Primary Progressive Aphasia Lacking Core Features of Nonfluent and Semantic Variants: Clinical, Neuroimaging, and Neuropathologic Features.

Abstract

Evidence has accumulated that the 2011 consensus criteria for primary progressive aphasia (PPA) do not fully capture features of logopenic variant PPA (lvPPA/LPA). We aimed to examine clinical, neuroimaging, and neuropathologic features of PPA lacking features of nonfluent/semantic variants and to provide practical additions to the 2011 consensus criteria. This was a retrospective examination of data from 2 observational cohort studies where patients with PPA were prospectively recruited at Mayo Clinic. Based on performance on 2 cardinal features (repetition and comprehension), patients were classified as: pure-LPA (poor repetition, acceptable comprehension), Wernicke-like (poor in both), anomic-like (acceptable in both), and transcortical sensory aphasia-like (TCSA-like) (acceptable repetition, poor comprehension). The entire cohort consisted of 102 patients with PPA lacking features of nonfluent/semantic variants (median age at onset 63.5 years, 56% female). Thirty-one patients were followed up at 1 year. Twenty-three patients were included in a neuropathologic cohort. The proportion of repetition-preserved PPA (anomic-like and TCSA-like) was more than double that of repetition-impaired PPA (pure-LPA and Wernicke-like) (73% vs 27%). Regarding clinical course, the anomic-like subgroup was a prodromal state of the pure-LPA or TCSA-like subgroup, whereas the pure-LPA and TCSA-like subgroups were a prodromal state of the Wernicke-like subgroup. There was left temporoparietal atrophy on MRI and/or hypometabolism on ¹⁸F-fluorodeoxyglucose-PET in all groups. Furthermore, repetition-impaired PPA showed severe hypometabolism in the left superior temporal lobe associated with repetition ability. Regarding pathologic diagnoses, 70% had Alzheimer disease (AD). The pure-LPA, Wernicke-like, and TCSA-like subgroups all showed AD pathology. Only 53% of the anomic-like subgroup had AD. The remaining 47% showed Pick disease (7%), frontotemporal lobar degeneration with TDP-43-immunoreactive pathology (20%), and Lewy body disease (20%). This observed clinical heterogeneity reflects different time points/severities of the same disease process and hence can be reconceptualized as an AD-related aphasia spectrum, incorporating lvPPA and the 4 subgroups (pure-LPA, Wernicke-like, anomic-like, and TCSA-like). Specifying moderate/severe repetition deficits as a core feature of lvPPA in the 2011 consensus criteria can enhance its pathologic correlations. Recognizing progressive anomic aphasia (anomic-like) as an additional PPA variant could lessen pathologic heterogeneity of lvPPA.