Data-driven classification of patients with primary progressive aphasia

Data-driven subtypes of mixed semantic-logopenic primary progressive aphasia: Linguistic features, biomarker profiles and brain metabolic patterns

Mixed primary progressive aphasia (mPPA) accounts for a substantial proportion of primary progressive aphasia (PPA) cases. However, the lack of a standardised definition of this condition has resulted in misclassification of PPA cases. In this study, we enrolled 55 patients diagnosed with PPA, comprising 12 semantic variant (svPPA), 23 logopenic variant (lvPPA), and 20 mPPA cases with linguistic characteristics consistent with both svPPA and lvPPA (s/lvPPA). All patients underwent language assessments, evaluation of Alzheimer's disease biomarkers (via cerebrospinal fluid analysis or Amyloid-PET), and 18F-FDG-PET brain scans. An agglomerative hierarchical clustering (AHC) analysis based on linguistic characteristics revealed two distinct clusters within the s/lvPPA group: cluster k1 (n = 10) displayed an AD-like biomarker profile, with lower levels of Aβ42 and Aβ42/Aβ40 ratio, along with higher levels of t-tau and p-tau compared to cluster k2 (n = 10). Interestingly, k1 exhibited linguistic features that were similar to those of svPPA. Both clusters exhibited extensive temporoparietal hypometabolism. These findings support the hypothesis that a subgroup of s/lvPPA may represent a clinical manifestation of AD-related PPA.

Clinical dimensions along the non-fluent variant primary progressive aphasia spectrum

It is debated whether primary progressive apraxia of speech (PPAOS) and progressive agrammatic aphasia (PAA) belong to the same clinical spectrum, traditionally termed non-fluent/agrammatic variant primary progressive aphasia (nfvPPA), or exist as two completely distinct syndromic entities with specific pathologic/prognostic correlates. We analysed speech, language and disease severity features in a comprehensive cohort of patients with progressive motor speech impairment and/or agrammatism to ascertain evidence of naturally occurring, clinically meaningful non-overlapping syndromic entities (e.g. PPAOS and PAA) in our data. We also assessed if data-driven latent clinical dimensions with aetiologic/prognostic value could be identified. We included 98 participants, 43 of whom had an autopsy-confirmed neuropathological diagnosis. Speech pathologists assessed motor speech features indicative of dysarthria and apraxia of speech (AOS). Quantitative expressive/receptive agrammatism measures were obtained and compared with healthy controls. Baseline and longitudinal disease severity was evaluated using the Clinical Dementia Rating Sum of Boxes (CDR-SB). We investigated the data's clustering tendency and cluster stability to form robust symptom clusters and employed principal component analysis to extract data-driven latent clinical dimensions (LCD). The longitudinal CDR-SB change was estimated using linear mixed-effects models. Of the participants included in this study, 93 conformed to previously reported clinical profiles (75 with AOS and agrammatism, 12 PPAOS and six PAA). The remaining five participants were characterized by non-fluent speech, executive dysfunction and dysarthria without apraxia of speech or frank agrammatism. No baseline clinical features differentiated between frontotemporal lobar degeneration neuropathological subgroups. The Hopkins statistic demonstrated a low cluster tendency in the entire sample (0.45 with values near 0.5 indicating random data). Cluster stability analyses showed that only two robust subgroups (differing in agrammatism, executive dysfunction and overall disease severity) could be identified. Three data-driven components accounted for 71% of the variance [(i) severity-agrammatism; (ii) prominent AOS; and (iii) prominent dysarthria]. None of these data-driven LCDs allowed an accurate prediction of neuropathology. The severity-agrammatism component was an independent predictor of a faster CDR-SB increase in all the participants. Higher dysarthria severity, reduced words per minute and expressive and receptive agrammatism severity at baseline independently predicted accelerated disease progression. Our findings indicate that PPAOS and PAA, rather than exist as completely distinct syndromic entities, constitute a clinical continuum. In our cohort, splitting the nfvPPA spectrum into separate clinical phenotypes did not improve clinical-pathological correlations, stressing the need for new biological markers and consensus regarding updated terminology and clinical classification.

