Chapter

Computer-Aided Diagnosis of Parkinson’s Disease, Based on SPECT Scans of the Dopamine Transporter

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Computer-aided diagnosis (CAD) tools have been, more and more frequently, proposed as a complement to the visual evaluation of dopamine transporter (DaT) single-photon emission computed tomography (SPECT) scans. Classification accuracies up to 98% have been reported in the differentiation between healthy control (HC) subjects and patients with Parkinson’s disease (PD) based on DaT SPECT. CAD systems have also been used to differentiate between different types of Parkinsonism, but the accuracies are not as high as for the classification of PD patients versus HC subjects. When compared with cutoff-based classification techniques or visual classification, the CAD originated higher accuracy. The works reported in the literature on CAD systems of DaT SPECT have some limitations, with the lack of postmortem diagnosis the most important of them. The lack of consistency in the definition of the groups of patients and overlap of datasets used are also significant limitations. Despite all limitations and inconsistencies found in the published works, there is evidence that CAD based on DaT SPECT can effectively help in the clinics and is surely a better solution than the cutoff-based techniques only. Therefore, we believe CAD should be used in clinics as a complement to the visual and quantitative evaluation.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Purpose This study investigated the potential of deep convolutional neural networks (CNN) for automatic classification of FP-CIT SPECT in multi-site or multi-camera settings with variable image characteristics. Methods The study included FP-CIT SPECT of 645 subjects from the Parkinson’s Progression Marker Initiative (PPMI), 207 healthy controls, and 438 Parkinson’s disease patients. SPECT images were smoothed with an isotropic 18-mm Gaussian kernel resulting in 3 different PPMI settings: (i) original (unsmoothed), (ii) smoothed, and (iii) mixed setting comprising all original and all smoothed images. A deep CNN with 2,872,642 parameters was trained, validated, and tested separately for each setting using 10 random splits with 60/20/20% allocation to training/validation/test sample. The putaminal specific binding ratio (SBR) was computed using a standard anatomical ROI predefined in MNI space (AAL atlas) or using the hottest voxels (HV) analysis. Both SBR measures were trained (ROC analysis, Youden criterion) using the same random splits as for the CNN. CNN and SBR trained in the mixed PPMI setting were also tested in an independent sample from clinical routine patient care (149 with non-neurodegenerative and 149 with neurodegenerative parkinsonian syndrome). Results Both SBR measures performed worse in the mixed PPMI setting compared to the pure PPMI settings (e.g., AAL-SBR accuracy = 0.900 ± 0.029 in the mixed setting versus 0.957 ± 0.017 and 0.952 ± 0.015 in original and smoothed setting, both p < 0.01). In contrast, the CNN showed similar accuracy in all PPMI settings (0.967 ± 0.018, 0.972 ± 0.014, and 0.955 ± 0.009 in mixed, original, and smoothed setting). Similar results were obtained in the clinical sample. After training in the mixed PPMI setting, only the CNN provided acceptable performance in the clinical sample. Conclusions These findings provide proof of concept that a deep CNN can be trained to be robust with respect to variable site-, camera-, or scan-specific image characteristics without a large loss of diagnostic accuracy compared with mono-site/mono-camera settings. We hypothesize that a single CNN can be used to support the interpretation of FP-CIT SPECT at many different sites using different acquisition hardware and/or reconstruction software with only minor harmonization of acquisition and reconstruction protocols.
Article
Full-text available
Objectives: To provide an automated classification method for degenerative parkinsonian syndromes (PS) based on semiquantitative 123I-FP-CIT SPECT striatal indices and support-vector-machine (SVM) analysis. Methods: 123I-FP-CIT SPECT was performed at a single-center level on 370 individuals with PS, including 280 patients with Parkinson’s disease (PD), 21 with multiple-system atrophy-parkinsonian type (MSA-P), 41 with progressive supranuclear palsy (PSP) and 28 with corticobasal syndrome (CBS) (mean age 70.3 years, 47% female, mean disease duration at scan 1.4 year), as well as 208 age- and gender-matched control subjects. Striatal volumes-of-interest (VOIs) uptake, VOIs asymmetry indices (AIs) and caudate/putamen (C/P) ratio were used as input for SVM individual classification using 5-fold cross-validation. Results: Univariate analyses showed significantly lower VOIs uptake, higher striatal AI and C/P ratio for each PS in comparison to controls (all p<0.001). Among PS, higher degree of striatal impairment was observed in MSA-P and PSP, while CBS showed moderate uptake reduction and higher AI. Binary SVM classification showed 92.9% accuracy in distinguishing PS from controls. Classification based on each binary combination of PS ranged 62.9-83.7%-accuracy with the most satisfactory results when separating CBS from the other PS. Sensitivity and specificity values were high and balanced ranging from 60 to 80% for all analyses with >70% accuracy. Overall, striatal AI and C/P ratio on the more affected side had the highest weighting factors. Conclusion: Semiquantitative 123I-FP-CIT SPECT striatal evaluation combined with SVM represents a promising approach to disentangle PD from non-degenerative conditions and from atypical PS at the early stage. Keywords: Parkinson’s disease; degenerative parkinsonism; SPECT; classification
Article
Full-text available
Objective The Parkinson's Progression Markers Initiative (PPMI) is an observational, international study designed to establish biomarker‐defined cohorts and identify clinical, imaging, genetic, and biospecimen Parkinson's disease (PD) progression markers to accelerate disease‐modifying therapeutic trials. Methods A total of 423 untreated PD, 196 Healthy Control (HC) and 64 SWEDD (scans without evidence of dopaminergic deficit) subjects were enrolled at 24 sites. To enroll PD subjects as early as possible following diagnosis, subjects were eligible with only asymmetric bradykinesia or tremor plus a dopamine transporter (DAT) binding deficit on SPECT imaging. Acquisition of data was standardized as detailed at www.ppmi-info.org. Results Approximately 9% of enrolled subjects had a single PD sign at baseline. DAT imaging excluded 16% of potential PD subjects with SWEDD. The total MDS‐UPDRS for PD was 32.4 compared to 4.6 for HC and 28.2 for SWEDD. On average, PD subjects demonstrated 45% and 68% reduction in mean striatal and contralateral putamen Specific Binding Ratios (SBR), respectively. Cerebrospinal fluid (CSF) was acquired from >97% of all subjects. CSF (PD/HC/SWEDD pg/mL) α‐synuclein (1845/2204/2141) was reduced in PD vs HC or SWEDD (P < 0.03). Similarly, t‐tau (45/53) and p‐tau (16/18) were reduced in PD versus HC (P < 0.01), Interpretation PPMI has detailed the biomarker signature for an early PD cohort defined by clinical features and imaging biomarkers. This strategy provides the framework to establish biomarker cohorts and to define longitudinal progression biomarkers to support future PD treatment trials.
Article
Full-text available
Purpose This work aimed to assess the potential of a set of features extracted from [123I]FP-CIT SPECT brain images to be used in the computer-aided “in vivo” confirmation of dopaminergic degeneration and therefore to assist clinical decision to diagnose Parkinson’s disease. Methods Seven features were computed from each brain hemisphere: five standard features related to uptake ratios on the striatum and two features related to the estimated volume and length of striatal region with normal uptake. The features were tested on a dataset of 652 [123I]FP-CIT SPECT brain images from the Parkinson’s Progression Markers Initiative. The discrimination capacities of each feature individually and groups of features were assessed using three different machine learning techniques: support vector machines (SVM), k-nearest neighbours and logistic regression. Results Cross-validation results based on SVM have shown that, individually, the features that generated the highest accuracies were the length of the striatal region (96.5%), the putaminal binding potential (95.4%) and the striatal binding potential (93.9%) with no statistically significant differences among them. The highest classification accuracy was obtained using all features simultaneously (accuracy 97.9%, sensitivity 98%, and specificity 97.6%). Generally, slightly better results were obtained using the SVM with no statistically significant difference to the other classifiers for most of the features. Conclusions The length of the striatal region uptake is clinically useful and highly valuable to confirm dopaminergic degeneration "in vivo" as an aid to the diagnosis of Parkinson’s disease. It compares fairly well to the standard uptake ratio-based features, reaching, at least, similar accuracies and is easier to obtain automatically. Thus, we propose its day to day clinical use, jointly with the uptake ratio-based features, in the computer-aided diagnosis of dopaminergic degeneration in Parkinson’s disease.
