DOI 10.1007/s00330-015-3667-6 NUCLEAR MEDICINE Robust, fully automatic delineation of the head contour by stereotactical normalization for attenuation correction according to Chang in dopamine transporter scintigraphy Catharina Lange & Jens Kurth & Anita Seese & Sarah Schwarzenböck & Karen Steinhoff & Bert Umland-Seidler & Bernd J. Krause & Winfried Brenner & Osama Sabri & Swen Hesse & Ralph Buchert Received: 29 September 2014 /Revised: 27 January 2015 /Accepted: 12 February 2015 # European Society of Radiology 2015 Abstract Objectives Chang s method, the most widely used attenuation correction (AC) in brain single-photon emission computed tomography (SPECT), requires delineation of the outer contour of the head. Manual and automatic threshold-based methods are prone to errors due to variability of tracer uptake in the scalp. The present study proposes a new method for fully automated delineation of the head based on stereotactical normalization. The method was validated for SPECT with I-123-ioflupane. Methods The new method was compared to threshold-based delineation in 62 unselected patients who had received I-123- ioflupane SPECT at one of 3 centres. The impact on diagnostic power was tested for semi-quantitative analysis and visual reading of the SPECT images (six independent readers). Results The two delineation methods produced highly consistent semi-quantitative results. This was confirmed by receiver operating characteristic analyses in which the putamen specific-to-background ratio achieved highest area under the curve with negligible effect of the delineation method: 0.935 versus 0.938 for stereotactical normalization and thresholdbased delineation, respectively. Visual interpretation of DVR images was also not affected by the delineation method. Conclusions Delineation of the head contour by stereotactical normalization appears useful for Chang AC in I-123-ioflupane SPECT. It is robust and does not require user interaction. Key Points Chang attenuation correction in brain SPECT requires delineation of the head contour. Manual and threshold-based methods are prone to errors. The study proposes a fully-automated method for delineation based on stereotactical normalization. The method is shown to work reliably in I-123-ioflupane SPECT. It might improve the workflow of I-123-ioflupane SPECT in everyday patient care. Keywords I-123-ioflupane. Attenuation correction. Chang. Contour delineation. Stereotactical normalization C. Lange: W. Brenner : R. Buchert (*) Department of Nuclear Medicine, Charité - Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany e-mail: ralph.buchert@charite.de J. Kurth: S. Schwarzenböck : B. J. Krause Department of Nuclear Medicine, Universitätsmedizin Rostock, 18057 Rostock, Germany A. Seese: K. Steinhoff : O. Sabri : S. Hesse Department of Nuclear Medicine, Universitätsklinikum Leipzig, 04103 Leipzig, Germany B. Umland-Seidler GE Healthcare Buchler GmbH & Co. KG, 80807 Munich, Germany Introduction Single photon emission computed tomography (SPECT) with I-123-labelled cocaine ligands for the presynaptic dopamine transporter (DAT) is widely used for the diagnosis of Parkinsonian syndromes [1 9]. The interpretation of DAT SPECT is based on visual image evaluation supported by semiquantitative analysis [10 14]. Both visual and semiquantitative analyses are affected by various physical factors, including photon attenuation in the head [15]. These factors not only cause distortion of the SPECT images but also
additional, inter-subject variability of tracer uptake which might reduce the ability to detect disease-related alterations. There are several methods for attenuation correction (AC) in brain SPECT. The most widely used method, proposed by Chang [16], is based on post-processing of images reconstructed without AC. The major limitation of Chang s method is related to the fact that the outer contour of the head must be determined from the SPECT image. This can be done by semiautomatic thresholding or by manual delineation in every single transversal slice [17 20]. Both approaches are prone to error, particularly in case of highly specific tracers providing highly variable anatomical delineation of the scalp, for example, the DAT ligand N-ω-fluoropropyl-2β-carbomethoxy- 3β-(4-I-123-iodophenyl)nortropane (I-123-ioflupane). I-123- ioflupane is registered in both the United States and in Europe. In 