The term frontotemporal lobar degeneration (FTLD) refers to a group of progressive neurodegenerative disorders that are clinically, pathologically and genetically heterogeneous, but share selective frontal and/or temporal cortical atrophy. FTLD is a common cause of early-onset dementia, with an average age of onset between 45 and 65 years. Among its clinical subtypes, frontotemporal dementia (FTD) is mainly characterized by behavioral symptoms, while aphasic variants comprehend the non-fluent/agrammatic variant primary progressive aphasia (nfvPPA) and the semantic variant (svPPA). Besides behavioral and language dysfunctions, FTLD frequently comprises motor manifestations, including atypical parkinsonian syndromes and motor neuron disorders (Bang et al., 2015; Rabinovici and Miller, 2010).

Although many FTLD cases are sporadic, up to 50% show a positive family history. The most frequent genetic causes of familial FTD are mutations in chromosome 9 open reading frame (C9orf72), microtubule associated protein tau (MAPT) and progranulin gene (GRN), which together account for approximately 60% of inherited cases (Olszewska et al., 2016).

The most prevalent clinical presentation is the behavioral variant of FTD (bvFTD), accounting for approximately half of FTLD cases (Johnson et al., 2005). Typical symptoms include disinhibition, loss of empathy, apathy, changes in eating habits, stereotyped behaviors, and executive dysfunctions affecting planning, cognitive flexibility and judgment (Boeve et al., 2022; Rascovsky et al., 2011). Neuroimaging studies show involvement of orbitofrontal, fronto-insular, dorsolateral prefrontal, anterior cingulate and anterior temporal cortices with regional patterns of atrophy correlating with specific behavioral symptoms (Rabinovici and Miller, 2010).

nfvPPA is characterized by apraxia of speech, agrammatism, progressively reduced verbal output, and impaired grammatical comprehension with relative preservation of single word meaning (Boeve et al., 2022; Gorno-Tempini et al., 2011; Thompson and Mack, 2014). Neuroimaging studies typically show left perisylvian atrophy, involving the frontal operculum, premotor and supplementary motor areas and anterior insula (Rabinovici and Miller, 2010).

Instead, svPPA is characterized by impaired single-word comprehension and naming deficits, reflecting a loss of semantic knowledge, while sentence-level comprehension is relatively preserved in early stages. Speech remains fluent but semantically empty, often appearing structurally intact despite reduced informational content (Gorno-Tempini et al., 2011). As the disease progresses, subjects develop severe anomia and comprehension deficit even for highly familiar concepts (Meteyard and Patterson, 2009). Neuroimaging reveals predominant atrophy of the anterior temporal lobe (Rabinovici and Miller, 2010).

FTLD is neuropathologically heterogeneous, including tauopathies (FTLD-TAU), TAR DNA-binding protein 43 (TDP-43) pathologies (FTLD-TDP), and, less commonly, Alzheimer’s disease (AD)-related pathology (Bang et al., 2015; Gorno-Tempini et al., 2011; Rabinovici and Miller, 2010; Rascovsky et al., 2011). In this complex clinical and anatomical setting, neuropathological characterization is required to define the nature and distribution of pathological protein aggregates and to support clinicopathological correlations.

This case series presents a selection of neuropathologically examined cases from our cohort, showing clinical phenotypes within the frontotemporal spectrum at disease onset. This series aims to illustrate the heterogeneity within the FTLD spectrum and the importance of neuropathological characterization to clarify the underlying pathophysiology of these disorders.

The results of this study are reported according to the Case Report Checklist CARE guidelines for case series (Gagnier et al., 2013). This retrospective case-series was compiled using longitudinal clinical and neuropathological data collected at the Golgi Cenci Foundation (Abbiategrasso, Italy) within the brain donation program in the frame of the InveCe.Ab cohort (Invecchiamento Cerebrale Abbiategrasso study), which currently comprises 60 cases (Guaita et al., 2013). The clinical neurocognitive diagnosis was formulated based on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (American Psychiatric Association [APA], 2013).

Eligible cases met the following criteria: (I) a clinical phenotype within the frontotemporal spectrum disorder, including behavioral or language presentations (bvFTD, svFTD, nfvPPA), diagnosed according to established clinical criteria (Gorno-Tempini et al., 2011; Rascovsky et al., 2011); (II) availability of a complete neuropathological characterization; (III) provision of written informed consent for brain donation. Cases were included without additional selection beyond eligibility criteria.

