Magnetization transfer can predict clinical evolution in patients with multiple sclerosis

Transcription

1 J Neurol (2002) 249 : /s ORIGINAL COMMUNICATION A. Carlos Santos Sridar Narayanan Nicola de Stefano M. Carmela Tartaglia Simon J. Francis Rozie Arnaoutelis Zografos Caramanos Jack P. Antel G. Bruce Pike Douglas L. Arnold Magnetization transfer can predict clinical evolution in patients with multiple sclerosis Abstract The clinical course of multiple sclerosis (MS) is highly variable ranging from benign to aggressive, and is difficult to predict. Since magnetization transfer (MT) imaging can detect focal abnormalities in normal-appearing Received: 3 May 2001 Received in revised form: 11 October 2001 Accepted: 22 October 2001 A. C. Santos, MD S. Narayanan, MSc N. de Stefano, MD M. C. Tartaglia, BSc S. J. Francis, BSc R. Arnaoutelis, BSc Z. Caramanos, MA J. P. Antel, MD G. B. Pike, PhD Prof. D. L. Arnold Montreal Neurological Institute 3801 University Street, WB 321 Montreal, Quebec, H3A 2B4, Canada Tel.: / Fax: / doug@mrs.mni.mcgill.ca A. C. Santos, MD Department of Clinical Medicine Imaging Center Ribeirao Preto Medical School University of Sao Paulo, Brazil N. de Stefano, MD Institute of Neurological Sciences Neurometabolic Unit University of Siena, Italy white matter (NAWM) before the appearance of lesions on conventional MRI, we hypothesized that changes in MT might be able to predict the clinical evolution of MS. We assessed MR data from MS patients who were subsequently followed clinically for 5 years. We computed the mean MT ratio (MTr) in gray matter, in lesions identified on T2-weighted MRI, and in NAWM, as well as in a thick central brain slice for each patient. Patients were divided into stable and worsening groups according to their change in Expanded Disability Status Scale (EDSS) scores over 5 years. We calculated the sensitivity, specificity, predictive value, and odds ratio of the baseline MTr measures in order to assess their prognostic utility. We found significant differences in baseline MTr values in NAWM (p = 0.005) and brain slice (p = 0.03) between clinically stable and worsening MS patients. When these MTr values were compared with changes in EDSS over 5 years, a strong correlation was found between the EDSS changes and MTr values in both NAWM (SRCC = 0.76, p < 0.001) and in the brain slice (SRCC = 0.59, p = 0.01). Baseline NAWM MTr correctly predicted clinical evolution in 15/18 patients (1 false positive and 2 false negatives), yielding a positive predictive value of %, a negative predictive value of %, and an odds ratio of 28. The relationship between 5-year changes in EDSS and MTr values in T2 weighted MRI lesions was weaker (SRCC = 0.43, p = 0.07). Our data support the notion that the quantification of MTr in the NAWM can predict the clinical evolution of MS. Lower MTr values predict poorer long-term clinical outcome. Abnormalities of MTr values in the NAWM are more relevant to the development of future patient disability than those in the T2-weighted MRI lesions. Key words Multiple sclerosis magnetic resonance imaging magnetization transfer magnetic resonance spectroscopy prognosis. JON 686 Introduction Multiple sclerosis (MS) is a chronic neurological disorder with highly variable progression and severity. A number of recent studies [23, 28, 29, 38 42] have attempted to predict progression by means of statistical models, but the high variability between individuals makes it extremely difficult to predict a patient s outcome. Conventional magnetic resonance (MR) imaging (MRI) has been shown to be useful in predicting the conversion from a clinically isolated syndrome sugges-