Article
Full-text available
Background: Semi-quantification methods are well established in the clinic for assisted reporting of (I123) Ioflupane images. Arguably, these are limited diagnostic tools. Recent research has demonstrated the potential for improved classification performance offered by machine learning algorithms. A direct comparison between methods is required to establish whether a move towards widespread clinical adoption of machine learning algorithms is justified. This study compared three machine learning algorithms with that of a range of semi-quantification methods, using the Parkinson's Progression Markers Initiative (PPMI) research database and a locally derived clinical database for validation. Machine learning algorithms were based on support vector machine classifiers with three different sets of features: Voxel intensities Principal components of image voxel intensities Striatal binding radios from the putamen and caudate. Semi-quantification methods were based on striatal binding ratios (SBRs) from both putamina, with and without consideration of the caudates. Normal limits for the SBRs were defined through four different methods: Minimum of age-matched controls Mean minus 1/1.5/2 standard deviations from age-matched controls Linear regression of normal patient data against age (minus 1/1.5/2 standard errors) Selection of the optimum operating point on the receiver operator characteristic curve from normal and abnormal training data Each machine learning and semi-quantification technique was evaluated with stratified, nested 10-fold cross-validation, repeated 10 times. Results: The mean accuracy of the semi-quantitative methods for classification of local data into Parkinsonian and non-Parkinsonian groups varied from 0.78 to 0.87, contrasting with 0.89 to 0.95 for classifying PPMI data into healthy controls and Parkinson's disease groups. The machine learning algorithms gave mean accuracies between 0.88 to 0.92 and 0.95 to 0.97 for local and PPMI data respectively. Conclusions: Classification performance was lower for the local database than the research database for both semi-quantitative and machine learning algorithms. However, for both databases, the machine learning methods generated equal or higher mean accuracies (with lower variance) than any of the semi-quantification approaches. The gain in performance from using machine learning algorithms as compared to semi-quantification was relatively small and may be insufficient, when considered in isolation, to offer significant advantages in the clinical context.
Article
Full-text available
Dopaminergic degeneration is a pathologic hallmark of Parkinson's disease (PD), which can be assessed by dopamine transporter imaging such as FP-CIT SPECT. Until now, imaging has been routinely interpreted by human though it can show interobserver variability and result in inconsistent diagnosis. In this study, we developed a deep learning-based FP-CIT SPECT interpretation system to refine the imaging diagnosis of Parkinson's disease. This system trained by SPECT images of PD patients and normal controls shows high classification accuracy comparable with the experts’ evaluation referring quantification results. Its high accuracy was validated in an independent cohort composed of patients with PD and nonparkinsonian tremor. In addition, we showed that some patients clinically diagnosed as PD who have scans without evidence of dopaminergic deficit (SWEDD), an atypical subgroup of PD, could be reclassified by our automated system. Our results suggested that the deep learning-based model could accurately interpret FP-CIT SPECT and overcome variability of human evaluation. It could help imaging diagnosis of patients with uncertain Parkinsonism and provide objective patient group classification, particularly for SWEDD, in further clinical studies.
Article
Full-text available
Early and accurate identification of parkinsonian syndromes (PS) involving presynaptic degeneration from non-degenerative variants such as Scans Without Evidence of Dopaminergic Deficit (SWEDD) and tremor disorders, is important for effective patient management as the course, therapy and prognosis differ substantially between the two groups. In this study, we use Single Photon Emission Computed Tomography (SPECT) images from healthy normal, early PD and SWEDD subjects, as obtained from the Parkinson's Progression Markers Initiative (PPMI) database, and process them to compute shape- and surface fitting-based features for the three groups. We use these features to develop and compare various classification models that can discriminate between scans showing dopaminergic deficit, as in PD, from scans without the deficit, as in healthy normal or SWEDD. Along with it, we also compare these features with Striatal Binding Ratio (SBR)-based features, which are well-established and clinically used, by computing a feature importance score using Random forests technique. We observe that the Support Vector Machine (SVM) classifier gave the best performance with an accuracy of 97.29%. These features also showed higher importance than the SBR-based features. We infer from the study that shape analysis and surface fitting are useful and promising methods for extracting discriminatory features that can be used to develop diagnostic models that might have the potential to help clinicians in the diagnostic process.
Article
Full-text available
Background: Dopamine transporter (DaT) imaging (DaTSCAN) is useful for the differential diagnosis of parkinsonian syndromes. Visual evaluation of DaTSCAN images represents the generally accepted diagnostic method, but it is strongly dependent on the observer's experience and shows inter- and intra-observer variability. A reliable and automatic method for DaTSCAN evaluation can provide objective quantification; it is desirable for longitudinal studies, and it allows for a better follow-up control. Moreover, it is crucial for an automated method to produce coherent measures related to the severity of motor symptoms. Methods: In this work, we propose a novel fully automated technique for DaTSCAN analysis that generates quantitative measures based on striatal intensity, shape, symmetry and extent. We tested these measures using a support vector machine (SVM) classifier. Results: The proposed measures reached 100 % accuracy in distinguishing between patients with Parkinson's disease (PD) and control subjects. We also demonstrate the existence of a linear relationship and an exponential trend between pooled structural and functional striatal characteristics and the Unified Parkinson's disease Rating Scale (UPDRS) motor score. Conclusions: We present a novel, highly reproducible, user-independent technique for DaTSCAN analysis producing quantitative measures directly connected to striatum uptake and shape. In our method, no a priori assumption is required on the spatial conformation and localization of striatum, and both uptake and symmetry contribute to the index quantification. These measures can reliably support a computer-assisted decision system.
Article
Full-text available
Intensity normalization is an important pre-processing step in the study and analysis of DaTSCAN SPECT imaging. As most automatic supervised image segmentation and classification methods base their assumptions regarding the intensity distributions on a standardized intensity range, intensity normalization takes on a very significant role. In this work, a comparison between different novel intensity normalization methods is presented. These proposed methodologies are based on Gaussian Mixture Model (GMM) image filtering and mean-squared error (MSE) optimization. The GMM-based image filtering method is achieved according to a probability threshold that removes the clusters whose likelihood are negligible in the non-specific regions. The MSE optimization method consists of a linear transformation that is obtained by minimizing the MSE in the non-specific region between the intensity normalized image and the template. The proposed intensity normalization methods are compared to: i ) a standard approach based on the specific-to-non-specific binding ratio that is widely used, and ii ) a linear approach based on the α-stable distribution. This comparison is performed on a DaTSCAN image database comprising analysis and classification stages for the development of a computer aided diagnosis (CAD) system for Parkinsonian syndrome (PS) detection. In addition, these proposed methods correct spatially varying artifacts that modulate the intensity of the images. Finally, using the leave-one-out cross-validation technique over these two approaches, the system achieves results up to a 92.91% of accuracy, 94.64% of sensitivity and 92.65 % of specificity, outperforming previous approaches based on a standard and a linear approach, which are used as a reference. The use of advanced intensity normalization techniques, such as the GMM-based image filtering and the MSE optimization improves the diagnosis of PS.