During life, brain donors underwent standardized neurological and neuropsychological assessments as established in the Abbiategrasso Brain Bank protocol (Poloni et al., 2020). After death, brain tissue was processed following the same protocol: alternate coronal sections of the cerebral hemisphere, brainstem, and cerebellum were either snap-frozen at −80 °C or fixed in phosphate-buffered formalin and subsequently paraffin-embedded. Paraffin sections then underwent the histological and immunohistochemical (IHC) staining procedures required for diagnostic assessment. Neuropathological assessment was performed following international consensus criteria, including Alzheimer-type pathology (Braak et al., 2006; Mirra et al., 1991; Montine et al., 2013; Thal et al., 2002), LTS (Beach et al., 2009), TDP-43 proteinopathies (Josephs et al., 2014; Mackenzie et al., 2011; Nelson et al., 2019), hippocampal sclerosis (HS) (Rauramaa et al., 2013) and cerebrovascular pathology (Deramecourt et al., 2012; Skrobot et al., 2016). A detailed description of the tissue processing steps, histological and IHC staining, and diagnostic consensus criteria is reported in the Abbiategrasso Brain Bank protocol (Poloni et al., 2020). Neuropathological diagnoses were formulated based on the type and anatomical distribution of lesions and assigned in accordance with validated diagnostic criteria.

Ethical approval for all study procedures was obtained from the Ethics Committee of the University of Pavia (Committee report 3/2009), within the framework of the InveCe.Ab project (Guaita et al., 2013). Written informed consent for clinical evaluation, brain donation, neuropathological examination, and research use of data was obtained from all donors in accordance with the ethical principles governing the brain donation program of the Fondazione Golgi Cenci (Klioueva et al., 2015). All procedures were conducted in full compliance with the Declaration of Helsinki and subsequent amendments, and oversight by “Federazione Alzheimer Italia” ensured adherence to relevant ethical and professional standards.

Of our cohort of 60 cases, five cases met the eligibility criteria for inclusion in the case series. Of these, two were women and three were men. The mean age at onset was 65.4 years (SD ± 8.1), and the mean age at death was 71.8 years (SD ± 7.4). At the first clinical evaluation, all five cases had a clinical diagnosis of major-NCD due to FTD; the diagnosis of one case was later revised to a possible frontal variant Alzheimer’s disease (fvAD), according to a subsequent biomarkers assessment.

A summary of clinical and neuropathological profiles of the five cases is reported in Tables 1, 2.

This case series includes emblematic examples of the heterogeneity of FTLD, highlighting the variability of clinical presentations and the range of underlying proteinopathies encountered in the frontotemporal dementia spectrum.

The first case represents the most “classical” presentation within the series, with an early behavioral syndrome consistent with bvFTD associated with a pathogenic GRN mutation, and a neuropathological diagnosis of FTLD-TDP-A. Neuroimaging revealed asymmetric atrophy and hypometabolism, and the patient developed aphasia and hemiparkinsonism, features classically associated with tauopathies. However, pTDP-43 pathology may also be associated with aphasia and parkinsonism (Mackenzie et al., 2011; Spinelli et al., 2017). Moreover, hemispheric asymmetry is a well-recognized feature of GRN-associated FTLD, described as a characteristic rather than an exception (Borrego-Ecija et al., 2024). Our findings corroborate these observations. This was the only case in the series with a relatively typical bvFTD presentation, whereas the remaining cases showed atypical phenotypes, reflecting that non-canonical presentations are common within the FTLD spectrum.

In the second case, the clinical presentation was characterized by predominant posterior and right-sided brain involvement, with early visuospatial disorientation and left-sided visuo-perceptual deficits, followed by behavioral symptoms and logopenic language impairment, while memory remained preserved. Although a PCA phenotype could initially be considered, current consensus classification requires the absence of early behavioral and language disturbances (Crutch et al., 2017), making this diagnosis unlikely. The appearance of behavioral symptom, and the subsequent identification of a pathogenic GRN Thr272fs mutation allowed a diagnosis of probable bvFTD with atypical onset. Rare bvFTD cases with posterior-predominant involvement have been reported, particularly in genetically determined forms associated with C9orf72 or GRN mutations (Caroppo et al., 2015; Floris et al., 2012), often accompanied by hemispheric asymmetry (Borrego-Ecija et al., 2024; Rohrer, 2012). Neuropathology confirmed asymmetric FTLD-TDP-A with an posterior distribution extending to parieto-occipital regions. This case highlights the phenotypic and anatomical heterogeneity of GRN-associated FTLD, and illustrates that posterior cortical involvement, although uncommon, can occur within the FTLD spectrum.

The third case presented with nfvPPA, associated with bucco-facial apraxia and behavioral disturbances, followed by rapidly progressive right-sided hemiparkinsonism with dystonia. Neuroimaging showed predominant left frontal involvement. While this presentation would classically suggest a nfvPPA with cortico-basal syndrome associated with an underlying tauopathy, neuropathology demonstrated predominant FTLD-TDP-A with extensive fronto-temporal and limbic involvement. Although less frequent, FTLD-TDP-A has been reported in nfvPPA (Mackenzie et al., 2011; Spinelli et al., 2017). Additional ADNC was detected, a finding previously described in nfvPPA (Spinelli et al., 2017). Another notable feature was the relatively late age at onset with a rapid disease progression. Older age increases the likelihood of amyloid co-pathology, and FTLD cases with mixed neuropathological changes tend to present later than “pure” forms (Bergeron et al., 2018; Seo et al., 2018). Unlike previous reports of longer disease duration in FTLD-TDP-A with co-pathology (Gefen et al., 2018), this case showed rapid evolution, underscoring the potential detrimental impact of co-pathologies and the variability of individual disease trajectories.