Characterization of the logopenic variant of Primary Progressive Aphasia: A systematic review and meta-analysis

The linguistic and anatomical variability of the logopenic variant of Primary Progressive Aphasia (lv-PPA) as defined by current diagnostic criteria has been the topic of an intense debate. The present review and meta-analysis aims at characterizing the profile of lv-PPA, by a comprehensive analysis of the available literature on the neuropsychological, neuroimaging, electrophysiological, pathological, and genetic features of lv-PPA. We conducted a systematic bibliographic search, leading to the inclusion of 207 papers. Of them, 12 were used for the Anatomical Likelihood Estimation meta-analysis on grey matter revealed by magnetic resonance imaging data. The results suggest that the current guidelines outline a relatively consistent syndrome, characterized by a core set of linguistic and, to a lesser extent, non-linguistic deficits, mirroring the involvement of left temporal and parietal regions typically affected by Alzheimer Disease pathology. Variations of the lv-PPA profile are discussed in terms of heterogeneity of the neuropsychological instruments and the diagnostic criteria adopted.

Utility of the Addenbrooke’s Cognitive Examination III online calculator to differentiate the primary progressive aphasia variants

The Addenbrooke’s Cognitive Examination III is a brief cognitive screening tool that is widely used for the detection and monitoring of dementia. Recent findings suggest that the three variants of primary progressive aphasia can be distinguished based on their distinct profiles on the five subdomain scores of this test. Here, we investigated the utility of the Addenbrooke’s Cognitive Examination III to differentiate the primary progressive aphasia variants based on their item-by-item performance profiles on this test. From these results, we created an interactive primary progressive aphasia Addenbrooke’s Cognitive Examination III calculator which predicts the variant based on a patient’s unique item-by-item profile. Twenty-eight logopenic variant, 25 non-fluent variant and 37 semantic variant primary progressive aphasia patients and 104 healthy controls completed the Addenbrooke’s Cognitive Examination III at first clinical presentation. Multinomial regression analyses were conducted to establish performance profiles among groups, and R Shiny from RStudio was used to create the interactive Addenbrooke’s Cognitive Examination III diagnostic calculator. To verify its accuracy, probability values of the regression model were derived based on a 5-fold cross-validation of cases. The calculator’s accuracy was then verified in an independent sample of 17 logopenic, 19 non-fluent and 13 semantic variant primary progressive aphasia patients and 68 Alzheimer’s disease patients who had completed the Addenbrooke’s Cognitive Examination III (or an older version of this test: Revised) and had in vivo amyloid-PET imaging and/or brain autopsy pathological confirmation. Cross-validation of cases in the calculator model revealed different rates of sensitivity in classifying variants: semantic = 100%, non-fluent = 80.6% and logopenic = 79.9%; healthy controls were distinguished from primary progressive aphasia patients with 100% sensitivity. Verification of in vivo amyloid and/or autopsy-confirmed patients showed that the calculator correctly classified 10/13 (77%) semantic variant, 3/19 (16%) non-fluent variant and 4/17 (24%) logopenic variant patients. Importantly, for patients who were not classified, diagnostic probability values mostly pointed toward the correct clinical diagnosis. Furthermore, misclassified diagnoses of the primary progressive aphasia cohort were rare (1/49; 2%). Although 22 of the 68 Alzheimer’s disease patients (32%) were misclassified with primary progressive aphasia, 19/22 were misclassified with the logopenic variant (i.e. falling within the same neuropathological entity). The Addenbrooke’s Cognitive Examination III primary progressive aphasia diagnostic calculator demonstrates sound accuracy in differentiating the variants based on an item-by-item Addenbrooke’s Cognitive Examination III profile. This calculator represents a new frontier in using data-driven approaches to differentiate the primary progressive aphasia variants.

Learning From Mistakes: Cognitive and Metabolic Correlates of Errors on Picture Naming in the Alzheimer’s Disease Spectrum