Article
Full-text available
The Parkinson Progression Marker Initiative (PPMI) is a comprehensive observational, international, multi-center study designed to identify PD progression biomarkers both to improve understanding of disease etiology and course and to provide crucial tools to enhance the likelihood of success of PD modifying therapeutic trials. The PPMI cohort will comprise 400 recently diagnosed PD and 200 healthy subjects followed longitudinally for clinical, imaging and biospecimen biomarker assessment using standardized data acquisition protocols at twenty-one clinical sites. All study data will be integrated in the PPMI study database and will be rapidly and publically available through the PPMI web site-www.ppmi-info.org. Biological samples including longitudinal collection of blood, cerebrospinal fluid (CSF) and urine will be available to scientists by application to an independent PPMI biospecimen review committee also through the PPMI web site. PPMI will rely on a partnership of government, PD foundations, industry and academics working cooperatively. This approach is crucial to enhance the potential for success of this ambitious strategy to develop PD progression biomarkers that will accelerate research in disease modifying therapeutics.
Article
Full-text available
Objective. The aim of the present study was to develop a fully-automated computational solution for computer-aided diagnosis in Parkinson syndrome based on [¹²³I]FP-CIT single photon emission computed tomography (SPECT) images. Approach. A dataset of 654 [¹²³I]FP-CIT SPECT brain images from the Parkinson’s Progression Markers Initiative were used. Of these, 445 images were of patients with Parkinson’s disease at an early stage and the remainder formed a control group. The images were pre-processed using automated template-based registration followed by the computation of the binding potential at a voxel level. Then, the binding potential images were used for classification, based on the voxel-as-feature approach and using the support vector machines paradigm. Main results. The obtained estimated classification accuracy was 97.86%, the sensitivity was 97.75% and the specificity 98.09%. Significance. The achieved classification accuracy was very high and, in fact, higher than accuracies found in previous studies reported in the literature. In addition, results were obtained on a large dataset of early Parkinson’s disease subjects. In summation, the information provided by the developed computational solution potentially supports clinical decision-making in nuclear medicine, using important additional information beyond the commonly used uptake ratios and respective statistical comparisons. (ClinicalTrials.gov Identifier: NCT01141023)
Article
Full-text available
Central dopaminergic and overall brain functional activity (OBFA) are altered to different degrees in idiopathic Parkinson's disease and atypical parkinsonian disorders, which cause difficulties in achieving an unambiguous clinical diagnosis. A dual examination using (123)I-FP-CIT SPECT (DaTSCAN) and(18)F-FDG PET provides complementary information regarding dopamine transporter (DAT) availability and OBFA, respectively. Parametric images based on a single, dynamic (11)C-PE2I scan, supply potentially both DAT availability (BPND) and relative cerebral blood flow (R1) at voxel-level. This study aimed to evaluate the validity of (11)C-PE2I PET against the dual modality approach using (123)I-FP-CIT SPECT and (18)F-FDG PET. Sixteen patients with parkinsonian disorders had a dual examination with (18)F-FDG PET and (123)I-FP-CIT SPECT following clinical routines and additionally one experimental (11)C-PE2I PET scan. Parametric BPND and R1 images were generated using receptor parametric mapping with cerebellum as reference. T1-weighted MRI was used for automated definition of volumes of interest (VOI). The DAT VOIs included basal ganglia, whereas the OBFA was examined using VOIs across the brain. BPND and R1 values were compared to normalized (123)I-FP-CIT and (18)F-FDG uptake values, respectively, using Pearson correlations and regression analyses. In addition, two blinded readers evaluated the images visually, in both the routine and experimental dataset, for comparison of patient diagnoses. Parametric (11)C-PE2I BPND and R1 images showed high consistency with (123)IFP- CIT SPECT and (18)F-FDG PET images. Correlations between (11)C-PE2I BPND and (123)I-FP-CIT uptake ratio values were 0.97 and 0.76 in putamen and caudate nucleus, respectively. Regional (11)C-PE2I R1 values were moderately to highly correlated with normalized (18)F-FDG values (range 0.59-0.94). Visual assessment of DAT availability showed a high consistency between (11)C-PE2I BPND and (123)I-FP-CIT images, whereas the consistency was somewhat lower for appraisal of OBFA using (123)I-FP-CIT and (18)FFDG images. Substantial differences were found between clinical diagnosis and both neuroimaging diagnoses. One single, dynamic (11)C-PE2I PET investigation is a powerful alternative to a dual examination with (123)I-FP-CIT SPECT and (18)F-FDG PET for a differential diagnosis of parkinsonian disorders. A large-scale patient study is, however, needed to investigate further distinct pathological patterns in OBFA for various parkinsonian disorders. Copyright © 2015 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
Article
Full-text available
Brain single-photon-emission-computerized tomography (SPECT) with 123I-ioflupane (123I-FP-CIT) is useful to diagnose Parkinson disease (PD). To investigate the diagnostic performance of 123I-FP-CIT brain SPECT with semiquantitative analysis by Basal Ganglia V2 software (BasGan), we evaluated semiquantitative data of patients with suspect of PD by a support vector machine classifier (SVM), a powerful supervised classification algorithm. 123I-FP-CIT SPECT with BasGan analysis was performed in 90 patients with suspect of PD showing mild symptoms (bradykinesia-rigidity and mild tremor). PD was confirmed in 56 patients, 34 resulted non-PD (essential tremor and drug-induced Parkinsonism). A clinical follow-up of at least 6 months confirmed diagnosis. To investigate BasGan diagnostic performance we trained SVM classification models featuring different descriptors using both a “leave-one-out” and a “five-fold” method. In the first study we used as class descriptors the semiquantitative radiopharmaceutical uptake values in the left (L) and right (R) putamen (P) and in the L and R caudate nucleus (C) for a total of 4 descriptors (CL, CR, PL, PR). In the second study each patient was described only by CL and CR, while in the third by PL and PR descriptors. Age was added as a further descriptor to evaluate its influence in the classification performance. 123I-FP-CIT SPECT with BasGan analysis reached a classification performance higher than 73.9% in all the models. Considering the “Leave-one-out” method, PL and PR were better predictors (accuracy of 91% for all patients) than CL and CR descriptors; using PL, PR, CL, and CR diagnostic accuracy was similar to that of PL and PR descriptors in the different groups. Adding age as a further descriptor accuracy improved in all the models. The best results were obtained by using all the 5 descriptors both in PD and non-PD subjects (CR and CL + PR and PL + age = 96.4% and 94.1%, respectively). Similar results were observed for the “five-fold” method. 123I-FP-CIT SPECT with BasGan analysis using SVM classifier was able to diagnose PD. Putamen was the most discriminative descriptor for PD and the patient age influenced the classification accuracy.
Article
Full-text available
Early and accurate diagnosis of Parkinson’s disease (PD) is important for early management, proper prognostication and for initiating neuroprotective therapies once they become available. Recent neuroimaging techniques such as dopaminergic imaging using single photon emission computed tomography (SPECT) with 123I-Ioflupane (DaTSCAN) have shown to detect even early stages of the disease. In this paper, we use the striatal binding ratio (SBR) values that are calculated from the 123I-Ioflupane SPECT scans (as obtained from the Parkinson’s progression markers initiative (PPMI) database) for developing automatic classification and prediction/prognostic models for early PD. We used support vector machine (SVM) and logistic regression in the model building process. We observe that the SVM classifier with RBF kernel produced a high accuracy of more than 96% in classifying subjects into early PD and healthy normal; and the logistic model for estimating the risk of PD also produced high degree of fitting with statistical significance indicating its usefulness in PD risk estimation. Hence, we infer that such models have the potential to aid the clinicians in the PD diagnostic process.
