Identifying and measuring axonal loss in patients with MS using magnetic resonance imaging (MRI) methods is challenging. Multi-compartment microscopic diffusion imaging using SMT provides the apparent axon volume fraction (V_ax), an index immune from fiber orientation and proxy of axonal content.

To examine the capacity of diffusion tensor imaging (DTI), spherical mean technique (SMT) and Neurite Orientation Dispersion and Density Imaging (NODDI) derived metrics in distinguishing lesional vs. non-lesional white matter (WM) tissue in multiple sclerosis (MS) brains.

Eighteen patients with MS underwent a 3.0 Tesla brain MRI, including T1-weighted and T2-weighted sequences, DTI, SMT and NODDI. Parametric maps of axial diffusivity (AD), V_ax, intracellular volume fraction (ficvf), isotropic volume fraction (fiso), and orientation dispersion index (ODI) were reconstructed. Anatomically matched regions of interest were used to quantify differences in chronic black holes (cBHs), T2-lesions, and normal appearing white matter (NAWM).

V_ax(p=0.0005), ficvf(p=0.02) and ODI(p=0.05) were lower in cBHs compared to T2-lesions. V_ax decreased by 60% in cBHs vs. T2-lesions, significantly different(p=0.004) from what was observed in ODI(-14%) but not in ficvf(-35%).  AD(p=0.001) and fiso(p=0.002) were higher in cBHs vs. T2-lesions. fiso increased by 37% in cBHs vs. T2-lesions whilst AD increased by 15%(p=0.04).

V_ax(p=0.0001) and ficvf(p=1.1E-07) but not ODI were lower in T2-lesions vs. NAWM. ficvf decreased by 21% in T2-lesions vs. NAWM, significantly different(p=0.006) from what was observed in ODI(-7%) but not V_ax(-20%). Only AD(p=0.0003) measured significantly higher (11%) in T2-lesions compared to NAWM, but was not significantly different from what was observed in fiso(+6%).

Our data prove that DTI, SMT and NODDI are able to distinguish lesional vs. non-lesional WM tissue in brains of MS patients. V_ax appears to have a larger effect size. By indirectly reflecting axonal count rather than micro-environmental integrity, V_ax generates a new conceptual approach to axonal imaging.