[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Survival curves demonstrating conversion rates to clinically definite multiple sclerosis (CDMS) in patients receiving early vs delayed treatment (Tx) and having 4 vs fewer than 4 Barkhof criteria (B crit) at baseline.

Survival curves demonstrating conversion rates to clinically definite multiple sclerosis (CDMS) in patients receiving early vs delayed treatment (Tx) and having 4 vs fewer than 4 Barkhof criteria (B crit) at baseline.

Figure 2.
Dissemination in time (DIT) at the 9-month (M9) magnetic resonance imaging was associated with increased risk for conversion to clinically definite multiple sclerosis (CDMS) in patients without dissemination in space (DIS) at baseline (left panel). In patients with DIS at baseline, DIT was not associated with increased risk for conversion to CDMS (right panel).

Dissemination in time (DIT) at the 9-month (M9) magnetic resonance imaging was associated with increased risk for conversion to clinically definite multiple sclerosis (CDMS) in patients without dissemination in space (DIS) at baseline (left panel). In patients with DIS at baseline, DIT was not associated with increased risk for conversion to CDMS (right panel).

Table 1. 
Baseline Magnetic Resonance (MR) Imaging Characteristics
Baseline Magnetic Resonance (MR) Imaging Characteristics
Table 2. 
Conversion to Clinically Definite Multiple Sclerosis (CDMS) After 3 Years by Baseline Magnetic Resonance (MR) Imaging Characteristic
Conversion to Clinically Definite Multiple Sclerosis (CDMS) After 3 Years by Baseline Magnetic Resonance (MR) Imaging Characteristic
Table 3. 
Conversion to Clinically Definite Multiple Sclerosis (CDMS) After 3 Years by Number of Barkhof Criteria
Conversion to Clinically Definite Multiple Sclerosis (CDMS) After 3 Years by Number of Barkhof Criteria
Table 4. 
Conversion to Clinically Definite Multiple Sclerosis (CDMS) After 3 Years Among Patients Having New Lesions Observed at the 3-, 6-, and 9-Month Magnetic Resonance (MR) Imaging With or Without Dissemination in Space (DIS) at Baseline
Conversion to Clinically Definite Multiple Sclerosis (CDMS) After 3 Years Among Patients Having New Lesions Observed at the 3-, 6-, and 9-Month Magnetic Resonance (MR) Imaging With or Without Dissemination in Space (DIS) at Baseline
1.
Miller  DBarkhof  FMontalban  XThompson  AFilippi  M Clinically isolated syndromes suggestive of multiple sclerosis, part I: natural history, pathogenesis, diagnosis, and prognosis. Lancet Neurol 2005;4 (5) 281- 288
PubMedArticle
2.
Poser  CMPaty  DWScheinberg  L  et al.  New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13 (3) 227- 231
PubMedArticle
3.
Brex  PACiccarelli  OO’Riordan  JISailer  MThompson  AJMiller  DH A longitudinal study of abnormalities on MRI and disability from multiple sclerosis. N Engl J Med 2002;346 (3) 158- 164
PubMedArticle
4.
Beck  RWTrobe  JDMoke  PS  et al. Optic Neuritis Study Group, High- and low-risk profiles for the development of multiple sclerosis within 10 years after optic neuritis: experience of the Optic Neuritis Treatment Trial. Arch Ophthalmol 2003;121 (7) 944- 949
PubMedArticle
5.
Minneboo  ABarkhof  FPolman  CHUitdehaag  BMKnol  DLCastelijns  JA Infratentorial lesions predict long-term disability in patients with initial findings suggestive of multiple sclerosis. Arch Neurol 2004;61 (2) 217- 221
PubMedArticle
6.
Fisniku  LKBrex  PAAltmann  DR  et al.  Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain 2008;131 (pt 3) 808- 817
PubMedArticle
7.
Barkhof  FFilippi  MMiller  DH  et al.  Comparison of MRI criteria at first presentation to predict conversion to clinically definite multiple sclerosis. Brain 1997;120 (pt 11) 2059- 2069
PubMedArticle
8.
Tintoré  MRovira  AMartínez  MJ  et al.  Isolated demyelinating syndromes: comparison of different MR imaging criteria to predict conversion to clinically definite multiple sclerosis. AJNR Am J Neuroradiol 2000;21 (4) 702- 706
PubMed
9.
McDonald  WICompston  AEdan  G  et al.  Recommended diagnostic criteria for multiple sclerosis: guidelines from the international panel on the diagnosis of multiple sclerosis. Ann Neurol 2001;50 (1) 121- 127
PubMedArticle
10.
CHAMPS Study Group, MRI predictors of early conversion to clinically definite MS in the CHAMPS placebo group. Neurology 2002;59 (7) 998- 1005
PubMedArticle
11.
Korteweg  TTintoré  MUitdehaag  B  et al.  MRI criteria for dissemination in space in patients with clinically isolated syndromes: a multicentre follow-up study. Lancet Neurol 2006;5 (3) 221- 227
