Copyright # Munksgaard 2001 Acta Neurol Scand 2001: 104: 24–30 Printed in UK. All rights reserved ACTA NEUROLOGICA SCANDINAVICA ISSN 0001-6314 Correlations of brain MRI parameters to disability in multiple sclerosis Schreiber K, Sørensen PS, Koch-Henriksen N, Wagner A, Blinkenberg M, Svarer C, Petersen HC. Correlations of brain MRI parameters to disability in multiple sclerosis. Acta Neurol Scand 2001: 104: 24–30. # Munksgaard 2001. Objectives – The objective was to correlate magnetic resonance imaging (MRI) T2-weighted lesion load and measures of white matter atrophy in the brain to disability in a population-based sample of patients with multiple sclerosis (MS). Material and methods – A well defined cohort of patients was drawn at random from the general MS population by using the Danish Multiple Sclerosis Reigistry. A semi-automated local thresholding technique was used to quantify T2-weighted lesions on MRI; whereas manual tracing was applied to measure the corpus callosum brain ratio (CCR) and the ventricle brain ratio (VBR). Results – A sample of 86 patients with a mean age of 43.3 years (SD 4.3), mean disease duration of 13.6 years (SD 4.4) and a median Expanded Disability Status Score (EDSS) of 6.0 was identified. The correlation between total lesion area of the brain (TLA) and disability (EDSS) for the whole sample was moderate (Spearman rank correlation coefficient r=0.48, P<0.001). Also correlations of CCR and VBR to disability (r=0.32–0.46) were significant. Conclusions – Correlations of TLA and disability in this study were rather strong. Hence, T2-weighted MRI lesion load in the brain still plays an important role as a surrogate marker of disease and as a secondary outcome measure in phase III treatment trials. Magnetic resonance imaging (MRI) has been established as the most sensitive paraclinical tool to demonstrate lesions disseminated in time and space in multiple sclerosis (MS) (1). Furthermore, autopsy studies have confirmed that hyperintense lesions on T2-weighted MR images represent the pathological abnormalities of MS (2–4). However, the strength of correlations between T2-weighted lesion load and disability has most often been poor (5–7). Nevertheless, the measurement of lesion load and especially counts and measurements of new and active lesions on MRI, have proven invaluable as secondary outcome measures in phase III trials (28, 29, 43) and as primary outcome measures in exploratory phase II trials (8–11). The clinicoradiological paradox, that is the discrepancy between MRI findings and disability in established MS might to some degree be explained by biological 24 K. Schreiber1, P. S. Sørensen1, N. Koch-Henriksen2, A. Wagner3, M. Blinkenberg1, C. Svarer1, H. C. Petersen4 Copenhagen University Hospital, Rigshospital, 1 Department of Neurology, 2The Danish Multiple Sclerosis Registry, 3Department of Radiology and 4The Danish Institute of Clinical Epidemiology Key words: multiple sclerosis; MRI; disability; epidemiology P. S. Sørensen/K. Schreiber, Copenhagen MS Clinic, Department of Neurology, Copenhagen University Hospital, Rigshospital, DK-2100 Copenhagen, Denmark Fax: +45 35452626 Accepted for publication January 19, 2001 reasons, where the anatomic site and pathological heterogeneity of lesions may be more relevant to disability than the total lesion load. However, technical and methodological factors may also play some role. Only the most recent studies have applied semi-automated segmentation techniques to measure MRI lesion load, which have proven to have high intra-rater reliability (12–15). Furthermore, all previous studies have included patients selected from MS hospital clinics, often as participants in clinical trials where particularly severely disabled patients have been underrepresented, thereby providing a highly selected patient sample and possibly biasing the results regarding the natural manifestation of the disease. Thus the objective of this study was to examine a large population-based sample of MS patients, with a balanced