Prognostic Value of MR and Magnetization Transfer Imaging Findings in Patients withClinically Isolated Syndromes Suggestive ofMultiple Sclerosis at Presentation

Patients

We performed MR imaging of the brain in all patients presenting with CIS suggestive of MS between August 1995 and February 1997. In addition to the criteria previously used to select patients with CIS at presentation for prognostic studies (12, 14, 15, 17), we required that the patients had to have at least four T2 lesions of the type seen in cases of CDMS or only three of such lesions if one of them was either subcortical or infratentorial. This was to minimize the chance of recruiting patients with neurologic conditions other than MS, as shown by a recent prospective study (25). When MR images were obtained, a full neurologic assessment was conducted by a single observer who was unaware of the MR findings, and the disability was rated using the Expanded Disability Status Scale score (26). Patients were then prospectively followed up by the same neurologists, with visits scheduled every 6 months, or when the patients experienced new symptoms. The follow-up duration was a mean of 33 months (range, 25–42 months). During the follow-up period, MS was diagnosed solely on clinical grounds by using the Poser criteria (27); paraclinical data were not used. Twenty sex- and age-matched control volunteers (eight men and 12 women; mean age, 33.8 years; SD, 5.0 years) with no history of neurologic diseases and with normal results of their neurologic examinations were also studied. Local Ethical Committee approval and written informed consent from all the patients and control volunteers were obtained before study initiation.

Image Acquisition

MR images of the brain were obtained using an imager operating at 1.5 T. During a single session, the following images were obtained: dual-echo spin-echo (2400/30, 80/1 [TR/TE/excitations]); T1-weighted spin-echo (768/15/2); and 2D gradient-echo (600/12, α = 20°), with and without an MT saturation pulse. The RF saturation pulse was 1.5 kHz below the water frequency, with a gaussian envelope of duration of 16.4 ms, a bandwidth of 250 Hz, and an amplitude of 3.4 × 10⁻⁶ T. For the dual-echo and T1-weighted images, 24 contiguous, interleaved, axial sections were acquired with 5-mm section thickness, a 256 × 256 matrix, and a 250 × 250-mm field of view. MT images were obtained using the same acquisition parameters, except for the number of sections, which was 20. The set of sections for the MT images was positioned to obtain the same central 20 sections as for the dual-echo and T1-weighted images. The sections were positioned to run parallel to a line that joins the most inferoanterior and inferoposterior parts of the corpus callosum, according to published guidelines (28).

Image Analysis

Lesions were first identified by agreement between two experienced observers who did not know to whom the images belonged, on the dual-echo and T1-weighted hard copies, following a procedure described elsewhere (29). Using four contiguous T1-weighted sections, with the most caudal section at the level of the velum interpositum cerebri (30), we calculated an MR measure of cerebral volume. Brain and lesion volume measurements were obtained by a single observer, who did not know to whom the images belonged, using a semiautomated segmentation technique based on local thresholding (29).

From the two gradient-echo images, with and without the saturation pulse, and after their coregistration, MTR maps were derived pixel-by-pixel, as previously described (29). Lesion outlines on proton density–weighted images were superimposed onto the coregistered MTR maps and average lesion MTR was calculated. We also derived MTR histograms (with bins 1% in width) for the NABT (ie, the overall brain tissue not involved by lesions visible on the dual-echo images). First, we removed the skull and other extracranial tissues from the proton density–and gradient-echo images without and with the MT pulse, using the same local thresholding technique as that used for lesion segmentation (29). Second, the scalp-stripped gradient-echo images were coregistered and MTR maps obtained. The MTR maps were then coregistered with the corresponding scalp-stripped proton density–weighted images. Coregistration of images was performed using a surface-matching technique that fits the contours corresponding to the CSF-dura interface of the two MR images (31, 32). Finally, the macroscopic lesions segmented on proton density–weighted images were superimposed automatically onto the coregistered MTR map, and the areas corresponding to the segmented lesions were nulled. For comparability of results, the entire procedure, except the removal from the MTR maps of the tissue belonging to macroscopically visible T2 lesions, was performed also for the images of the control volunteers. To minimize the inclusion of pixels with partial volume averaging from the CSF, we excluded from the analysis all the pixels with MTR values lower than 10%. To correct for the between-patient differences in brain volume, each histogram was normalized by dividing it by the total number of pixels included. For each histogram, several parameters were analyzed: the height and the position of the histogram peak, with respect to the x axis, the average MTR value, and the MTR values corresponding to the 25th, 50th, and 75th percentiles of the histogram (ie, MTR₂₅, MTR₅₀, and MTR₇₅) that indicate the MTR value at which the integral of the histogram is 25%, 50%, and 75% of the total, respectively.