2 663 tive of MS to definite MS [10, 21, 30]. However, in patients with established MS, there has been only weak correlation demonstrated between MRI-defined lesion load and subsequent clinical status [11, 32, 37]. This may be due to the lower pathological specificity of conventional MRI markers that are not able to differentiate multiple pathological changes affecting water relaxation (inflammation or demyelination) coexisting in MS lesions. Furthermore, there is pathologic evidence [1, 2] that there can be focal microscopic changes and diffuse molecular pathology of myelin outside the macroscopic lesions of MS visible on MRI. In the last few years, new quantitative MR techniques have been shown to have the potential to provide more pathologically specific information about both macroscopic and microscopic disease affecting axons and myelin [3]. Magnetization transfer (MT) imaging appears to be among the most promising. Based on the interactions between protons in bulk water and those attached to macromolecules [8, 43], brain MT is particularly sensitive to pathological changes in myelin [7] and axons [36] and can therefore provide a reliable measure of the structural changes occurring in brain tissues. MT is usually quantified by the MT ratio (MTr), which is calculated voxel-by-voxel from the difference in signal intensity before and after the application of a radio frequency pulse that saturates the macromolecular magnetization. Decreases of MTr in MS lesions [6, 14, 18, 22, 31] and in white matter (WM) that appears normal on conventional MRI [7, 11, 14, 19, 25, 34, 35] have been reported in several studies. In addition, it has been demonstrated that focal MTr decrease can precede enhancing [9] or T2-weighted lesions [13, 26] by months to years. These latter findings imply that MTr changes in normal-appearing WM (NAWM) are relevant to disease progression in MS. Thus, in the present study we evaluated the potential of MTr to predict future clinical disability, retrospectively assessing the MTr values of patients with clinically definite MS who were either stable or significantly more disabled 5 years later. Subjects and methods Subjects We selected from our historical MRI study cohorts those patients with definite MS who were followed clinically at the Multiple Sclerosis Clinic of the Montreal Neurological Hospital for at least 5 years after they had undergone MR examination that included an MT acquisition. Eighteen patients (seven women and eleven men) met these criteria. Eight patients were classified as relapsing-remitting (RR) and ten as secondary progressive (SP) [27] at the time of the MR examination. The mean age of the whole group was ± years (range 31 63) and the mean disease duration was ± 8.9 years (range 9 39). Patients clinical status was quantified using the Kurtzke ex- panded disability status scale (EDSS) [17]. According to their clinical evolution after the MRI examination, MS patients were divided into two groups: i) one that showed no or minimal changes in disability (increase of EDSS 0.5 point from baseline) during the 5 years of follow-up (stable group,n = 9) and ii) one that showed an increase in disability (increase of EDSS 1 from baseline) over 5 years (worsening group, n = 9). Twelve adult healthy subjects were studied using the same MT protocol and were used for comparison. The MNI Ethics Review Committee had approved the study and informed consent was obtained from all subjects. MRI method MRI examinations were performed on a Philips Gyroscan 1.5 T system. The MRI protocol included a sagittal spin-echo (SE) T1- weighted sequence; an axial double-spin-echo proton density/t2- weighted (PD/T2) sequence parallel to the anterior-posterior commissural line; and a pair of T1-weighted SE sequences, without (nosat) and with (Sat) an MT pulse. The PD/T2 parameters were a TR of 2010 ms and a TE1/TE2 of 30/80 ms, field of view (FOV) of 250 mm, and 20 slices with 5.0-mm thickness, a gap of 0.5 mm, and a 256x256 matrix. The MT sequences had the same parameters except for a TR of 940 and a single TE of 20 ms. The MT pulse was a 1.2-millisecond on-resonance, 121 binomial pulse (radio-frequency field strength = 20 µt) placed just before each slice-selective excitation [24]. Twenty slices with and without the MT pre-pulse were acquired in the same position. The stability of this sequence has been previously described [25]. After thresholding above the background noise, a percent difference image, hereafter called the MTr image, was calculated on a voxelby-voxel basis according to the equation: MTr = (1 Sat /nosat) x 100 [7] (Fig. 1A). Mean MTr values were