Advantages of QBI in TBSS analyses

Abstract

Diffusion-weighted magnetic resonance imaging (DWMRI) is used to study white matter (WM) in normal and clinical populations. In DWMRI studies, diffusion tensor imaging (DTI) models the WM anisotropy with one dominant direction, detecting possible pathway abnormalities only in large and highly coherent fiber tracts. However, more general anisotropy models like Q-ball imaging (QBI) may provide more sensitive WM descriptors in single patients. The present study aimed to compare DTI and QBI models in a group-level population analysis, using Amyotrophic Lateral Sclerosis (ALS) as a pathological case model of WM tract degeneration. DWMRI was performed in 19 ALS patients and 19 age and sex-matched healthy controls. DTI and QBI estimates were compared in whole-brain tract-based spatial statistics (TBSS) and volume of interest (VOI) analyses, and correlated with ALS clinical scores of disability. A significant decrease of the QBI-derived generalized fractional anisotropy (GFA) was observed in both motor and extramotor fibers of ALS patients compared to controls. Homologue DTI-derived FA maps were only partially overlapping with GFA maps. Particularly, the left corticospinal tracts resulted more markedly depicted by the QBI than by the DTI model, with GFA predicting ALS disability better than FA. The present findings demonstrate that QBI model is suitable for studying WM tract degeneration in population-level clinical studies. Particularly, group-level studies of fiber integrity may benefit from QBI when DTI is biased towards low values, such as in cases of fiber degeneration, and in regions with more than one dominant fiber direction.

Keywords: ALS; DTI; Fractional Anisotropy; Generalized Fractional Anisotropy; QBI; TBSS.