The fourth patient resembled svPPA, a syndrome most commonly associated with FTLD-TDP-C (Mackenzie et al., 2011). However, neuropathological examination revealed high ADNC, with predominant left parietal, anterior and mesial temporal involvement, without pTDP-43 pathology. Similar clinicopathological mismatches have been reported, as AD can rarely present with a semantic dementia-like phenotype and asymmetric involvement (Chow et al., 2010; Davies et al., 2005; Whitwell et al., 2019). This case supports prior reports, confirming that AD may rarely present with a svPPA clinical syndrome when neurodegeneration preferentially targets semantic networks. Thus, the clinical phenotype appears to depend more on anatomical distribution than on protein type, which cannot be reliably inferred from phenotype alone. Alongside neurodegenerative causes, an FTD-like phenotype may also reflect non-degenerative and sometimes treatable conditions, which should be considered in the differential diagnosis (Desmarais et al., 2021; Mez et al., 2013; Sugiyama et al., 2022). Finally, behavioral disturbances observed in this case may also be consistent with AD, where agitation, aggressivity, wandering, and delusions are frequently reported as a consequence of limbic involvement (Cerejeira et al., 2012; Negro et al., 2025).

The fifth patient initially presented with behavioral symptoms mimicking bvFTD, supported by 18-FDG-PET evidence of fronto-temporal and anterior cingulate hypometabolism. The subsequent development of logopenic aphasia and memory deficits suggested a more diffuse process, prompting further investigations, including CSF biomarkers consistent with AD. Neuropathology confirmed high ADNC with extensive frontal involvement which, along with severe limbic-predominant LTS, provided a plausible substrate for behavioral and parkinsonian features observed (Beach et al., 2009). Additional LATE-NC stage 1 and vascular pathology further highlighted the frequent coexistence of multiple proteinopathies in AD, particularly in cases with severe behavioral symptoms (Negro et al., 2025; Nelson et al., 2023). Although fvAD is typically characterized mainly by executive dysfunction, severe behavioral symptoms may occur in the presence of extensive frontal involvement and coexisting proteinopathies (Ossenkoppele et al., 2015). This case highlights the importance of longitudinal clinical follow-up and biomarker integration to improve diagnostic accuracy. Behavioral assessment in suspected FTD is often non-standardized, with potential delayed diagnoses; structured instruments such as the Mild Behavioral Impairment Checklist, the Frontal Systems Behavior Scale or the Frontal Behavioral Inventory can help mitigate this gap (Ismail et al., 2017; Kertesz et al., 2000; Malloy et al., 2007; Zapata-Restrepo et al., 2021). Further, functional imaging, such as diffusion MRI/DTI or FDG-PET, holds promise for improving syndromic and etiologic discrimination (Chandrasekar et al., 2025; Hoffmann, 2013; Santillo et al., 2013). However, definitive diagnosis ultimately relies on neuropathological examination, which often reveals clinically relevant pathologies.

Across the series, the clinicopathological variability can be summarized as follows: (I) the clinical presentation generally reflected the anatomical topography of pathology, irrespective of different underlying proteinopathies; (II) hemispheric asymmetry appeared to shape clinical expression, sometimes contributing to presentations that diverged from canonical frontotemporal patterns; (III) coexisting pathologies may contribute to more complex trajectories over time, particularly in older individuals. Taken together, these patterns help explain why similar clinical syndromes may emerge when different proteinopathies affect the same neural networks, whereas less typical or asymmetric pathological involvement may lead to misleading clinical presentations (Borrego-Ecija et al., 2024; Caroppo et al., 2015). Another recurring feature is the coexistence of multiple proteinopathies, an increasingly recognized finding at autopsy that may contribute to clinical variability of dementia (Gefen et al., 2018; Karanth et al., 2020; Nelson et al., 2023). Experimental studies suggest that misfolded proteins, such as amyloid-β, tau, pTDP-43, and α-synuclein, may interact through cross-seeding mechanisms, amplifying neurodegenerative processes beyond the effect of a single dominant pathology (Nonaka et al., 2018). These observations underscore the central role of neuropathological examination in establishing the etiological diagnosis of neurodegenerative diseases. Despite advances in neuroimaging and biomarkers, in vivo prediction of the underlying proteinopathy remains limited in atypical and mixed cases. Brain banking protocols based on unilateral hemispheric sampling may underestimate the contribution of hemispheric asymmetry (Vonsattel et al., 2008), whereas bilateral neuropathological analysis may provide a more complete view of clinicopathological heterogeneity (Bell et al., 2008; Poloni et al., 2020).

This case series highlights the clinical, anatomical, and molecular heterogeneity of FTLD, showing that individual cases may deviate from classical clinicopathological expectations and primarily reflect lesion topography over the molecular substrate. By providing neuropathologically confirmed examples of clinicopathological mismatch, this series offers reference points for interpretating atypical frontotemporal presentations and supports a multimodal diagnostic work-up when available.