2013, about 67,500 patients received brain SPECT with I-123-ioflupane in Europe, making SPECT with I-123- ioflupane one of the most frequent nuclear medicine brain imaging procedures. With the increasing availability of SPECT/CT hybrid systems, AC of I-123-ioflupane SPECTcan also be based on lowdose transmission CT. However, in a recent study we found that the impact of CT-based AC versus Chang AC on the interpretation of I-123-ioflupane SPECT is negligible [21]. Therefore, CT-based AC cannot be recommended for routine use in clinical patient care because of the additional radiation exposure caused by the additional low-dose CT [22]. This underlines the clinical relevance of methods for simplifying Chang AC. Therefore, the aim of the present study was to propose and validate a robust method for fully automatic delineation of the head for use in Chang AC of I-123-ioflupane SPECT. The novel method is based on stereotactical normalization of the patient s brain SPECT image, which is increasingly used not only in preparation of voxel-based statistical testing but also for standardized semi-quantitative analysis [23]. Materials and methods Delineation of the head by stereotactical normalization The method is based on stereotactical normalization of the patient's I-123-ioflupane SPECT into the anatomical space of the Montreal Neurological Institute (MNI) using the normalization tool provided by the freely available Statistical Parametric Mapping software package SPM8 [24] anda custom-made tracer-specific template (Figs. 1 and 2). Stereotactical normalization is performed using SPM8 s default settings, except for source smoothing = 20 mm, nonlinear frequency cut-off = 50, non-linear iterations = 12, and bounding box = [ 90-126 -72; 91 91 109]. The inverse of the normalization transformation is used to map a predefined binary head mask from MNI space to the patient s anatomical space. The standard head mask was generated from the avg152t1 T1-weighted magnetic resonance image in MNI space provided by SPM8. First, the avg152t1 was smoothed with a Gaussian filter of 8 8 8 mm 3 fullwidth-at-half-maximum. Then, the binary mask was obtained by thresholding the smoothed avg152t1 image at 0.15. Small gaps inside the head mask were filled manually using PMOD (PMOD Technologies Ltd., Zurich, Switzerland) [25]. The edges of the transformed mask are assumed to delineate the patient s head and, thus, are used for Chang AC. The 0th order Chang AC using the individually transformed standard head mask was implemented in MATLAB (The Mathworks, Inc., Nattick, MA, USA). Subjects The study included 62 unselected subjects from routine clinical patient care who had been referred to SPECT with I-123- ioflupane because of suspected neurodegenerative etiology of a clinically uncertain Parkinsonian symptom. Imaging had been performed in one of three centres (Berlin, n=21; Leipzig, n=22; Rostock, n=19). The protocol of this retrospective study was approved by the local Ethics Committees. The same sample of patients had previously been used to evaluate AC based on low-dose transmission CT in I-123-ioflupane SPECT [21]. The new method for fully automated delineation of the head based on stereotactical normalization of the patient s I-123-ioflupane SPECT image has not been described before. SPECT imaging SPECT imaging was performed with a Symbia T6 dual-head SPECT/CT equipped with low-energy high-resolution parallel-hole collimators at each centre. 150 to 200 MBq I-123-ioflupane was injected intravenously after blocking the thyroid. SPECT acquisition of 30 min duration was started between 3 and 4 hours post injection. Anonymized raw data were transferred to one centre (Berlin) for centralized image reconstruction and processing. The iterative reconstruction algorithm Flash-3D of the scanner software was used with 15 subsets and 6 iterations [26 28]. Scatter correction was performed using the triple-energywindow approach implemented in the scanner software [29]. For each patient, SPECT images were reconstructed with Chang AC using either the threshold-based delineation of the head contour implemented in the scanner software (in the following denoted as tchang ) or the novel delineation method based on stereotactical normalization ( snchang ), both with the narrow-beam attenuation coefficient μ= 0.148 cm 1.Thisresultedin2 62=124SPECTimages. Semi-automatic, threshold-based delineation of the head
