphetamine (23) facilitates clinical use. Errors caused by ambiguity in the absolute scaling, and possible changes in physiologic status of the subjects between scans, can be reduced substantially with the DTARG protocol. The quantitative reconstruction program enabled the compartment model–based kinetic analysis to compensate for the residual radioactivity concentration during the second session of the dynamic scan.

Major error sources in SPECT, namely attenuation and scatter, are only object-dependent (14) and not γ-camera– or collimator-dependent, and thus SPECT images obtained by this quantitative reconstruction package should be consistent across systems. Septal penetration of high-energy photons for 123I is, however, collimator-dependent (24) but could be compensated as part of the TDCS scatter correction algorithm (11), as demonstrated in Supplemental Figure 2. The radioactivity concentration of the uniform cylinder phantom estimated in units of Bq/mL was consistent and showed variation within ±5.1% (Fig. 2; Supplemental Table 1), though a systematic underestimation by 12.5%, which is attributed to the BCFs being derived from a line source in air, reconstructed without scatter, attenuation, and septal penetration corrections. However, this underestimation does not affect the CBF estimation, because it relies on the direct cross-calibration between the γ-counter used to count the blood sample and the SPECT measurements.

This phantom study also highlighted the importance of proper calibration and quality control of the γ-camera to avoid artifacts and bias in the reconstructed images. These corrections were applied, as for other clinical studies, by the vendors’ software rather than as part of the QSPECT system, because these corrections are typically performed online and on-the-fly, with only the corrected data being stored. The nonuniformities seen on some phantom images should improve with more rigorous quality-control procedures.

The previously validated population-based input function requiring only a single arterial blood sample for scaling (1,25–28) has been incorporated in the software package. Blood from this single arterial sample is also used to measure arterial blood gases, which are relevant and of interest clinically in these patients. The timing of the single blood sample (∼10 min after iodoamphetamine injection) was optimized previously (1,25–28) to minimize the errors associated with individual differences in shape of the arterial input function. In addition, absolute global CBF was estimated from SPECT images taken at an optimized mid scan time of approximately 30 min (24–28 min), rather than from the initial part of the study, to maximize the accuracy of using the population-based input function (1,25–28).

Partial-volume correction has not been implemented as part of this processing protocol. Partial-volume effects can potentially lead to underestimation of flow values in gray matter regions because of the limited resolution of SPECT. The small underestimation of 6.1 mL/100 g/min by the DTARG method, compared with 15O-water PET (Fig. 4B), is attributed to the partial-volume effects due to differences in resolution between PET and SPECT. The underestimation can also lead to variations in CBF values obtained with different-resolution collimators. However, consistent postreconstruction filtering, as applied in this study, can reduce this effect.

Only the reproducibility within an institution was assessed. Hence, the reproducibility of measurements between institutions cannot be gleaned from these data, particularly because patients with vascular disease were studied. Thus, unlike estimates from healthy volunteers, flow values and vascular reactivity are expected to vary from patient to patient, and flow values determined at one institution with one group of patients are therefore not directly comparable with flow values from another group of patients in another institution. A realistic brain phantom, such as recently developed by our group, simulating head contour with bone attenuation, could be used to assess the consistency of brain images between institutions.

CONCLUSION

The developed QSPECT package allows absolute CBF and CVR to be estimated in routine clinical studies. This multicenter study has demonstrated the applicability of QSPECT for a variety of clinical settings and equipment. Results from the studies suggest that a change of approximately 10% or 5 mL/min/100 g can be readily detected in follow-up studies. The graphical user interface for easily controlling the in-built sophisticated programs and tools ensures that routine use does not require dedicated support from scientific or computing staff. The package is now successfully used in over 130 institutions in Japan, and more than 25,000 patient studies have been analyzed with the QSPECT package.

