| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Neuroradiology |
1 From the McConnell Brain Imaging Center, Room WB-315, Montreal Neurological Institute, 3801 University St, Montreal, Québec, Canada H3A 2B4 (G.B.P., S.N., K.J.W., D.P., G.S.F., J.P.A., D.L.A.), and the Department of Neurology, University of Siena, Italy (N.D.S.). Received December 8, 1998; revision requested February 9, 1999; final revision received September 29; accepted October 6. Supported in part by the Medical Research Council of Canada and Fonds de la Recherche en Santé du Québec. G.B.P., G.S.F., and D.L.A. are Killam Scholars. Address correspondence to G.B.P. (e-mail: bruce@bic.mni.mcgill.ca).
 |
Abstract |
|---|
 TOP Abstract IntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
MATERIALS AND METHODS: In a 5-year longitudinal evaluation, 30 patients with MS underwent conventional magnetic resonance (MR) imaging, MT MR imaging, and clinical assessment. Cross-sectional data in 12 healthy subjects were also collected. Semiautomated lesion classification with use of T2-weighted MR images was used to measure the time course of the MT ratio (calculated with MR data acquired without and with MT saturation) in every voxel and to help analyze the relationship with the status of lesions depicted on T2-weighted images.
RESULTS: There was a significant (P < .001) temporal decline in lesion MT ratio after lesion appearance on T2-weighted images. A significant (P < .001) progressive decline in MT ratio was also present in voxels that later became lesions, prior to initial detection on T2-weighted images. Even 1
years prior to lesion appearance, the MT ratio (33.3%) in regions destined to become such lesions was significantly (P < .001) lower than that in both white matter in healthy subjects (41.3%) and other normal-appearing white matter in patients with MS (38.1%).
CONCLUSION: The MT ratio reveals progressive focal abnormalities in MS that antedate by up to 2 years the appearance of lesions on T2-weighted MR images.
Index terms: Brain, white matter • Magnetic resonance (MR), magnetization transfer, 13.121417 • Magnetic resonance (MR), tissue characterization, 13.121417 • Sclerosis, multiple, 13.871
|
Introduction |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Hydrogen nuclei associated with various semisolid (macromolecular) components such as the lipids of white matter (WM) have an extremely short T2 (<100 µsec) and are not directly detectable with MR imagers. However, interactions between semisolid and bulk water protons result in a continuous exchange of magnetization, which is referred to as cross-relaxation or MT (30,31). MT MR imaging can be used to detect this exchange by means of selective saturation of the semisolid magnetization pool and measurement of the resultant decrease in water signal intensity due to transfer of this saturation in regions undergoing exchange (26,32–35).
In brain, the largest MT effect is observed in WM. Despite uncertainties regarding the macromolecules primarily responsible for MT in WM (36–39), breakdown and loss of myelin will result in a diminished MT, and such reductions have been observed by several groups (23–25,40–42) in MS lesions and normal-appearing WM in patients with MS. The reduction in MT cannot be attributed solely to demyelination, because changes in water content and other membranes may also modify the exchange characteristics of tissue. However, using an animal model of acute experimental allergic encephalomyelitis, Dousset et al (23) observed that purely edematous lesions, which were easily visible on T2-weighted images, experienced only small decreases (5%–8%) in MT, whereas the plaques in patients with clinically definite MS showed a much larger reduction. Thus MT should still provide a measure that more closely reflects demyelination than do high-signal-intensity findings on conventional T2-weighted images.
Given the potential for profiling demyelination, MT MR imaging should constitute a valuable tool for help in assessing the burden of disease in MS and in the study of its natural history. We present here the results of a longitudinal study of a group of patients with MS. The purpose of our study was to determine the evolution of MT properties in lesions seen on T2-weighted images and in normal-appearing brain that later evolve into such lesions.
|
MATERIALS AND METHODS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
–2 years after the fifth study). MT data were acquired only during the last three time points (covering approximately 3 years). Healthy subjects were examined once. After excluding studies lost due to logistic or technical difficulties, a total of 187 conventional MR imaging and 80 MT data sets were collected. This study was approved by the institutional review board of McGill University (Montreal, Québec, Canada), and informed consent was obtained from all subjects.