Background: Analysis of subtypes of picture naming errors produced by patients with Alzheimer’s disease (AD) have seldom been investigated yet may clarify the cognitive and neural underpinnings of naming in the AD spectrum. Objective: To elucidate the neurocognitive bases of picture naming in AD through a qualitative analysis of errors. Methods: Over 1000 naming errors produced by 70 patients with amnestic, visuospatial, linguistic, or frontal AD were correlated with general cognitive tests and with distribution of hypometabolism on FDG-PET. Results: Principal component analysis identified 1) a Visual processing factor clustering visuospatial tests and unrecognized stimuli, pure visual errors and visual-semantic errors, associated with right parieto-occipital hypometabolism; 2) a Concept-Lemma factor grouping language tests and anomias, circumlocutions, superordinates, and coordinates, correlated with left basal temporal hypometabolism; 3) a Lemma-Phonology factor including the digit span and phonological errors, linked with left temporo-parietal hypometabolism. Regression of brain metabolism on individual errors showed that errors due to impairment of basic and higher-order processing of object visual attributes or of their interaction with semantics, were related with bilateral occipital and left occipito-temporal dysfunction. Omissions and superordinates were linked to degradation of broad and basic concepts in the left basal temporal cortex. Semantic-lexical errors derived from faulty semantically- and phonologically-driven lexical retrieval in the left superior and middle temporal gyri. Generation of nonwords was underpinned by of phonological impairment within the left inferior parietal cortex. Conclusion: Analysis of individual naming errors allowed to outline a comprehensive anatomo-functional model of picture naming in classical and atypical AD.

Natural language processing in psychiatry: the promises and perils of a transformative approach

A person's everyday language can indicate patterns of thought and emotion predictive of mental illness. Here, we discuss how natural language processing methods can be used to extract indicators of mental health from language to help address long-standing problems in psychiatry, along with the potential hazards of this new technology.

'Moderate global aphasia': A generalized decline of language processing caused by glioma surgery but not stroke

Unlike stroke, neurosurgical removal of left-hemisphere gliomas acts upon a reorganized language network and involves brain areas rarely damaged by stroke. We addressed whether this causes the profiles of neurosurgery- and stroke-induced language impairments to be distinct. K-means clustering of language assessment data (neurosurgery cohort: N = 88, stroke cohort: N = 95) identified similar profiles in both cohorts. But critically, a cluster of individuals with specific phonological deficits was only evident in the stroke but not in the neurosurgery cohort. Thus, phonological deficits are less clearly distinguished from other language deficits after glioma surgery compared to stroke. Furthermore, the correlations between language production and comprehension scores at different linguistic levels were more extensive in the neurosurgery than in the stroke cohort. Our findings suggest that neurosurgery-induced language impairments do not correspond to those caused by stroke, but rather manifest as a 'moderate global aphasia' - a generalized decline of language processing abilities.

Utility of the Repeat and Point Test for Subtyping Patients With Primary Progressive Aphasia

BACKGROUND

Primary progressive aphasia (PPA) may present with three distinct clinical sybtypes: semantic variant PPA (svPPA), nonfluent/agrammatic variant PPA (nfvPPA), and logopenic variant PPA (lvPPA).

OBJECTIVE

The aim was to examine the utility of the German version of the Repeat and Point (R&P) Test for subtyping patients with PPA.

METHOD

During the R&P Test, the examiner reads out aloud a noun and the participants are asked to repeat the word and subsequently point to the corresponding picture. Data from 204 patients (68 svPPA, 85 nfvPPA, and 51 lvPPA) and 33 healthy controls were analyzed.

RESULTS

Controls completed both tasks with >90% accuracy. Patients with svPPA had high scores in repetition (mean=9.2±1.32) but low scores in pointing (mean=6±2.52). In contrast, patients with nfvPPA and lvPPA performed comparably in both tasks with lower scores in repetition (mean=7.4±2.7 for nfvPPA and 8.2±2.34 for lvPPA) but higher scores in pointing (mean=8.9±1.41 for nfvPPA and 8.6±1.62 for lvPPA). The R&P Test had high accuracy discriminating svPPA from nfvPPA (83% accuracy) and lvPPA (79% accuracy). However, there was low accuracy discriminating nfvPPA from lvPPA (<60%).

CONCLUSION

The R&P Test helps to differentiate svPPA from 2 nonsemantic variants (nfvPPA and lvPPA). However, additional tests are required for the differentiation of nfvPPA and lvPPA.

Clinical and neuroimaging characteristics of primary progressive aphasia

The chapter covers the clinical syndrome of a primary progressive aphasia (PPA), the demographics of this rare neurodegenerative disease, defining clinical and neuroanatomic characteristics of each PPA variant, disease progression, and behavioral features. The chapter begins with a brief introduction that includes references to seminal papers that defined this clinical syndrome and its three variants. The classic PPA subtypes discussed in the chapter are semantic variant PPA (svPPA), nonfluent/agrammatic PPA (nfaPPA), and logopenic variant PPA (lvPPA). The key language and cognitive characteristics, and language tasks that can elicit these language impairments, are detailed. Overlap in the clinical profiles of the PPA variants, which make differential diagnosis challenging, are explained. Disease progression is described, revealing that the PPA variants become more similar over time. Although PPA is language-predominant dementia, there are behavioral manifestations, particularly in svPPA. Changes in behavior in this variant are addressed as well as behavioral changes in nfaPPA and lvPPA that are less well recognized. The patterns of atrophy in the left temporal, parietal, and/or frontal cortices unique to each PPA variant are described. The underlying neuropathologies of the PPA variants are discussed, specifically tauopathies and non-tauopathies associated with svPPA and nfaPPA and Alzheimer's disease pathology in lvPPA.