Acknowledgments

We thank the staff of each of the following institutions that participated in this project for their invaluable help with supporting the SPECT studies: Azabu Neurosurgical Hospital, Sapporo City; Asahikawa Red Cross Hospital, Asahikawa City; Handa City Hospital, Handa City; Ichinomiya Municipal Hospital, Ichinomiya City; Kashiwaba Neurosurgical Hospital, Sapporo City; Japanese Red Cross Kobe Hospital, Kobe City; Nakamura Memorial Hospital, Sapporo City; National Cardiovascular Center, Osaka; Ogori Daiichi General Hospital, Yamaguchi City; Oji General Hospital, Tomakomai City; Sunagawa City Medical Center, Sunagawa City; and Teine Keijinkai Hospital, Sapporo City. The present study was supported by the Japan Cardiovascular Research Foundation and a grant for Translational Research from the Ministry of Health, Labor and Welfare (MHLW), Japan.

REFERENCES

  1. 1.↵
    1. Iida H,
    2. Akutsu T,
    3. Endo K,
    4. et al
    . A multicenter validation of regional cerebral blood flow quantitation using [123I]iodoamphetamine and single photon emission computed tomography. J Cereb Blood Flow Metab. 1996;16:781–793.
    OpenUrlCrossRefPubMed
  2. 2.
    1. Hatazawa J,
    2. Iida H,
    3. Shimosegawa E,
    4. Sato T,
    5. Murakami M,
    6. Miura Y
    . Regional cerebral blood flow measurement with iodine-123-IMP autoradiography: normal values, reproducibility and sensitivity to hypoperfusion. J Nucl Med. 1997;38:1102–1108.
    OpenUrlAbstract/FREE Full Text
  3. 3.↵
    1. Yamaguchi T,
    2. Kanno I,
    3. Uemura K,
    4. et al
    . Reduction in regional cerebral metabolic rate of oxygen during human aging. Stroke. 1986;17:1220–1228.
    OpenUrlAbstract/FREE Full Text
  4. 4.↵
    1. Ogasawara K,
    2. Ito H,
    3. Sasoh M,
    4. et al
    . Quantitative measurement of regional cerebrovascular reactivity to acetazolamide using 123I-N-isopropyl-p-iodoamphetamine autoradiography with SPECT: validation study using H215O with PET. J Nucl Med. 2003;44:520–525.
    OpenUrlAbstract/FREE Full Text
  5. 5.↵
    1. Ogasawara K,
    2. Ogawa A,
    3. Terasaki K,
    4. Shimizu H,
    5. Tominaga T,
    6. Yoshimoto T
    . Use of cerebrovascular reactivity in patients with symptomatic major cerebral artery occlusion to predict 5-year outcome: comparison of xenon-133 and iodine-123-IMP single-photon emission computed tomography. J Cereb Blood Flow Metab. 2002;22:1142–1148.
    OpenUrlCrossRefPubMed
  6. 6.↵
    1. Ogasawara K,
    2. Ogawa A,
    3. Yoshimoto T
    . Cerebrovascular reactivity to acetazolamide and outcome in patients with symptomatic internal carotid or middle cerebral artery occlusion: a xenon-133 single-photon emission computed tomography study. Stroke. 2002;33:1857–1862.
    OpenUrlAbstract/FREE Full Text
  7. 7.↵
    1. Ogasawara K,
    2. Ogawa A
    . JET study (Japanese EC-IC Bypass Trial) [in Japanese]. Nippon Rinsho. 2006;64(suppl 7):524–527.
    OpenUrl
  8. 8.↵
    1. Fujita M,
    2. Ichise M,
    3. Zoghbi SS,
    4. et al
    . Widespread decrease of nicotinic acetylcholine receptors in Parkinson's disease. Ann Neurol. 2006;59:174–177.
    OpenUrlCrossRefPubMed
  9. 9.↵
    1. Deloar HM,
    2. Watabe H,
    3. Kudomi N,
    4. Kim KM,
    5. Aoi T,
    6. Iida H
    . Dependency of energy and spatial distributions of photons on edge of object in brain SPECT. Ann Nucl Med. 2003;17:99–106.
    OpenUrlCrossRefPubMed
  10. 10.↵
    1. Iida H,
    2. Eberl S,
    3. Kim KM,