obtained within rectangular regions of interest (ROI) manually defined by a trained neuroradiologist (A. C. S). Values of MTr in NAWM were calculated by taking consistent samples from five regions (corona radiata and centrum semiovale, frontal lobe, genu of the corpus callosum, splenium of the corpus callosum and occipital lobe). The mean for each region was calculated from at least 10 rectangular samples. The MTr for NAWM was then obtained by averaging the mean MTr from each region.values of MTr for normal white matter in the control group were obtained in the same way. The mean MTr in gray matter (GM) of cortex, putamen and caudate nuclei was measured in controls and patients with the same ROI analysis. Mean lesion MTr and lesion volume was obtained using masks generated by semi-automated tissue segmentation [16] performed by one experienced observer (S. N), as previously described [20]. In addition to MTr values in lesions and NAWM, a brain MTr was calculated from a thick slab representative of the brain including six consecutive slices above the bi-commissural line [34] (Fig. 1B). In that slab the brain parenchyma was separated from ventricles and extra cerebral tissue by semi-automated segmentation using locally developed software (Display, David MacDonald, McConnell Brain Imaging Centre, McGill University). The mean MTr of the thick slab, hereafter called brain MTr, was calculated according to the equation given above. Central brain N-acetylaspartate (NAA) and creatine (Cr) resonance intensities were also measured as previously described [5]. NAA was measured relative to the intravoxel creatine and then averaged over all voxels in the spectroscopic volume of interest. Cr can be used as internal standard here due to its relative stability in brain tissue of chronic MS patients. Statistics Baseline data from both groups were compared using Student s twosample separate-variance t-tests, with Bonferroni correction to adjust for multiple comparisons. Kruskal-Wallis one way analysis variance

3 664 Fig. 1 Shown in the top row (A) is an example of the calculation of a single slice of an MTR image. On the left is a T1-weighted, spin-echo image with no MT pulse; the centre image is from an identical T1- weighted acquisition with an MT saturation pulse; the rightmost image is the percent difference image computed from the previous two. Shown in the bottom row (B) are coronal (left), sagittal (center) and transverse (right) views of the 6 consecutive slices above the bi-commissural line (blue) used to compute the brain slab MTR. was used to compare the two groups EDSS at baseline and at followup because these values are only ordinal in nature. Fisher exact test (two-tailed) was used to assess group differences in gender and MS subtype. The correlation between MR metrics at baseline and the change in EDSS after 5 years was assessed using Spearman rank order correlation (SRCC). Data were considered significant at p < To test the prognostic utility of MTr measurements, we divided the patient cohort into two groups based on their initial, baseline NAWM MTr using a value of two standard deviations below the mean WM MTr in normal controls as the threshold between groups. We then calculated the sensitivity, specificity, predictive value, and odds ratio to evaluate the clinical usefulness of the MTr measurements at predicting clinical evolution. Results The clinically stable and worsening MS groups were comparable at baseline as there were no significant differences between them in age, disease duration, gender, type of MS, or EDSS (see Table 1), although there was a trend toward lower age and disease duration in the worsening group. In the stable MS group the EDSS was 5.94 ± 1.59 (mean ± standard deviation) at baseline and 5.83 ± 1.94 five years later, whereas in the clinically worsening MS group EDSS was 5.67 ± 1.41 at baseline and 7.33 ± 1.22 five years later. In our group of MS patients, there was no significant difference between the worsening and stable groups in T2-weighted lesion load (40.78±22.86 versus 26.29±13.68, p=1.00) and NAA/Cr (2.67±0.23 versus 2.88±0.25,p=0.90).As previously reported