A magnetic resonance imaging study of the cervical cord of patients with CADASIL

Patients and methods.

We studied 25 patients with CADASIL (15 women and 10 men; mean age = 48.0 years, range = 24 to 63 years; median disease duration = 10 years, range = 2 to 28 years; mean Rankin scale score6 = 2, range = 0 to 4). All of the patients were symptomatic, having one or more of the following manifestations: ischemic episodes (21), cognitive deficits (13), psychiatric symptoms (6), and migraine with aura (9). The diagnosis was confirmed by either the demonstration of a mutation within the Notch3 gene1 (17 cases) or a skin biopsy7 (8 cases). Fourteen subjects with no history of neurologic dysfunctions and with a fully normal neurologic examination at study entry served as controls (seven women and seven men; mean age = 46.7 years; range = 22 to 61 years). Written informed consent was obtained from all the subjects before study initiation.

MRI scans were obtained from all subjects by using a 1.5 T system. The following pulse sequences were acquired: 1) T2-weighted turbo spin-echo (TSE) (repetition time [TR]/echo time [TE] = 4,700/112); 2) fast short tau inversion recovery (fast-STIR) (TR/TE/inversion time [TI] = 2,288/60/110); 3) T1-weighted spin-echo (TR/TE = 500/12,); and 4) two-dimensional gradient-echo (GE) (TR/TE = 792/10, flip angle = 20°) without and with the application of an MT saturation pulse (off-resonance radiofrequency pulse centered 1.5 kHz below the water frequency with a gaussian envelope of duration of 7.68 ms and a flip angle of 500°). To reduce misregistration artifacts between non-MT and MT scans caused by patient motion during the acquisition, we used a semi-interleaved GE sequence, which collects one third of the k-space lines without the saturation pulse, followed by the collection of one third of the k-space lines with the saturation pulse. This is then repeated for the second and third portions of the k-space. For sequences 1, 2, and 3, eight 3-mm-thick sagittal slices were obtained. For the GE scans, 24 interleaved, 5-mm-thick, axial slices were acquired. We also obtained a dual-echo TSE scan of the brain (TR/TE = 3,300/16 to 98, 24 axial, contiguous, 5-mm-thick images).

Any abnormality present on the MRI scans of the brain (including those in the basal ganglia) and cervical cord of all subjects were identified consensually by two observers, who were unaware to whom the scans belonged. Brain lesion volumes were calculated as previously described.8 Cervical cord MT ratio (MTR) histograms were obtained by a single observer, as follows. First, from the two GE images, without and with the saturation pulse, quantitative MTR images were derived pixel-by-pixel.9 Then, the entire cervical cord was segmented from the MTR images by using a technique based on local thresholding.9 Finally, MTR histograms were created. We excluded from the analysis all of the pixels with MTR values lower than 10%, to reduce partial volume effect from the CSF. To correct for the between-subject differences in cervical cord volume, each histogram was normalized by dividing it by the total number of pixels included. For each histogram, the following measures were derived: the average MTR and the height and position of the peak.

The Mann–Whitney test was used to compare MTR histogram metrics between controls and patients with CADASIL. The correlations between MTR metrics and age, disability, and brain lesion volume were assessed by using the Spearman Rank Correlation Coefficient. A multivariable logistic model adjusted for age was used to assess the effect of brain lesion volume and average cervical cord MTR on the probability that patients belonged to the group with no or mild disability (Rankin scores between 0 and 1) or to the group with moderate/severe disability (Rankin score equal or greater than 2).

Results.

No abnormalities were seen on the MRI scans of the brain and cervical cord of healthy volunteers. All individuals with CADASIL showed brain hyperintense T2-weighted abnormalities. Mean lesion volume was 73.3 mL (range, 14.7 to 144.3 mL). Brain T2 lesion volume was correlated with patients’ age (r = 0.7, p = 0.0001), and disability (r = 0.6, p = 0.001).

No MRI-visible abnormalities were detected in the cervical cord of patient with CADASIL. In the table, the mean values of the cervical cord MTR metrics from healthy controls and patients with CADASIL are reported. In the figure, the MTR histograms of the two groups are shown. The peak height of the MTR histogram from patients with CADASIL was significantly lower than that from controls. Cervical cord MTR was inversely correlated with patients’ age (r = −0.5, p = 0.02), disability (r = −0.4, p = 0.05), and brain lesion volume (r = −0.5, p = 0.02). The logistic regression model showed that brain lesion volume and cervical cord MTR in isolation explained 46% (p = 0.002) and 15% (p = 0.09) of the variance in disability scores, respectively. These two quantities taken together explained 47% of the disability variance.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Cervical cord magnetization transfer ratio histograms from healthy controls (black line) and patients with CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) (gray line).

Discussion.

We assessed the presence and extent of cervical cord abnormalities in a relatively large group of patients with CADASIL. We did not detect any lesion on conventional MRI scans confirming the concept that, although CADASIL is a systemic arterial disease, it preferentially affects the brain. This observation fits well with the results of previous brain MRI studies that show a relative sparing of the cerebellum and the lower brainstem.4,5⇓ It also agrees with the concept that a major determinant of lesion distribution in CADASIL is the length of the perforating arteries, with shorter arteries being less affected than longer ones.4 The absence of macroscopic lesions in the cervical cord from patients with CADASIL might be helpful in the diagnostic workup out of patients with multiple white matter abnormalities of the brain. The clinical and brain MRI manifestations of patients with CADASIL may mimic those of other neurologic conditions, such as MS. In MS, however, multiple abnormalities of the brain white matter are typically associated with cord abnormalities.9

We also assessed the integrity of the cervical cord tissue of patients with CADASIL by using MTR histogram analysis. We found that the peak height of the MTR histogram was lower in CADASIL patients than in healthy subjects. Low MTR indicates a reduced capacity of the macromolecules to exchange magnetization with the surrounding water molecules, thus reflecting reduced tissue integrity.8,9⇓ A recent postmortem study of patients with MS has shown that low MTR values correspond to areas where severe axonal and myelin loss has occurred.10 Because there were no macroscopic cord lesions in our patients, subtle changes beyond the resolution of conventional MRI are likely to be the substrate of the reduced peak height of the cervical cord MTR histogram. The observation that average cord MTR and peak position of the histogram were not different between patients and controls suggests that these changes are widespread enough to reduce the amount of “truly” normal tissue8 but not severe enough to cause a broadening of the peak to its left-hand side, as shown in MS.9 Although we can only speculate about the nature of these abnormalities, the observation of a significant correlation between the extent of brain lesions and MTR changes in the cervical cord suggests that CADASIL damage in the cord might be the result of wallerian degeneration of fibers passing through brain lesions. Nevertheless, albeit less likely, other potential factors responsible for the reduced MTR histogram peak height, such as subtle ischemic damage of the cord secondary to impaired microcirculation, cannot be ruled out. That cord abnormalities in CADASIL are not independent from the extent of brain damage is also suggested by the results of the multivariable analysis, showing that the contribution of CADASIL brain damage to patients’ disability is much higher than that of cord damage and that a combined assessment of CADASIL brain and cord damage adds negligible information compared with that provided by assessment of brain damage alone.