distribution of age and disease durations, Correlations of brain MRI parameters to estimate the association of brain T2-weighted lesion load and atrophy of white matter to disability. Material and methods The experimental design was unique. Patients were recruited from the Danish Multiple Sclerosis Registry (DMSR) which has a completeness of case ascertainment of 90% (16). They were randomly recruited according to the stratified age groups of 35–39, 40–44 and 45–49 years, and according to disease durations stratified to 5–9, 10–14 and 15–19 years. This design provided a patient sample with a well-defined range of age and disease duration, thus attempting to identify the possible confounding effects of age and disease duration in MS. The examinations took place in 1993–95, in the catchment area of the city and county of Copenhagen and its surrounding counties, a predominantly urban population of 2 million people. Cases in the DMSR were registered and classified according to the criteria of Allison & Millar (17) (clinically definite MS, probable, latent probable and possible MS) and further modified to include laboratory (IgG index and oligoclonal bands in the cerebrospinalfluid) and paraclinical data (evoked potentials). A total of 279 patients were registered according to the criteria of age and disease duration. In addition, 108 observational cases were revised, by scrutinizing the medical records and consulting general practitioners for updated information, which yielded 24 cases meeting the criteria of the Registry. They were included in the candidate population; a total of 303 cases. A study sample of 164 patients was randomly drawn from this population. All patients were evaluated by the same neurologist (KS) and after examination, only patients who fulfilled the Poser criteria of clinically definite or probable MS were included (18). Nine patients had died after having been drawn from the Registry. Furthermore, 17 patients were excluded for the following reasons: 2 had moved from the area; 7 patients had only possible MS; 7 patients proved not to have MS but on examination had other organic diseases; and 1 patient had MS and a large internal hydrocephalus. The remaining 138 patients were eligible for the study but 19 patients (i.e. 14%) were non-responders. Thus the MS study sample included 119 patients; 86 had MRI of the brain done, whereas 33 patients were unwilling or excluded. Reasons for not having MRI done were: claustrophobia (12 patients), metal implants (2), obesity (2), severe disability with dyspnea (EDSS 7.5–9–5; n=8) and unwillingness although not severely disabled (6). Furthermore 3 patients did have MR scans done, but the scans were discarded because their poor technical quality could not allow for measurements. All the included patients had clinically definite MS and were assessed using the Kurztke Expanded Disability Status Score (EDSS) (19). All patients were clinically stable, i.e. did not have an acute exacerbation at the time of assessment, except for 1 patient who suffered a spinal cord relapse. Only 2 patients had recently received steroid treatment (2 months prior to the examination) and none had ever received long-term immunosuppressive treatment; thus this patient sample reflects the natural history of MS, including patients with long disease duration and severe disability. The MRI protocol All scans of the brain and brainstem were performed on a 0.3 T Fonar equipment, using a 256r256 image matrix, a field of view of 25.2 cm and a double spin-echo technique yielding 13 axial scans of 7 mm thickness with an interslice gap of 3 mm. T1-weighted images (TR/TE=441/16 ms) were obtained in the sagittal plane. Using the midline image as a scout; the proton- and T2weighted axial images (TR/TE=2000/20/85 ms) were obtained in the transverse plane parallel to a line located from the glabella to the bottom of the pons. The scans were transformed to a computer workstation and quantified using the Dispimage program developed by Mr D. Plummer (Dept