in other studies [32, 35], MTr values in T2-weighted MRI lesions (27.59±3.03) and NAWM (38.45 ± 1.30) in the whole group of MS patients were significantly lower than the Table 1 Patient demographic and MR data Stable group Worsening group p value Mean ± SD Mean ± SD Gender 3 F and 6 M 4 F and 5 M 1.00 Age 52.78± ± Disease duration 25.56± ± Type of MS 3 RR and 6 SP 5 RR and 4 SP 0.63 T2-W lesion volume 26.29± ± brain MTr 32.22± ± lesion MTr 29.19± ± NAWM MTr 39.42± ± NAA/Cr 2.88± ± baseline EDSS 5.94± ± follow-up EDSS 5.83± ± Delta EDSS 0.11± ±0.75 < Data considered significant at p < The t test was used for the analysis of all variables, except for EDSS (Kruskal-Wallis one-way analysis of variance), gender, and type of MS (Fisher exact test). MTr values in WM of normal controls (41.32 ± 1.4, p < 0.001). The gray matter (GM) MTr values of MS patients were not different from those of the normal controls (28.1 ± 2.90 versus 28.5 ± 1.7, respectively, p = 0.40). The difference in MTr values within T2-weighted lesions in the worsening group of MS patients compared to the stable group did not reach significance (p = 0.25). In contrast, there were highly significant differences in NAWM MTr between the stable (39.4±0.6) and worsening patient groups (37.4±1.1, p = 0.005) (Fig. 2A). The brain MTr values (from the brain slab which included GM, T2-w lesions, and NAWM) also were significantly different between the patient groups (30.6 ± 1.1 versus

4 ±0.8, respectively, p = 0.03), but to a lesser degree than the NAWM values. To assess whether our brain MRI markers could predict future clinical disability, values of MTr, total T2-W lesion volume, and NAA/Cr of all the MS patients were related to changes in clinical disability 5 years later ( EDSS). We found a very close relationship between EDSS and both NAWM MTr (SRCC = 0.76, p < 0.001, see Fig. 2B) and the brain MTr values (SRCC = 0.59, p = 0.01). In contrast, we just found a trend between EDSS and MTr values in T2-weighted MRI lesions (SRCC = 0.43, p = 0.07), and we did not find a significant correlation between EDSS and T2-weighted MRI lesion volume (SRCC = 0.41, p = 0.10). The correlation between EDSS and NAA/Cr in central brain was also significant (SRCC = 0.55, p=0.02). We assessed the potential for NAWM MTr to prospectively predict a worsening clinical course for an individual patient by segregating the cohort based on baseline NAWM MTr.We made the assumption that patients with NAWM MTr values more than two standard deviations (SD) below the mean WM MTr of controls would have an elevated risk of progression on EDSS at 5 years,while patients with NAWM MTr within 2 SD of the control WM value would be stable over the following 5 years. Using this criterion, 10 patients were deemed to be at risk for progression while 8 were predicted to have a stable course (Table 2). When compared with their actual clinical courses, this test yielded one false positive (predicted progression vs. an observed stable course) and 2 false negative result (predicted stable course vs. observed progression). Thus, the positive predictive value of the test was % and negative predictive value was %. The sensitivity and specificity were computed to be % and %, respectively, and the Odds- Ratio was 28 (Table 2). Table 2 Utilization of NAWM MTr for predicting disability progression of MS Increase in EDSS 5 years later Yes No total MTr NAWM < MTr NAWM > Total Positive predictive value = 77.78%. Negative predictive value = 88.89%. Sensitivity = 87.5%. Specificity = 80%. Odds ratio = 28 (95% CI: , 245). Fig. 2 Statistical analysis of the MTr results in normal-appearing white matter (NAWM). A: box-and-whiskers plots of the NAWM MTr difference between stable and worsening groups. The center horizontal line marks the median of the sample, the upper and lower edges of the box (the hinges) mark the 25 th and 75 th percentiles (i. e. the central 50 % of the values fall within the box). The empty circles represent individual subjects and the full circles the group mean. B: Spearman rank order correlation between mean MTr values in NAWM obtained at baseline and EDSS change 5 years later.

