Classification of degenerative parkinsonism subtypes by support-vector-machine analysis and striatal 123I-FP-CIT indices

Abstract

Objectives To provide an automated classification method for degenerative parkinsonian syndromes (PS) based on semiquantitative 123 I-FP-CIT SPECT striatal indices and support-vector-machine (SVM) analysis. Methods 123 I-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 fivefold 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 123 I-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.

Using Artificial Neural Network to Discriminate Parkinson’s Disease from Other Parkinsonisms by Focusing on Putamen of Dopamine Transporter SPECT Images

“…The performance of the ANN classifier, with sensitivity and specificity both above 80%, was comparable to that of quantitative olfactory examinations and MIBG myocardiac scintigraphy suggested by diagnostic guidelines. Furthermore, this method is promising because of several advantages: (1) as the sample size of the dataset increases, training results can be further improved; (2) with an adequate number of images taken during the earlier phase of disease (PD or atypical parkinsonian syndromes), the ANN classifier may be trained to identify PD at an early phase [ 24 ] or even possibly at a preclinical phase; (3) medical centers and hospitals can train a site-specific ANN classifier using SPECT images based on their own existing dataset without developing new diagnostic modalities or purchasing expensive machines, especially for places where MIBG is not available; (4) SPECT is more widely available, so that when the diagnosis is not straightforward, physicians tend to order SPECT imaging first to confirm striatal neuron loss, such as to differentiate essential tremors from PD, but not MIBG myocardial scintigraphy before proving a neurodegenerative disease in the early phase; and (5) PD can be differentiated from many disease types of parkinsonism, not just a few Parkinson-plus syndromes or other Lewy body diseases [ 19 , 30 ]. Therefore, this classifier is more applicable when facing uncertain types of parkinsonism in clinical practice.…”

“…To classify parkinsonism based on DAT-SPECT images, advanced engineering techniques with semi-quantitative analysis have been applied [ 18 ]. In addition, images or signals from striatal regions (SRs) alone can provide adequate differentiating information [ 19 ]. One research group differentiated degenerative parkinsonism using a computer-aided automatic algorithm and SR and whole-brain uptake patterns.…”

Imaging of dopamine transporters in Parkinson disease: a meta‐analysis of 18F/123I‐FP‐CIT studies

“…We included the remaining 33 articles (six articles reported PET studies involving 779 patients with PD and 124 controls, and 27 reported SPECT studies involving 1244 PD patients and 859 controls) in this meta‐analysis. These studies comprised cross‐sectional surveys (10 studies), case‐control studies (19 studies), and cohort studies (four studies) 17,22–26,35–61 . Figure 1 illustrates the protocol for the literature search and study selection process.…”

A Shape Approximation for Medical Imaging Data

“…The selected images are superimposed into a single 2D image, which is called a 2D-combined image hereafter, for disease diagnosis. For example, Prashanth et al [ 1 ], Taylor et al [ 2 ], Oliveira et al [ 3 ], Nicastro et al [ 4 ], and Iwabuchi et al [ 5 ] constructed classifiers for identifying Parkinson’s disease (PD) by extracting features from 2D-combined SPECT images. However, this approach may lose some useful information or features for classification since a 2D-combined image is a projection of the original 3D structure on a specific 2D space.…”