Unique combinations of shape morphometric features improves discriminability of ftld phenotypes

Background Frontotemporal lobar degeneration (FTLD) is associated with diverse clinical phenotypes underlain by multiple disease pathologies and genetic mutations. As such, traditional structural MRI analyses lack sensitivity specificity for discriminating FTLD syndromes. Here, we use data-driven methods to extract a concise set of MRI-derived shape morphometric features examine the discriminatory capability their unique combinations in four phenotypes. Method 190 patients sporadic or familial spectrum disorders (i.e., behavioral variant (bvFTD, n = 107), non-fluent primary progressive aphasia (nfvPPA, 27), semantic (svPPA, 12) supranuclear palsy (PSP, 44)) 27 controls without pathogenic mutations from ALLFTD cohort were evaluated. data preprocessed FreeSurfer software cortical measures extracted indexed normalized surface atlas: thickness (CT), area (SA), curvature (SC) jacobian white matter metric distortion (JW). For all phenotypes, each feature was contrasted using linear models adjusted age gender. Principal component analysis (PCA) then applied individual measure power based on combined assessed logistic regression 10-fold cross-validation. Result Figure 1 displays FDR-corrected T-scores differences bvFTD. Results reveal complementary patterns CT, SA, SC JW phenotype, greatest observed CT across groups. In 2, display derived PCA, weighted eigenvector coefficients, Figures 3-6, receiver operating characteristic curves alongside areas under curve (AUCs) are shown phenotype. bvFTD, nfvPPA PSP, superior model included at AUCs 88.3, 87, 78.6, respectively. svPPA, an AUC 85.6. Conclusion Integrating additional improved classification performance The principal analysis-based approach indicated distinct brain regions contribute discrimination feature, suggesting they may reflect aspects neurodegeneration