Article
Full-text available
Aim: The purpose of the current paper is to present a computational solution to accurately quantify a specific to a non-specific uptake ratio in [123I]fP-CIT single photon emission computed tomography (SPECT) images and simultaneously measure the spatial dimensions of the basal ganglia, also known as basal nuclei. A statistical analysis based on a reference dataset selected by the user is also automatically performed. Methods: The quantification of the specific to non-specific uptake ratio here is based on regions of interest defined after the registration of the image under study with a template image. The computational solution was tested on a dataset of 38 [123I]FP-CIT SPECT images: 28 images were from patients with Parkinson's disease and the remainder from normal patients, and the results of the automated quantification were compared to the ones obtained by three well-known semi-automated quantification methods. Results: The results revealed a high correlation coefficient between the developed automated method and the three semi-automated methods used for comparison (r ≥0.975). The solution also showed good robustness against different positions of the patient, as an almost perfect agreement between the specific to non-specific uptake ratio was found (ICC=1.000). The mean processing time was around 6 seconds per study using a common notebook PC. Conclusion: The solution developed can be useful for clinicians to evaluate [123I]FP-CIT SPECT images due to its accuracy, robustness and speed. Also, the comparison between case studies and the follow-up of patients can be done more accurately and proficiently since the intra- and inter-observer variability of the semi-automated calculation does not exist in automated solutions. The dimensions of the basal ganglia and their automatic comparison with the values of the population selected as reference are also important for professionals in this area.
Article
Full-text available
Purpose In this work, an approach to computer aided diagnosis (CAD) system is proposed as a decision‐making aid in Parkinsonian syndrome (PS) detection. This tool, intended for physicians, entails fully automatic preprocessing, normalization, and classification procedures for brain single‐photon emission computed tomography images. Methods Ioflupane[¹²³I]FP‐CIT images are used to provide in vivo information of the dopamine transporter density. These images are preprocessed using an automated template‐based registration followed by two proposed approaches for intensity normalization. A support vector machine (SVM) is used and compared to other statistical classifiers in order to achieve an effective diagnosis using whole brain images in combination with voxel selection masks. Results The CAD system is evaluated using a database consisting of 208 DaTSCAN images (100 controls, 108 PS). SVM‐based classification is the most efficient choice when masked brain images are used. The generalization performance is estimated to be 89.02 (90.41–87.62)% sensitivity and 93.21 (92.24–94.18)% specificity. The area under the curve can take values of 0.9681 (0.9641–0.9722) when the image intensity is normalized to a maximum value, as derived from the receiver operating characteristics curves. Conclusions The present analysis allows to evaluate the impact of the design elements for the development of a CAD‐system when all the information encoded in the scans is considered. In this way, the proposed CAD‐system shows interesting properties for clinical use, such as being fast, automatic, and robust.
Article
Full-text available
To contribute to the differentiation of Parkinson's disease (PD) and essential tremor (ET), we compared two different artificial neural network classifiers using (123)I-FP-CIT SPECT data, a probabilistic neural network (PNN) and a classification tree (ClT). (123)I-FP-CIT brain SPECT with semiquantitative analysis was performed in 216 patients: 89 with ET, 64 with PD with a Hoehn and Yahr (H&Y) score of ≤2 (early PD), and 63 with PD with a H&Y score of ≥2.5 (advanced PD). For each of the 1,000 experiments carried out, 108 patients were randomly selected as the PNN training set, while the remaining 108 validated the trained PNN, and the percentage of the validation data correctly classified in the three groups of patients was computed. The expected performance of an "average performance PNN" was evaluated. In analogy, for ClT 1,000 classification trees with similar structures were generated. For PNN, the probability of correct classification in patients with early PD was 81.9±8.1% (mean±SD), in patients with advanced PD 78.9±8.1%, and in ET patients 96.6±2.6%. For ClT, the first decision rule gave a mean value for the putamen of 5.99, which resulted in a probability of correct classification of 93.5±3.4%. This means that patients with putamen values >5.99 were classified as having ET, while patients with putamen values <5.99 were classified as having PD. Furthermore, if the caudate nucleus value was higher than 6.97 patients were classified as having early PD (probability 69.8±5.3%), and if the value was <6.97 patients were classified as having advanced PD (probability 88.1%±8.8%). These results confirm that PNN achieved valid classification results. Furthermore, ClT provided reliable cut-off values able to differentiate ET and PD of different severities.
Article
Full-text available
These guidelines summarize the current views of the European Association of Nuclear Medicine Neuroimaging Committee (ENC). The aim of the guidelines is to assist nuclear medicine practitioners when making recommendations, performing, interpreting, and reporting the results of clinical dopamine transporter (DAT) single photon emission computed tomography (SPECT) studies using (123)I-labelled radiopharmaceuticals. The aim is to achieve a high-quality standard of DAT SPECT imaging, which will increase the diagnostic impact of this technique in neurological practice. The present document is an update of the 2002 guidelines [1] and has been guided by the views of various national societies: the Task Group Neuro-Nuclear-Medicine of the German Society of Nuclear Medicine [2], a consensus statement of the imaging centres included in the "Kompetenznetz-Parkinson" sponsored by the German Federal Ministry of Education, and the Task Group of Neuro-Nuclear-Medicine of the French Society of Nuclear Medicine [3]. The guidelines reflect the individual experience of experts in European countries. The guidelines are intended to present information specifically adapted to European practice. The information provided should be taken in the context of local conditions and regulations.
Article
Full-text available
In SPECT, the binding of radiotracers in brain areas is usually assessed by manual positioning of regions of interest (ROIs). The disadvantages of this method are that it is an observer-dependent procedure and that it may not be sensitive for assessing defects significantly smaller than the ROI. To circumvent these limitations, we developed a fully automatic three-dimensional technique that quantifies neuronal radiotracer binding on a voxel-by-voxel basis. To build a model of normal 123I-labeled N-omega-fluoropropyl-2beta-carbomethoxy-3beta-(4-iodophenyl) nortropane (FPCIT) binding, 17 studies of healthy volunteers were registered to the same orientation. After registration, the specific-to-nonspecific binding ratio was calculated for each voxel of the striatal volumes of interest (VOIs). The mean and SD of that binding ratio were then calculated on a voxel-by-voxel basis. For the analysis of 10 healthy volunteer studies (control group) and 21 studies of drug-naive patients with Parkinson's disease, the registration and calculation of the specific-to-nonspecific [123I]FPCIT binding ratio were performed by the same method. Subsequently, a voxel of the striata was classified as a diminished [123I]FPCIT binding ratio if its value was lower than the mean -2 x SD. For each subject, the defect size, the relative number of voxels with a diminished binding ratio and the binding ratio of the whole striatal VOIs were calculated and compared with the binding ratio as assessed by the traditional ROI method. The results of the automatic method correlated significantly with the results of the traditional ROI method. Furthermore, for the ipsilateral side, the automatically calculated defect size had less overlap between the patient and the control group than the traditionally calculated binding ratio. The method presented quantifies [123I]FPCIT binding ratio automatically on a voxel-by-voxel basis, by comparison with a model of healthy volunteers. We have shown that it is appropriate to use the automatic method as a replacement for the traditional manual method, which enables us to study the localized dopaminergic degeneration process in Parkinson's disease more precisely and without any inter- or intraobserver variability.