Patients with MS underwent clinical examination by the same neurologist prior to each imaging session, and the level of their disability was graded according to the Expanded Disability Status Scale (EDSS) of Kurtzke (44). At the time of entry in the study, the relapsing-remitting subgroup consisted of five women and six men aged 24–43 years (mean ± SD, 31.1 years ± 6.3) with an EDSS score of 3.0–6.5 (mean ± SD, 4.8 ± 1.1). The primary-progressive subgroup consisted of one women and four men aged 27–54 years (mean, 43.2 years ± 9.8) with an EDSS score of 3.5–6.5 (mean, 5.6 ± 1.2). The secondary-progressive subgroup consisted of five women and nine men aged 30–56 years (mean, 47.9 years ± 7.5) with an EDSS score of 4.0–7.0 (mean, 5.9 ± 0.9). The healthy subjects were five women and seven men aged 31.0 years ± 7.1.
MR data were acquired using a 1.5-T imager (ACS III; Philips Medical Systems, Best, the Netherlands). Imaging was performed in oblique transverse planes parallel to the anterior commissure–posterior commissure line and were positioned on the superior margin of the corpus callosum. A multisection double-echo spin-echo sequence (repetition time msec/echo times msec = 2,100/30, 80; section thickness, 5.5 mm with 0.5-mm intersection gap; 20 sections) was used to acquire the intermediate- and T2-weighted images.
T1-weighted and MT MR images were obtained by using a pair of spin-echo acquisitions (1,000/20), without and with MT saturation pulses. Semisolid spin saturation was achieved by using 1.2-msec on-resonance 1
1 binomial pulses (radio-frequency field strength, 20 µT) placed just before each section-selective excitation (26). Twenty sections, coincident with the T2- and intermediate-weighted images, were acquired per repetition, thus giving an effective saturation repetition period of 50 msec. Percentage difference MT ratio (MTR) data were calculated, after thresholding above the background noise, by using the following formula: [(NoSat - Sat)/NoSat] x 100, where Sat and NoSat are image data obtained with MT saturation and without MT saturation, respectively.
To determine whether MTR values would be stable, we performed MR imaging with an agar gel phantom. These phantom studies were performed before and after acquisition of the longitudinal data. The protocol described in the preceding paragraph was used for the phantom studies.
With the aid of a semiautomated tissue segmentation and analysis software package (45), one experienced user (S.N.) classified all MS lesions and the ventricles, as seen on T2-weighted and intermediate-weighted images, for each time point. Automated software (46) was then used to spatially coregister all data to a single time point for each patient, to permit temporal tracking of voxels. The binary lesion and ventricle maps for all time points were collected and stored in a single (encoded) data volume for each patient. This permitted easy extraction of all voxels with a common feature (eg, all voxels identified as "lesion" at every time point). To allow automated differentiation between normal-appearing WM voxels at the periphery of lesions and those further removed from the lesions, the pathologic characteristics of which might differ, the binary lesion maps were enlarged by varying amounts by using a simple "N nearest-neighbor dilation" algorithm (47). The difference between the original and enlarged lesion maps provided a map of perilesional voxels.
To characterize the evolution of MTR values within lesions seen on T2-weighted images, each voxel identified as "lesion" was assigned an age. The time of lesion appearance was defined to be the midpoint between the time of the first study on which it was identified and that of the previous study. Those voxels classified as lesion on all studies were assigned an age corresponding to the time of the first study.
An example data set of T1-, T2-, and intermediate-weighted and MTR images for one time point is shown in Figure 1. Also shown are the lesion classification maps for this and a subsequent examination obtained 31 months later, registered to images obtained at the first time point.
|
Cross-sectional analysis of MTR values was performed by using the Student t test and a one-way analysis of variance. Group EDSS values were compared by using the Wilcoxon rank sum test. Cross-sectional relations between EDSS and MTR values, as well as between EDSS scores and lesion volume as measured on T2-weighted images, were evaluated by using the Spearman rank correlation analysis. The longitudinal data acquired in this study constituted repeated measures from the same group of patients, with some missing data. To account for repeated measures in our statistical analysis of temporal trends in MTRs and lesion volume while accommodating missing data, we first performed a linear regression for each patient individually. The regression parameters (slopes and intercepts) were then analyzed for statistical significance by using the Student t test and analysis of variance.