Automatic Subtyping of Individuals with Primary Progressive Aphasia

BACKGROUND

The classification of patients with primary progressive aphasia (PPA) into variants is time-consuming, costly, and requires combined expertise by clinical neurologists, neuropsychologists, speech pathologists, and radiologists.

OBJECTIVE

The aim of the present study is to determine whether acoustic and linguistic variables provide accurate classification of PPA patients into one of three variants: nonfluent PPA, semantic PPA, and logopenic PPA.

METHODS

In this paper, we present a machine learning model based on deep neural networks (DNN) for the subtyping of patients with PPA into three main variants, using combined acoustic and linguistic information elicited automatically via acoustic and linguistic analysis. The performance of the DNN was compared to the classification accuracy of Random Forests, Support Vector Machines, and Decision Trees, as well as to expert clinicians' classifications.

RESULTS

The DNN model outperformed the other machine learning models as well as expert clinicians' classifications with 80% classification accuracy. Importantly, 90% of patients with nfvPPA and 95% of patients with lvPPA was identified correctly, providing reliable subtyping of these patients into their corresponding PPA variants.

CONCLUSION

We show that the combined speech and language markers from connected speech productions can inform variant subtyping in patients with PPA. The end-to-end automated machine learning approach we present can enable clinicians and researchers to provide an easy, quick, and inexpensive classification of patients with PPA.

Description of connected speech across different elicitation tasks in the logopenic variant of primary progressive aphasia

BACKGROUND

Despite its importance, in-depth analysis of connected speech is often neglected in the diagnosis of primary progressive aphasia (PPA) - especially for the logopenic variant (lvPPA) for which unreliable differential diagnosis has been documented. Only a few studies have been conducted on this topic in lvPPA.

AIMS

The aim of this study was to describe and compare lexico-semantic and morphosyntactic features of connected speech in participants with lvPPA, in comparison with healthy controls, using three different elicitation tasks (i.e., picture description, story narration and semi-structured interviews). In addition to a number of discourse features, we were particularly interested in the presence or absence of syntactic deficits in this PPA variant in line with recent findings.

METHODS & PROCEDURES

A prospective group study was conducted to compare lvPPA participants (n = 13) to age- and education-matched healthy controls (n = 13). For each individual, connected speech was obtained using three tasks: (1) The Cookie Theft picture description; (2) Cinderella Story; (3) Topic-directed interview. Production on each task was recorded, transcribed and analysed according to the Quantitative Production Analysis (QPA) protocol, a tool developed by Berndt et al. (2000) for the analysis of sentence production in aphasia. Differences between lvPPA and healthy controls and among elicitation tasks were analysed using repeated measures multilevel mixed-effects regression, separately for each outcome.

OUTCOMES & RESULTS

Four measures were significantly different between lvPPA participants and healthy controls across all elicitation tasks. Specifically, lvPPA participants produced a reduced proportion of open-class words, a higher proportion of verbs, a higher proportion of pronouns and fewer well-formed sentences. For these measures, the difference between lvPPA and healthy controls was consistent among elicitation tasks, except for the proportion of well-formed sentences, where the difference between the two groups was significantly greater in the story narration task than in the other tasks.

CONCLUSIONS & IMPLICATIONS

Across elicitation tasks that used the same analysis protocol (i.e., QPA), a similar pattern of deficits in connected speech emerged in lvPPA patients. Importantly, the findings replicate previous studies, which used different elicitation tasks and analysis protocols. Especially in relation to the documented syntactic deficits, these findings provide implications for differential diagnosis in PPA.

