Characterization of FTLD spectrum through advanced diffusion-weighted MRI metrics: a connectome approach
Federica Agosta1, Silvia Basaia1, Federica Facente1, Camilla Cividini1, Edoardo Spinelli1, Elisa Canu1, Veronica Castelnovo1, Giuseppe Magnani1, Francesca Caso1, Paola Caroppo2, Sara Prioni2, Cristina Villa2, Lucio Tremolizzo3, Ildebrando Appollonio3, Vincenzo Silani4, Massimo Filippi5
1IRCCS San Raffaele Scientific Institute, 2Fondazione IRCCS Istituto Neurologico Carlo Besta, 3“San Gerardo” Hospital and University of Milano-Bicocca, 4University of Milan Medical School - IRCCS Istituto Auxologico Italiano, 5Ospedale San Raffaele, Neuroimaging Research Unit
Objective:
To investigate structural alterations in brain network of patients within the frontotemporal lobar degeneration (FTLD) spectrum using connectome analysis with advanced diffusion-weighted MRI metrics.
Background:
MRI connectomics is a promising tool to describe network-based degeneration in neurodegenerative diseases. Neurite orientation dispersion and density imaging (NODDI) has shown to provide novel insights on structural connectivity alterations.
Design/Methods:
Thirty-four behavioral-variant frontotemporal dementia (bvFTD), 11 semantic variant primary progressive aphasia (svPPA), 11 nonfluent (nfvPPA) patients and 48 healthy controls underwent multi-shell diffusion MRI. Fractional anisotropy (FA) maps were computed. Intra-cellular Volume Fraction (ICVF) maps were also estimated using NODDI, providing a direct quantification of neurite integrity. Graph analysis and connectomics assessed global and local structural topological network properties and regional structural connectivity.
Results:
Overall, widespread structural changes were observed in bvFTD patients relative to healthy controls and svPPA patients in terms of FA-derived network properties. nfvPPA patients also showed altered FA topological properties at global level compared to healthy controls. Sum of node weights (SN) in the sensorimotor lobe revealed that nfvPPA patients showed an altered connection strength compared to svPPA. Moreover, nfvPPA patients showed a significant decreased FA in the left hemispheric fronto-temporal-insular regions relative to controls. Considering ICVF, more severe alterations compared to FA maps allowed to find differences also between svPPA and nfvPPA patients. ICVF graph analysis measures also showed that svPPA had a lower degree centrality and SN in the temporal lobe compared to controls.
Conclusions:
These findings suggest that conventional diffusion-tensor measures might be sensitive enough to highlight vulnerable connections in the FTLD spectrum. However, ICVF demonstrated to be a more specific biomarker to differentiate FTLD syndromes. Specifically, the benefits emerged in the differentiation between svPPA patients and other groups.