PubMedArticle
12.
Tintoré  MRovira  ARio  J  et al.  Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology 2006;67 (6) 968- 972
PubMedArticle
13.
Barkhof  FRocca  MFrancis  G  et al. Early Treatment of Multiple Sclerosis Study Group, Validation of diagnostic magnetic resonance imaging criteria for multiple sclerosis and response to interferon β1a. Ann Neurol 2003;53 (6) 718- 724
PubMedArticle
14.
Kappos  LPolman  CHFreedman  MS  et al.  Treatment with interferon beta-1b delays conversion to clinically definite and McDonald MS in patients with clinically isolated syndromes. Neurology 2006;67 (7) 1242- 1249
PubMedArticle
15.
Kurtzke  JF Rating neurologic impairment in multiple sclerosis: an Expanded Disability Status Scale (EDSS). Neurology 1983;33 (11) 1444- 1452
PubMedArticle
16.
Barkhof  FFilippi  Mvan Waesberghe  JH  et al.  Improving interobserver variation in reporting gadolinium-enhanced MRI lesions in multiple sclerosis. Neurology 1997;49 (6) 1682- 1688
PubMedArticle
17.
van Walderveen  MABarkhof  FHommes  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 (9) 1684- 1690
PubMedArticle
18.
Barkhof  FFilippi  Mvan Waesberghe  JHCampi  AMiller  DHAdèr  HJ Interobserver agreement for diagnostic MRI criteria in suspected multiple sclerosis. Neuroradiology 1999;41 (5) 347- 350
PubMedArticle
19.
Uitdehaag  BMKappos  LBauer  L  et al.  Discrepancies in the interpretation of clinical symptoms and signs in the diagnosis of multiple sclerosis: a proposal for standardization. Mult Scler 2005;11 (2) 227- 231
PubMedArticle
20.
Molyneux  PDTubridy  NParker  GJ  et al.  The effect of section thickness on MR lesion detection and quantification in multiple sclerosis. AJNR Am J Neuroradiol 1998;19 (9) 1715- 1720
PubMed
21.
Dolezal  ODwyer  MGHorakova  D  et al.  Detection of cortical lesions is dependent on choice of slice thickness in patients with multiple sclerosis. Int Rev Neurobiol 2007;79475- 489
PubMed
22.
Barkhof  FPolman  CHRadue  EW  et al.  Magnetic resonance imaging effects of interferon beta-1b in the BENEFIT study: integrated 2-year results. Arch Neurol 2007;64 (9) 1292- 1298
PubMedArticle
23.
Swanton  JKRovira  ATintore  M  et al.  MRI criteria for multiple sclerosis in patients presenting with clinically isolated syndromes: a multicentre retrospective study. Lancet Neurol 2007;6 (8) 677- 686
PubMedArticle
24.
Polman  CHReingold  SCEdan  G  et al.  Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald criteria.” Ann Neurol 2005;58 (6) 840- 846
PubMedArticle
25.
Bink  ASchmitt  MGaa  JMugler  JP  IIILanfermann  HZanella  FE Detection of lesions in multiple sclerosis by 2D FLAIR and single-slab 3D FLAIR sequences at 3.0 T: initial results. Eur Radiol 2006;16 (5) 1104- 1110
PubMedArticle
26.
Moraal  BRoosendaal  SDPouwels  PJ  et al.  Multi-contrast, isotropic, single-slab 3D MR imaging in multiple sclerosis. Eur Radiol 2008;18 (10) 2311- 2320
PubMedArticle
27.
Wattjes  MPLutterbey  GGHarzheim  M  et al.  Higher sensitivity in the detection of inflammatory brain lesions in patients with clinically isolated syndromes suggestive of multiple sclerosis using high field MRI: an intraindividual comparison of 1.5 T with 3.0 T. Eur Radiol 2006;16 (9) 2067- 2073
PubMedArticle
28.
Bedell  BJNarayana  PA Implementation and evaluation of a new pulse sequence for rapid acquisition of double inversion recovery images for simultaneous suppression of white matter and CSF. J Magn Reson Imaging 1998;8 (3) 544- 547
PubMedArticle
29.
Geurts  JJPouwels  PJUitdehaag  BMPolman  CHBarkhof  FCastelijns  JA Intracortical lesions in multiple sclerosis: improved detection with 3D double inversion-recovery MR imaging. Radiology 2005;236 (1) 254- 260
PubMedArticle
30.
Nelson  FPoonawalla  AHHou  PHuang  FWolinsky  JSNarayana  PA Improved identification of intracortical lesions in multiple sclerosis with phase-sensitive inversion recovery in combination with fast double inversion recovery MR imaging. AJNR Am J Neuroradiol 2007;28 (9) 1645- 1649
PubMedArticle
31.
Tan  ILvan Schijndel  RAFazekas  F  et al.  Image registration and subtraction to detect active T2 lesions in MS: an interobserver study. J Neurol 2002;249 (6) 767- 773
PubMedArticle
32.
Moraal  BMeier  DSPoppe  PA  et al.  Subtraction MR images in a multiple sclerosis multicenter clinical trial setting. Radiology 2009;250 (2) 506- 514
PubMedArticle
Original Contribution
November 2009