5 666 Discussion Recent MS studies showing that focal decreases in NAWM MTr can precede lesion detection by conventional MRI [9, 13, 26], suggested that appropriate MTr measures could predict disease progression [26].The results reported in this paper confirm the potential of MTr imaging to provide a predictive index of disease progression in patients with MS. The predictive value of MTr in NAWM may be stronger than that of T2 lesion or central brain NAA because the MTr abnormalities might detect an anticipation of focal inflammation plus diffuse axonal pathology in progressive as opposed to stable disease. In contrast, lesion MTr was not significantly different between the worsening and stable MS patients and did not show a significant correlation with the 5-year changes in clinical scores. This suggests that diffuse microscopic white matter pathology may be more relevant to the accrual of subsequent disability than the pathology within focal lesions, especially later in the disease. Disability may be related more to diffuse or distributed damage in the NAWM than with focal inflammatory events. The latter suggestion is supported by the recent observations of Confavreux et al. [4] that, once a level of irreversible disability has been reached (EDSS = 4), the further progression of disability is independent of relapses. The cerebral NAWM changes in MS include diffuse gliosis and diffuse inflammatory activity [1, 2], small foci of demyelination [1] and axon alterations [33, 36]. The presence in MS of widespread brain pathology heavily involving the NAWM is in agreement with the strong correlation found in several studies between clinical disability and MR markers of axonal damage [12] or tissue destruction [15, 51 53] in NAWM. Thus, diffuse or distributed demyelination or inflammation may be the substrate for subsequent axonal damage, which is most likely the ultimate cause of irreversible clinical disability. The strong correlation between reduction in MTr and axonal density in the NAWM of post-mortem brain [36] supports this hypothesis and is in agreement with our present data. The MTr values within lesions can fluctuate over time depending on lesion activity. On average, MTr drops substantially when a lesion starts to enhance, and then recovers to a varying extent over a period of months [6, 9,13,22,26,31].Thus,at a given point in time,lesion MTr is strongly influenced by recent inflammatory activity. It should not be surprising that the mean lesion MTr that is, in principle, independent of lesion volume and location, does not have a strong relationship to late permanent clinical disability. In addition to MTr in NAWM, we computed the mean MTr from a central slab of tissue that is representative of the whole brain. Following the work of van Buchem [34], we measured MTr in a brain slab that included six consecutive 5mm slices (including lesions, NAWM and gray matter) located in and above the bi-commissural line, and automatically calculated MTr values after separating the brain parenchyma from the extra cerebral tissue and ventricles. Results of this analysis showed a correlation between the 5-year changes in EDSS and the brain MTr values at baseline that was nearly, but not quite, as strong as the correlation between the 5-year EDSS change and NAWM MTr. While proper segmentation and calculation of mean MTr in multiple tissue classes provides more specific information, brain MTr measurement may provide a useful, simple and more easily automated index of disease burden than the NAWM MTr. In conclusion, our data support the notion that the focal lesions of MS are only one aspect of the disease, and that understanding the pathological changes occurring in the NAWM may be crucial to understanding the complex mechanisms responsible for disease progression. The quantification of MTr in NAWM appears to have potential as a prognostic indicator of how patients will progress clinically over the next five years. Given the potential implications of this preliminary observation for the management of patients with MS, we believe that larger, prospective studies of the predictive value of NAWM MTr are warranted. Acknowledgements The study was supported by grants from the Multiple Sclerosis