Multiple-pinhole collimators improve intra- and between-rater agreement and the certainty of the visual interpretation in dopamine transporter SPECT

Patients

The prospective study included 71 patients (62.1 ± 12.7 years, range 34–85 years, 41% females) referred to DAT-SPECT in clinical routine to support the etiological diagnosis of a clinically uncertain Parkinsonian syndrome [1, 3, 17]. The study included only patients who were able to give informed consent and had sufficiently good health status so that a second SPECT acquisition immediately after the first was considered an acceptable burden to the patient. There were no further eligibility criteria in order to guarantee that the included patient sample was representative of clinical routine at our site.

There was only very limited clinical information available for the vast majority of the patients, because 56 out of 71 (79%) patients were referred to our department for DAT-SPECT by external neurologists. However, it can be assumed that the patient sample is representative of everyday clinical routine at the Department of Nuclear Medicine of the University Medical Center Hamburg-Eppendorf, since no specific eligibility criteria were imposed for this study. From previous studies in our department that included patients in whom clinical follow-up was available [18], it might be assumed that among the patients with reduced DAT-SPECT about 90% had a disease from the spectrum of Lewy body diseases (Parkinson’s disease without or with cognitive impairment, dementia with Lewy bodies), whereas the remaining 10% suffered from an atypical neurodegenerative Parkinsonian syndrome including multiple systems atrophy, progressive supranuclear palsy, and corticobasal degeneration. The diagnoses of the patients with normal DAT-SPECT most likely included essential tremor, drug-induced Parkinsonism, various types of dystonia, psychogenic Parkinsonism, and various other diagnoses not associated with nigrostriatal degeneration.

LEHRHS-SPECT and MPH-SPECT were performed in randomized order after a single injection of 182 ± 9 MBq 123I-FP-CIT (range 163–200 MBq). Both SPECT acquisitions were performed with the same general purpose triple-head SPECT camera (AnyScan Trio SC, Mediso Medical Imaging Systems, Budapest, Hungary) in order to avoid ‘contamination’ of collimator effects by camera effects.

MPH-SPECT

The general purpose brain MPH collimator tested in this study was designed for high count sensitivity at the center of the field of view with a rather broad peak of the sensitivity profile for improved stability with respect to off-center positioning (e.g., of the striatum). The collimator features a solid tungsten aperture of 18 mm thickness with 30 pinholes arranged in 5 axially oriented columns and 6 transaxially oriented rows. The MPH aperture is mounted on a lead blind that defines the orthogonal distance between the pinhole focal plane and the detector surface to 145 mm.

A total of 90 projection views (30 per head, 120° scan arc) at angular steps of 4° were acquired in a 256 × 256 matrix with 2.13 mm × 2.13 mm pixel size. The energy window was set to 143–175 keV. The distance between the center-of-rotation axis and the pinhole focal plane was fixed to 140 mm. Helical acquisition mode was used to avoid axial undersampling [19, 20]. Helical acquisition was achieved by moving the patient table at each angular gantry step out of the gantry. The total table displacement during the SPECT acquisition was 40 mm. The total (net) duration of the MPH acquisition was 30 min.

MPH projection data were reconstructed to transaxial SPECT images with the Monte Carlo photon simulation engine and iterative one-step-late maximum-a-posteriori expectation–maximization implemented in the camera software (30 iterations, 3 subsets). A more detailed description of the reconstruction method has been given previously [16, 21]. Chang’s order zero method with linear broad-beam attenuation coefficient μ = 0.12/cm was used for post-reconstruction attenuation correction [22]. Scatter correction was not performed.

LEHRHS-SPECT

LEHRHS-SPECT was performed in double-head mode, that is, using only two of the three detector heads. The third head was switched off and moved away from the center of rotation in order to allow about 140 mm radius of rotation with the two remaining detectors. Interlaced triple-head mode [23] was not available for the SPECT camera used in this study. A total of 120 projection views (60 per head, 180° scan arc) at angular steps of 3° were acquired in a 128 × 128 matrix with 2.43 mm × 2.43 mm pixel size. The energy window was set to 143–175 keV. The radius of rotation was 146 ± 7 mm. The total (net) duration of the LEHRHS acquisition was 40 min.

Two different algorithms were used for the reconstruction of the LEHRHS projection data. First, transaxial SPECT images were obtained by filtered backprojection (FBP) implemented in the SPECT camera software (Butterworth window of 6th order and 2.3 cycles/cm cutoff). Uniform post-reconstruction attenuation correction was performed according to order zero Chang (µ = 0.12/cm), no scatter correction. Second, SPECT images were reconstructed using the iterative ordered-subsets expectation–maximization algorithm with resolution recovery implemented in the HybridRecon-Neurology tool of the Hermes SMART workstation v1.6 with parameter settings recommended for FP-CIT SPECT by Hermes (5 iterations, 15 subsets, post-filtering with 3-dimensional Gaussian kernel of 7 mm full width at half maximum, uniform attenuation correction with narrow-beam attenuation coefficient 0.146/cm, simulation-based scatter correction, resolution recovery with a Gaussian model).

Representative SPECT images are shown in Fig. 1.