WHAT THIS PAPER ADDS

What is already known on the subject Connected speech analysis can provide an important contribution to the language assessment for the logopenic variant of primary progressive aphasia (lvPPA). However, only a few studies have been conducted with this population. What this paper adds to existing knowledge This study highlights differences between patients with lvPPA and healthy controls regarding the proportion of open-class words, nouns, verbs and well-formed sentences. What are the potential or actual clinical implications of this work? Importantly, our results highlight syntactic deficits in the same group of individuals with lvPPA, using the same analysis protocol and across various elicitation tasks, which has implications for differential diagnosis.

Contribution of the Cognitive Approach to Language Assessment to the Differential Diagnosis of Primary Progressive Aphasia

Diagnosis of primary progressive aphasia (PPA) is essentially based on the identification of progressive impairment of language abilities while other cognitive functions are preserved. The three variants of PPA are characterized by core and supportive clinical features related to the presence or absence of language impairment in different linguistic domains. In this article, we review the cognitive neuropsychological approach to the assessment of PPA and its contribution to the differential diagnosis of the three variants. The main advantage of this assessment approach is that it goes beyond the mere description and classification of clinical syndromes and identifies impaired and preserved cognitive and linguistic components and processes. The article is structured according to the main language domains: spoken production, language comprehension, and written language. Each section includes a brief description of the cognitive processes involved in the assessment tasks, followed by a discussion of typical characteristics for each PPA variant and common pitfalls in the interpretation of the results. In addition, the clinical benefit of the cognitive neuropsychological approach for the behavioral management of PPA is briefly sketched out in the conclusion.

Longitudinal decline in spoken word recognition and object knowledge in primary progressive aphasia

The premise of this study is that spoken word recognition and object knowledge are impaired in semantic variant primary progressive aphasia (PPA) (svPPA) and are spared in logopenic variant (lvPPA) and nonfluent agrammatic primary progressive aphasia (nfaPPA) at disease onset. Over time, however, there may be heterogeneity in these abilities in lvPPA and nfaPPA. We hypothesized that individuals with svPPA would demonstrate poorer performance on baseline spoken word recognition and object knowledge than those with lvPPA and nfaPPA) as documented in the literature, but that rates of decline over time on spoken word recognition and object knowledge would be similar in all 3 PPA variants because these become less distinguishable with disease progression.The aim of this study was to investigate longitudinal patterns of decline in spoken word recognition and object knowledge across PPA variants.Ninety-five individuals with PPA completed the Semantic Word Picture Matching and Semantic Associates tests at baseline to establish expected performance in these areas. Thirty-five individuals completed follow-up testing.The distributions of trichotomized mean rates of decline in object knowledge were similar for lvPPA and svPPA (P = .05). There were weak negative correlations between symptom duration and baseline scores on Semantic Word Picture Matching (r[37] = -0.399, P = .01), and baseline scores on Semantic Associates (r[37] = -0.394, P = .01) in lvPPA.Degradation of spoken word recognition and object knowledge occurs over time in lvPPA. Further investigation of the receptive language deficits in PPA is warranted to characterize language changes that lessen the distinctions between PPA variants with disease progression.

Graded, multidimensional intra- and intergroup variations in primary progressive aphasia and post-stroke aphasia

Language impairments caused by stroke (post-stroke aphasia, PSA) and neurodegeneration (primary progressive aphasia, PPA) have overlapping symptomatology, nomenclature and are classically divided into categorical subtypes. Surprisingly, PPA and PSA have rarely been directly compared in detail. Rather, previous studies have compared certain subtypes (e.g. semantic variants) or have focused on a specific cognitive/linguistic task (e.g. reading). This study assessed a large range of linguistic and cognitive tasks across the full spectra of PSA and PPA. We applied varimax-rotated principal component analysis to explore the underlying structure of the variance in the assessment scores. Similar phonological, semantic and fluency-related components were found for PSA and PPA. A combined principal component analysis across the two aetiologies revealed graded intra- and intergroup variations on all four extracted components. Classification analysis was used to test, formally, whether there were any categorical boundaries for any subtypes of PPA or PSA. Semantic dementia formed a true diagnostic category (i.e. within group homogeneity and distinct between-group differences), whereas there was considerable overlap and graded variations within and between other subtypes of PPA and PSA. These results suggest that (i) a multidimensional rather than categorical classification system may be a better conceptualization of aphasia from both causes; and (ii) despite the very different types of pathology, these broad classes of aphasia have considerable features in common.