Despite the fact that observations in the same subject are correlated, it is straightforward to show that the least squares slope is unbiased, even in the presence of correlation, because the correlation affects the variance of the estimate, not its expectation. When combining slopes from separate individuals to estimate a common slope, the sample mean is consistent and unbiased. A consistent unbiased estimator of variance is the usual sample variance of the slopes, even if the slopes have unequal variances, because the slopes are independent. Thus, the standard t test will be valid asymptotically, by reason of the central limit theorem, even if the slopes are not normally distributed. This method is, in fact, a simplified implementation of the method of Liang and Zeger (48).
|
RESULTS |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Because gray matter MTR values were both reproducible and unchanged between healthy subjects and patients with MS, we used these values to monitor MR imager performance throughout this longitudinal study. Mean MTR values measured in regions of interest in the head of the caudate nucleus (left and right sides averaged) were plotted as a function of time (relative to the time of the first MTR image) for all patients with MS (Fig 2). No significant variation in MTR with time was observed.
|
The mean MTRs for each lesion age range, extracted from all MT data sets, were plotted against lesion age (Fig 3). For the entire MS cohort, there was a significant decline in lesion MTR with lesion age (P < .001). The mean change (slope) in lesion MTR was -1.7%/y, and the mean intercept, which indicates the mean MTR for new lesions, was 30.1%. The lesion MTR decline was significant for each patient subgroup (for relapsing-remitting, P < .001; for primary-progressive, P = .03; for secondary-progressive, P < .001), although the differences in slopes and intercepts were not significant.
|
years before detection on T2-weighted MR images to approximately 1 year after detection. As in the lesion age analysis described previously, newly identified lesions were assumed to have appeared midway between MR studies. Thus, for example, the cluster of points around 1 year in Figure 4 represent lesions first identified on the 2-year follow-up study (sixth time point). Focal MTR was observed to decline significantly (P < .001) with time prior to lesion appearance on T2-weighted images. The mean slope was -1.9%/y, and the mean intercept was 31.4%. A significant decline in MTR was also present in the relapsing-remitting and secondary-progressive subgroups, but this decrease was not significant in the primary-progressive subgroup. There were no significant differences in slopes and intercepts between patient subgroups. Approximately 1–1 years before lesions were detected on T2-weighted images (ie, the two leftmost clusters of data points in Fig 4), the mean MTR was 33.3% in normal-appearing WM regions that eventually became lesions. This was significantly (P < .001) lower than both the mean WM MTR in healthy subjects (41.3%) and in the hand-drawn normal-appearing WM regions of interest in the patients with MS (38.1%).
|
|
|
|
DISCUSSION |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Our observed lack of correlation between cross-sectional lesion volume on T2-weighted images and EDSS score is not inconsistent with the results of other investigators (2,6,24,51), who have found, at best, a weak correlation. This is likely due to four primary factors: the pathologic heterogeneity of lesions that are demonstrated on T2-weighted images and that are all classified as equivalent; the presence of such lesions in relatively "silent" brain regions and that do not have a marked effect on EDSS score; spinal cord disease, which was not measured in this study; and the nonlinear and complex nature of the EDSS assessment itself (7). The highly nonuniform distribution of EDSS scores (mean, 5.4 ± 1.1) for this group of patients also complicated the detection of a correlation with EDSS score. The small range of EDSS scores might also explain our failure to detect the correlation between lesion MTR and EDSS score that has been reported by others (24). However, our observations of significant increases in lesion volume and decreases in mean lesion MTR with increases in EDSS score indicate that changes in these measures are correlated with accumulation of clinical disability.