Fig. 1 Workflow of 0th order Chang attenuation correction (AC) with delineation of the head by stereotactical normalization ( snchang ). The procedure shown is performed twice. The first iteration starts with the original patient image without AC. In the second iteration, the ACcorrected image obtained from the first iteration is used as input. The rationale for the iteration is that the original I-123-ioflupane image without AC has slightly different image characteristics than the ACcorrected template, which might result in suboptimal stereotactical normalization. This potential error is reduced by the second iteration. (SPM8: version 8 of the Statistical Parametric Mapping software package; MNI: anatomical space of the Montreal Neurological Institute; μ: attenuation coefficient) contour started with the default parameter settings, i.e. projection angle=2, edge thickness=0.5, background threshold=15 and smoothing kernel=5 5 mm 2. The resulting head contour was examined in each slice and all parameters were adjusted manually until the contour fitted the head in all slices according to visual inspection. Data evaluation Impact on semi-quantitative analysis For semi-quantitative assessment of DAT availability, the stereotactically normalized I-123-ioflupane uptake image was scaled to the 75th percentile of the voxel intensities in the whole brain without striata as a reference region, i.e. each voxel value was divided by the 75th percentile of voxel values in the reference region. The reference region was predefined as a binary mask in MNI space by cutting the regions-ofinterest (ROIs) for the left and right caudate and putamen from the standard SPM8 brain mask (Fig. 2). The rationale for using the 75th percentile rather than the mean (or median) to characterize the I-123-ioflupane uptake in the reference region is that the 75th percentile mainly represents cortical gray matter, excluding white matter and cerebrospinal fluid [30]. The voxel intensities of the scaled SPECT image represent the distribution volume ratio (DVR).
Fig. 2 Top: Transversal slices of the custom-made I-123-ioflupane template. Middle: Regions of interest (ROI) for left/right caudate (green/orange) and putamen (yellow/red) used for hottest voxel analysis and ROI for the reference region used for intensity scaling (blue), all defined in the standard space. Bottom: Fusion image. The DVR in the caudate and putamen was obtained by hottest voxel analysis in large ROIs predefined in MNI space (Fig. 2) which guarantee that the striatum is completely included, independent of some residual anatomical inter-subject variability after stereotactical normalization. The ROIs do not include other brain structures with high tracer uptake, such as the thalamus or midbrain. In order to restrict averaging of voxel intensities in the ROIs to the striatum (to avoid dilution effects), only the voxels with the highest DVR were included. The number of hottest voxels to be averaged was fixed to a total volume of 5 ml for the caudate and 10 ml for the putamen. The DVR of the whole striatum was obtained by averaging over the 15 ml hottest voxels in the union of the caudate ROI and the putamen ROI. DVRs were converted to specific binding ratios (SBRs) accordingtosbr=dvr 1. The SBR estimates the nondisplaceable binding potential that is proportional to the density of DAT available for binding of I-123-ioflupane [31]. The caudate-to-putamen ratio (= SBR caudate /SBR putamen ) in both hemispheres and left/right asymmetry (asym(%)= 200 abs[(sbr left SBR right )/(SBR left +SBR right )]) were also considered. Impact on visual scoring For retrospective visual evaluation, a portable document format (pdf) document with 124 pages was prepared, one page for each SPECT image (Fig. 3a). The SPECT images were anonymized and presented in randomized order. To guarantee comparable display conditions, the upper threshold of the colour table was adjusted separately for each delineation method. For tchang, the upper threshold of the colour table was set to 5.50. For snchang, the upper threshold of the colour table was scaled by the DVR of the caudate (mean over left and right hemisphere) averaged over all patients, i.e. threshold(snchang) = avgdvr(snchang)/ avgdvr(tchang) threshold(tchang). The lower threshold of the colour table was set to one-tenth of the upper threshold in both cases.
