Morphometric features enhance phenotype discrimination in frontotemporal lobar degeneration

Area of Science:

• Neuroimaging
• Neurology
• Biostatistics

Background:

• Frontotemporal lobar degeneration (FTLD) presents diverse clinical phenotypes, challenging traditional structural MRI analysis due to limited sensitivity and specificity.
• Multiple underlying pathologies and genetic mutations contribute to FTLD, necessitating advanced diagnostic methods.

Purpose of the Study:

• To evaluate the discriminatory capability of MRI-derived shape morphometric features for three FTLD clinical phenotypes.
• To identify unique combinations of features that can differentiate between FTLD subtypes and healthy controls.

Main Methods:

• Utilized a data-driven multivariate approach (sparse partial least squares discriminatory analysis - sPLS-DA) on MRI data from FTLD patients (behavioral variant, non-fluent variant PPA, semantic variant PPA) and controls.
• Extracted cortical morphometry measures including thickness, surface curvature, and metric distortion.
• Validated discriminatory power on independent, age-matched test data.

Main Results:

• Each cortical morphometric feature exhibited significant, spatially distinct differences between FTLD clinical syndromes.
• The combination of cortical thickness and surface curvature achieved the best discrimination between behavioral variant and non-fluent variant PPA patients and controls.
• Including cortical thickness in models maximized performance for semantic variant PPA discrimination.

Conclusions:

• MRI-derived shape morphometric features, particularly cortical thickness and surface curvature, show high discriminatory power for FTLD phenotypes.
• The sPLS-DA method highlights unique neurodegenerative patterns across FTLD subtypes.
• This data-driven approach holds promise for early detection and differential diagnosis of FTLD phenotypes.