In amyotrophic lateral sclerosis (ALS), conventional MRI of the brain is currently used in the differential diagnosis process.1 2 In an attempt to quantify in vivo the extent of the critical features of ALS pathology (ie, loss of motor neurons and corticospinal tract (CST) degeneration), brain proton MR spectroscopy3 and diffusion tensor (DT) MRI4^–7 have been used. However, longitudinal studies of ALS patients with these quantitative MR techniques gave conflicting results.8^–14 More recently, DT MRI has been successfully used to grade the extent of cervical cord damage associated with ALS,7 but the temporal evolution of such damage has not been investigated yet.

We performed this longitudinal DT MRI study: (a) to define the temporal evolution of intrinsic tissue damage and atrophy in the cervical cord of ALS patients, and (b) to assess the correlation between cord MRI measurements and concomitant brain damage (in terms of T2-visible abnormalities and DT-derived measurements from the brain portion of the CST).

PATIENTS AND METHODS

Patients

We enrolled 28 patients with probable or definite ALS2 in a 6–12-month longitudinal study.7 Disease severity was assessed using the ALS functional rating scale (ALSFRS).15 Twenty-four of these patients (10 women; mean age = 56.2 years, range = 27–73; median disease duration = 24–months, range = 6–58; mean ALSFRS at baseline = 28, range = 7–38) had a clinical follow-up, and 17 (seven women; mean age = 55.2 years, range = 27–73; median disease duration = 27–months, range = 12–58; mean ALSFRS at baseline = 27, range = 7–33) also had an MRI follow-up (mean follow-up = 9 months, range = 6–12). Three patients refused any follow-up examination; seven patients had follow-up clinical assessment but not MRI because of respiratory impairment (two), severe dysphagia (two) or severe locomotor disability (three); and one patient died by the time follow-up was done.

At follow-up, the mean ALSFRS scores were 21 (range = 3–33) in the 24 patients with clinical follow-up, and 21 (range = 6–33) in those with MRI. In this latter group, the median rate of clinical progression during follow-up (calculated as [(baseline ALSFRS–follow-up ALSFRS)/follow-up duration]) was 0.5 units/month. Seven patients with progression rate more than the median value were classified as “rapidly” progressing.5

Twenty (nine female, mean age = 53.2, range = 28–73) healthy controls were included. Local Ethical Committee approval and written informed consent from all subjects were obtained.

MRI acquisition

Using a 1.5-Tesla scanner, the following sequences were acquired at study entry and follow-up:7 (1) cervical cord: (a) dual-echo (DE) turbo spin echo (TSE); (b) T1-weighted 3D magnetisation-prepared rapid acquisition gradient echo; and (c) single shot SE echo planar imaging (EPI); (2) brain: (a) axial DE TSE; (b) coronal T2-weighted TSE; and (c) axial single shot SE EPI.

Image analysis

Cervical cord abnormalities were identified on DE scans. The cord average mean diffusivity (MD) and fractional anisotropy (FA) were derived,7 and the cross-sectional area was measured.16 The presence of T2 hyperintensities along the brain CST was assessed on DE scans. Brain DT MRI data were reconstructed, as for the cord. A tractography algorithm was used to reconstruct the anatomy of brain portion of the CST.17

Statistical analysis

Baseline MRI variables were compared between ALS patients and controls using a Student t test. Baseline and follow-up MRI measurements were compared using a mixed effect linear model (accounting for repeated measures within each subject not equally spaced in time), including follow-up duration as independent variable. The same model was used to compare cord MRI changes between patients with and without hyperintensities along the brain CST at baseline, and between rapidly and non-rapidly progressing patients. Univariate correlations were assessed using the Spearman rank correlation coefficient. A univariate logistic regression model, adjusted for follow-up duration, was used to investigate whether clinical and cord MRI quantities at baseline independently influenced the probability to return for the follow-up visit.

RESULTS

No abnormalities were seen on the conventional brain and cervical cord MR images from healthy controls. During the study period, all DT MRI measurements from healthy controls remained stable (data not shown). Baseline cord MRI measurements did not differ between patients who returned for follow-up visit and those who did not (data not shown).