Fig. 3 (a): Example page from the pdf document for visual scoring. The pdf document comprised one page for each I-123-ioflupane image showing a 12-mm thick slab (left) and 4 4 slices of 4-mm thickness (right). (b): Example I-123-ioflupane images used as reference images for the visual scoring. (R: right; L: left) Visual scoring of DAT availability was performed on patient base referring to Benamer et al. [32], using the following five-point score: normal : clear delineation of both striata, minor global reduction and minor left/right asymmetry allowed; reduced type 1 : distinct reduction in one putamen ( big effect); reduced type 2 : distinct reduction in both putamina; reduced type 3 : essentially no uptake in both striata; reduced other : clear reduction of tracer uptake, but atypical pattern not matching type 1, 2 or 3. The examples shown in Fig. 3b were provided to the raters. Certainty with respect to the differentiation between normal and reduced (including all types of reduction) was scored from 1= very sure to 5= very unsure. Image quality (delineation of the striata and statistical noise in the background) was scored from 1 = very good to 5 = very bad. Visual scoring was performed by two independent raters at each centre. In addition, the two raters at each centre reached a consensus with respect to DAT availability in those cases in which they had disagreed. Image quality was scored in Leipzig only. Statistics Statistical analyses were performed with SPSS (version 21, IBM Corp., Armonk, NY, USA). The effect of the delineation method on the SBR, left/right asymmetry and caudate-to-putamen ratio was assessed by Bland-Altman plots [33]. The impact on diagnostic power was tested by receiver operating characteristic (ROC) analysis. The area under the curve (AUC) was used as a performance measure. Classification of DAT availability in the written report of I-123-ioflupane SPECT in the patient s file served as a gold standard (22 patients with normal, 40 patients with reduced DAT availability). Multiple binary logistic regression was performed to evaluate the diagnostic power of combinations of SBR, asymmetry, and caudate-to-putamen ratio. The logistic model was estimated iteratively using the forward conditional approach. The effect of the delineation method on the certainty of the visual differentiation between normal and reduced DAT availability as well as on image quality was tested by the general linear model for repeated measures. DAT availability (normal or reduced) according to the report (gold standard) was added to the model as a within-subject factor. Results Impact on semi-quantitative analysis Results of the semi-quantitative analysis are summarized in Table 1. Bland-Altman plots for the comparison of SBRs between snchang and tchang are shown in Fig. 4. The mean difference
Table 1 Results (mean values±one standard deviation) of the semiquantitative analysis after Chang with delineation of the head contour by stereotactical normalization (snchang) or a threshold-based approach (tchang). The minimum over both hemispheres is given for reduced DAT availability (n=40) the specific binding ratios (SBRs), and the maximum for the caudateto-putamen ratio. Subjects were categorized as reduced or normal dopamine transporter (DAT) availability according to the written report in the patient s file. normal DAT availability (n=22) snchang tchang snchang tchang SBR caudate 2.44±0.77 2.42±0.83 3.72±0.60 3.67±0.65 SBR putamen 1.07±0.46 1.07±0.46 2.04±0.28 2.00±0.25 SBR striatum 1.79±0.61 1.78±0.65 2.95±0.42 2.91±0.45 caudate-to-putamen ratio 2.56±0.66 2.53±0.64 1.91±0.28 1.91±0.28 asymmetry caudate (%) 11.02±10.15 10.72±10.16 9.92±8.72 9.94±9.20 asymmetry 17.07±13.64 17.78±14.76 10.47±16.40 10.94±14.96 putamen (%) asymmetry striatum (%) 11.90±10.76 11.56±10.83 8.06±7.45 8.53±7.55 of the SBR in the caudate was 0.030±0.114 (one sample t-test for zero mean: p=0.004), and in the putamen it was 0.008± 0.079 (p=0.276). The mean difference of left/right asymmetry (in %) in the caudate was 0.19±2.12 (p=0.480), in the putamen, it was 0.63±5.53 (p=0.376), and in the whole striatum 0.06±2.32 (p=0.851). The mean