of Medical Physics, University College, London) (20). This semi-automated local thresholding technique has been assessed by others and proven to have a high degree of reproducibility (12, 13, 21). All scans were evaluated by the same neurologist (KS) who was blinded to the clinical data (patients were assigned code numbers) and 40 scans were reviewed together with an experienced neuroradiologist in order to reach consensus. T2-weighted images were evaluated for the presence of hyperintensities greater than 3 mm, taking care not to include cortical sulci which could be mistaken for MS lesions. Hyperintense signals at the frontal horns were only classified as lesions if they were larger than 5 mm or if they were asymmetrical and had a difference of more than 2 mm between the sides. A semi-automated contouring method, with a manual selection of perimeter threshold in lesions was applied to calculate the lesion area. However, manual correction was often necessary to modify the boundary of poorly defined or confluent lesions. The size of the lateral ventricles was measured on one axial slice where they were contiguous and largest, by manual tracing. The corresponding brain 25 Schreiber et al. Table 1. Clinical data on the study sample and the non-responding sample Study sample (n=119) Female/male ratio Mean age at onset-years (range) Mean disease duration-years (range) Mean age at assessment (range) DSS (1993) median (range) Statistical analysis: a Fisher’s exact test; 1.5 29.2 13.8 42.9 4.0 b (18–42) (5–24) (35–50) (0–8) Non-resp. (n=19) 1.1 28.9 13.7 42.6 6.0 (17–39) (7–21) (35–51) (1–9) Table 2. Clinical characteristics of the examined patient samples Study sample (n=119) P 0.62 0.93 0.87 0.81 0.09 Female/male ratio Mean age at onset (y) (SD) Mean disease duration (y) (range) Mean age at assessment (SD) Median EDSS (range) 1.5 29.2 13.8 42.9 6.0 MRI (n=86) (5.5) (5–24) (4.5) (1–9.5) 1.6 29.8 13.6 43.3 6.0 (5.6) (5–24) (4.3) (1–9) Non-MRI (n=33) 1.4 27.6 14.3 41.9 6.5 (4.7) (6–22) (4.9) (1–9.5) MRI/ non-MRI P 0.68 0.05 0.46 0.12 0.17 Mann–Whitney U-test. area on this slice was measured by the contouring method; providing the ventricle/brain ratio (VBR). The size of the corpus callosum was measured by manual tracing on one mid-sagittal T1-weighted slice, and the corresponding sagittal brain area was traced manually, outlining the hemisphere excluding the cerebellum and continuing over the top of the mesencephalon to reach the medial boundary of the frontal lobe. Thus the corpus callosum/brain ratio (CCR) was calculated. To assess the intrarater agreement of measurements, 10 random MR scans were re-evaluated by the author, blind to the first evaluation with a 1-week interval. Statistical analysis Non-parametric analyses were used (Mann– Whitney U-test corrected for ties) to evaluate the mean differences in the patient samples regarding age, disease duration and disability. Fisher’s exact test was used to evaluate the difference in female/ male ratios. The Kruskal–Wallis one-way ANOVA was used to evaluate the mean differences between clinical and MRI parameters among the clinical groups. If significant differences were found, Dunn’s method for post hoc multiple comparisons was used. Correlations were calculated by the Spearman rank correlation coefficient (r). Because several correlations were done, a two-sided significance level of <0.01 was considered statistically significant. Results The Registry (DMSR) provided clinical data on the non-responding group, which could therefore be compared to the study sample at the latest updated status of 1992. Table 1 shows that there was no Statistical analysis: a Fisher’s exact Test; b Mann–Whitney U-test. significant difference between the two groups regarding female/male ratio, age at onset, disease duration, age at assessment and Disability Status Score (DSS) although there was a trend towards more severe disability in the non-responding group (P=0.09). As shown in Table 2, the group of patients that had MRI done (N=86) was