Magnetic Resonance Imaging Predictors of Conversion to Multiple Sclerosis in the BENEFIT Study

Author Affiliations

Author Affiliations: Departments of Diagnostic Radiology (Drs Moraal and Barkhof), Clinical Epidemiology and Biostatistics (Dr Uitdehaag), and Neurology (Drs Uitdehaag and Polman), Multiple Sclerosis Center Amsterdam, Vrije University Medical Center, Amsterdam, the Netherlands; Bayer Schering Pharma AG, Berlin (Drs Pohl, Lanius, and Sandbrink), Department of Neurology, University Hospital of Bonn, Bonn (Dr Pohl), and Heinrich-Heine University, Düsseldorf (Drs Hartung and Sandbrink), Germany; Department of Neurology, Centre Hospitalier Universitaire, Rennes, France (Dr Edan); Multiple Sclerosis Research Unit, The Ottawa Hospital, Ottawa, Ontario, Canada (Dr Freedman); Departments of Neurology and Neurosurgery, University Hospital Basel, Basel, Switzerland (Dr Kappos); Institute of Neurology, University College London, London, England (Dr Miller); and Unit of Clinica Neuroimmunology, Hospital Vall d’Hebron, Barcelona, Spain (Dr Montalban).

Arch Neurol. 2009;66(11):1345-1352. doi:10.1001/archneurol.2009.243
Abstract

Background  Several studies have confirmed the predictive value of baseline and follow-up magnetic resonance (MR) imaging variables for conversion to clinically definite multiple sclerosis (CDMS), depending on the population, follow-up duration, and treatment intervention. However, the timing of follow-up imaging and the effect of treatment intervention on the predictive value of baseline MR imaging variables require further elucidation.

Objectives  To assess the prognostic value of baseline MR imaging variables for conversion to CDMS over 3 years and whether this was affected by treatment intervention and (2) to assess the increased risk for conversion posed by dissemination in time on follow-up MR imaging.

Design  Cohort study.

Setting  Multicenter randomized clinical trial.

Patients  Four hundred sixty-eight patients with a clinically isolated syndrome who had an initial clinical demyelinating event within the past 60 days who received early treatment (3 years of interferon beta-1b) or delayed treatment (placebo first, followed by ≥1 year of interferon beta-1b).

Intervention  Magnetic resonance imaging.

Main Outcome Measure  Time to CDMS.

Results  The overall conversion rate to CDMS was 42%. Barkhof criteria with the strongest prognostic value were the presence at baseline of at least 9 T2-weighted lesions (hazard ratio [HR], 1.64; 95% confidence interval [CI], 1.15-2.33; P = .006) and at least 3 periventricular lesions (1.66; 1.14-2.41; P = .009). No specific advantage was noted in using a fixed cutoff of at least 3 Barkhof criteria (HR, 1.31; 95% CI, 0.95-1.79; P = .10). The prognostic value of all MR imaging criteria was unaffected by treatment intervention (P ≥ .20 for all). Dissemination in time resulted in increased risk for CDMS only in patients without dissemination in space at baseline and was most informative at the 9-month MR imaging (HR, 2.72; 95% CI, 1.26-5.87; P = .01).

Conclusions  The modified Barkhof criteria showed moderate predictive value for conversion to CDMS, although all patients had received interferon beta-1b therapy for at least 1 year. The predictive value was unaffected by treatment intervention. Follow-up MR imaging was most informative after 9 months in patients without dissemination in space at baseline.

In 85% to 90% of patients with multiple sclerosis (MS), the onset of symptoms occurs as a clinically isolated syndrome (CIS).1 Long-term follow-up studies of patients with a CIS have shown that the syndrome eventually converts to clinically definite MS (CDMS) in only 38% to 68% of patients2 and that abnormal magnetic resonance (MR) images at disease presentation are a powerful prognostic tool, as 56% to 88% of patients having a CIS with abnormal images progress to CDMS compared with 19% to 24% with normal images.36 In 2001, the modified Barkhof criteria7,8 were incorporated into the diagnostic McDonald criteria,9 allowing earlier diagnosis of MS based on MR imaging–derived measures. The MR imaging criteria included locations in the brain where demyelination typically occurs in MS (periventricular, juxtacortical, and infratentorial) and yielded high specificity and adequate sensitivity in predicting conversion to CDMS.

Several studies1013 have since confirmed the predictive value of baseline and follow-up MR imaging variables for conversion to CDMS, depending on the population, follow-up duration, and treatment intervention. However, the timing of follow-up imaging and the effect of treatment intervention on the predictive value of baseline MR imaging variables require further elucidation. In 2001, the Betaferon/Betaseron in Newly Emerging Multiple Sclerosis for Initial Treatment (BENEFIT) study was started, in which 468 patients with a CIS were initially randomized to interferon beta-1b therapy or placebo. The objectives of the present study were (1) to assess the prognostic value of baseline MR imaging variables for conversion to CDMS over 3 years and whether this was affected by treatment intervention and (2) to assess the increased risk for conversion posed by dissemination in time (DIT) on follow-up MR imaging.