difference of the caudate-toputamen ratio was 0.027±0.115 (p=0.011). ROC curves for the differentiation between reduced and normal DAT availability by the SBRs are shown in Fig. 5. The putamen provided a larger AUC than the caudate. The impact of the delineation method on the AUC was very small: for the caudate 0.897 versus 0.887, and for the putamen 0.935 versus 0.938 for snchang and tchang, respectively. Asymmetries and caudate-to-putamen ratio provided much less diagnostic power than the SBR (AUC 0.836). The binary regression model included only the putamen SBR (minimum over both hemispheres). The asymmetry of the SBR in the whole striatum and the caudate-to-putamen ratio (maximum over both hemispheres) were not included, independent of the delineation method (asymmetry: p 0.415, caudate-to-putamen ratio: p 0.179). The number of falsely classified patients was six and nine for snchang and tchang, respectively. Impact on visual scoring Fig. 4 Bland-Altman plots comparing the specific binding ratio (SBR) of the caudate (a) and the putamen (b) between snchang and tchang (SBRs of both hemispheres were included independently). Different scales were chosen for abscissae and ordinates in (a) and (b) for display purposes. The horizontal continuous line represents the mean difference, the dashed lines indicate the 95 % confidence interval. The given p-value corresponds to the one-sample t-test for zero mean. (snchang: head delineation by stereotactical normalization; tchang: threshold-based head delineation) Agreement of the visual score for DAT availability between the two delineation methods was 78.2±2.8 % for the full fivepoint score and 88.2±4.8 % for the binary differentiation between normal and reduced (averaged over all 6 raters). Agreement was improved by consensus between the two raters at each centre to 81.7±3.7 % for the five-point score and 89.8± 4.1 % for the binary decision. Diagnostic accuracy averaged over all raters was 79.6±3.3 % and 79.6±1.7 % for snchang and tchang, respectively. Diagnostic accuracy of the consensus was 82.3±3.2 % and 79.6±2.5 %.
Fig. 5 Receiver operating characteristic (ROC) curves for the differentiation between reduced and normal dopamine transporter (DAT) availability by the specific binding ratio of the caudate (a) and the putamen (b) (minimum over both hemispheres). (snchang: head delineation by stereotactical normalization; tchang: thresholdbased head delineation) The delineation method had no effect on the rater s certainty in the differentiation between reduced and normal DAT availability (p=0.956). However, there was a significant effect of the status of DAT availability: the certainty of the visual scoring was lower in patients with normal DAT availability (score averaged over all raters and both delineation methods 2.44±0.89 versus 1.23±0.60 in patients with normal and reduced DAT availability, p=0.000). Image quality was also not affected by the delineation method (image quality score of 2.16±0.49 and 2.23±0.53 with snchang and tchang, respectively, p=0.213, Fig. 6). the putamen SBR with the left/right asymmetries and/or the caudate-to-putamen ratio did not improve diagnostic power compared to the putamen SBR alone. This reflects the fact that reduction of putaminal DAT availability occurs more consistently in the course of the disease than specific patterns of the reduction, such as left/right asymmetry or a caudate-toputamen gradient. To some extent, this is explained by the heterogeneity of neurodegenerative Parkinsonian syndromes, including Morbus Parkinson, multiple systems atrophy, progressive supranuclear palsy and corticobasal degeneration. Discussion Delineation of the head by stereotactical normalization worked properly in all subjects, i.e. visual inspection of the standard head mask after transformation into the patient s anatomical space side-by-side with the patient s SPECT did not show any outliers, although there was no exclusion criterion with respect to image quality of the SPECTs. This demonstrates the robustness of the method, an important prerequisite for use in everyday clinical routine. Bland-Altman plots demonstrated Chang AC with head