Semantic variant of primary progressive aphasia

The core features of the svPPA in the early stage are a prominent word-finding impairment in spontaneous speech, with a tendency to produce nouns of high frequency, and severe anomia in confrontation naming tasks, with apparently preserved non-verbal semantics. The progressive semantic impairment subsequently affects single-word comprehension, and it selectively distorts intra-category differentiations leading to overgeneralized concepts [47]. As the disease progresses the impairment interferes also with inter-category discrimination, as defective single-word processing becomes clinically more severe. Impaired reading of words with irregular spelling is also typical of this subtype, and may be a consequence of semantic impairment [16]. Syntactic processing is typically spared and syntactic errors, mostly paragrammatic (substitutions of closed-class words and bound morphemes), are relatively scarce [48]. The presence of many embedded sentences in the discourse reflects attempts to cope with anomia [14]. Non-verbal skills are usually moderately affected in the early stages of the typical left-sided presentation, suggesting a prominent impairment of lexical semantics [47]. When object, people, and environmental sound identification deficits are prominent, they suggest the presence of a less common clinical presentation, associated with prominent involvement of the right anterior temporal lobe [49]. The anterior temporal lobe involvement is typically bilateral, usually more extensive in the left hemisphere [19, 36]. Preserved motor speech and syntactic function reflect the integrity of the dorsal language regions and of the corresponding white matter connections [19, 50]. The atrophy predominantly involves inferior and middle temporal gyri, anterior fusiform gyrus, amygdala and hippocampus, and entorhinal/perirhinal cortices [51–53]. Left anterior and inferior temporal atrophy accounts for the typical lexical retrieval deficits [47]. The degree of semantic memory impairment correlates with the atrophy of ventral and lateral portions of temporal lobe, while amygdala atrophy is related to emotional processing impairment [54]. Imaging findings are quite characteristic in the intermediate and advanced stages (Figure 5.3), showing a severe “knife-edge”-type atrophy of the anterior temporal lobes with a volume loss of 50% or greater, making svPPA the easiest syndrome to diagnose on the basis of simple visual inspection. At the very beginning of disease, coronal MRI acquisitions might help to better detect the size changes of left anterior temporal lobe. Cerebral glucose metabolism is selectively reduced in the temporal lobes with a highly typical pattern of hypometabolism (Figure 5.4) [55]. As the degeneration progresses, both anterior temporal lobes as well as the ventromedial and posterior orbital frontal cortices, the insula bilaterally, and the left anterior cingulate cortex might be involved, overlapping with the imaging features of bvFTD patients [56], and accounting for the prominent behavioral symptoms that can be observed in most of these patients. In contrast to nfvPPA, fractional anisotropy is mostly reduced in the “ventral” pathway, i.e., in the left inferior longitudinal fasciculus and uncinate fasciculus [41, 42, 50], rather than in the SLF/arcuate, which is affected only in its temporal component. Almost 20% of FTLD subjects showed a right temporal dominant atrophy pattern in one study [57]. While the majority of these cases (12 out of 20) presented as bvFTD, 8 subjects had a diagnosis of svPPA.

Figure 5.3

Axial T1-weighted MRI images showing a severe left temporopolar atrophy in a 58-year-old man affected by svPPA. Images are in radiologic convention (i.e., left = right; right = left).

Courtesy of Professor A. Falini, Neuroradiology – CERMAC, Vita-Salute San Raffaele University and San Raffaele Scientific Institute, Milan, Italy.

Figure 5.4

[¹⁸F]FDG-PET findings in a 60-year-old man affected by svPPA. (A) FDG-uptake distribution (displayed in radiologic convention – i.e., left = right; right = left) and (B) the relative SPM-t map of hypometabolism (displayed in neurologic convention – i.e., left = left; right = right) of the single-patient [¹⁸F]FDG-PET scan compared to 112 normal scans (p = 0.05, FWE; minimum cluster size = 100 voxels) (Della Rosa PA, Cerami C, Gallivanone F, et al. (2014). A standardized [18F]FDG-PET template for spatial normalization in statistical parametric mapping of dementia. Neuroinformatics, 12(4), 575–93.), showing a pattern of involvement of temporal poles, mostly on the left side, accompanied by left orbitofrontal cortex, parahippocampal cortex, and anterior temporal regions hypometabolism.