MTR versus Lesion Age
The -1.7%/y change in lesion MTR with lesion age likely reflected progression of demyelination in MS lesions associated with chronic activity. The 30.1% intercept of the regression lines for MTR versus lesion age reflected the mean MTR of new lesions. Thus, when a new lesion, which persisted on subsequent MR studies, first appeared, its MTR was reduced by approximately one-fourth of that for normal WM. Two possible (non–mutually exclusive) explanations for the substantial reduction in MTR at initial lesion appearance are that (a) an increase in the free water pool, associated with early inflammation, decreased the relative amount of MT or (b) there was substantial demyelination or other macromolecular abnormality prior to detection of typical lesions on T2-weighted images. Although both factors likely are contributors, support for the predominance of the latter can be provided by examining MTR values in normal-appearing WM regions that later evolve into lesions visible on T2-weighted images.
Prelesional MTR Abnormalities
Perhaps the most intriguing results of this study are related to the prelesional analysis. We observed substantial focal reductions in WM MTR as long as 2 years before the detection of lesions in these regions on T2-weighted images. This reduction was significant in comparison with the MTRs in both WM in healthy subjects and normal-appearing WM in patients with MS, which was identified in the absence of any knowledge of future lesion locations (ie, in hand-drawn normal-appearing WM regions of interest). Furthermore, there was a significant decline in MTR over time prior to lesion detection (Fig 4). These data indicate that, in cases of MS, the MTR helps reveal progressive focal disease that antedates by up to 2 years the appearance of a lesion on T2-weighted images.
Given the importance of this finding and its potential effect on our understanding of the natural history of MS, we examined our data closely to ensure that our findings were not artifactual. Although studies with an agar gel phantom, which we performed before and after the longitudinal study, showed stable MTR values, calibration phantom data were not acquired throughout the study. However, since MTR values were equivalent in the gray matter of both healthy subjects and patients with MS, we monitored the MTR in the head of the caudate nuclei on all studies in all patients with MS. These data (Fig 2) showed no statistically significant change over time and thus provided strong inferential support to eliminate machine drift as a possible explanation for our results.
Another possible explanation for our observation of focally abnormal and declining MTR in prelesional normal-appearing WM was that we were simply measuring MTR values in regions immediately adjacent to identified lesions. These regions might represent small differences in lesion boundary identification or slight errors in serial image registration. To check for this possibility, we repeated the analysis with all classified lesion boundaries, except that of the last time point, dilated radially by approximately 2 mm. This resulted in a substantial increase in lesion volume at each time point (eg, lesion volume was increased by a mean of 85.7% for the fourth time point). Furthermore, because classification of a voxel as a lesion on any previous study prevented such a voxel from being categorized as a "new" lesion, random misregistration errors tended to increase the total prior lesion volume and make the classification as "new" even more conservative. For the 2-mm dilation used in this analysis, the mean increase in lesion volume defined on all previous studies was 120%. This was substantially greater than the error in lesion volume reproducibility (<5%) achieved by using our semiautomated lesion-classification technique (52). Thus, subdivision of the prelesional analysis factors into voxels that fall inside the expanded prior-lesion maps (40.4%) and those that fall outside (59.6%) provided a robust method for differentiation between perilesional voxels and remote voxels. Results of this analysis verified that our observation of significantly decreased MTR prior to the appearance of a lesion on T2-weighted images was not artifactual.
In addition to our initial report (53) of focal prelesional MTR abnormalities, Filippi et al (29) and Goodkin et al (54) have made similar observations but over shorter time periods. The results of our present study, which spanned several years, suggest the presence of important, undetected, focal abnormalities long before the appearance of lesions on T2-weighted MR images. One possibility that cannot be completely eliminated is that lesions appeared on MR images and completely resolved in these regions between MR studies or before enrollment in this investigation. Our experience indicates, however, that complete development and resolution of lesions is rarely seen within 6 months. In addition, we attempted to minimize this contribution by excluding from the lesion classification all voxels that "wax and wane." The presence of a correlation in the secondary-progressive subgroup, where new lesions were less frequent, also suggested that resolved lesions were not an important factor. An alternative, and more probable, explanation of our results was that marked gradual demyelination and/or other tissue changes occurred prior to lesion demonstration on conventional T2-weighted MR images.