delineation based on stereotactical normalization ( snchang ) provides essentially the same semi-quantitative results as the threshold-based delineation with manual optimization of the thresholding parameters in each individual patient ( tchang ) (Fig. 4). Because of the rather large sample size, the difference reached statistical significance for some of the tested semiquantitative parameters. However, the relative difference was smaller than test-retest variability due to physiological variability, i.e. normal intra-subject variability of DAT availability for I-123-ioflupane [34, 35]. The putamen provided the highest power for differentiation between normal and reduced DAT availability amongst the tested ROIs. This is in agreement with the time course of neurodegeneration in Parkinson s disease[36, 37]. Combining Fig. 6 Top: Slab of 12-mm thickness of the scaled, stereotactically normalized I-123-ioflupane (DVR) image averaged over all patients with normal dopamine transporter (DAT) availability (n=22). Bottom: Slab displaying the coefficient of variation (%) of the DVR over all patients with normal DAT availability. (DVR: distribution volume ratio; snchang: head delineation by stereotactical normalization; tchang: threshold-based head delineation)
Semi-quantitative analyses were confirmed by the results of visual reading. Differentation between normal and reduced DAT availability based on the visual reading of SPECT images was in agreement between the two delination methods in about 90 % of the cases. Discrepancy in about 10 % of the cases most likely was caused by intra-rater (test-retest) variability in the interpretation of the SPECT images. This was suggested by retrospective visual evaluation of the cases with discrepant visual scores: the images were indistinguishable in almost all cases. Adjustment of the colour table separately for each delineation method, although small, might have contributed to the lack of an effect of the delineation method on visual scoring of DAT availability, and, therefore, is important (Fig. 6). The delineation method also did not affect the certainty of the visual reading. However, the status of DAT availability showed a significant effect: the certainty was considerably higher in patients with reduced DAT availability than in patients with normal DAT availability. This bias is caused by the rules for visual scoring: the diagnosis of reduced DAT availability was based on the detection of a big effect to account for the fact that clinical symptoms occur only when 50 % or more of the dopaminergic neurons are lost [38, 39]. A small reduction of DAT availability might be physiological variation or due to non-neurodegenerative causes [34]. This conservative interpretation of I-123-ioflupane images results in an increased rate of uncertain cases in the group of normal DAT availability. A limitation of the present study is that the diagnosis in the original report was used as a gold standard for the evaluation of diagnostic accuracy, since a final clinical diagnosis on the basis of follow-ups was not available. However, we do not consider this a major limitation, because limitations of the gold standard mainly affect the absolute value of the diagnostic accuracy. The relative diagnostic accuracy with snchang compared to tchang ( smaller, equal or larger accuracy) most likely is not affected. Therefore, the main result of the study, namely, that the impact of snchang versus tchang on diagnostic accuracy is negligible, is not affected. In addition, semi-quantitative analyses (Bland-Altman plots) as well as percent agreement, certainty and image quality according to the visual reading are also not affected by the gold standard. Conclusion Chang AC with delineation of the outer contour of the head based on stereotactical normalization was found to provide essentially the same results as Chang AC using semiautomatic threshold-based delineation with manual optimization in each patient. Delineation of the head by stereotactical normalization is fully automatic, i.e. it does not require any input from the user. Therefore, there is no inter- and intraoperator variability with this method, results are fully reproducible. This is in contrast to manual or semi-automatic threshold-based methods that require user input and, therefore, are prone to inter- and intra-user variability. In addition, fully automatic processing saves time for the technician or physician. Thus, the novel method for fully automatic and robust delineation of the head has the potential to improve the workflow of I-123-ioflupane SPECT in clinical routine patient care. Acknowledgments The scientific guarantor of this publication is Ralph Buchert. The authors of this manuscript declare relationships with the following companies: Bert Umland-Seidler is an employee of GE Healthcare Buchler GmbH & Co. KG. The authors state that this work has not received any funding. No complex statistical methods were necessary for this paper. Institutional Review Board approval was obtained. Written informed consent was waived by the Institutional Review Board. Study subjects were previously reported in Lange C, Seese A, Schwarzenbock S, et al. (2014) CT-Based Attenuation Correction in I- 123-Ioflupane SPECT. PLoS One 9:e108328. Methodology: retrospective, diagnostic study, multicenter study. Conflict of interest interest. References The authors declare that they have no conflict of 1. Bartenstein P, Grunwald F, Kuwert T et al (2000) Clinical applications of single photon emission tomography in neuromedicine. 1. Neuro-oncology, epilepsy, movement disorders, cerebrovascular disease. Nuklearmedizin 39:180 195 2. Booij J, Speelman JD, Horstink MW, Wolters EC (2001) The clinical benefit of imaging striatal dopamine transporters with [123I]FP-CIT SPET in differentiating patients with presynaptic parkinsonism from those with other forms of parkinsonism. Eur J Nucl Med 28:266 272 3. Booij J, Tissingh G, Boer GJ et al (1997) [123I]FP-CIT SPECT shows a pronounced decline of striatal dopamine transporter labelling in early and advanced Parkinson's disease. J Neurol Neurosurg Psychiatry 62:133 140 4. Hesse S, Oehlwein C, Meyer PT et al (2003) Is there a role for I-123- FP-CIT SPECT in the management of suspected Parkinson's disease? J Nucl Med 44:234p 235p 5. Innis RB, Seibyl JP, Scanley BE et al (1993) Single photon emission computed tomographic imaging demonstrates loss of striatal dopamine transporters in Parkinson disease. Proc Natl Acad Sci U S A 90: 11965 11969 6. Seibyl JP, Marek KL, Quinlan D et al (1995) Decreased singlephoton emission computed tomographic [123I]beta-CIT striatal uptake correlates with symptom severity in Parkinson's disease. Ann Neurol 38:589 598 7. Tatsch K, Poepperl G (2013) Nigrostriatal dopamine terminal imaging with dopamine transporter SPECT: an update. J Nucl Med 54: 1331 1338 8. Van Laere K, Everaert L, Annemans L, Gonce M, Vandenberghe W, Vander Borght T (2008) The cost effectiveness of 123I-FP-CIT SPECT imaging in patients with an uncertain clinical diagnosis of parkinsonism. Eur J Nucl Med Mol Imaging 35:1367 1376 9. Walker Z, Costa DC, Walker RW et al (2002) Differentiation of dementia with Lewy bodies from Alzheimer's disease using a
dopaminergic presynaptic ligand. J Neurol Neurosurg Psychiatry 73: 134 140 10. Darcourt J, Booij J, Tatsch K et al (2010) EANM procedure guidelines for brain neurotransmission SPECT using (123)I-labelled dopamine transporter ligands, version 2. Eur J Nucl Med Mol Imaging 37: 443 450 11. Djang DS, Janssen MJ, Bohnen N et al (2012) SNM practice guideline for dopamine transporter imaging with 123I-ioflupane SPECT 1.0. J Nucl Med 53:154 163 12. Soderlund TA, Dickson JC, Prvulovich E et al (2013) Value of semiquantitative analysis for clinical reporting of 123I-2-betacarbomethoxy-3beta-(4-iodophenyl)-N-(3-fluoropropyl)nortropane SPECT studies. J Nucl Med 54:714 722 13. Tatsch K, Poepperl G (2012) Quantitative approaches to dopaminergic brain imaging. Q J Nucl Med Mol Imaging 56:27 38 14. Zubal IG, Early M, Yuan O, Jennings D, Marek K, Seibyl JP (2007) Optimized, automated striatal uptake analysis applied to SPECT brain scans of Parkinson's disease patients. J Nucl Med 48:857 864 15. Soret M, Koulibaly PM, Darcourt J, Hapdey S, Buvat I (2003) Quantitative accuracy of dopaminergic