Eight of the 17 ALS patients who returned for MRI follow-up had bilateral hyperintensities of the CST at baseline. At follow-up, none of the patients showed new areas of increased signal intensity in the CST. No macroscopic cervical cord abnormalities were seen on the DE scans from ALS patients at both time points.

During the follow-up, ALS patients showed a significant decrease in cord area (p = 0.003) and average FA (p = 0.01), and a significant increase in cord average MD (p = 0.01) (table 1). Cord MRI measurements at baseline were not associated with their changes over time. No correlation was found between the longitudinal changes of cord average MD and FA, and the concomitant change in cord area. Cord MRI changes did not differ between ALS patients with and without hyperintensities along the brain portion of the CST and between rapidly and non-rapidly progressing patients. They did not correlate with ALSFRS changes (r values ranging from −0.33 to 0.22). At baseline, ALS patients had a significantly higher brain CST average MD compared with controls (p = 0.001). In ALS patients, CST diffusivity measurements remained stable during the follow-up (table 1).

DISCUSSION

This study shows that ALS patients experience progressive cord atrophy and accumulation of structural changes in the residual cervical cord tissue, as measured by DT MRI. These findings suggest that longitudinal assessment of cervical cord damage using MRI is sensitive to short-term changes related to the disease.

The pathological hallmarks of cord injury in ALS are pronounced CST degeneration and reduction in the number of lower motorneurons and dendritic spines in the anterior horn.18^–21 Neuronal cytoskeleton abnormalities, neuronal inclusions and intracytoplasmic spheroids, as well as anterior horn astrocytosis,20 21 have also been described. All these pathological changes have the potential to alter the diffusivity properties of the cord. In particular, an increased average MD can result from the enlargement of extracellular spaces, due to neuronal and axonal loss, which can promote water diffusion. Conversely, the reduction in average FA might reflect both intracellular abnormalities of surviving axons,22 and formation of “new” barriers, due to the presence of cell debris resulting from partially degenerated or disintegrated CST fibres, inflammatory infiltrates23 and astrocytsis. The progressive increase in MD and decrease in FA in the cord of ALS patients might reflect a net loss of structural barriers to water molecular motion and fibre bundle coherence, which in turn are likely to be the result of loss of motor neurons and axonal degeneration.

The present study showed a lack of association between progressive accumulation of cord tissue damage and involvement of the brain portion of the CST. We also did not find any cord MRI difference between patients with and without hyperintensities along the brain CST, and DT-derived measurements from brain CST did not change significantly over time, despite the fact that at baseline, in ALS patients who returned for follow-up, brain CST MD was significantly higher than in controls. These results are in agreement with those of previous longitudinal regions of interest11 12 and tractography13 studies, which failed to show a significant evolution of brain CST damage in ALS patients. They also agree with findings of pathological studies which have demonstrated that ALS-related intrinsic cord injury is independent of brain changes.18 19 21 Taken together, these studies support the notion of a heterogeneous distribution and progression of the various components of the ALS pathological process across the brain and the spinal cord.7

Despite the progressive accumulation of cervical cord damage over the follow-up, cord atrophy and diffusivity measurements were not associated with worsening of disability and rate of disease progression in ALS patients. Admittedly, the drop-out rate (39%) of the study was relatively high. However, it was at the lower limit of the range of previous longitudinal studies in ALS patients (from 27% to 75%).8^–14 At least two factors might contribute to explain our findings. First, longitudinal assessment might have been obtained mainly in patients with a more stable disease. However, we found no difference in clinical and MRI characteristics between patients with and without follow-up and this argues, at least partially, against this hypothesis. Second, pathological changes occurring at a given stage of the disease may result in an immediate change of MR properties of injured tissues, but cause measurable clinical impact at a later stage. Finally, the scale used for measuring clinical impairment is not specifically tailored to detect disability due to cervical cord damage. These considerations would call for future longitudinal DT MRI studies of brain and cord with a longer follow-up duration, possibly on patients recruited from the onset of the disease.