METHODS
PATIENTS AND MR IMAGING ACQUISITION

The BENEFIT study was a multicenter, randomized, double-blind, placebo-controlled, phase 3 trial, described in more detail by Kappos et al.14 Briefly, 468 patients aged 18 to 45 years having an initial clinical demyelinating event within the past 60 days were randomized to interferon beta-1b (250 μg subcutaneously every other day) (n = 292) therapy or to placebo (n = 176) until CDMS was diagnosed or they had been followed up for 24 months. In total, 209 patients (71.6%) in the treatment group and 123 patients (69.9%) in the placebo group received corticosteroid treatment for the initial clinical event. The double-blind study period was extended by an open-label follow-up study in which all patients were offered interferon beta-1b therapy for at least 3 additional years. In total, 418 patients (261 receiving interferon beta-1b and 157 receiving placebo) entered the follow-up study, as 31 patients receiving interferon beta-1b (4 of whom had reached CDMS) and 19 patients receiving placebo (2 of whom had reached CDMS) had dropped out. In the present study, we report the integrated 3-year data for patients who received delayed treatment (placebo first, followed by interferon beta-1b after converting to CDMS or completing 24 months of follow-up) or early treatment (interferon beta-1b for 3 years from the start of the double-blind study). Magnetic resonance imaging criteria for study entry were at least 2 clinically silent lesions of at least 3 mm on T2-weighted images, at least 1 of which was ovoid, periventricular, or infratentorial. The main outcome measure was time to CDMS, defined as a validated second clinical demyelinating event or as progression of at least 1.5 points on the Expanded Disability Status Scale.15

Baseline MR images were obtained as part of the prestudy screening using an oblique axial acquisition plane and gadolinium diethylenetriamine pentaacetic acid (0.1 mmol/kg of body weight) as a contrast agent. The MR imaging protocol included a T1-weighted spin echo before and after contrast administration with repetition times of 400 to 700 milliseconds and echo times of 5 to 25 milliseconds, in addition to a dual-echo T2-weighted spin echo with repetition times of 2000 to 3000 milliseconds and echo times of 20 to 40 milliseconds (first echo) and 60 to 100 milliseconds (second echo). The field of view for all examinations was 25 cm with a 256 × 256–pixel matrix, resulting in a pixel size of roughly 1 × 1 mm2. Images were acquired in 2 interleaved sets with a 3-mm gap, resulting in whole-brain coverage with contiguous 3-mm-thick sections.

IMAGE ANALYSIS

Images were centrally evaluated by the Image Analysis Center, Amsterdam, the Netherlands, by raters blinded to treatment allocation. Gadolinium-enhancing lesions were identified according to criteria described by Barkhof et al.16 Hypointense T1-weighted lesions (black holes) were identified after contrast administration according to criteria derived by van Walderveen et al,17 and hyperintense T2-weighted lesions were identified using short- and long-echo T2-weighted images according to the criteria described by Barkhof et al.18 Supratentorial T2-weighted lesions were subclassified as periventricular lesions (in direct contact with the ventricular system), juxtacortical lesions (directly abutting the cortical gray matter), or deep white matter lesions (in neither of the other 2 subclasses). Infratentorial lesions comprised lesions in the brainstem and cerebellum.

STATISTICAL ANALYSIS

Descriptive statistics (number and percentage) of patients fulfilling specific baseline MR imaging variables were calculated for each treatment arm. For time to CDMS, Kaplan-Meier estimates were calculated beginning at study randomization, and corresponding survival curves were plotted. For patients who converted to CDMS within 1080 days, the number and proportion (Kaplan-Meier estimates) were calculated for all dichotomized baseline MR imaging variables, including the cumulative number of positive criteria and the modified Barkhof criteria (≥1 gadolinium-enhancing lesions or ≥9 T2-weighted lesions, ≥1 infratentorial lesions, ≥1 juxtacortical lesions, and ≥3 periventricular lesions). Cox proportional hazards regression models for all patients were used to assess the effect of each baseline MR imaging variable on time to CDMS (adjusted for age, sex, treatment intervention, corticosteroid use, clinical presentation [optic neuritis vs other], and disease onset type [monofocal vs multifocal as described by Uitdehaag et al19]) by calculating the hazard ratios (HRs) (95% confidence intervals [CIs]) and corresponding P values. Possible interactions with treatment regimen (early vs delayed) and disease onset type (monofocal vs multifocal) were explored using Cox proportional hazards models (adjusted for treatment intervention, disease onset type, and MR imaging by specific interaction term). Furthermore, Cox proportional hazards models were applied to evaluate the increased risk for conversion to CDMS among patients developing new lesions by 3, 6, and 9 months (all compared with baseline MR images) and were stratified for patients with and without dissemination in space (DIS) at baseline. The presence of at least 1 new lesion on follow-up MR imaging is referred to as DIT. In general, P < .05 was considered statistically significant; however, because all statistical analyses were exploratory in nature, P values were not corrected for multiple testing and should be interpreted descriptively. All statistical calculations were performed using commercially available software (SAS version 9.1.3; SAS Institute Inc, Cary, North Carolina; or SPSS version 12.0; SPSS Inc, Chicago, Illinois).