Statistical Analysis

Differences in MR and MT imaging metrics between patients and control volunteers and between patients who developed CDMS and those who did not were evaluated using a two-tailed Student's t test for nonpaired data. A multivariate analysis with a logistic model was conducted to evaluate the MR and MT imaging variables that independently influenced the probability of patients developing CDMS. The independent variables included in the model were the hyperintense T2 lesion load, the hypointense T1 lesion load, the average lesion MTR, and the average MTR of the NABT.

Results

Twenty-four patients (11 men and 13 women) were included in the study. The mean patient age was 27.2 years (SD, 5.6 years), the mean duration between clinical onset and MR acquisition was 16 days (range, 3–38 days), and the median Expanded Disability Status Scale score was 1.0 (range, score of 0.0–1.5). Ten patients had isolated syndrome of the brain, nine had unilateral acute optic neuritis, and five had isolated spinal cord syndrome. In all patients, CSF analysis was also conducted. CSF oligoclonal bands were found in 20 (83%) patients. At follow-up, 10 (41%) patients (four men and six women) developed CDMS. There were no significant differences in any of the baseline clinical characteristics among patients with different clinical presentations or between those who did and did not develop CDMS during the follow-up period. At follow-up, the median Expanded Disability Status Scale score was 0.5 for patients who did not develop CDMS and 1.5 for patients who did (P = not significant).

No abnormalities were found on the images of the control volunteers. In the entire cohort of patients with CIS, the mean hyperintense T2 lesion volume was 6.3 mL (range, 0.6–25.4 mL), the mean hypointense T1 lesion volume was 0.2 mL (range, 0.0–1.0 mL), and the average lesion MTR was 40.5% (SD, 2.0%). No significant differences in any of the MR or MT imaging–derived metrics studied were found among patients with different clinical presentations. In Table 1, the hyperintense T2 and hypointense T1 lesion volumes together with average lesion MTR for the patients who developed CDMS during the follow-up period and those who did not are reported. Patients who developed CDMS during the follow-up period had significantly higher T2 and T1 lesion volumes and significantly lower average lesion MTR than did those who did not develop CDMS. The mean brain volume was 386 mL (SD, 20 mL) for control volunteers, 384 mL (SD, 36 mL) for patients with CIS and no further clinical evolution, and 356 mL (SD, 20 mL) for patients with CIS who subsequently developed CDMS (these differences were not statistically significant).

In Table 2, NABT-MT metrics from control volunteers and the entire cohort of patients with CIS are reported. Patients with CIS had significantly lower average NABT-MTR, MTR₂₅, MTR₅₀, MTR₇₅, and peak position than did control volunteers. In Table 3, NABT-MT metrics from patients who developed CDMS during the follow-up period and from those who did not are reported. In Figure 1, the mean NABT-MTR histograms for the two groups of patients are presented. Patients developing CDMS had significantly lower average NABT-MTR, MTR₂₅, MTR₅₀, MTR₇₅, and peak position than did those who did not develop CDMS. The multivariate analysis showed that the T2 lesion volume (odd ratio, 3.54; P = .0005) and the average NABT-MTR (odd ratio, 0.81; P = .01) were independent predictors of the subsequent development of CDMS. All the other factors were excluded from the multiple regression model.

• Download figure
• Open in new tab
• Download powerpoint

fig 1.

MTR histograms of the NABT of patients with CIS at presentation who developed CDMS during the follow-up period (continuous line) and of patients with CIS at presentation who did not develop CDMS during the follow-up period (dotted line)