Society of Canada and the Medical Research Council of Canada, and a pilot grant from the National Multiple Sclerosis Society.ACS was supported by a grant from the FAPESP Brazil (Proc No 99/ ). References 1. Allen IV, McKeown SR (1979) A histological, histochemical and biochemical study of the macroscopically normal white matter in multiple sclerosis. J Neurol Sci 41: Allen IV, Glover G, Anderson R (1981) Abnormalities in the macroscopically normal white matter in cases of mild or spinal multiple sclerosis (MS). Acta Neuropathol Suppl (Berl) 7: Arnold DLN, Narayanan S, De Stefano N, Reddy H, Matthews P (2001) Magnetic Resonance Spectroscopy of multiple sclerosis: imaging axonal damage. In: Kappos LJ, Kesselring J, Radu EW (ed) Multiple sclerosis: tissue destruction and repair. 1st edn. London: Martin Dunitz, pp Confavreux C, Vukusic S, Moreau T, Adeleine P (2000) Relapses and Progression of Disability in Multiple Sclerosis. N Engl J Med 343: De Stefano N, Narayanan S, Francis GS, Arnaoutelis R, Tartaglia MC, Antel JP, et al. (2001) Evidence of axonal damage in the early stages of multiple sclerosis and its relevance to disability. Arch Neurol 58:65 70

6 Dousset V, Gayou A, Brochet B, Caille JM (1998) Early structural changes in acute MS lesions assessed by serial magnetization transfer studies. Neurology 51: Dousset V, Grossman RI, Ramer KN, Schnall MD, Young LH, Gonzalez- Scarano F, et al. (1992) Experimental allergic encephalomyelitis and multiple sclerosis: lesion characterization with magnetization transfer imaging [published erratum appears in Radiology 1992 Jun, 183 (3):878]. Radiology 182: Edzes HT, Samulski ET (1977) Cross relaxation and spin diffusion in the proton NMR or hydrated collagen. Nature 265: Filippi M, Rocca MA, Martino G, Horsfield MA, Comi G (1998) Magnetization transfer changes in the normal appearing white matter precede the appearance of enhancing lesions in patients with multiple sclerosis. Ann Neurol 43: Filippi M, Horsfield MA, Morrissey SP, MacManus DG, Rudge P, McDonald WI, et al. (1994) Quantitative brain MRI lesion load predicts the course of clinically isolated syndromes suggestive of multiple sclerosis. Neurology 44: Filippi M, Paty DW, Kappos L, Barkhof F, Compston DA, Thompson AJ, et al. (1995) Correlations between changes in disability and T2-weighted brain MRI activity in multiple sclerosis: a follow-up study. Neurology 45: Fu L, Matthews PM, De Stefano N, Worsley KJ, Narayanan S, Francis GS, et al. (1998) Imaging axonal damage of normal-appearing white matter in multiple sclerosis. Brain 121: Goodkin DE, Rooney WD, Sloan R, Bacchetti P, Gee L, Vermathen M, et al. (1998) A serial study of new MS lesions and the white matter from which they arise. Neurology 51: Hiehle JF Jr., Lenkinski RE, Grossman RI, Dousset V, Ramer KN, Schnall MD, et al. (1994) Correlation of spectroscopy and magnetization transfer imaging in the evaluation of demyelinating lesions and normal appearing white matter in multiple sclerosis. Magn Reson Med 32: Iannucci G, Minicucci L, Rodegher M, Sormani MP, Comi G, Filippi M (1999) Correlations between clinical and MRI involvement in multiple sclerosis: assessment using T(1), T(2) and MT histograms. J Neurol Sci 171: Kamber M (1991) Automated detection of multiple sclerosis lesions in magnetic resonance images of the human brain. Montreal, Quebec, Canada: Concordia University 17. Kurtzke JF (1983) Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 33: Loevner LA, Grossman RI, McGowan JC, Ramer KN, Cohen JA (1995) Characterization of multiple sclerosis plaques with T1-weighted MR and quantitative magnetization transfer. AJNR Am J Neuroradiol 16: Loevner LA, Grossman RI, Cohen JA, Lexa FJ, Kessler D, Kolson DL (1995) Microscopic disease in normal-appearing white matter on conventional MR images in patients with multiple sclerosis: assessment with magnetization-transfer measurements. Radiology 196: Narayanan S, Fu L, Pioro E, De Stefano N, Collins DL, Francis GS, et al. (1997) Imaging of axonal damage in multiple sclerosis: spatial distribution of magnetic resonance imaging lesions. Ann Neurol 41: O Riordan JI, Thompson AJ, Kingsley DP, MacManus DG, Kendall BE, Rudge P, et al. (1998) The prognostic value of brain MRI in clinically