RESULTS
BASELINE MR IMAGING CHARACTERISTICS

At baseline, 64.3% of all patients had DIS (≥3 Barkhof criteria). These results are summarized in Table 1.

PREDICTIVE VALUE OF BASELINE MR IMAGING VARIABLES FOR CONVERSION TO CDMS

After 3 years, the overall conversion rate to CDMS was 42% (51% for patients receiving delayed treatment and 37% for patients receiving early treatment). Table 2 gives the number and percentage of patients converting to CDMS per baseline MR imaging characteristic. Strong Barkhof criteria predictors of conversion to CDMS were the presence at baseline of at least 9 T2-weighted lesions (46% vs 31% without; HR, 1.64, 95% CI, 1.15-2.33; P = .006) and at least 1 gadolinium-enhancing lesion (48% vs 37% without; 1.56; 1.15-2.10; P = .004). The predictive value of gadolinium-enhancing lesions increased slightly with the presence of more enhancing lesions (54% for ≥2 gadolinium-enhancing lesions vs 40% without; HR, 1.74; 95% CI, 1.26-2.41; P = .001). Furthermore, there was predictive value in the spatial distribution of T2-weighted lesions at baseline, as patients with at least 3 periventricular lesions were more likely to convert to CDMS (46% vs 29% without; HR, 1.66; 95% CI, 1.14-2.41; P = .009). Patients fulfilling more Barkhof criteria at baseline were more likely to convert to CDMS than patients fulfilling few Barkhof criteria at baseline (Table 3 and Figure 1). Because the numbers of patients fulfilling 0 or 1 Barkhof criterion are few in the BENEFIT study (Table 1), we only report the comparisons that produced more evenly balanced subgroup sizes. The McDonald criteria cutoff for DIS (≥3 criteria fulfilled) did not show significantly greater predictive value than fulfillment of 1 to 2 criteria (45% vs 37% conversion to CDMS; HR, 1.31; 95% CI, 0.95-1.79; P = .10) (Table 3). Patients fulfilling at least 4 criteria at baseline showed higher conversion rates than patients fulfilling 1 to 3 criteria (48% vs 39%; HR, 1.32; 95% CI, 0.97-1.78; P = .08). Significant covariates for conversion to CDMS in all models (in addition to treatment regimen) were age (P < .001 for all) and corticosteroid use (range, P = .02 to .03) at the initial clinical demyelinating event, whereas no other covariates (sex, clinical presentation, or disease onset type) were significant in any model (range, P = .41 to .96). Data regarding conversion to CDMS per the McDonald criteria are given in the (eAppendix.

PREDICTIVE VALUE OF FOLLOW-UP MR IMAGING VARIABLES FOR CONVERSION TO CDMS

In patients with DIS on MR imaging at baseline (and without CDMS at the time of follow-up MR imaging), DIT on MR imaging was not associated with increased risk of conversion to CDMS regardless of the timing of follow-up MR imaging (P > .18 for all) (Table 4). In patients without DIS on MR imaging at baseline, DIT on MR imaging was associated with increased risk of conversion to CDMS at the 9-month MR imaging (HR, 2.72; 95% CI, 1.26-5.87; P = .01) (Figure 2) and at the 6-month MR imaging (2.24; 1.34-4.42; P = .02) but not at the 3-month MR imaging (1.25; 0.69-2.25; P = .46).

IS THE PROGNOSTIC VALUE OF BASELINE MR IMAGING VARIABLES FOR CONVERSION TO CDMS AFFECTED BY TREATMENT INTERVENTION OR DISEASE ONSET TYPE?

The predictive value of all baseline MR imaging variables for conversion to CDMS was unaffected by disease onset type (P ≥ .24 for all). Similarly, it was unaffected by treatment intervention (P ≥ .20 for all).

COMMENT

In this study, we examined data from a large clinical trial (the BENEFIT study) to evaluate the predictive value of baseline and follow-up MR imaging variables in patients with a CIS. This had certain advantages. First, the population of patients with a CIS studied is one of the largest so far (N = 468) and was recruited from 98 centers worldwide. Second, the MR imaging protocol comprised standardized acquisition variables, which, combined with centralized analysis, prevented center bias in interpretation of the MR imaging data. Third, considering the use of 5-mm-thick MR imaging sequences in previous clinical trials,10,13 the 3-mm thickness used in this study allowed for ample detection of MS lesions in the brain, as lesion detection increases with decreasing section thickness.20,21 It is important to consider the design of the BENEFIT study when comparing the results of our study with those of other large patient cohorts having a CIS. The trial started with a double-blind phase of 2 years in which all patients were randomized to placebo or treatment, followed by an open-label phase in which all patients were offered interferon beta-1b therapy. Our study reports the integrated 3-year data, meaning that all patients received at least 1 year of open-label treatment. This likely lowered the overall conversion rates to the clinical end point (CDMS) and to the radiologic end point (McDonald criteria for MS), as interferon beta-1b exerts an inhibitory effect on the occurrence of a second clinical demyelinating event13 and on the formation of new lesions.22 An overall lower event rate will lead to a lower predictive value for any prognostic variable studied.