Our data suggest that abnormalities in normal-appearing WM occur both focally, as new independent lesions form, and more diffusely, as wallerian degeneration occurs secondarily to active lesions located elsewhere. Using a feline model, Lexa et al (50) showed that MTR values decrease 2–4 weeks after injury in WM areas undergoing wallerian degeneration. Thus, axonal injury in existent lesions visible on T2-weighted images may be resulting in connected normal-appearing WM undergoing undetected myelin breakdown associated with wallerian degeneration. In either case, a global evaluation of MTR in WM such as that suggested by van Buchem et al (27,55) should allow a more complete assessment of disease burden than one based solely on MTR in lesions depicted on T2-weighted images. Furthermore, because focal MTR abnormalities also predate detection of lesions on T2-weighted images, the MTR may provide a predictive index of lesion development and disease progression in patients with MS.
In the entire MS group, the slopes of the MTR lines before and after lesion appearance on T2-weighted images were remarkably consistent. The difference between the intercepts of these lines before and after lesion appearance also was small (1.3%), which indicated that a large sustained discontinuity in MTR does not, in general, occur at lesion appearance; rather, the decline in MTR simply continues. In fact, results of more frequent imaging indicate that acute lesions with severe edema show a transient decrease in MTR followed by partial recovery (56), as well as a persistent gradual decline due to continued demyelination.
The precise pathologic interpretation of MTR abnormalities remains difficult. Although further studies with animal models and histopathologic correlation may provide valuable information, additional developments in MT methods are also needed. Current percentage difference MTR techniques provide a semiquantitative measure that reflects a complex combination of sequence and relaxation parameters in addition to cross-relaxation terms (26,35). This has made the direct comparison of MTR data acquired at multiple centers with different protocols difficult. The more fundamental parameters of interest include the relative sizes of water and semisolid component pools, individual relaxation times, and the cross-relaxation rate. Methods for complete model characterization have been developed for in vitro studies (57) but are not directly transferable to in vivo human imaging. However, the results presented in this report provide impetus for such developments.
In conclusion, we performed a 5-year longitudinal study with 30 patients with MS in which we used a combination of conventional MR imaging, MT MR imaging, and clinical evaluation. Using basic image processing techniques, we were able to track the evolution of MTR in voxels both before and after their appearance as lesions on T2- and intermediate-weighted images. We observed a continuous gradual decline in MTR within lesions and measured important MTR abnormalities in regions that would become lesions long before they were identifiable as such on T2-weighted MR images. This result indicates that there is important disease activity in MS that goes undetected at conventional MR imaging. Extrapolation back in time suggests that focal MT abnormalities were present in regions destined to become active plaques years before such lesions could be detected on conventional MR images. The early detection capability facilitated by the MTR method used in this study may also provide a predictive index of disease progression in patients with MS.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Author contributions: Guarantors of integrity of entire study, G.B.P., D.L.A.; study concepts and design, G.B.P., D.L.A.; definition of intellectual content, G.B.P., D.L.A.; literature research, G.B.P., N.D.S.; clinical studies, D.L.A., N.D.S., G.S.F., J.P.A.; experimental studies, G.B.P., S.N., N.D.S.; data acquisition, G.B.P., S.N., N.D.S.; data analysis, G.B.P., S.N., D.P., K.J.W.; statistical analysis, G.B.P., K.J.W.; manuscript preparation, G.B.P.; manuscript editing, G.B.P., N.D.S., D.L.A.; manuscript review, all authors.