Courtesy of Professor D. Perani, Nuclear Medicine Department, Vita-Salute San Raffaele University and San Raffaele Scientific Institute, Milan, Italy.

Considerable variability in the speed of progression has been reported [58], but prognostic variables are still incompletely known. At the opposite of nfvPPA cases, svPPA is characterized by a relatively homogeneous clinical picture and by consistent imaging/neuropathologic correlates. In particular, approximately 75% of cases are associated with TDP-43 pathology type C [46].

Logopenic/phonologic variant of primary progressive aphasia

This is the most recently described PPA syndrome (Table 5.1). Gorno-Tempini and colleagues reported 10 patients with slow, hesitant speech, without articulation deficits, but with many false starts, long word-finding pauses, and filled pauses and constant rewording, giving rise to an overall impression of non-fluency [11]. Impaired sentence comprehension and naming, and spared single-word comprehension and non-verbal semantics completed the clinical picture. A further analysis of the neuropsychological and imaging features of this clinical phenotype, based on an extensive cognitive assessment [59], included phonologic errors and defective repetition as further hallmarks of the syndrome. Investigation of phonologic loop functions showed that patients were severely impaired in digit, letter, and word span tasks. The performance did not improve with pointing, was influenced by word length, and did not show the normal phonologic similarity effect. These features pointed to defective phonologic memory as a crucial determinant of the clinical picture, including the impaired understanding of grammatically complex sentences [32].

It is probably the case that PPA patients have been variably considered “fluent” or “non-fluent” by different researchers, depending on which aspects of fluency (e.g., speech rate, phrase length, articulatory agility, syntactic structure, and prosody) were considered. The lvPPA language production is characterized by an “intermediate” pattern of fluency, distinct from the other two variants [59]. The patients usually produce few or no errors involving speech sounds (e.g., misarticulations), whereas a subset present phonemic errors, reflecting phonologic retrieval and assembly problems, rather than motor speech impairments. Continuous rephrasing in speech production leads to phonologic paraphasias or neologisms, as seen in vascular conduction aphasia [60]. Although these patients are not agrammatic, an in-depth analysis of the connected speech has shown that they can produce paragrammatic errors [14].

Structural analyses showed a distinctive pattern of gray- and white-matter damage, involving the left posterior superior and middle temporal gyri, and inferior parietal lobule [11, 59], consistent with the hypothesis of a phonologic short-term memory impairment as the core cognitive deficit (Figure 5.5). The same pattern of damage has been shown by cortical thickness [21, 38] as well as brain metabolism studies [61, 62]. lvPPA patients have reduced fractional anisotropy largely restricted to the temporo-parietal branch of the SLF/arcuate [41, 42, 50].

Figure 5.5

Axial T2-weighted fluid-attenuated inversion recovery (FLAIR) scans of a 65-year-old woman fulfilling criteria for lvPPA showing a bilateral widening of the sylvian fissure and the parietal subarachnoid spaces (left > right). Images are displayed in radiologic convention (i.e., left = right; right = left).

Courtesy of Professor A. Falini, Neuroradiology – CERMAC, Vita-Salute San Raffaele University and San Raffaele Scientific Institute, Milan, Italy.

Recent findings show that AD is probably the most common underlying pathology of lvPPA [62, 63]. A study with [¹¹C]PiB-PET (Pittsburgh compound B-PET) and [¹⁸F]FDG-PET indicated that logopenic subjects present with a higher cortical uptake of [¹¹C]PiB radioligand than patients classified with the other variants. In addition, the distribution pattern of [¹¹C]PiB-positive PPA was diffuse and comparable to those of matched AD subjects. In contrast with this diffuse pattern of [¹¹C]PiB uptake in the PiB-positive PPA cases, the location of hypometabolism on [¹⁸F]FDG-PET scans reflected the features of the PPA subtypes. In particular, lvPPA subjects presented the greatest metabolic changes in the left parietal and posterolateral temporal lobes (Figure 5.6). A recent study of a large sample of patients fulfilling the lvPPA criteria found that about one third are not positive for AD biomarkers [64]. The two thirds with probable AD pathology had a more severe clinical picture, with defective performance extending beyond language tasks and extensive involvement of the left temporo-parietal areas.