Fig. 1figure 1

Representative DAT-SPECT images. Representative DAT-SPECT images from four different patients (columns) in the three different settings (rows)

The SPECT acquisition was performed first with the MPH collimators in 46 of the 71 patients (65%). The mean delay of the start of the subsequent acquisition with LEHRHS collimators relative to the start of the MPH acquisition was 53 ± 8 min. The SPECT acquisition was performed first with the LEHRHS collimators in the remaining 25 patients (35%). The mean delay of the start of the subsequent acquisition with MPH collimators relative to the start of the LEHRHS acquisition was 63 ± 7 min. (Note that the net duration of the acquisition was 10 min longer with the LEHRHS collimators, 40 min versus 30 min.) The delay between i.v. administration of 123I-FP-CIT and start of the acquisition was 212 ± 39 min for the MPH acquisitions and 224 ± 37 min for the LEHRHS acquisitions (paired t test p = 0.067).

Visual interpretation

Visual interpretation of the DAT-SPECT images was based on a standardized display (Fig. 2), similar to the display used in everyday clinical routine in our department. For generation of the display, each individual DAT-SPECT image was stereotactically normalized (affine transformation) to the anatomical standard space of the Montreal Neurological Institute (MNI) using the Normalization tool of the Statistical Parametric Mapping software package (version SPM12) and a custom FP-CIT template. The stereotactically normalized DAT-SPECT image was smoothed by convolution with an isotropic Gaussian kernel with 4 mm full width at half maximum, independent of the collimator used for the SPECT acquisition. Distribution volume ratio (DVR) images were obtained by voxelwise scaling of the smoothed image to the 75th percentile of the intensity values in a reference region comprising the whole brain parenchyma without striata, thalamus and brainstem [24]. The display comprised ten transversal DVR image slices of 4 mm thickness and one transversal DVR image slab of 12 mm thickness (Fig. 2) [25].

Fig. 2figure 2

Standard display for visual interpretation of the DAT-SPECT images. Standard display for visual interpretation of the DAT-SPECT images presenting ten transversal distribution volume ratio (DVR) image slices of 4 mm thickness from the superior to the inferior edge of the striatum with the maximum of the colortable individually scaled to the maximum intensity in the ten images. In addition, the display presents a transversal DVR image slab of 12 mm thickness through the center of the striatum with the maximum of the colortable scaled to a fixed DVR threshold [25]. The DVR threshold had been optimized previously, separately for each of the three settings (MPH, LEHRH-FBP, LEHRHS-OSEM)

Visual interpretation of the SPECT images was performed independently by three raters with different experience in clinical reading of DAT-SPECT (about 200, about 2000, and about 5000 cases). The raters were blinded for all clinical data. The raters were asked to score each SPECT image with respect to Parkinson-typical reduction of striatal 123I-FP-CIT uptake using the following Likert 6-score: − 3 = clearly normal, − 2 = probably normal, − 1 = more likely normal than reduced, 1 = more likely reduced than normal, 2 = probably reduced, 3 = clearly reduced. This was performed twice for each of the three settings (MPH, LEHRHS-FBP, LEHRHS-OSEM); that is, each rater performed six reading sessions. A different randomization of the patients was used in each session. Images with a discrepant Likert score between the two reading sessions of a given setting by a given rater were rated a third time by the same rater to obtain an intra-rater consensus.

Semiquantitative analysis

The specific binding ratio (SBR) of 123I-FP-CIT in left and right putamen was obtained by hottest voxels analysis of the stereotactically normalized DVR image using large unilateral putamen masks predefined in MNI space [18]. The putamen masks were much bigger than the actual putamen volume to guarantee that all putaminal counts were included. The number of hottest voxels within a unilateral putamen mask to be averaged was fixed to a total volume 10 ml. The putamen SBR was computed as mean DVR in the hottest voxels minus 1. The minimum putamen SBR of both hemispheres was used for the analysis.

Statistical analysis

Continuous variables are reported as mean ± standard deviation of the sample. Intra- and between-rater agreement with respect to the Likert score and with respect to the dichotomized Likert score (< 0: normal, > 0: reduced) was characterized by Cohen’s kappa. Cohen’s effect size d was used to characterize the difference of the putamen SBR between visually normal and visually reduced DAT-SPECT.

Statistical analyses were conducted using SPSS version 27 (SPSS Inc., Chicago, Illinois). All p-values are given two-sided. Statistical significance was defined as p < 0.05.

Technical performance characteristics of MPH-SPECT

The system count sensitivity profile of the triple-head SPECT system equipped with the novel general purpose brain imaging MPH collimator was measured as described previously by our group for a DAT-SPECT-specific MPH collimator [16]. In brief, a point source (5–10 MBq 99mTc in about 5 µl in an Eppendorf tube) was placed on the lattice points of a 1 cm grid covering the whole field of view. A full SPECT acquisition was performed for each localization of the point source using the same acquisition parameters as for clinical DAT-SPECT described in ‘MPH-SPECT’ section. The total number of counts acquired during the SPECT acquisition was obtained by summing the counts over all views. Uniformity correction was turned on. Dead time correction was negligible. The system count sensitivity at the localization of the point source was obtained by the following formula: sensitivity = total number of counts/total net scan duration/activity of the point source decay-corrected to the start time of the SPECT acquisition.

Spatial resolution of MPH-SPECT was assessed with a Derenzo-type hot rod phantom with rod diameter of 2 to 7 mm filled with 123I solution. Acquisition and image reconstruction were performed using exactly the same protocol and parameter settings as for clinical MPH-SPECT except that no attenuation correction was performed.

留言 (0)

沒有登入
gif