neurotransmission imaging with (123)I SPECT. J Nucl Med 44:1184 1193 16. Chang LT (1978) Method for attenuation correction in radionuclide computed tomography. Ieee Transactions on Nuclear Science 25: 638 643 17. Zaidi H, Hasegawa B (2003) Determination of the attenuation map in emission tomography. J Nucl Med 44:291 315 18. Hosoba M, Wani H, Toyama H, Murata H, Tanaka E (1986) Automated body contour detection in SPECT: effects on quantitative studies. J Nucl Med 27:1184 1191 19. Larsson SA (1980) Gamma camera emission tomography. Development and properties of a multi-sectional emission computed tomography system. Acta Radiol Suppl 363:1 75 20. Macey DJ, DeNardo GL, DeNardo SJ (1988) Comparison of three boundary detection methods for SPECT using Compton scattered photons. J Nucl Med 29:203 207 21. Lange C, Seese A, Schwarzenbock S et al (2014) CT-based attenuation correction in I-123-ioflupane SPECT. PLoS One 9:e108328 22. Hulme KW, Kappadath SC (2014) Implications of CT noise and artifacts for quantitative 99mTc SPECT/CT imaging. Med Phys 41: 042502 23. Koch W, Radau PE, Hamann C, Tatsch K (2005) Clinical testing of an optimized software solution for an automated, observerindependent evaluation of dopamine transporter SPECT studies. J Nucl Med 46:1109 1118 24. Frackowiak RSJ, Friston KJ, Frith CD et al (2004) Human brain function. Academic Press, San Diego 25. Mikolajczyk K, Szabatin M, Rudnicki P, Grodzki M, Burger C (1998) A JAVA environment for medical image data analysis: initial application for brain PET quantitation. Med Inform (Lond) 23:207 214 26. Dickson JC, Tossici-Bolt L, Sera T et al (2010) The impact of reconstruction method on the quantification of DaTSCAN images. Eur J Nucl Med Mol Imaging 37:23 35 27. Koch W, Hamann C, Welsch J, Popperl G, Radau PE, Tatsch K (2005) Is iterative reconstruction an alternative to filtered backprojection in routine processing of dopamine transporter SPECT studies? J Nucl Med 46:1804 1811 28. Winz OH, Hellwig S, Mix M et al (2012) Image quality and data quantification in dopamine transporter SPECT: advantage of 3- dimensional OSEM reconstruction? Clin Nucl Med 37:866 871 29. Ichihara T, Ogawa K, Motomura N, Kubo A, Hashimoto S (1993) Compton scatter compensation using the triple-energy window method for single-isotope and dual-isotope spect. J Nucl Med 34:2216 2221 30. Buchert R, Berding G, Wilke F et al (2006) IBZM tool: a fully automated expert system for the evaluation of IBZM SPECT studies. Eur J Nucl Med Mol Imaging 33:1073 1083 31. Innis RB, Cunningham VJ, Delforge J et al (2007) Consensus nomenclature for in vivo imaging of reversibly binding radioligands. J Cereb Blood Flow Metab 27:1533 1539 32. Benamer HTS, Patterson J, Grosset DG et al (2000) Accurate differentiation of parkinsonism and essential tremor using visual assessment of [I-123]-FP-CIT SPECT imaging: The [I-123]-FP-CIT study group. Mov Disord 15:503 510 33. Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1:307 310 34. Booij J, Habraken JB, Bergmans P et al (1998) Imaging of dopamine transporters with iodine-123-fp-cit SPECT in healthy controls and patients with Parkinson's disease. J Nucl Med 39:1879 1884 35. Tsuchida T, Ballinger JR, Vines D et al (2004) Reproducibility of dopamine transporter density measured with 123I-FPCIT SPECT in normal control and Parkinson's disease patients. Ann Nucl Med 18: 609 616 36. Bernheimer H, Birkmayer W, Hornykiewicz O, Jellinger K, Seitelberger F (1973) Brain dopamine and the syndromes of Parkinson and Huntington. Clinical, morphological and neurochemical correlations. J Neurol Sci 20:415 455 37. Oh M, Kim JS, Kim JY et al (2012) Subregional patterns of preferential striatal dopamine transporter loss differ in Parkinson disease, progressive supranuclear palsy, and multiple-system atrophy. J Nucl Med 53:399 406 38. Scherfler C, Schwarz J, Antonini A et al (2007) Role of DAT-SPECT in the diagnostic work up of parkinsonism. Mov Disord 22:1229 1238 39. Schwarz J, Storch A, Koch W, Pogarell O, Radau PE, Tatsch K (2004) Loss of dopamine transporter binding in Parkinson's disease follows a single exponential rather than linear decline. J Nucl Med 45:1694 1697