Frontotemporal dementia: does structural MRI-based clustering match clinical syndromes?

Area of Science:

• Neuroimaging
• Neurology
• Data Science

Background:

• Frontotemporal dementia (FTD) is a group of disorders affecting behavior, language, and motor skills.
• FTD syndromes often share clinical and imaging features, necessitating unbiased analysis.
• Investigating FTD syndromic heterogeneity is crucial for accurate diagnosis and treatment.

Purpose of the Study:

• To explore the underlying syndromic heterogeneity of frontotemporal dementia (FTD).
• To apply data-driven clustering analysis to structural MRI data for FTD classification.
• To identify distinct patterns of brain atrophy associated with different FTD clinical presentations.

Main Methods:

• Structural MRI data from 400 FTD patients were analyzed using data-driven clustering.
• Principal Component Analysis (PCA) was performed on w-scored MRI data.
• Hierarchical clustering algorithms were employed to group patients based on imaging patterns.

Main Results:

• A three-cluster solution provided the most clinically meaningful separation of FTD syndromes.
• Cluster 1 showed frontal atrophy linked to speech, language, and behavioral deficits.
• Cluster 2 revealed temporal atrophy, primarily in semantic variant primary progressive aphasia (svPPA) and right temporal variant FTD (rtvFTD).
• Cluster 3 identified midbrain and subcortical atrophy, encompassing progressive supranuclear palsy (PSP) and corticobasal syndrome (CBS).

Conclusions:

• FTD heterogeneity is characterized by three main axes: frontal, temporal, and midbrain/subcortical atrophy.
• These axes correspond to distinct patterns of speech, language, behavioral, and motor deficits.
• The findings underscore the importance of data-driven approaches in understanding complex neurodegenerative disorders like FTD.