Article
Full-text available
Dementia with Lewy bodies (DLB) is one of the main differential diagnoses of Alzheimer's disease (AD). Key pathological features of patients with DLB are not only the presence of cerebral cortical neuronal loss, with Lewy bodies in surviving neurones, but also loss of nigrostriatal dopaminergic neurones, similar to that of Parkinson's disease (PD). In DLB there is 40-70% loss of striatal dopamine. To determine if detection of this dopaminergic degeneration can help to distinguish DLB from AD during life. The integrity of the nigrostriatal metabolism in 27 patients with DLB, 17 with AD, 19 drug naive patients with PD, and 16 controls was assessed using a dopaminergic presynaptic ligand, (123)I-labelled 2beta-carbomethoxy-3beta-(4-iodophenyl)-N-(3-fluoropropyl)nortropane (FP-CIT), and single photon emission tomography (SPET). A SPET scan was carried out with a single slice, brain dedicated tomograph (SME 810) 3.5 hours after intravenous injection of 185 MBq FP-CIT. With occipital cortex used as a radioactivity uptake reference, ratios for the caudate nucleus and the anterior and posterior putamen of both hemispheres were calculated. All scans were also rated by a simple visual method. Both DLB and PD patients had significantly lower uptake of radioactivity than patients with AD (p<0.001) and controls (p<0.001) in the caudate nucleus and the anterior and posterior putamen. FP-CIT SPET provides a means of distinguishing DLB from AD during life.
Article
Spatial and intensity normalizations are nowadays a prerequisite for neuroimaging analysis. Influenced by voxel-wise and other univariate comparisons, where these corrections are key, they are commonly applied to any type of analysis and imaging modalities. Nuclear imaging modalities such as PET-FDG or FP-CIT SPECT, a common modality used in Parkinson's disease diagnosis, are especially dependent on intensity normalization. However, these steps are computationally expensive and furthermore, they may introduce deformations in the images, altering the information contained in them. Convolutional neural networks (CNNs), for their part, introduce position invariance to pattern recognition, and have been proven to classify objects regardless of their orientation, size, angle, etc. Therefore, a question arises: how well can CNNs account for spatial and intensity differences when analyzing nuclear brain imaging? Are spatial and intensity normalizations still needed? To answer this question, we have trained four different CNN models based on well-established architectures, using or not different spatial and intensity normalization preprocessings. The results show that a sufficiently complex model such as our three-dimensional version of the ALEXNET can effectively account for spatial differences, achieving a diagnosis accuracy of 94.1% with an area under the ROC curve of 0.984. The visualization of the differences via saliency maps shows that these models are correctly finding patterns that match those found in the literature, without the need of applying any complex spatial normalization procedure. However, the intensity normalization - and its type - is revealed as very influential in the results and accuracy of the trained model, and therefore must be well accounted.
Article
PurposeQuantification of the tracer distribution would add objectivity to the visual assessments of dopamine transporter (DAT) single photon emission computed tomography (SPECT) data. Our study aimed to evaluate the diagnostic utility of fractal dimension (FD) as a quantitative indicator of tracer distribution and compared with the conventional quantitative value: specific binding ratio (SBR). We also evaluated the utility of the combined index SBR/FD (SBR divided by FD). Materials and methodsWe conducted both clinical and phantom studies. In the clinical study, 150 patients including 110 patients with Parkinsonian syndrome (PS) and 40 without PS were enrolled. In the phantom study, we used a striatal phantom with the striatum chamber divided into two spaces, representing the caudate nucleus and putamen. The SBR, FD, and SBR/FD were calculated and compared between datasets for evaluating the diagnostic utility. Mann–Whitney test and receiver-operating characteristics (ROC) analysis were used for analysis. ResultsROC analysis revealed that the FD value had high diagnostic performance [the areas under the curve (AUC) = 0.943] and the combined use of SBR and FD (SBR/FD) delivered better results than the SBR alone (AUC, 0.964 vs 0.899; p < 0.001). The sensitivity, specificity, and accuracy, respectively, were 79.1, 85.0, and 80.7% with SBR, 84.5, 97.5, and 88.0% with FD, and 92.7, 87.5, and 91.3% with SBR/FD. Conclusion Our results confirmed that the FD value is a useful diagnostic index, which reflects the tracer distribution in DAT SPECT images. The combined use of SBR and FD was more useful than either used alone.
Article
Objective: We aimed to assess whether a combined analysis of dopamine transporter (DAT)- and perfusion-SPECT images (or either) could: (1) distinguish atypical parkinsonian syndromes (APS) from Lewy body diseases (LBD; majority Parkinson disease [PD]), and (2) differentiate among APS subgroups (progressive supranuclear palsy [PSP], corticobasal syndrome [CBS], and multiple system atrophy [MSA]). Methods: We recruited consecutive patients with neurodegenerative parkinsonian syndromes (LBD, n = 46; APS, n = 33). Individual [(123)I]FP-CIT- and [(123)I]iodoamphetamine-SPECT images were coregistered onto anatomical MRI segmented into brain regions. Striatal DAT activity and regional perfusion were extracted from each brain region for each patient and submitted to logistic regression analyses. Stepwise procedures were used to select predictors that should be included in the models to distinguish APS from LBD, and differentiate among the APS subgroups. Receiver-operating characteristic (ROC) analyses were performed to measure diagnostic power. Leave-one-out cross-validation (LOOCV) was performed to evaluate the diagnostic accuracy. Results: The model to discriminate APS from LBD showed that the area under the ROC curve (AUC) was 0.923, while the total diagnostic accuracy (TDA) was 86.1% in LOOCV. In the model to distinguish PSP, CBS, and MSA from LBD, the AUC/TDA values were 0.978/94.6%, 0.978/87.0%, and 0.880/80.3%, respectively. In the model to differentiate between CBS and MSA, MSA and PSP, and PSP and CBS, the AUC/TDA values were 0.967/91.3%, 0.920/88.0%, 0.875/77.8%, respectively. Conclusion: An image-based automated classification using striatal DAT activity and regional perfusion patterns provided a good performance in the differential diagnosis of neurodegenerative parkinsonian syndromes without clinical information.
Article
A comprehensive analysis of the Parkinson's Progression Markers Initiative (PPMI) Dopamine Transporter Single Photon Emission Computed Tomography (DaTscan) images is carried out using a voxel-based logistic lasso model. The model reveals that sub-regional voxels in the caudate, the putamen, as well as in the globus pallidus are informative for classifying images into control and PD classes. Further, a new technique called logistic component analysis is developed, and this technique reveals that intra-population differences in dopamine transporter concentration and imperfect normalization are significant factors influencing logistic analysis. The interactions with handedness, sex, and age are also evaluated.
Article
TensorFlow is a second-generation open-source machine learning software library with a built-in framework for implementing neural networks in wide variety of perceptual tasks. Although TensorFlow usage is well established with computer vision datasets, the TensorFlow interface with DICOM formats for medical imaging remains to be established. Our goal is to extend the TensorFlow API to accept raw DICOM images as input; 1513 DaTscan DICOM images were obtained from the Parkinson’s Progression Markers Initiative (PPMI) database. DICOM pixel intensities were extracted and shaped into tensors, or n-dimensional arrays, to populate the training, validation, and test input datasets for machine learning. A simple neural network was constructed in TensorFlow to classify images into normal or Parkinson’s disease groups. Training was executed over 1000 iterations for each cross-validation set. The gradient descent optimization and Adagrad optimization algorithms were used to minimize cross-entropy between the predicted and ground-truth labels. Cross-validation was performed ten times to produce a mean accuracy of 0.938 ± 0.047 (95 % CI 0.908–0.967). The mean sensitivity was 0.974 ± 0.043 (95 % CI 0.947–1.00) and mean specificity was 0.822 ± 0.207 (95 % CI 0.694–0.950). We extended the TensorFlow API to enable DICOM compatibility in the context of DaTscan image analysis. We implemented a neural network classifier that produces diagnostic accuracies on par with excellent results from previous machine learning models. These results indicate the potential role of TensorFlow as a useful adjunct diagnostic tool in the clinical setting.