Establishing two principal dimensions of cognitive variation in logopenic progressive aphasia

Logopenic progressive aphasia is a neurodegenerative syndrome characterized by sentence repetition and naming difficulties arising from left-lateralized temporoparietal atrophy. Clinical descriptions of logopenic progressive aphasia largely concentrate on profiling language deficits, however, accumulating evidence points to the presence of cognitive deficits even on tasks with minimal language demands. Although non-linguistic cognitive deficits in logopenic progressive aphasia are thought to scale with disease severity, patients at discrete stages of language dysfunction display overlapping cognitive profiles, suggesting individual-level variation in cognitive performance, independent of primary language dysfunction. To address this issue, we used principal component analysis to decompose the individual-level variation in cognitive performance in 43 well-characterized logopenic progressive aphasia patients who underwent multi-domain neuropsychological assessments and structural neuroimaging. The principal component analysis solution revealed the presence of two, statistically independent factors, providing stable and clinically intuitive explanations for the majority of variance in cognitive performance in the syndrome. Factor 1 reflected 'speech production and verbal memory' deficits which typify logopenic progressive aphasia. Systematic variations were also confirmed on a second, orthogonal factor mainly comprising visuospatial and executive processes. Adopting a case-comparison approach, we further demonstrate that pairs of patients with comparable Factor 1 scores, regardless of their severity, diverge considerably on visuo-executive test performance, underscoring the inter-individual variability in cognitive profiles in comparably 'logopenic' patients. Whole-brain voxel-based morphometry analyses revealed that speech production and verbal memory factor scores correlated with left middle frontal gyrus, while visuospatial and executive factor scores were associated with grey matter intensity of right-lateralized temporoparietal, middle frontal regions and their underlying white matter connectivity. Importantly, logopenic progressive aphasia patients with poorer visuospatial and executive factor scores demonstrated greater right-lateralized temporoparietal and frontal atrophy. Our findings demonstrate the inherent variation in cognitive performance at an individual- and group-level in logopenic progressive aphasia, suggesting the presence of a genuine co-occurring cognitive impairment that is statistically independent of language function and disease severity.

Clinical Assessment of Characteristics of Apraxia of Speech in Primary Progressive Aphasia

Purpose We sought to examine interrater reliability in clinical assessment of apraxia of speech (AOS) in individuals with primary progressive aphasia and to identify speech characteristics predictive of AOS diagnosis. Method Fifty-two individuals with primary progressive aphasia were recorded performing a variety of speech tasks. These recordings were viewed by 2 experienced speech-language pathologists, who independently rated them on the presence and severity of AOS as well as 14 associated speech characteristics. We calculated interrater reliability (percent agreement and Cohen's kappa) for these ratings. For each rater, we used stepwise regression to identify speech characteristics significantly predictive of AOS diagnosis. We used the overlap between raters to create a more parsimonious model, which we evaluated with multiple linear regression. Results Results yielded high agreement on the presence (90%) and severity of AOS (weighted Cohen's κ = .834) but lower agreement for specific speech characteristics (weighted Cohen's κ ranging from .036 to .582). Stepwise regression identified 2 speech characteristics predictive of AOS diagnosis for both raters (articulatory groping and increased errors with increased length/complexity). These alone accounted for ≥ 50% of the variance of AOS severity in the constrained model. Conclusions Our study adds to a growing body of research that highlights the difficulty in objective clinical characterization of AOS and perceptual characterization of speech features. It further supports the need for consensus diagnostic criteria with standardized testing tools and for the identification and validation of objective markers of AOS. Additionally, these findings underscore the need for a training protocol if diagnostic tools are to be effective when shared beyond the research teams that develop and test them and disseminated to practicing speech-language pathologists, in order to ensure consistent application.

Classification of primary progressive aphasia: challenges and complexities

Primary progressive aphasia (PPA) is classified into three variants, logopenic variant PPA (lvPPA), nonfluent agrammatic PPA (nfaPPA), and semantic variant PPA (svPPA), based on clinical (syndromic) characteristics with support from neuroimaging and/or underlying neuropathology. Classification of PPA variants provides information valuable to disease management. International consensus criteria are widely employed to identify PPA subtypes; however, classification is complex, and some individuals do not fit neatly into the subtyping scheme. In this review, diagnostic challenges and their implications are discussed, possible explanations for these challenges are explored, and approaches to address PPA classification are considered.