|
References |
|---|
|
TOPAbstractIntroductionMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  I. Kezele, D. Arnold, and D. Collins Atrophy in white matter fiber tracts in multiple sclerosis is not dependent on tract length or local white matter lesions Multiple Sclerosis, July 1, 2008; 14(6): 779 - 785. [Abstract] [PDF]
|
|
|
|
 |
|
 C. Di Perri, M. Battaglini, M. L. Stromillo, M. L. Bartolozzi, L. Guidi, A. Federico, and N. De Stefano Voxel-Based Assessment of Differences in Damage and Distribution of White Matter Lesions Between Patients With Primary Progressive and Relapsing-Remitting Multiple Sclerosis Arch Neurol, February 1, 2008; 65(2): 236 - 243. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Arnold Evidence for neuroprotection and remyelination using imaging techniques Neurology, May 29, 2007; 68(22_suppl_3): S83 - S90. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. De Stefano, M. Battaglini, M. L. Stromillo, V. Zipoli, M. L. Bartolozzi, L. Guidi, G. Siracusa, E. Portaccio, A. Giorgio, S. Sorbi, et al. Brain damage as detected by magnetization transfer imaging is less pronounced in benign than in early relapsing multiple sclerosis Brain, August 1, 2006; 129(8): 2008 - 2016. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ge Multiple Sclerosis: The Role of MR Imaging AJNR Am. J. Neuroradiol., June 1, 2006; 27(6): 1165 - 1176. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. T. M. Fernando, M. A. McLean, D. T. Chard, D. G. MacManus, C. M. Dalton, K. A. Miszkiel, R. M. Gordon, G. T. Plant, A. J. Thompson, and D. H. Miller Elevated white matter myo-inositol in clinically isolated syndromes suggestive of multiple sclerosis Brain, June 1, 2004; 127(6): 1361 - 1369. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Ge, R. I. Grossman, J. S. Babb, J. He, and L. J. Mannon Dirty-Appearing White Matter in Multiple Sclerosis: Volumetric MR Imaging and Magnetization Transfer Ratio Histogram Analysis AJNR Am. J. Neuroradiol., November 1, 2003; 24(10): 1935 - 1940. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F Fazekas, S Ropele, C Enzinger, T Seifert, and S Strasser-Fuchs Quantitative magnetization transfer imaging of pre-lesional white-matter changes in multiple sclerosis Multiple Sclerosis, December 1, 2002; 8(6): 479 - 484. [Abstract] [PDF]
|
|
|
|
|
|
N. De Stefano, S. Narayanan, S. J. Francis, S. Smith, M. Mortilla, M. C. Tartaglia, M. L. Bartolozzi, L. Guidi, A. Federico, and D. L. Arnold Diffuse Axonal and Tissue Injury in Patients With Multiple Sclerosis With Low Cerebral Lesion Load and No Disability Arch Neurol, October 1, 2002; 59(10): 1565 - 1571. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. L. Arnold and P.M. Matthews MRI in the diagnosis and management of multiple sclerosis Neurology, April 23, 2002; 58(90084): S23 - 31. [Abstract] [Full Text]
|
|
|
|
|
|
A. C. Guo, V. L. Jewells, and J. M. Provenzale Analysis of Normal-Appearing White Matter in Multiple Sclerosis: Comparison of Diffusion Tensor MR Imaging and Magnetization Transfer Imaging AJNR Am. J. Neuroradiol., November 1, 2001; 22(10): 1893 - 1900. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Honkaniemi, P. Dastidar, V. Kahara, and H. Haapasalo Delayed MR Imaging Changes in Acute Disseminated Encephalomyelitis AJNR Am. J. Neuroradiol., June 1, 2001; 22(6): 1117 - 1124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Iannucci, M. Dichgans, M. Rovaris, R. Bruning, T. Gasser, L. Giacomotti, T.A. Yousry, and M. Filippi Correlations Between Clinical Findings and Magnetization Transfer Imaging Metrics of Tissue Damage in Individuals With Cerebral Autosomal Dominant Arteriopathy With Subcortical Infarcts and Leukoencephalopathy Stroke, March 1, 2001; 32(3): 643 - 648. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N D Richert, J L Ostuni, C N Bash, T P Leist, H F McFarland, and J A Frank Interferon beta-1b and intravenous methylprednisolone promote lesion recovery in multiple sclerosis Multiple Sclerosis, February 1, 2001; 7(1): 49 - 58. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||
= relapsing-remitting subgroup, 
= secondary-progressive subgroup.