Article
The aims of this study were: (1) to cross-compare data from semiquantitative, software-assisted, and phantom-corrected evaluations of (123) I-ioflupane [(123) I]N-ω-fluoropropyl-2β-carbomethoxy-3β-{4-iodophenyl}nortropane FP-CIT brain single-photon emission computed tomography (SPECT) acquired in three centers; (2) to assess the accuracy of semiquantitative evaluation; and (3) to identify the threshold with the best accuracy, sensitivity, and specificity in patients with suspected Parkinsonian Syndrome. Two-hundred-twenty patients, acquired in three centers, were included. All of them underwent (123) I-FP-CIT brain SPECT. All examinations were analyzed with the freely available software, BasGan, and semiquantitative data were used to predict disease. Analysis was based on the values from the most deteriorated putamen and caudate, normalized for age, and corrected by anthropomorphic phantom data. Receiver operating characteristic (ROC) analysis was performed and areas under the curve (AUC) were estimated. Analysis showed high AUCs (.880, .866, .920, and .882 for each center and multicenter setting). Best thresholds were 1.53 and 1.56 for putamen and caudate, respectively. Thresholds of putamen data showed sensitivity and specificity of 86% and 89%, respectively, in the multicenter setting. Neither sensibility nor specificity showed significant differences among centers. A unique, accurate threshold for all centers, with high sensitivity and specificity was identified. Semiquantitative assessment of (123) I-FP-CIT brain SPECT among different centers resulted reliable, accurate, and potentially useful in clinical trials. Copyright © 2015 by the American Society of Neuroimaging.
Article
Purpose: The study's objective was to develop diagnostic predictive models using data from two commonly used [(123)I]FP-CIT SPECT assessment methods: region-of-interest (ROI) analysis and whole-brain voxel-based analysis. Methods: We included retrospectively 80 patients with vascular parkinsonism (VP) and 164 patients with Parkinson's disease (PD) who underwent [(123)I]FP-CIT SPECT. Nuclear-medicine specialists evaluated the scans and calculated bilateral caudate and putamen [(123)I]FP-CIT uptake and asymmetry indices using BRASS software. Statistical parametric mapping (SPM) was used to compare the radioligand uptake between the two diseases at the voxel level. Quantitative data from these two methods, together with potential confounding factors for dopamine transporter availability (sex, age, disease duration and severity), were used to build predictive models following a tenfold cross-validation scheme. The performance of logistic regression (LR), linear discriminant analysis and support vector machine (SVM) algorithms for ROI data, and their penalized versions for SPM data (penalized LR, penalized discriminant analysis and SVM), were assessed. Results: Significant differences were found in the ROI analysis after covariate correction between VP and PD patients in [(123)I]FP-CIT uptake in the more affected side of the putamen and the ipsilateral caudate. Age, disease duration and severity were also found to be informative in feeding the statistical model. SPM localized significant reductions in [(123)I]FP-CIT uptake in PD with respect to VP in two specular clusters comprising areas corresponding to the left and right striatum. The diagnostic predictive accuracy of the LR model using ROI data was 90.3 % and of the SVM model using SPM data was 90.4 %. Conclusion: The predictive models built with ROI data and SPM data from [(123)I]FP-CIT SPECT provide great discrimination accuracy between VP and PD. External validation of these methods is necessary to confirm their applicability across centres.
Article
The study of neurodegenerative diseases has been based for some time on visual and semi-quantitative analysis of medical imaging. This is the case of Parkinsonian Syndrome (PS) or Parkinsonism, which is the second most common neurodegenerative disorder, where 123I-ioflupane (better known by its tradename, DaTSCAN) images have been of great help. Recently, new developments in machine learning methods and statistics have been applied to the analysis of medical images, yielding to a more operator-independent, objective analysis of them, and thus, setting the Computer Aided Diagnosis (CAD) paradigm. In this work, a new CAD system based on preprocessing, voxel selection, feature extraction and classification of the images is proposed. After preprocessing the images, voxels are ranked by means of their significance in class discrimination, and the first N are selected. Then, these voxels are modelled using Independent Component Analysis (ICA), obtaining a few components that represent each image, which will be used later to train a classifier. The proposed system has been tested on two databases: a 208-DaTSCAN image database from the “Virgen de la Victoria” Hospital in Málaga (VV), Spain and a 289-DaTSCAN image database from the Parkinson Progression Markers Initiative (PPMI). Values of accuracy up to 94.7% and 91.3% for VV and PPMI databases are achieved by the proposed system, which has proved its robustness in PS pattern detection, and significantly improves the baseline Voxels-as-Features (VAF) approach, used as an approximation of the visual analysis.
Article
BACKGROUND AND PURPOSEOver the last two decades I-123-FP-CIT-SPECT, has been used to discriminate neurodegenerative Parkinsonian syndrome from other diseases. BasGan is a freely available software that assists I-123-FP-CIT-SPECT evaluation by estimating semiquantitative values for each basal nucleus and compares the results to a database of healthy subjects. The aims of this study were: (1) to assess the accuracy of qualitative analysis and of semiquantitative, BasGan-assisted evaluations of I-123-FP-CIT-SPECT; (2) to compare the accuracy of both methods when applied to doubtful cases; (3) to appreciate the reproducibility of the BasGan-assisted evaluations. MATERIALS AND METHODS Seventy-eight patients were included in this 4-year follow-up study. The diagnostic cut-off for semiquantitative uptake values of each basal nucleus was determined based on ROC curves analysis. Accuracy scores were calculated for the entire population and for doubtful cases. Intra- and interoperator reproducibility was assessed. RESULTSAccuracy of the software-assisted analyses was high for data from each nucleus. In doubtful exams accuracy was higher when using BasGan as opposed to relying solely on visual assessment. Intra- and interoperator reproducibility of the BasGan-assisted evaluations was good to excellent. CONCLUSION BasGan-assisted evaluations of I-123-FP-CIT-SPECT were very useful, particularly in doubtful cases. Multicenter studies are mandatory before routine use of BasGan.
Article
An accurate and early diagnosis of Parkinsonian syndrome (PS) is nowadays a challenge. This syndrome includes several pathologies with similar symptoms (Parkinson's disease, multisystem atrophy, progressive supranuclear palsy, corticobasal degeneration and others) which make the diagnosis more difficult. (123)I-ioflupane allows to obtain in vivo images of the brain that can be used to assist the PS diagnosis and provides a way to improve its accuracy. In this paper, we show a novel method to automatically classify (123)I-ioflupane images into two groups: controls or PS. The proposed methodology analyzes separately each hemisphere of the brain by means of a novel approach based on partial least squares (PLS) and support vector machine. A database with 189 (123)I-ioflupane images (94 controls and 95 pathological images) was used for evaluation purposes. The application of the proposed method based on PLS yields high accuracy rates up to 94.7% with sensitivity = 93.7% and specificity = 95.7%, outperforming previous approaches based on singular value decomposition, which are used as a reference. The use of advanced techniques based on classical signal analysis and their application to each hemisphere of the brain separately improves the (assisted) diagnosis of PS.
Article
We present a method of automatic classification of I-fluoropropyl-carbomethoxy-3β-4-iodophenyltropane (FP-CIT) images. This technique uses singular value decomposition (SVD) to reduce a training set of patient image data into vectors in feature space (D space). The automatic classification techniques use the distribution of the training data in D space to define classification boundaries. Subsequent patients can be mapped into D space, and their classification can be automatically given. The technique has been tested using 116 patients for whom the diagnosis of either Parkinsonian syndrome or non-Parkinsonian syndrome has been confirmed from post I-FP-CIT imaging follow-up. The first three components were used to define D space. Two automatic classification tools were used, naïve Bayes (NB) and group prototype. A leave-one-out cross-validation was performed to repeatedly train and test the automatic classification system. Four commercially available systems for the classification were tested using the same clinical database. The proposed technique combining SVD and NB correctly classified 110 of 116 patients (94.8%), with a sensitivity of 93.7% and specificity of 97.3%. The combination of SVD and an automatic classifier performed as well or better than the commercially available systems. The combination of data reduction by SVD with automatic classifiers such as NB can provide good diagnostic accuracy and may be a useful adjunct to clinical reporting.
Article
Single photon emission computed tomography (SPECT) imaging with (123)I-FP-CIT is of great value in differentiating patients suffering from Parkinson's disease (PD) from those suffering from essential tremor (ET). Moreover, SPECT with (123)I-IBZM can differentiate PD from Parkinson's "plus" syndromes. Diagnosis is still mainly based on experienced observers' visual assessment of the resulting images while many quantitative methods have been developed in order to assist diagnosis since the early days of neuroimaging. The aim of this work is to attempt to categorize, briefly present and comment on a number of semi-quantification methods used in nuclear medicine neuroimaging. Various arithmetic indices have been introduced with region of interest (ROI) manual drawing methods giving their place to automated procedures, while advancing computer technology has allowed automated image registration, fusion and segmentation to bring quantification closer to the final diagnosis based on the whole of the patient's examinations results, clinical condition and response to therapy. The search for absolute quantification has passed through neuroreceptor quantification models, which are invasive methods that involve tracer kinetic modelling and arterial blood sampling, a practice that is not commonly used in a clinical environment. On the other hand, semi-quantification methods relying on computers and dedicated software try to elicit numerical information out of SPECT images. The application of semi-quantification methods aims at separating the different patient categories solving the main problem of finding the uptake in the structures of interest. The semi-quantification methods which were studied fall roughly into three categories, which are described as classic methods, advanced automated methods and pixel-based statistical analysis methods. All these methods can be further divided into various subcategories. The plethora of the existing semi-quantitative methods reinforces the feeling that visual assessment is still the base of image interpretation and that the unambiguous numerical results that will allow the absolute differentiation between the known diseases have not been standardized yet. Switching to a commonly agreed-ideally PC-based-automated software that may take raw or mildly processed data (checked for consistency and maybe corrected for attenuation and/or scatter and septal penetration) as input, work with basic operator's inference and produce validated numerical results that will support the diagnosis is in our view the aim towards which efforts should be directed. After all, semi-quantification can improve sensitivity, strengthen diagnosis, aid patient's follow-up and assess the response to therapy. Objective diagnosis, altered diagnosis in marginal cases and a common approach to multicentre trials are other benefits and future applications of semi-quantification.
Article
The purpose of this study was to show the viability and performance of a shape-based pattern recognition technique applied to I-N-omega-fluoropropyl-2-beta-carbomethoxy-3beta-(4-iodophenyl) nortropane single-photon emission computed tomography (FP-CIT SPECT) in patients with parkinsonism. A fully automated pattern recognition tool, based on the shape of FP-CIT SPECT images, was written using Java. Its performance was evaluated and compared with QuantiSPECT, a region-of-interest-based quantitation tool, and observer performance using receiver operating characteristic analysis and kappa statistics. The techniques were compared using a sample of patients and controls recruited from a prospective community-based study of first presentation of parkinsonian symptoms with longitudinal follow up (median 3 years). The shape-based technique as well as the conventional semiquantitative approach was performed by experienced observers. The technique had a high level of automation, thereby avoiding observer/operator variability. A pattern recognition approach is a viable alternative to traditional methods of analysis in FP-CIT SPECT and has additional advantages.
Article
Neuroimaging studies of the striatal dopamine transporter (DAT) are useful in the assessment of the dopaminergic system in Parkinson's disease (PD). We used positron emisson tomography (PET) and the tracer [11C]FE-CIT to measure DAT binding in the caudate nucleus and putamen of 31 patients with PD, 5 with essential tremor and 8 healthy control subjects. Of the patients with PD, 17 were drug naive, while the others were either on levodopa or dopamine agonist monotherapy. DAT binding was significantly reduced in the caudate nucleus and to a greater extent in the putamen of PD patients compared to both healthy controls and essential tremor individuals. No overlap was observed between putamen values in PD and normals. No differences were found between controls and essential tremor subjects. These data confirm that measurements of DAT binding can provide an accurate and highly sensitive measure of degeneration in the dopamine system in PD.
Article
A comparative study was carried out on two promising presynaptic dopamine transporter single-photon emission tomography (SPECT) radioligands with a fast pharmacokinetic profile, 123I-FP-beta-CIT (FP) and 99mTc-TRODAT-1 (TR), in order to assess their differential diagnostic power in early parkinsonism and their sensitivity for detection of disease progression. This cross-sectional study was conducted on 96 patients with early-stage parkinsonism referred in a tertiary clinical setting. Mean disease duration was 2.0+/-1.3 years, and patients had a modified Hoehn and Yahr (H&Y) stage of 1-2 (average 1.2). Forty-seven patients received TR, and 49 received FP. In both groups, ten patients with normal presynaptic function were included as a control population; all other patients were clinically diagnosed as having idiopathic Parkinson's disease. Groups were matched for gender, age, disease duration and modified H&Y stage. Triple-head gamma camera SPECT was analysed using a semiquantitative index of transporter binding (BI). Discriminant analysis with cross-validation resulted in a maximal classification accuracy for FP of 93% (sensitivity 95% and specificity 86%) for the contralateral putamen BI. For TR, the corresponding values were 87% accuracy, 92% sensitivity and 70% specificity. For FP, disease duration was correlated with both the putamen BI (-8.8%/year, rho=-0.41, P=0.025) and the putamen/caudate ratio (-7.4%/year, rho=-0.51, P=0.004), but for TR no significant correlation was found (all P values >0.5). In conclusion, both FP and TR show high sensitivity in a clinically relevant setting, but FP has superior accuracy for early differential diagnosis of idiopathic parkinsonism and non-degenerative extrapyramidal disorders, as well as better sensitivity for disease follow-up.
Article
The quantification of DaTSCAN images can be used as an adjunct to visual assessment to differentiate between Parkinson's syndrome and essential tremor. Many programs have been written to assess the relative uptake in the striatum. To compare two of the commercially available programs: QuantiSPECT, which analyses isolated data in two dimensions, and BRASS, which performs three-dimensional processing referencing a normal image template. Twenty-two patients (11 with Parkinson's syndrome and 11 with essential tremor) were visually assessed by two nuclear medicine consultants. The patient data were then processed using two commercial programs to determine the relative uptake in the striatum. A comparison of the results from the programs was performed, together with a comparison with the visual assessment. The inter-operator and intra-operator variabilities were also ascertained. All programs and processing methods could distinguish between Parkinson's syndrome and essential tremor. There was also a good correlation between the results from the three- and two-dimensional methods. The intra-operator and inter-operator variabilities were dependent on the amount of operator intervention. Both programs allowed statistical differentiation between Parkinson's syndrome and essential tremor. Strict operator protocols are needed with QuantiSPECT to reduce inter- and intra-operator variation. The three-dimensional method (BRASS) gave greater concordance than the two-dimensional method (QuantiSPECT) with the visual assessment, but at a cost of increased operator time.