isolated syndromes of the CNS. A 10-year followup. Brain 121: Petrella JR, Grossman RI, McGowan JC, Campbell G, Cohen JA (1996) Multiple sclerosis lesions: relationship between MR enhancement pattern and magnetization transfer effect. AJNR Am J Neuroradiol 17: Phadke JG (1990) Clinical aspects of multiple sclerosis in north-east Scotland with particular reference to its course and prognosis. Brain 113: Pike GB, Glover GH, Hu BS, Enzmann DR (1993) Pulsed magnetization transfer spin-echo MR imaging. J Magn Reson Imaging 3: Pike GB, de Stefano N, Narayanan S, Francis GS, Antel JP, Arnold DL (1999) Combined magnetization transfer and proton spectroscopic imaging in the assessment of pathologic brain lesions in multiple sclerosis. AJNR Am J Neuroradiol 20: Pike GB, De Stefano N, Narayanan S, Worsley KJ, Pelletier D, Francis GS, et al. (2000) Multiple sclerosis: magnetization transfer MR imaging of white matter before lesion appearance on T2-weighted images. Radiology 215: Poser CM, Paty DW, Scheinberg L, Mc- Donald WI, Davis FA, Ebers GC, et al. (1983) New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 13: Riise TGM, Fernandez O, Lauer K, Midgard R, Minderhoud JM (1992) Early prognostic factors for disability in multiple sclerosis, a European multicenter study. Acta Neurol Scand 85: Runmarker B, Andersen O (1993) Prognostic factors in a multiple sclerosis incidence cohort with twenty-five years of follow-up. Brain 116: Sailer M, O Riordan JI, Thompson AJ, Kingsley DP, MacManus DG, McDonald WI, et al. (1999) Quantitative MRI in patients with clinically isolated syndromes suggestive of demyelination [see comments]. Neurology 52: Silver NC, Lai M, Symms MR, Barker GJ, McDonald WI, Miller DH (1998) Serial magnetization transfer imaging to characterize the early evolution of new MS lesions [see comments]. Neurology 51: Thompson AJ, Kermode AG, Mac- Manus DG, Kendall BE, Kingsley DP, Moseley IF, et al. (1990) Patterns of disease activity in multiple sclerosis: clinical and magnetic resonance imaging study. BMJ 300: Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mork S, Bo L (1998) Axonal transection in the lesions of multiple sclerosis [see comments]. N Engl J Med 338: van Buchem MA, McGowan JC, Kolson DL, Polansky M, Grossman RI (1996) Quantitative volumetric magnetization transfer analysis in multiple sclerosis: estimation of macroscopic and microscopic disease burden. Magn Reson Med 36: van Buchem MA, Udupa JK, McGowan JC, Miki Y, Heyning FH, Boncoeur- Martel MP, et al. (1997) Global volumetric estimation of disease burden in multiple sclerosis based on magnetization transfer imaging. AJNR Am J Neuroradiol 18: van Waesberghe JH, Kamphorst W, De Groot CJ, van Walderveen MA, Castelijns JA, Ravid R, et al. (1999) Axonal loss in multiple sclerosis lesions: magnetic resonance imaging insights into substrates of disability. Ann Neurol 46: van Walderveen MA, Kamphorst W, Scheltens P, van Waesberghe JH, Ravid R, Valk J, et al. (1998) Histopathologic correlate of hypointense lesions on T1- weighted spin-echo MRI in multiple sclerosis. Neurology 50: Weinshenker BG, Issa M, Baskerville J (1996) Long-term and short-term outcome of multiple sclerosis: a 3-year follow-up study. Arch Neurol 53:

7 Weinshenker BG, Rice GP, Noseworthy JH, Carriere W, Baskerville J, Ebers GC (1991) The natural history of multiple sclerosis: a geographically based study. 3. Multivariate analysis of predictive factors and models of outcome. Brain 114: Weinshenker BG, Bass BB, Rice GP, Noseworthy J, Carriere W, Baskerville J, Ebers GC (1989) The natural history of multiple sclerosis: a geographically based study. I. Clinical course and disability. Brain 112: Weinshenker BG, Bass BB, Rice GP, Noseworthy J, Carriere W, Baskerville J, Ebers GC (1989) The natural history of multiple sclerosis: a geographically based study. 2. Predictive value of the early clinical course. Brain 112: Weinshenker BG, Rice RG, Noseworthy JH, Carriere W, Baskerville J, Ebers GC (1991) The natural history of multiple sclerosis: a geographically based study. 4. Applications to planning and interpretation of clinical therapeutic trials. Brain 114: Wolff SD, Balaban RS (1989) Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn Reson Med 10: