A longitudinal diffusion tensor imaging study on Wallerian degeneration of corticospinal tract after motor pathway stroke

Abstract

Wallerian degeneration of the corticospinal tract (CST) after motor pathway ischemic stroke can be characterized by diffusion tensor imaging (DTI). However, the dynamic evolution of the diffusion indices in the degenerated CST has not previously been completely identified. We investigated this dynamic evolution and the relationship between early changes of the diffusion indices in the degenerated CST and long-term clinical outcomes. DTI and neurological examinations were performed repeatedly in 9 patients with first-onset motor pathway subcortical infarction at 5 consecutive time points, i.e. within 1 week, at 2 weeks, 1 month, 3 months and 1 year. Using a region of interest method, we analyzed the ratios of the fractional anisotropy (rFA), mean diffusivity (rMD), primary eigenvalue (rλ₁) and transverse eigenvalue (rλ₂₃) between the affected and unaffected sides of the CSTs. We did not find any significant changes in the diffusion indices of the contralesional CSTs across time points. The rFA decreased monotonously during the first 3 months and then stabilized. The rMD increased after 2 weeks and stabilized after the third month. The rλ₁ decreased during the first 2 weeks and then remained unchanged. The rλ₂₃ increased during the first 3 months and then stabilized. We also found that the changes in the rFA between the first 2 time points were correlated with the NIHSS (P = 0.00003) and the Motricity Indices (P = 0.0004) after 1 year. Our results suggest that for patients with motor pathway stroke the diffusion indices in the degenerated CST stabilize within 3 months and that early changes in the rFA of the CST may predict long-term clinical outcomes.

Introduction

Degeneration of the distal parts of nerves after injury to the proximal axon or cell body is referred to as Wallerian degeneration (WD), which occurs in both peripheral and central nervous systems. WD was first reported by Waller (1850) who found this pattern of degeneration after cutting the glossopharyngeal and hypoglossal nerves of the frogs. In the central nervous system, WD is characterized by a highly stereotypical course, starting with disintegration of the axonal skeleton and membrane within days after injury, followed by degradation of the myelin sheath and infiltration by macrophages and microglia, with subsequent atrophy of the affected fiber tracts (Johnson et al., 1950, Lampert and Cressman, 1966). WD of the corticospinal tract (CST) after motor pathway ischemic stroke is a well-known phenomenon, which can be characterized by diffusion tensor imaging (DTI).

DTI, a non-invasive MRI technique, measures the random motion of water molecules and provides information about cellular integrity and pathology (Le Bihan 2003). Within a highly ordered white matter tract, water molecules diffuse faster in the direction parallel to the tract than in the perpendicular directions because axonal membranes and myelin sheaths restrict transverse diffusion (Beaulieu and Allen, 1994, Wimberger et al., 1995). Pathological processes which change the microstructural environment, such as neuronal size, extracellular space and tissue integrity, result in altered diffusion (Anderson et al., 1996, Sevick et al., 1992). Molecular diffusion can be measured by several indices derived from DTI data. Mean diffusivity (MD) and fractional anisotropy (FA) are commonly used indices that reflect the average amplitude and the directionality of molecular motion, respectively (Pierpaoli and Basser, 1996). They are calculated from 3 diffusion tensor eigenvalues, which are the diffusion coefficients along the major, medium and minor axes of the diffusion ellipsoid. The primary eigenvalue (λ₁) is the diffusion coefficient along the direction of maximum diffusion. The transverse eigenvalue (λ₂₃) is generated by averaging the medium (λ₂) and minimum eigenvalues (λ₃), thus avoiding sorting bias due to similar magnitudes of λ₂ and λ₃ (Basser and Pajevic, 2003). This measure reflects the average diffusivity perpendicular to the direction of maximum diffusion.

Two popular analytic approaches for detecting changes in the diffusion indices of a specific fiber tract are the pure region of interest (ROI) analysis and the tract ROI analysis. In the pure ROI analysis the ROIs are manually defined on MR images. This method is easy to perform, but the definition of the ROI must be carefully designed to improve the reproducibility, because the boundary of a fiber tract is often ill-defined. In tract ROI analysis defining the ROI of a fiber tract is based on tractography, which can either be performed by considering a fiber tract as an entire ROI or by defining a segment of the fiber tract as a ROI (Berman et al., 2005, Lin et al., 2006, Pagani et al., 2005, Partridge et al., 2005, Yu et al., 2007, Yu et al., 2008). In this study, we analyzed the CST using both methods.

Using a non-invasive approach to characterize and stage WD in the CST is clinically important for patients with motor pathway stroke. A few previous studies have delineated the process of WD using DTI (Pierpaoli et al., 2001, Thomalla et al., 2004, Thomalla et al., 2005). At the chronic stage (more than 1 year) after stroke, the degenerated CST showed a sharp decrease in FA, a slight increase in MD, a decrease in λ₁, and an increase in λ₂₃ (Pierpaoli et al., 2001). At the early stage (within 2 weeks), the degenerated CST demonstrated a decrease in FA and λ₁, an unchanged MD, and an increased λ₂₃ (Thomalla et al., 2004). A single pioneering study on 2 stroke patients who were examined at three time points (TP), is the only one to date that has attempted to describe the time course of WD (Thomalla et al., 2005). Thus, the detailed time course of WD after stroke still needs to be investigated.

In previous DTI studies of WD of the CST, the ratios of the diffusion indices (rFA, rMD, rλ₁ and rλ₂₃) between the affected and unaffected sides of the CSTs have commonly been used based on the hypothesis that diffusion indices of the contralesional CST are unchanged after stroke. However, a recent study has reported that the FA of the normal-appearing white matter of the whole brain was increased within 2 years after stroke, apparently suggesting a long-term improvement in apparent white matter integrity following ischemic stroke (Wang et al., 2006). Therefore, it is critically important to study the possible plasticity in the contralesional CST before using the ratios (rFA, rMD, rλ₁ and rλ₂₃) to study the dynamic changes of the degenerated CST.

Impairment of motor function is one of the most serious disabling sequelae of ischemic stroke. Thus an early prediction of the long-term outcome of motor function is critically important for stroke treatment and rehabilitation. Two previous studies (Maeda et al., 2005, Thomalla et al., 2004) attempted to predict the motor outcome at 3 months after stroke using the ratios of the diffusion indices of the CSTs from an earlier time point (TP). However, this method may be confounded by individual side differences in the diffusion indices of the CSTs before the stroke. Thus we propose that it seems to be more appropriate to use the changes in the ratios of the diffusion indices between 2 earlier TPs (i.e. within 1 week and at 2 weeks) to predict the long-term motor outcomes.

Here, we aimed to investigate (1) whether the contralesional CST shows plastic changes following motor pathway ischemic stroke; (2) the dynamic evolution of the diffusion indices in the degenerated CST; and (3) whether the clinical outcomes could be predicted by the early changes in the diffusion indices.