highly comparable to the study sample of 119 patients. The statistical analyses of non-participants refers only to the comparison between patients who had MRI done to the group who did not have MRI done (non-MRI, N=33). The inclusion of data from the 119 sample is merely illustrative. The non-MRI patients did not differ significantly from the MRI group regarding the distribution of female/male ratios, disease duration, and age at assessment or disability. However, the non-MRI group did have a slightly younger age at onset than the MRI group, which cannot readily be explained. The non-MRI group did tend to be more disabled (median EDSS 6.5 contra 6.0 for the MRI group; P=0.17) although the difference was not significant. Table 3 shows the clinical characteristics of the MRI group, according to the clinical course. The following definitions were used: the relapsing/ remitting (RRMS) course was characterized by complete or incomplete remissions being stable between relapses, whereas the secondary progressive course (SPMS) after an initial relapsing/ remitting course, had developed progressive disability with or without super-imposed relapses during a 6 month period. The primary progressive course (PPMS) was defined as progressive deterioration from onset without relapses or remissions. In accordance with epidemiological studies (22–24), the PPMS group comprised 12.8% of the sample, Table 3. Clinical characteristics of the 86 patients with brain MRI No. of patients Mean age at assessment years (SD) Mean disease duration years (range) Median EDSS (range) Statistical analysis: Kruskal–Wallis test. 26 86 (all) 23 (RR) 52 (SP) 11 (PP) P 43.3 (4.3) 13.6 (5–24) 6.0 (1–9) 41.5 (4.6) 12.5 (6–19) 3.0 (1–6.5) 43.9 (4.2) 14.4 (5–24) 6.5 (1.5–9) 44.6 (3.6) 11.6 (5–20) 8.0 (4–9) 0.0705 0.0726 <0.0001 Correlations of brain MRI parameters Fig. 1. Expanded disability status scale versus total lesion load. whereas 87.2% had a relapsing/remitting onset (RR+SPMS). At the time of examination 52 patients (60.5%) were in the secondary progressive phase. The distribution of the MRI parameters TLA, VBR and CCR regarding the clinical courses are shown in Table 4. A significant difference between the clinical courses was found regarding TLA (P=0.018), and the difference arose between the RR and SPMS groups (Dunn’s test for Multiple Comparisons, P<0.02), reflecting the more advanced stage of the SPMS group. The total lesion area in the PPMS group was high compared to the SPMS group, which is quite contrary to the findings in other studies. In the literature, the reported lesions in PPMS are typically small, mostly non-inflammatory and TLA is much lower than in the SPMS group (13, 25, 26). The authors acknowledge difficulties in discriminating the SPMS from PPMS courses with few remissions, as the medical records were not originally designed for scientific purposes and interviews of patients and family members regarding events years ago could not always verify the clinical course. Undoubtedly, some patients classified as PPMS were probably SPMS if they had entered the progressive phase early after onset of the disease. Furthermore, the definition of PPMS has been very variable and only Table 4. Distribution of brain MRI lesion area in 86 patients No. of patients Periventricular mean (range) Discrete cerebral white matter Internal capsule Cerebellum and Brainstem VBR CCR Total lesion area 86 all 1630.6 408.1 24.3 23.8 0.147 0.055 (0–6731) (0–1829) (0–270) (0–257) (0.08–0.250) (0.03–0.076) 2087 (0–8414) 23 RR 756.0 324.3 9.1 33.1 0.139 0.059 (0–2754) (0–1116) (0–115) (0–207) (0.097–0.183) (0.044–0.076) 1123 (0–4001) 52 SP 1995.5 450.3 34.6 16.5 0.150 0.052 (0–6731) (0–1829) (0–270) (0–257) (0.08–0.250) (0.03–0.075) 2497 (0–8414) 11 PP 1734.2 383.8 7.3 38.9 0.146 0.055 (300–5129) (33–1656) (0–59) (0–177) (0.115–0.184) (0.038–0.075) 2164 (416–6785) P 0.008 0.559 0.086 0.135 0.359 0.045 0.018 Statistical analysis: Kruskal–Wallis test. Values in mean square mm. (range). Total lesion area rounded to the whole