    4. et al
    . Absolute quantitation of myocardial blood flow with 201Tl and dynamic SPECT in canine: optimisation and validation of kinetic modelling. Eur J Nucl Med Mol Imaging. 2008;35:896–905.
    OpenUrlCrossRefPubMed
  11. 11.↵
    1. Iida H,
    2. Eberl S
    . Quantitative assessment of regional myocardial blood flow with thallium-201 and SPECT. J Nucl Cardiol. 1998;5:313–331.
    OpenUrlCrossRefPubMed
  12. 12.↵
    1. Eberl S,
    2. Chan HK,
    3. Daviskas E,
    4. Constable C,
    5. Young I
    . Aerosol deposition and clearance measurement: a novel technique using dynamic SPET. Eur J Nucl Med. 2001;28:1365–1372.
    OpenUrlCrossRefPubMed
  13. 13.↵
    1. Hapdey S,
    2. Soret M,
    3. Ferrer L
    . Quantification in SPECT: myth or reality? A multicentric study. IEEE Nucl Sci Symp Conf Rec. 2004;5:3170–3317.
    OpenUrlCrossRef
  14. 14.↵
    1. Graham LS,
    2. Fahey FH,
    3. Madsen MT,
    4. van Aswegen A,
    5. Yester MV
    . Quantitation of SPECT performance: report of Task Group 4, Nuclear Medicine Committee. Med Phys. 1995;22:401–409.
    OpenUrlCrossRefPubMed
  15. 15.↵
    1. Hendel RC,
    2. Corbett JR,
    3. Cullom SJ,
    4. DePuey EG,
    5. Garcia EV,
    6. Bateman TM
    . The value and practice of attenuation correction for myocardial perfusion SPECT imaging: a joint position statement from the American Society of Nuclear Cardiology and the Society of Nuclear Medicine. J Nucl Cardiol. 2002;9:135–143.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Zaidi H,
    2. Koral KF
    . Scatter modelling and compensation in emission tomography. Eur J Nucl Med Mol Imaging. 2004;31:761–782.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Hudson HM,
    2. Larkin RS
    . Accelerated image reconstruction using ordered subsets of projection data. IEEE Trans Med Imaging. 1994;13:601–609.
    OpenUrlCrossRefPubMed
  18. 18.↵
    1. Meikle SR,
    2. Hutton BF,
    3. Bailey DL
    . A transmission-dependent method for scatter correction in SPECT. J Nucl Med. 1994;35:360–367.
    OpenUrlAbstract/FREE Full Text
  19. 19.↵
    1. Iida H,
    2. Narita Y,
    3. Kado H,
    4. et al
    . Effects of scatter and attenuation correction on quantitative assessment of regional cerebral blood flow with SPECT. J Nucl Med. 1998;39:181–189.
    OpenUrlAbstract/FREE Full Text
  20. 20.↵
    1. Narita Y,
    2. Eberl S,
    3. Iida H,
    4. et al
    . Monte Carlo and experimental evaluation of accuracy and noise properties of two scatter correction methods for SPECT. Phys Med Biol. 1996;41:2481–2496.
    OpenUrlCrossRefPubMed
  21. 21.↵
    1. Narita Y,
    2. Iida H,
    3. Eberl S,
    4. Nakamura T
    . Monte Carlo evaluation of accuracy and noise properties of two scatter correction methods for 201Tl cardiac SPECT. IEEE Trans Nucl Sci. 1997;44:2465–2472.
    OpenUrlCrossRef
  22. 22.↵
    1. Willowson K,
    2. Bailey DL,
    3. Baldock C
    . Quantitative SPECT reconstruction using CT-derived corrections. Phys Med Biol. 2008;53:3099–3112.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Kim K,
    2. Watabe H,
    3. Hayashi T,
    4. et al
    . Quantitative mapping of basal and vasareactive cerebral blood flow using split-dose 123I-idoamphetamine and single photon emission computed tomography. Neuroimage. 2006;33:1126–1135.
    OpenUrlCrossRefPubMed
  24. 24.↵
    1. Kim KM,
    2. Watabe H,
    3. Shidahara M,
    4. Ishida Y,
    5. Iida H
    . SPECT collimator dependency of scatter and validation of transmission dependent scatter compensation methodologies. IEEE Trans Nucl Sci. 2001;48:689–696.
    OpenUrl
  25. 25.↵
    1. Iida H,
    2. Itoh H,
    3. Nakazawa M,
    4. et al
    . Quantitative mapping of regional cerebral blood flow using iodine-123-IMP and SPECT. J Nucl Med. 1994;35:2019–2030.
    OpenUrlAbstract/FREE Full Text
  26. 26.
    1. Iida H,
    2. Nakazawa M,
    3. Uemura K
    . Quantitation of regional cerebral blood flow using 123I-IMP from a single SPECT scan and a single blood sampling: analysis on statistical error source and optimal scan time [in Japanese]. Kaku Igaku. 1995;32:263–270.
    OpenUrlPubMed
  27. 27.
    1. Kurisu R,
    2. Ogura T,
    3. Takikawa S,
    4. Saito H,
    5. Nakazawa M,
    6. Iida H
    . Estimation and optimization of the use of standard arterial input function for split-dose administration of N-isopropyl-p[123I]iodoamphetamine [in Japanese]. Kaku Igaku. 2002;39:13–20.
    OpenUrlPubMed
  28. 28.↵
    1. Ogura T,
    2. Takikawa S,
    3. Saito H,
    4. Nakazawa M,
    5. Shidahara M,
    6. Iida H
    . Validation and optimization of the use of standardized arterial input function in N-isopropyl-p[123I]iodoamphetamine cerebral blood flow SPECT [in Japanese]. Kaku Igaku. 1999;36:879–890.
    OpenUrlPubMed
  29. 29.↵
    1. Shidahara M,
    2. Watabe H,
    3. Kim KM,
    4. et al
    . Evaluation of a commercial PET tomograph-based system for the quantitative assessment of rCBF, rOEF and rCMRO2 by using sequential administration of 15O-labeled compounds. Ann Nucl Med. 2002;16:317–327.
    OpenUrlCrossRefPubMed
  30. 30.↵
    1. Iida H,
    2. Higano S,
    3. Tomura N,
    4. et al
    . Evaluation of regional differences of tracer appearance time in cerebral tissues using [15O] water and dynamic positron emission tomography. J Cereb Blood Flow Metab. 1988;8:285–288.
    OpenUrlPubMed
  31. 31.
    1. Iida H,
    2. Kanno I,
    3. Miura S,
    4. Murakami M,
    5. Takahashi K,
    6. Uemura K
    . Error analysis of a quantitative cerebral blood flow measurement using H215O autoradiography and positron emission tomography, with respect to the dispersion of the input function. J Cereb Blood Flow Metab. 1986;6:536–545.
    OpenUrlPubMed
  32. 32.↵
    1. Iida H,
    2. Kanno I,
    3. Miura S,
    4. Murakami M,
    5. Takahashi K,
    6. Uemura K
    . A determination of the regional brain/blood partition coefficient of water using dynamic positron emission tomography. J Cereb Blood Flow Metab. 1989;9:874–885.
    OpenUrlPubMed
  33. 33.↵
    1. Iida H,
    2. Shoji Y,
    3. Sugawara S,
    4. et al
    . Design and experimental validation of a quantitative myocardial 201Tl SPECT System. IEEE Trans Nucl Sci. 1999;46:720–726.
    OpenUrlCrossRef
  34. 34.↵
    1. Narita Y,
    2. Iida H
    . Scatter correction in myocardial thallium SPECT: needs for optimization of energy window settings in the energy window-based scatter correction techniques [in Japanese]. Kaku Igaku. 1999;36:83–90.
    OpenUrlPubMed
  35. 35.↵
    1. Eberl S,
    2. Kanno I,
    3. Fulton RR,
    4. Ryan A,
    5. Hutton BF,
    6. Fulham MJ
    . Automated interstudy image registration technique for SPECT and PET. J Nucl Med. 1996;37:137–145.
    OpenUrlAbstract/FREE Full Text
  • Received for publication April 27, 2010.
  • Accepted for publication July 14, 2010.
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Bookmark this article

Jump to section