Clustering Analysis of FDG-PET Imaging in Primary Progressive Aphasia

Abstract

Background: Primary progressive aphasia (PPA) is a clinical syndrome characterized by the neurodegeneration of language brain systems. Three main clinical forms (non-fluent, semantic, and logopenic PPA) have been recognized, but applicability of the classification and the capacity to predict the underlying pathology is controversial. We aimed to study FDG-PET imaging data in a large consecutive case series of patients with PPA to cluster them into different subtypes according to regional brain metabolism.Methods: 122 FDG-PET imaging studies belonging to 91 PPA patients and 28 healthy controls were included. We developed a hierarchical agglomerative cluster analysis with Ward's linkage method, an unsupervised clustering algorithm. We conducted voxel-based brain mapping analysis to evaluate the patterns of hypometabolism of each identified cluster.Results: Cluster analysis confirmed the three current PPA variants, but the optimal number of clusters according to Davies-Bouldin index was 6 subtypes of PPA. This classification resulted from splitting non-fluent variant into three subtypes, while logopenic PPA was split into two subtypes. Voxel-brain mapping analysis displayed different patterns of hypometabolism for each PPA group. New subtypes also showed a different clinical course and were predictive of amyloid imaging results.Conclusion: Our study found that there are more than the three already recognized subtypes of PPA. These new subtypes were more predictive of clinical course and showed different neuroimaging patterns. Our results support the usefulness of FDG-PET in evaluating PPA, and the applicability of computational methods in the analysis of brain metabolism for improving the classification of neurodegenerative disorders.

An application of machine learning with feature selection to improve diagnosis and classification of neurodegenerative disorders

“…Nevertheless and once the feature selection algorithm was chosen and the best set of features were selected from PCA, we explored in the following section unsupervised learning process to compare with previous experiments carried out in [13].…”

“…aphasia) makes PPA a good model among neurodegenerative disorders to apply the machine learning techniques. Indeed, we recently have introduced the possibility of the existence of some additional subtypes within non-fluent and logopenic variants, with a better prediction of outcome using clustering analysis [13]…”

“…All PPA patients met the current consensus criteria [21] and were classified into the three clinical variants taking into account the diagnostic criteria and follow-up. Further details about clinical and neuropsychological assessments are shown in a previous work [13].…”

An update on blood-based biomarkers for non-Alzheimer neurodegenerative disorders

“…In bvFTD, MRI studies demonstrate prominent, usually symmetric, atrophy of the frontal lobes 8 . In contrast to the pattern typically seen in AD, involving the posterior temporal/parietal lobes and the posterior cingulate/precuneus [58][59][60][61][62] , three main patterns of glucose hypometabolism can classically be observed in FTD: precentral and inferior frontal in nfvPPA, anterior temporal lobes in svPPA (usually with marked leftward asymmetry) 63,64 , and frontal as well as temporo-limbic predominant patterns in bvFTD 65 . In case series that have examined Aβ status among FTD patients, low rates of Aβ positivity have been reported (0-15%), in line with Aβ plaques not being a feature characteristic of the FTLD pathology spectrum [66][67][68] .…”

Data-driven FDG-PET subtypes of Alzheimer’s disease-related neurodegeneration

“…The patients with AD dementia were classi ed into hypometabolic subtypes using agglomerative hierarchical clustering of voxel-wise FDG-PET data with Ward's linkage as implemented in MATLAB software (29,30).…”