Right Temporal Variant Frontotemporal Dementia Misdiagnosed as Schizophrenia

Background:

Schizophrenia-like presentations of frontotemporal dementia (FTD) are well documented in medical literature, especially during the initial stage of the illness. We present an interesting case of middle-aged white male, who was initially misdiagnosed with major depression at age 50 and then with schizophrenia due to florid psychosis typically seen in Mild Behavioral Impairment (MBI). It was not until over 4-years that a detailed clinical history along with specific investigations confirmed the diagnosis of right temporal variant frontotemporal dementia (rtv-FTD).

Methods:

The unusual presentation of this case warranted 2-deoxy-2-[fluorine-18]fluoro- D-glucose integrated with computed tomography (FDG-PET) and comprehensive neuropsychological testing to estblish a final diagnosis.

Results:

The results from neuroimaging and neuropsychological testing supported a diagnosis of rtv-FTD without any evidence for Alzheimer’s dementia.

Conclusion:

This case illustrates that late-onset atypical psychiatric symptoms should be evaluated for neurocognitive disorders, and the evaluation should include neuroimaging. It is preferable that the neuroimaging findings be personally interpreted by the provider, who therefore needs to be skilled in interpreting neuroimaging studies.

Clinical and neuroimaging characteristics of clinically unclassifiable primary progressive aphasia

Many patients who meet core/root criteria for Primary Progressive Aphasia (PPA) are not classifiable as a recognized variant and are often excluded from neuroimaging studies. Here, we detail neurological, neuropsychological, speech and language assessments, and anatomic and molecular neuroimaging (MRI, PiB-PET, and FDG-PET) for fifteen (8 female) clinically unclassifiable PPA patients. Median age of onset was 64 years old with median 3 years disease duration at exam. Three patients were amyloid positive on PiB-PET. 14/15 patients had abnormal FDG-PETs with left predominant hypometabolism, affecting frontal, temporal, parietal, and even occipital lobes. Patients had mild to severe clinical presentations. Visualization of the FDG-PETs principal component analysis revealed patterns of hypometabolism similar to those seen in the PPA variants and suggests the brain regions affected in unclassifiable PPA patients are no different from those who are more easily classifiable. These findings may inform future modifications to the diagnostic criteria to improve diagnostic classification.

Biomarker-guided clustering of Alzheimer's disease clinical syndromes

Alzheimer's disease (AD) neuropathology is extremely heterogeneous, and the evolution from preclinical to mild cognitive impairment until dementia is driven by interacting genetic/biological mechanisms not fully captured by current clinical/research criteria. We characterized the heterogeneous "construct" of AD through a cerebrospinal fluid biomarker-guided stratification approach. We analyzed 5 validated pathophysiological cerebrospinal fluid biomarkers (Aβ_1-42, t-tau, p-tau₁₈₁, NFL, YKL-40) in 113 participants (healthy controls [N = 20], subjective memory complainers [N = 36], mild cognitive impairment [N = 20], and AD dementia [N = 37], age: 66.7 ± 10.4, 70.4 ± 7.7, 71.7 ± 8.4, 76.2 ± 3.5 years [mean ± SD], respectively) using Density-Based Spatial Clustering of Applications with Noise, which does not require a priori determination of the number of clusters. We found 5 distinct clusters (sizes: N = 38, 16, 24, 14, and 21) whose composition was independent of phenotypical groups. Two clusters showed biomarker profiles linked to neurodegenerative processes not associated with classical AD-related pathophysiology. One cluster was characterized by the neuroinflammation biomarker YKL-40. Combining nonlinear data aggregation with informative biomarkers can generate novel patient strata which are representative of cellular/molecular pathophysiology and may aid in predicting disease evolution and mechanistic drug response.

Profiling sentence repetition deficits in primary progressive aphasia and Alzheimer's disease: Error patterns and association with digit span

The use of sentence repetition tasks to distinguish dementia syndromes, particularly variants of primary progressive aphasia (PPA), is receiving growing attention. Impaired sentence repetition is a core feature of logopenic variant PPA, although the underlying cognitive mechanisms of this impairment and its significance as a diagnostic criterion remain poorly understood. Sentence repetition abilities of 12 people with dementia, using an adapted error classification schema, were analyzed, along with digit span abilities, a measure frequently used to assess working memory capacity, to explore error patterns and correlations. Participants with semantic variant PPA showed the greatest performance on sentence repetition and digit span in comparison to the logopenic variant PPA and Alzheimer's disease participants. Sentence repetition errors were characterized by middle omissions for semantic variant PPA, ending omissions and phonological errors for logopenic variant PPA, and ending omissions and unrelated word substitutions for Alzheimer's disease. Significant correlations were found between sentence repetition and digit span abilities. Findings are discussed in relation to working memory capacity theories, specifically, the dual-component model.