mm. 27 Schreiber et al. Table 5. Correlations between clinical and MRI data for the whole group and for the SPMS sub-group Data correlated by Spearman rank EDSS vs TLA EDSS vs CCR EDSS vs VBR TLA vs CCR TLA vs VBR Duration vs TLA Duration vs CCR Duration vs VBR Whole group (n=86) r and P value 0.48 x0.40 0.32 x0.70 0.62 0.27 x0.33 0.37 (P<0.001) (P<0.001) (P=0.003) (P<0.001) (P<0.001) (P=0.01) (P=0.002) (P<0.001) SPMS group (n=52) r and P value 0.56 x0.39 0.46 x0.78 0.70 0.30 x0.41 0.42 (P<0.001) (P=0.005) (P=0.001) (P<0.001) (P<0.001) (P=0.029) (P=0.003) (P=0.002) Two-tailed P<0.01=significant. r=Spearman rank correlation coefficient. EDSS=Expanded disability status scale. TLA=Total lesion area. CCR=Corpus callosum brain ratio. VBR=Ventricle brain ratio. during the recent years been somewhat clarified (27). There was a significant difference between the courses on the CCR (P=0.045), where the difference arose between RR and SPMS groups (P<0.05; Dunn’s test). The VBR was not significantly different between the clinical courses. Correlations of MRI parameters to clinical data Table 5 shows the correlations for the whole sample and the subgroup SPMS (N=52). Fig. 1 shows the scatter of TLA versus EDSS (86 patients) which does not reveal an ideal elipsoid but no extreme values exert undue leverage. The TLA was significantly correlated to the EDSS (Spearman rank correlation coefficient was r=0.48, P<0.001). The larger SPMS subgroup was analyzed separately and the correlation remained quite the same; r=0.56, P<0.001. The parameters TLA, CCR and VBR were significantly and highly intercorrelated in the range of r=0.62–0.78; indicating that there was a strong association between TLA and measures of white matter atrophy. Furthermore, the correlations between EDSS and CCR and VBR were moderate but significant: in the whole sample the CCR vs EDSS was r=x0.40, P<0.001; and the VBR vs EDSS was r=0.32; P=0.003. The correlations in the SPMS showed somewhat the same association. The TLA correlated poorly but significantly to disease duration for the whole sample; r=0.27, P=0.01, and only a trend for the SPMS subgroup was found; r=0.30, P=0.029. Reproducibility In a sample of 10 random scans evaluated twice blindly, the median intra-rater agreement was: TLA 92.1% (range 55.1–98.9); for the VBR 94.0% (range 85.3–99.1) and for CCR 91.6% (range 85.7–99.6). 28 Discussion In most recent cross-sectional studies, the strength of correlation between brain T2-weighted lesion load and disability has been very modest (r=0.20–0.50) (13, 21, 28–30) and only few have demonstrated moderate correlations (r=0.6) (31, 32). In order to avoid the selection bias inherent in studies on patients attending MS clinics and trials, this study examined a large population-based sample of MS patients, having a balanced stratified distribution regarding age and disease duration, including patients with severe disability. The main objective was to estimate the association between brain T2-weighted total lesion area and atrophy of white matter to disability in MS patients. The nonparticipants amounted to 14% of the eligible patients, and the analyses showed that the study sample was not skewed by selection bias, although the non-participating patients tended to be more disabled (Table 1). The non-MRI group of 33 patients was unavoidably large mainly because of claustrophobia and severe disability, but did not differ from the MRI group (N=86) regarding main clinical characteristics (Table 2). Therefore it may be maintained that the MRI group studied is representative of the MS population within the defined age and disease duration limits, reflecting the natural manifestation of the disease. This patient sample represents a broader spectrum of disability, with a median EDSS of 6 whereas other MRI studies have had median EDSS scores of 3–5 (13, 14, 28, 29, 31). Likewise, the mean disease duration in this study is longer (mean 13.6 years), where others had mean durations of 3 to 5 years and only one study had 9.8 years (13). The MRI equipment (0.3 T) with 7 mm slices and 3 mm interslice gaps, might have underestimated the lesion load in this study compared with other studies using stronger magnetic field strengths and without interslice gaps. The intra-rater agreement for TLA was comparable to the results of others (12). The correlation of TLA and disability for the whole sample was moderate (r=0.48), but comparative to others (13, 21, 28–30). The highest correlation between TLA and disability was found in the largest subgroup, the SPMS, in accordance with others (13, 31). The TLA in SPMS was also the largest and probably reflects the pathological findings reported by Barnes and others, who showed that the expansion of extracellular space correlated to a reduction in axonal density, producing disability (4, 33). Also, Truyen et al. (34) illustrated that in the more disabled SPMS group, changes on T2 were more frequently accompanied by changes on T1 than in the RRMS group. Hypointense lesions on T1 images have low Correlations of brain MRI parameters magnetization transfer ratios, which indicate considerable myelin loss, resulting in irreversible disability (30). However, in the present study, only about 23% (r2) of the variability in the EDSS can be explained by the variability of TLA and this poor correlation may be due to a combination of biological and methodological factors (9, 15, 21, 30, 35–37). Due to the pathological heterogeneity of MS lesions, T2-weighted MRI cannot discriminate between lesions with inflammatory edema, demyelination, gliosis or axonal loss, which cause variable degrees of disability (3). The incomplete imaging of the CNS (i.e. the optic nerves and spinal cord were not imaged) might have weakened the correlation. Furthermore the variance of the clinical rating scale (EDSS) is limited and depends mainly on locomotor disability (38). The modest correlation of TLA to disease duration reflects the limitations of the crosssectional research design, but does underscore the well-known results of longitudinal studies, proving that lesion load does accumulate with time (14, 21, 29, 34, 39). The measures of atrophy (the CCR and VBR) correlated modestly but significantly to disease duration. Likewise, CCR and VBR, which were highly intercorrelated to TLA, also correlated modestly to EDSS. The corpus callosum and the periventricular regions are not specifically relevant to sensory-motor pathways, and the correlations to EDSS most likely arose because of the intercorrelations to TLA, where larger lesion loads increased the probability of plaques encroaching on these pathways. Atrophy implies axonal loss, and is not confounded to the same degree by the pathological heterogeneity of T2 lesions (3, 4, 40). Measures of atrophy may be valuable as measures of progressive neural degeneration (41). Atrophy of the spinal cord has been shown to correlate highly to disability (r=x0.70) and has been proposed as a measure of progressive neurological deterioration to monitor treatment trials (42). Furthermore, in the same study, the benign group had the same disability as the RRMS group but had longer disease duration and a decrease in cord area, which therefore led the authors to suggest that atrophy might also be a feature of disease duration. In conclusion, the correlations between T2weighted MRI lesion load and disability in this population-based sample were moderate, underscoring the limited pathological specificity of T2weighted lesions. Bearing these limitations in mind, T2-weighted lesion assessment is the most accessible MRI technology and may still serve as a surrogate marker of disease burden in phase III trials, where clinical disability remains the primary outcome measure. Newer, pathologically more specific tech- niques such as T1-weighted MRI (black holes), magnetization transfer imaging and magnetic resonance spectroscopy have shown better correlations (30, 34, 36, 36a, 37, 40) but are not yet available for clinical trials. Acknowledgements This study was supported by the Danish Multiple Sclerosis Society, the Bursary of Director Ejnar Jonasson (Johnsen) and spouse, and the Bursary of Lily Benthine Lund. References 1. GOODKIN DE, RUDICK RA, ROSS JS. The use of brain magnetic resonance imaging in multiple sclerosis. Arch Neurol 1994;51:505–16. 