Article
The correct diagnosis of Parkinson's disease is important for prognostic and therapeutic reasons and is essential for clinical research. Investigations of the diagnostic accuracy for the disease and other forms of parkinsonism in community-based samples of patients taking antiparkinsonian medication confirmed a diagnosis of parkinsonism in only 74% of patients and clinically probable Parkinson's disease in 53% of patients. Clinicopathological studies based on brain bank material from the UK and Canada have shown that clinicians diagnose the disease incorrectly in about 25% of patients. In these studies, the most common reasons for misdiagnosis were presence of essential tremor, vascular parkinsonism, and atypical parkinsonian syndromes. Infrequent diagnostic errors included Alzheimer's disease, dementia with Lewy bodies, and drug-induced parkinsonism. Increasing knowledge of the heterogeneous clinical presentation of the various parkinsonisms has resulted in improved diagnostic accuracy of the various parkinsonian syndromes in specialised movement-disorder units. Also genetic testing and various other ancillary tests, such as olfactory testing, MRI, and dopamine-transporter single-photon-emission computed-tomography imaging, help with clinical diagnostic decisions.
Article
Clinical differential diagnosis in parkinsonism can be difficult especially at early stages. We investigated whether combined perfusion and dopamine transporter (DAT) imaging can aid in the differential diagnosis of parkinsonian disorders: idiopathic Parkinson's disease (IPD), progressive supranuclear palsy (PSP), multiple system atrophy (MSA), dementia with Lewy bodies (LBD), and essential tremor (ET). One hundred twenty-nine patients were studied, retrospectively (69 males; 24 MSA, 12 PSP, 8 LBD, 27 ET, and 58 IPD; mean disease duration, 3.5 +/- 3.7 y). Diagnosis was based on established clinical criteria after follow-up of 5.5 +/- 3.8 y in a university specialist movement disorders clinic. Group characterization was done using a categoric voxel-based design and, second, a predefined volume-of-interest approach along Brodmann areas (BA) and subcortical structures, including striatal asymmetry and anteroposterior indices. Stepwise forward discriminant analysis was performed with cross-validation (CV) using the leave-one-out technique. Characteristic patterns for perfusion and DAT were found for all pathologies. In the parkinson-plus group, MSA, PSP, and LBD could be discriminated in 100% (+CV) of the cases. When including IPD, discrimination accuracy was 82.4% (99% without CV). 2beta-Carbomethoxy-3beta-(4-iodophenyl)nortropane imaging as a single technique was able to discriminate between ET and neurodegenerative forms with an accuracy of 93.0% (+CV); inclusion of perfusion information augmented this slightly to 97.4% (+CV). Dual-tracer DAT and perfusion SPECT in combination with discrimination analysis allows an automated, accurate differentiation between the most common forms of parkinsonism in a clinically relevant setting.
Article
A technique is described for accurate quantification of the specific binding ratio (SBR) in [(123)I]FP-CIT SPECT brain images. Using a region of interest (ROI) approach, the SBR is derived from a measure of total striatal counts that takes into account the partial volume effect. Operator intervention is limited to the placement of the striatal ROIs, a task facilitated by the use of geometrical template regions. The definition of the image for the analysis is automated and includes transaxial slices within a "slab" approximately 44 mm thick centred on the highest striatal signal. The reference region is automatically defined from the non-specific uptake in the whole brain enclosed in the slab, with exclusion of the striatal region. A retrospective study consisting of 25 normal and 30 abnormal scans-classified by the clinical diagnosis reached with the scan support-was carried out to assess intra- and inter-operator variability of the technique and its clinical usefulness. Three operators repeated the quantification twice and the variability was measured by the coefficient of variation (COV). The COVs for intra- and inter-operator variability were 3% and 4% respectively. A cutoff approximately 4.5 was identified that separated normal and abnormal groups with a sensitivity, specificity and diagnostic concordance of 97%, 92% and 95% respectively. The proposed technique provides a reproducible and sensitive index. It is hoped that its independence from the partial volume effect will improve consistency in quantitative measurements between centres with different imaging devices and analysis software.
Article
In the semi-quantitative assessment of DaTSCAN images, it has been suggested that the ratio of tracer accumulation in the putamen to that in the caudate nucleus may be helpful and could allow parkinsonian syndromes progression to be assessed. Separation of ratio values has been reported when early Parkinson's disease is compared with essential tremor. The separation is lost, however, when the Parkinson's disease is not early stage. To evaluate whether a two-stage analysis can differentiate between parkinsonian syndromes, of various stages, and essential tremor, and whether such a two-stage analysis can be undertaken in a single step using artificial neural networks (ANNs). Data from 18 patients were analysed. Quantification was undertaken by manually drawing irregular regions of interest (ROIs): over each caudate nucleus and putamen and over an occipital cortex area near the posterior edge of the brain. A two-stage analysis was undertaken and was repeated, in a single step, using an ANN. The first stage, of the two-stage analysis, identified 12 patients with non-early parkinsonian syndromes. The remaining six patients were then successfully classified into early parkinsonian syndromes and essential tremor. The ANN analysis successfully discriminated parkinsonian syndromes from essential tremor, in all patients, in a single step. The two-stage process provides a method for classifying early disease without being compromised by the noise from non-early disease. The results of the single stage ANN analysis were very definite and it may be considered to have potential in the quantification of DaTSCAN images for clinical use.
Article
To design a novel algorithm (BasGan) for automatic segmentation of striatal (123)I-FP-CIT SPECT. The BasGan algorithm is based on a high-definition, three-dimensional (3D) striatal template, derived from Talairach's atlas. A blurred template, obtained by convolving the former with a 3D Gaussian kernel (FWHM = 10 mm), approximates striatal activity distribution. The algorithm performs translations and scale transformation on the bicommissural aligned image to set the striatal templates with standard size in an appropriate initial position. An optimization protocol automatically performs fine adjustments in the positioning of blurred templates to best match the radioactive counts, and locates an occipital ROI for background evaluation. Partial volume effect correction is included in the process of uptake computation of caudate, putamen and background. Experimental validation was carried out by means of six acquisitions of an anthropomorphic striatal phantom. The BasGan software was applied to a first set of patients with Parkinson's disease (PD) versus patients affected by essential tremor. A highly significant correlation was achieved between true binding potential and measured (123)I activity from the phantom. (123)I-FP-CIT uptake was significantly lower in all basal ganglia in the PD group versus controls with both BasGan and a conventional ROI method used for comparison, but particularly with the former. Correlations with the motor UPDRS score were far more significant with the BasGan. The novel BasGan algorithm automatically performs the 3D segmentation of striata. Because co-registered MRI is not needed, it can be used by all nuclear medicine departments, since it is freely available on the Web.
EANM procedure guidelines for brain neurotransmission SPECT using
  • J Darcourt
  • J Booij
  • K Tatsch
  • A Varrone
Striatal binding of FP-CIT: a simple method to separate Parkinson’s disease patients and normal controls
  • D C Costa
  • Z Walker
  • S Dizdarevic
  • C Ionnides
Comparison of different methods of DatSCAN quantification
  • R J Morton
  • M J Guy
  • R Clauss
  • P J Hinton
  • RJ Morton
Computer-aided diagnosis of Parkinson’s disease based on
  • Fpm Oliveira
  • M Castelo-Branco
Striatal binding of FP-CIT: a simple method to separate Parkinson's disease patients and normal controls. Paper presented at the Joint Congress of the European Association of Nuclear Medicine and the World Federation of Nuclear Medicine and Biology
  • D C Costa
  • Z Walker
  • S Dizdarevic
  • C Ionnides
Costa DC, Walker Z, Dizdarevic S, Ionnides C et al (1988) Striatal binding of FP-CIT: a simple method to separate Parkinson's disease patients and normal controls. Paper presented at the Joint Congress of the European Association of Nuclear Medicine and the World Federation of Nuclear Medicine and Biology, Berlin, Germany, 30 Aug to 4 Sep Darcourt J, Booij J, Tatsch K, Varrone A 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(2):443-450. https://doi.org/10.1007/s00259-009-1267-x