Machine learning in the clinical and language characterisation of primary progressive aphasia variants

INTRODUCTION

Primary progressive aphasia (PPA) is a clinical syndrome of neurodegenerative origin with 3 main variants: non-fluent, semantic, and logopenic. However, there is some controversy about the existence of additional subtypes. Our aim was to study the language and cognitive features associated with a new proposed classification for PPA.

MATERIAL AND METHODS

Sixty-eight patients with PPA in early stages of the disease and 20 healthy controls were assessed with a comprehensive language and cognitive protocol. They were also evaluated with ¹⁸F-FDG positron emision tomography (PET). Patients were classified according to FDG PET regional metabolism, using our previously developed algorithm based on a hierarchical agglomerative cluster analysis with Ward's linkage method. Five variants were found, with both the non-fluent and logopenic variants being split into 2 subtypes. Machine learning techniques were used to predict each variant according to language assessment results.

RESULTS

Non-fluent type 1 was associated with poorer performance in repetition of sentences and reading of irregular words than non-fluent type 2. Conversely, the second group showed a higher degree of apraxia of speech. Patients with logopenic variant type 1 performed more poorly on action naming than patients with logopenic type 2. Language assessments were predictive of PET-based subtypes in 86%-89% of cases using clustering analysis and principal components analysis.

CONCLUSIONS

Our study supports the existence of 5 variants of PPA. These variants show some differences in language and FDG PET imaging characteristics. Machine learning algorithms using language test data were able to predict each of the 5 PPA variants with a relatively high degree of accuracy, and enable the possibility of automated, machine-aided diagnosis of PPA variants.

Clustering Analysis of FDG-PET Imaging in Primary Progressive Aphasia

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.

Single-word comprehension deficits in the nonfluent variant of primary progressive aphasia

BACKGROUND

A subset of patients with the nonfluent variant of primary progressive aphasia (PPA) exhibit concomitant single-word comprehension problems, constituting a 'mixed variant' phenotype. This phenotype is rare and currently not fully characterized. The aim of this study was twofold: to assess the prevalence and nature of single-word comprehension problems in the nonfluent variant and to study multimodal imaging characteristics of atrophy, tau, and amyloid burden associated with this mixed phenotype.

METHODS

A consecutive memory-clinic recruited series of 20 PPA patients (12 nonfluent, five semantic, and three logopenic variants) were studied on neurolinguistic and neuropsychological domains relative to 64 cognitively intact healthy older control subjects. The neuroimaging battery included high-resolution volumetric magnetic resonance imaging processed with voxel-based morphometry, and positron emission tomography with the tau-tracer [18F]-THK5351 and amyloid-tracer [11C]-Pittsburgh Compound B.

RESULTS

Seven out of 12 subjects who had been classified a priori with nonfluent variant PPA showed deficits on conventional single-word comprehension tasks along with speech apraxia and agrammatism, corresponding to a mixed variant phenotype. These mixed variant cases included three females and four males, with a mean age at onset of 65 years (range 44-77 years). Object knowledge and object recognition were additionally affected, although less severely compared with the semantic variant. The mixed variant was characterized by a distributed atrophy pattern in frontal and temporoparietal regions. A more focal pattern of elevated [18F]-THK5351 binding was present in the supplementary motor area, the left premotor cortex, midbrain, and basal ganglia. This pattern was closely similar to that seen in pure nonfluent variant PPA. At the individual patient level, elevated [18F]-THK5351 binding in the supplementary motor area and premotor cortex was present in six out of seven mixed variant cases and in five and four of these cases, respectively, in the thalamus and midbrain. Amyloid biomarker positivity was present in two out of seven mixed variant cases, compared with none of the five pure nonfluent cases.

CONCLUSIONS

A substantial proportion of PPA patients with speech apraxia and agrammatism also have single-word comprehension deficits. At the neurobiological level, the mixed variant shows a high degree of similarity with the pure nonfluent variant of PPA.

TRIAL REGISTRATION

EudraCT, 2014-002976-10 . Registered on 13-01-2015.