2. NEWCOMBE J. Histopathology of Multiple Sclerosis lesions detected by MRI in unfixed postmortem CNS tissue. Brain 1991;14:1013–23. 3. MCDONALD WI, MILLER DH, BARNES D. The pathological evolution of multiple sclerosis. Neuropathol Appl Neurobiol 1992;18:319–34. 4. BARNES D, MUNRO PM, YOUL BD, PRINEAS JW, MCDONALD WI. The longstanding MS lesion. A quantitative MRI and electron microscopic study. Brain 1991;114:1271–80. 5. KOOPMANS RA, LI DK, GROCHOWSKI E, CUTLER PJ, PATY DW. Benign versus chronic progressive multiple sclerosis: magnetic resonance imaging features. Ann Neurol 1989;25:74–81. 6. BAUMHEFNER RW, TOURTELLOTTE WW, SYNDULKO K et al. Quantitative multiple sclerosis plaque assessment with magnetic resonance imaging. Its correlation with clinical parameters, evoked potentials, and intra-blood–brain barrier IgG synthesis. Arch Neurol 1990;47:19–26. 7. THOMPSON AJ, KERMODE AG, MAcMANUS DG et al. Patterns of disease activity in multiple sclerosis: clinical and magnetic resonance imaging study [see comments]. BMJ 1990;300: 631–4. 8. STONE LA, FRANK JA, ALBERT PS et al. Characterization of MRI response to treatment with interferon beta-1b: contrast-enhancing MRI lesion frequency as a primary outcome measure. Neurology 1997;49:862–9. 9. MILLER DH. Guidelines for the use of magnetic resonance techniques in monitoring the treatment of multiple sclerosis. Ann Neurol 1996;39:6–16. 10. MOLYNEUX PD, FILIPPI M, BARKHOF F et al. Correlations between monthly enhanced MRI lesion rate and changes in T2 lesion volume in multiple sclerosis. Ann Neurol 1998; 43:332–9. 11. SORENSEN PS, WANSCHER B, JENSEN CV et al. Intravenous immunoglobulin G reduces MRI activity in relapsing multiple sclerosis. Neurology 1998;50:1273–81. 12. FILIPPI M, HORSFIELD MA, BRESSI S et al. Intra- and interobserver agreement of brain MRI lesion volume measurements in multiple sclerosis. A comparison of techniques. Brain 1995;118:1593–00. 13. MAMMI S. Correlation between brain MRI lesion volume and disability in patients with multiple sclerosis. Acta Neurol Scand 1996;94:93–6. 14. FILIPPI M, HORSFIELD MA, MORRISSEY SP et al. Quantitative brain MRI lesion load predicts the course of clinically isolated syndromes suggestive of multiple sclerosis. Neurology 1994;44:635–41. 15. FILIPPI M, HORSFIELD MA, TOFTS PS, BARKHOF F, THOMPSON 29 Schreiber et al. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 30 AJ, MILLER DH. Quantitative assessment of MRI lesion load in monitoring the evolution of multiple sclerosis. Brain 1995;118:1601–12. KOCH HENRIKSEN N, HYLLESTED K. Epidemiology of multiple sclerosis: incidence and prevalence rates in Denmark 1948–64 based on the Danish Multiple Sclerosis Registry. Acta Neurol Scand 1988;78:369–80. ALLISON RS, MILLAR JDH. Prevalence and familial incidence of disseminated sclerosis. Ulster Med J 1954;23(Suppl. 2):29–49. POSER CM, PATY DW, SCHEINBERG L et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227–31. KURTZKE JF. Rating neurologic impairment in multiple sclerosis: An expanded disability status scale (EDSS). Neurology 1983;33:1444–52. WICKS DA, TOFTS PS, MILLER DH et al. Volume measurement of multiple sclerosis lesions with magnetic resonance images. A preliminary study. Neuroradiology 1992;34: 475–9. vAN WALDERVEEN MA, BARKHOF F, HOMMES OR et al. Correlating MRI and clinical disease activity in multiple sclerosis: relevance of hypointense lesions on short-TR/ short-TE (T1-weighted) spin-echo images. Neurology 1995; 45:1684–90. WEINSHENKER BG, BASS B, RICE GP et al. The natural history of multiple sclerosis: a geographically based study. I. Clinical course and disability. Brain 1989;112:133–46. MILLER DH, HORNABROOK RW, PURDIE G. The natural history of multiple sclerosis: a regional study