The percentage of patients converting to the clinical end point (CDMS) was higher in the presence (compared with the absence) of all baseline MR imaging variables. This difference was significant for 2 of 4 Barkhof criteria, namely, the presence of at least 3 periventricular lesions and the presence of gadolinium enhancement or at least 9 T2-weighted lesions, which is consistent with results from the Early Treatment of Multiple Sclerosis Study (ETOMS).13 However, we found no significant increased risk associated with the presence of 1 or more infratentorial lesions, nor was there any specific advantage in using a fixed cutoff of at least 3 Barkhof criteria. In accord with the CHAMPS (Controlled High-Risk Subjects Avonex Multiple Sclerosis Prevention Study) data set,10 the presence of at least 2 gadolinium-enhancing lesions was a strong predictor of conversion to CDMS, although the predictive value was less distinct than that of the presence of at least 1 gadolinium-enhancing lesion. These conflicting results might be explained by the different trial designs, particularly the variations in inclusion criteria or population studied, treatment regimen, and follow-up duration. Furthermore, the present study used Cox proportional hazards models (with age, treatment intervention, and corticosteroid use as significant covariates in all models), while other analyses did not use covariate adjustment (eg, in the ETOMS data set). The BENEFIT data HRs were 1.31 (95% CI, 0.95-1.79; P = .10) for at least 3 Barkhof criteria and 1.32 (0.97-1.78; P = .08) for at least 4 Barkhof criteria. Analysis with the same statistical method as that used in the ETOMS data set over 2 years of follow-up results in substantially different HRs and P values, namely, 1.37 (95% CI, 0.97-1.94; P = .08) for at least 3 Barkhof criteria and 1.48 (1.08-2.04; P = .02) for at least 4 Barkhof criteria, showing a significant result for the latter. This indicates that covariate adjustment can have a substantial effect on the results of the statistical analyses.

Dissemination in time was associated with a significantly increased risk for conversion to CDMS only in patients without DIS at baseline (Figure 2). This suggests that follow-up MR imaging is most informative in patients without DIS at baseline and is less so in patients with DIS at baseline. This might be reflective of the interrelationship between DIS, DIT, and CDMS (ie, patients with DIS at baseline already have a higher risk of conversion to CDMS than patients without DIS, and the additional risk of new lesions on follow-up MR imaging might be lower). Clinical variables might influence this interrelation, although in this study both clinical presentation (optic neuritis vs other) and disease onset type (monofocal vs multifocal) showed no effect on time to CDMS. Furthermore, in patients without DIS (and without CDMS up to the respective time point), follow-up MR imaging to determine DIT was associated with increased risk for CDMS, most notably at the 9-month MR imaging and at the 6-month MR imaging but not at the 3-month MR imaging. Based on these data, follow-up MR imaging to determine DIT might best be performed after 9 months in patients without DIS at baseline; however, these results were not corrected for multiple comparisons and should be interpreted descriptively.

We also analyzed whether the predictive value of baseline MR imaging variables was affected by treatment regimen (early treatment vs delayed treatment) by calculating treatment intervention × MR imaging variable interaction. For conversion to CDMS, no variables were statistically significant, indicating that the predictive power of baseline MR imaging variables was unaffected by early treatment. As such, our data are consistent with results from the ETOMS data set, which also reported a lack of power among all MR imaging variables in predicting treatment response.13 Alternatively, treatment interactions can be interpreted as measures to identify which MR imaging variables predict treatment response (ie, which patients will benefit the most from early treatment). For conversion according to the McDonald criteria for MS, the predictive value of 4 Barkhof criteria was affected by treatment intervention, indicating that this MR imaging measure predicted treatment response.

Our study had several potential limitations. First, patients in our population may not be representative of those seen in daily practice, as only patients with abnormal brain MR images were included, and other neurologic diseases were ruled out by appropriate tests such as cerebrospinal fluid (CSF) analysis or laboratory measures. This is reflected by a slightly higher overall conversion rate to CDMS of 42% over 36 months compared with an overall conversion rate of 32.5% after 85.3 months in a large (N = 532) population of patients with a CIS at a less restricted referral center.11 Second, we did not investigate the full range of options in the McDonald criteria to ascertain DIS (eg, we did not investigate the combination of positive CSF status and 2 lesions on brain MR imaging). Furthermore, we did not include spinal cord MR imaging, although a recent study23 showed that (compared with brain MR imaging only) the increased value of spinal cord imaging might be limited, adding only 1% to the sensitivity and specificity of the 2001 McDonald criteria to predict CDMS. Third, we used the more conservative 2001 McDonald criteria9 (because the BENEFIT study was initiated in 2001) rather than the revised 2005 criteria.24 However, the criteria for lesions on brain MR imaging were unchanged; hence, application of the 2005 revisions would not have altered the results of the present study.

