Clinical Manifestations.

Area of Science:

• Neuroscience
• Psychiatry
• Medical Diagnostics

Background:

• Delays in diagnosing sporadic behavioral variant of frontotemporal dementia (s-bvFTD) impede patient care and research enrollment.
• Clinical heterogeneity and overlap with primary psychiatric disorders (PPD) contribute to diagnostic delays.
• The DIPPA-FTD consortium seeks to improve early s-bvFTD diagnosis by evaluating ambiguous late-onset behavioral changes.

Purpose of the Study:

• To predict the follow-up diagnosis of ambiguous cases using baseline clinical assessments.
• To apply principal component analysis (PCA) to identify distinct profiles for s-bvFTD and PPD.
• To evaluate the diagnostic accuracy of data-driven clusters.

Main Methods:

• PCA was applied to baseline clinical data from a subset of the DIPPA-FTD study (n=106).
• Data included ACE-III, BDI-II, Ekman-35, FTDvsPPD Checklist, SNQ, and TMT A+B.
• K-means clustering (k=2) and comparison of principal components (PCs) between diagnostic groups were performed.

Main Results:

• The first principal component (PC1) explained 43.8% of the variance and significantly differed between s-bvFTD, PPD, and ambiguous cases.
• Cluster 1 (high PC1) predominantly comprised s-bvFTD cases (78.1%), while Cluster 2 (low PC1) comprised PPD cases (73.0%).
• Higher PC1 scores correlated with diagnostic certainty for s-bvFTD (r=0.74), while lower scores indicated PPD certainty. Follow-up data showed ambiguous cases clustering with PPD were more likely to be diagnosed with PPD.

Conclusions:

• A data-driven PCA approach successfully identified baseline clinical profiles differentiating s-bvFTD from PPD.
• This method shows potential for aiding early and accurate diagnosis in individuals with late-life behavioral changes.
• Further validation is needed, but this pilot study demonstrates a promising tool for diagnostic improvement.