with some longitudinal data. J Neurol Neurosurg Psychiatry 1992;55:341–6. PHADKE JG. Clinical aspects of multiple sclerosis in northeast Scotland with particular reference to its course and prognosis. Brain 1990;113:1597–628. THOMPSON AJ, KERMODE AG, WICKS D et al. Major differences in the dynamics of primary and secondary progressive multiple sclerosis. Ann Neurol 1991;29:53–62. REVESZ T, KIDD D, THOMPSON AJ, BARNARD RO, MCDONALD WI. A comparison of the pathology of primary and secondary progressive multiple sclerosis. Brain 1994;117: 759–65. MCDONNELL GV. Primary progressive multiple sclerosis: a distinct syndrome? Multiple Sclerosis 1996;2:137–41. PATY DW, LI DK. Interferon beta-1b is effective in relapsing–remitting multiple sclerosis. II. MRI analysis results of a multicenter, randomized, double-blind, placebocontrolled trial. UBC MS/MRI Study Group and the IFNB Multiple Sclerosis Study Group [see comments]. Neurology 1993;43:662–7. The IFNB Multiple Sclerosis Study Group and The University of British Columbia MS/MRI Analysis Group. Interferon beta-1b in the treatment of multiple sclerosis: final outcome of the randomized controlled trial. Neurology 1995;45:1277–85. GASS A, BARKER GJ, KIDD D et al. Correlation of magnetization transfer ratio with clinical disability in multiple sclerosis. Ann Neurol 1994;36:62–7. 31. TRUYEN L, GHEUENS J, VAN DE VYVER FL, PARIZEL PM, PEERSMAN GV, MARTIN JJ. Improved correlation of magnetic resonance imaging (MRI) with clinical status in multiple sclerosis (MS) by use of an extensive standardized imaging-protocol. J Neurol Sci 1990;96:173–82. 32. RIAHI F, ZIJDENBOS A, NARAYANAN S et al. Improved correlation between scores on the expanded disability status scale and cerebral lesion load in relapsing–remitting multiple sclerosis. Results of the application of new imaging methods. Brain 1998;121:1305–12. 33. KATZ D, TAUBENBERGER JK, CANNELLA B, MCFARLIN DE, RAINE CS, MCFARLAND HF. Correlation between magnetic resonance imaging findings and lesion development in chronic, active multiple sclerosis. Ann Neurol 1993;34: 661–9. 34. TRUYEN L, vAN WAESBERGHE JH, vAN WALDERVEEN MA et al. Accumulation of hypointense lesions (‘‘black holes’’) on T1 spin-echo MRI correlates with disease progression in multiple sclerosis. Neurology 1996;47:1469–76. 35. Practice advisory on selection of patients with multiple sclerosis for treatment with Betaseron. Report of the Quality Standards Subcommittee of the American Academy of Neurology. Neurology 1994;44:1537–40. 36. MATTHEWS PM. Assessment of lesion pathology in multiple sclerosis using quantitative MRI morphometry and magnetic resonance spectroscopy. Brain 1996;119:715–22. 36a. vAN WALDERWEEN MA, KAMPHORST W, SCHELTENS P et al. Histopathological correlate of hypointense lesions on T1 weighted spin-echo MRI in multiple sclerosis. Neurology 1998;50:1282–8. 37. TOURBAH A, STIEVENART JL, GOUT O et al. Localized proton magnetic resonance spectroscopy in relapsing remitting versus secondary progressive multiple sclerosis. Neurology 1999;53:1091–7. 38. NOSEWORTHY JH. Clinical scoring methods for multiple sclerosis. Ann Neurol 1994;36(Suppl.):S80–5. 39. STONE LA, ALBERT PS, SMITH ME et al. Changes in the amount of diseased white matter over time in patients with relapsing–remitting multiple sclerosis. Neurology 1995;45: 1808–14. 40. ROVARIS M, BOZZALI M, RODEGHER M, TORTORELLA C, COMI G, FILIPPI M. Brain MRI correlates of magnetization transfer imaging metrics in patients with multiple sclerosis. J Neurol Sci 1999;166:58–63. 41. LOSSEFF NA, MILLER DH. Measures of brain and spinal cord atrophy in multiple sclerosis. J Neurol Neurosurg Psychiatry 1998;64(Suppl. 1):S102–5. 42. LOSSEFF NA. Spinal cord atrophy and disability in multiple sclerosis. Brain 1996;119:701–8. 43. PRISMS (Prevention of Relapses and Disability by Interferon beta-1-a Subcutaneously in MS) Study Group. Randomised double-blind placebo-controlled study of interferon beta-1-a in relapsing–remitting multiple sclerosis. Lancet 1998;352:1498–504.