Since introduction of the McDonald criteria in 2001, the diagnosis of MS in patients with a CIS has been made largely on radiologic grounds. New imaging techniques that cross sectionally and longitudinally detect more MS lesions in the brain will unequivocally have an effect on the number of patients fulfilling DIS and DIT criteria, respectively. Recent studies have shown increased detection of MS lesions in various parts of the brain using 3-dimensional (single-slab) sequences25,26 and higher field strength (3.0 T).27 Moreover, specific image contrasts such as double inversion recovery and (phase-sensitive) T1-weighted imaging are more sensitive than conventional methods in visualizing intracortical lesions.26,2830 Furthermore, image subtraction has been shown to significantly increase the detection of new T2-weighted lesions31,32 vs standard comparison of nonregistered images. Future work should examine the precise effects of new imaging sequences, contrasts, and hardware on the predictive value of MR imaging for conversion to CDMS.

In conclusion, MR imaging predictors of conversion to CDMS were assessed in a large population (N = 468) of patients having a CIS with abnormalities on brain MR imaging. In general, the percentage of patients converting to CDMS was higher in the presence (compared with the absence) of all baseline MR imaging variables. In particular, the modified Barkhof criteria showed moderate predictive value for conversion to CDMS, although all patients were treated for at least 1 year with interferon beta-1b. The predictive value of MR imaging variables was unaffected by treatment intervention. Follow-up MR imaging to determine DIT was most informative after 9 months in patients without DIS at baseline.

Back to top
Article Information

Correspondence: Bastiaan Moraal, MD, Department of Diagnostic Radiology, Multiple Sclerosis Center Amsterdam, Vrije University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, the Netherlands (b.moraal@vumc.nl).

Accepted for Publication: April 7, 2009.

Author Contributions:Study concept and design: Moraal, Pohl, Polman, Edan, Hartung, Kappos, Miller, Montalban, Sandbrink, and Barkhof. Acquisition of data: Pohl, Sandbrink, and Barkhof. Analysis and interpretation of data: Moraal, Pohl, Uitdehaag, Polman, Freedman, Hartung, Kappos, Miller, Lanius, Sandbrink, and Barkhof. Drafting of the manuscript: Moraal, Polman, and Freedman. Critical revision of the manuscript for important intellectual content: Moraal, Pohl, Uitdehaag, Polman, Edan, Freedman, Hartung, Kappos, Miller, Montalban, Lanius, Sandbrink, and Barkhof. Statistical analysis: Moraal, Uitdehaag, and Lanius. Obtained funding: Polman and Sandbrink. Administrative, technical, and material support: Moraal and Sandbrink. Study supervision: Uitdehaag, Polman, Edan, Freedman, Hartung, Montalban, Sandbrink, and Barkhof.

Financial Disclosure: Dr Polman has received consulting fees from Actelion, Biogen Idec, Bayer Schering Pharma AG, Teva, Merck-Serono, Novartis, GlaxoSmithKline, UCB, Roche, and Antisense Therapeutics; lecture fees from Biogen Idec, Bayer Schering Pharma AG, Novartis, and Teva; and grant support from Biogen Idec, Bayer Schering Pharma AG, GlaxoSmithKline, Novartis, UCB, Merck-Serono, and Teva. Dr Freedman has received honoraria from Bayer Schering Pharma AG, Biogen Idec, Bioms, Teva, Merck-Serono, Novartis, Sanofi-Aventis, and Pfizer and has served as a consultant on steering committees for Bayer Schering Pharma AG, Merck-Serono, Bioms, Sanofi-Aventis, and Novartis. Dr Hartung has received compensation for serving on a scientific advisory board and honoraria for lectures or consultancies from Biogen Idec, Teva, Sanofi-Aventis, Octapharma GmbH, Novartis, CSL Behring GmbH, and Merck-Serono. Dr Miller has received research grant support through his employer, University College London Institute of Neurology, for performing central MR imaging analysis in MS trials from GlaxoSmithKline, Biogen Idec, and Novartis; he has received honoraria through payments to University College London Institute of Neurology for advisory committee or consultancy advice in MS studies from Biogen Idec, GlaxoSmithKline, and Bayer Schering Pharma AG. Dr Miller served on the steering committee of the trial reported herein. Dr Montalban has received honoraria for speaking at scientific meetings and grants to travel to scientific events from Bayer Schering Pharma AG, Teva, Merck-Serono, Sanofi-Aventis, and Novartis; he has been involved in steering committees for Teva, Merck-Serono, Novartis, and Almirall. Dr Sandbrink is a salaried employee of Bayer Schering Pharma AG, which sponsored the clinical study reported herein, and he owns stock options in that company. Dr Barkhof has received compensation for consultancy or committee membership or lecture fees from Bayer Schering Pharma AG, Merck-Serono, Sanofi-Aventis, AstraZeneca, Genentech, Novartis, Biogen Idec, Active Biotech, European Charcot Foundation, Lundbeck, Talecris, Roche, UCB, Wyeth, and Medicinova Inc.