Figure 5.6

[¹⁸F]FDG-PET findings in a 70-year-old man affected by lvPPA. (A) FDG-uptake distribution (displayed in radiologic convention – i.e., left = right; right = left) and (B) the relative SPM-t map of hypometabolism (displayed in neurologic convention – i.e., left = left; right = right) of the single-patient [¹⁸F]FDG-PET scan compared to 112 normal scans (p = 0.05, FWE; minimum cluster size = 100 voxels) (Della Rosa PA, Cerami C, Gallivanone F, et al. (2014). A standardized [18F]FDG-PET template for spatial normalization in statistical parametric mapping of dementia. Neuroinformatics, 12(4), 575–93.), showing an involvement of superior parietal, supramarginal, angular, and superior temporal gyri, mostly on the left side.

Courtesy of Professor D. Perani, Nuclear Medicine Department, Vita-Salute San Raffaele University and San Raffaele Scientific Institute, Milan, Italy.

The differential diagnosis with the nfvPPA may be difficult, in particular in the case of patients with prominent phonologic errors, which can easily be mistaken for motor speech errors, especially in the context of decreased speech rate. This may be one of the reasons responsible for wide variations in the prevalence of this syndrome in different series [65]. The clinical picture may also have a substantial overlap with posterior cortical atrophy [66], in agreement with a common AD neuropathologic substrate. While both syndromes are considered as atypical AD presentations, the possibility of non-AD pathology should not be neglected.

Atypical and mixed cases of primary progressive aphasia

The current distinction in the three main variants, based on the most frequently observed symptom complexes, represents a gross oversimplification of the heterogeneity of the possible PPA phenotypes. Not surprisingly, variations in the location of neuropathologic damage within the large-scale language network are associated with unclassifiable, mixed clinical presentations. Among the variations, primary progressive AOS has already been mentioned [33]. Additional atypical presentations are a transcortical-type progressive aphasia, characterized by reduced speech output with preserved repetition, which is sometimes observed in the context of apathetic forms of bvFTD [44], and progressive fluent jargon aphasia, a possible evolution of the lvPPA [67]. An adynamic form of PPA has also recently been described [68].

Genetic studies

The syndromes belonging to the FTLD spectrum are often familial, and about half of cases have a positive family history [69]. Many PPA patients have a positive family history for degenerative diseases. Mutations of microtubule-associated protein tau (MAPT), progranulin (GRN), and chromosome 9 open reading frame 72 (C9orf72) genes are the most frequently found genetic abnormalities, accounting for a high proportion of familial cases with autosomal dominant transmission [69].

While the most common clinical presentation associated with this MAPT mutation is bvFTD, the clinical phenotype is widely heterogeneous, even within the same family pedigree. PPA can infrequently be one of the possible clinical presentations, and there are a few reports of nfvPPA cases carrying the MAPT mutation [70, 71]. The same phenotypical heterogeneity is also observed in the case of GRN and C9orf72 mutations.

GRN mutations are associated with many different phenotype presentations, including PPA [72–74]. Most of these GRN-mutated patients present clinical features of nfvPPA, but the phenotype can be atypical. Only a few single svPPA cases have been reported in association with GRN mutations [75, 76]. Data from the literature show a very low rate of inheritance in svPPA and lvPPA, even in studies with autopsy confirmation [58, 72, 77, 78]. A single case study described a GRN-mutated patient with a clinical profile closely resembling the logopenic variant, with some clinical features overlapping with the other variants [79].

More recently, FTLD cases have been reported in association with C9orf72 mutations [80–86], with a percentage nearly comparable to GRN mutations [83], and a pathology characterized by the deposition of the TDP-43 protein (i.e., FTLD-TDP type A and B [46]). Some nfvPPA C9orf72-mutated cases have been described (i.e., 6 out of 75 FTD cases of the Finnish cohort [81]). On the contrary, the association with other PPA varieties is extremely rare. In particular, there are a few reports of svPPA cases carrying C9orf72 mutations [76, 87].

Treatment and management

There is no effective pharmacologic treatment specifically designed for PPAs belonging to the FTLD spectrum. In the case of possible AD pathology, i.e., with the lvPPA presentation, treatment with acetylcholinesterase inhibitors and memantine is usually considered as an option. There is limited published evidence about the possible role of speech and language therapy. Some anecdotal reports indicate a positive effect of treatment programs focusing on communication strategies and alternative communication methods [88, 89]. The role of patient and caregiver information cannot be underestimated. The availability of an international registry (www.ppaconnection.org) is an important resource for patients and clinicians [5].