Funding/Support: The BENEFIT trial is sponsored by Bayer Schering Pharma AG. The Multiple Sclerosis Center Amsterdam is supported by the Dutch Multiple Sclerosis Research Foundation (Dr Moraal).

Additional Contributions: Cosima Klein helped with statistical programming, and Lars Bauer and Karola Beckmann assisted with the study. We thank the study investigators, the Image Analysis Center, and the patients with MS who participated in the study.

References
1.
Miller  DBarkhof  FMontalban  XThompson  AFilippi  M Clinically isolated syndromes suggestive of multiple sclerosis, part I: natural history, pathogenesis, diagnosis, and prognosis. Lancet Neurol 2005;4 (5) 281- 288
PubMedArticle
2.
Poser  CMPaty  DWScheinberg  L  et al.  New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13 (3) 227- 231
PubMedArticle
3.
Brex  PACiccarelli  OO’Riordan  JISailer  MThompson  AJMiller  DH A longitudinal study of abnormalities on MRI and disability from multiple sclerosis. N Engl J Med 2002;346 (3) 158- 164
PubMedArticle
4.
Beck  RWTrobe  JDMoke  PS  et al. Optic Neuritis Study Group, High- and low-risk profiles for the development of multiple sclerosis within 10 years after optic neuritis: experience of the Optic Neuritis Treatment Trial. Arch Ophthalmol 2003;121 (7) 944- 949
PubMedArticle
5.
Minneboo  ABarkhof  FPolman  CHUitdehaag  BMKnol  DLCastelijns  JA Infratentorial lesions predict long-term disability in patients with initial findings suggestive of multiple sclerosis. Arch Neurol 2004;61 (2) 217- 221
PubMedArticle
6.
Fisniku  LKBrex  PAAltmann  DR  et al.  Disability and T2 MRI lesions: a 20-year follow-up of patients with relapse onset of multiple sclerosis. Brain 2008;131 (pt 3) 808- 817
PubMedArticle
7.
Barkhof  FFilippi  MMiller  DH  et al.  Comparison of MRI criteria at first presentation to predict conversion to clinically definite multiple sclerosis. Brain 1997;120 (pt 11) 2059- 2069
PubMedArticle
8.
Tintoré  MRovira  AMartínez  MJ  et al.  Isolated demyelinating syndromes: comparison of different MR imaging criteria to predict conversion to clinically definite multiple sclerosis. AJNR Am J Neuroradiol 2000;21 (4) 702- 706
PubMed
9.
McDonald  WICompston  AEdan  G  et al.  Recommended diagnostic criteria for multiple sclerosis: guidelines from the international panel on the diagnosis of multiple sclerosis. Ann Neurol 2001;50 (1) 121- 127
PubMedArticle
10.
CHAMPS Study Group, MRI predictors of early conversion to clinically definite MS in the CHAMPS placebo group. Neurology 2002;59 (7) 998- 1005
PubMedArticle
11.
Korteweg  TTintoré  MUitdehaag  B  et al.  MRI criteria for dissemination in space in patients with clinically isolated syndromes: a multicentre follow-up study. Lancet Neurol 2006;5 (3) 221- 227
PubMedArticle
12.
Tintoré  MRovira  ARio  J  et al.  Baseline MRI predicts future attacks and disability in clinically isolated syndromes. Neurology 2006;67 (6) 968- 972
PubMedArticle
13.
Barkhof  FRocca  MFrancis  G  et al. Early Treatment of Multiple Sclerosis Study Group, Validation of diagnostic magnetic resonance imaging criteria for multiple sclerosis and response to interferon β1a. Ann Neurol 2003;53 (6) 718- 724
PubMedArticle
14.
Kappos  LPolman  CHFreedman  MS  et al.  Treatment with interferon beta-1b delays conversion to clinically definite and McDonald MS in patients with clinically isolated syndromes. Neurology 2006;67 (7) 1242- 1249
PubMedArticle
15.
Kurtzke  JF Rating neurologic impairment in multiple sclerosis: an Expanded Disability Status Scale (EDSS). Neurology 1983;33 (11) 1444- 1452
PubMedArticle
16.
Barkhof  FFilippi  Mvan Waesberghe  JH  et al.  Improving interobserver variation in reporting gadolinium-enhanced MRI lesions in multiple sclerosis. Neurology 1997;49 (6) 1682- 1688
PubMedArticle
17.
van Walderveen  MABarkhof  FHommes  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 (9) 1684- 1690
PubMedArticle
18.
Barkhof  FFilippi  Mvan Waesberghe  JHCampi  AMiller  DHAdèr  HJ Interobserver agreement for diagnostic MRI criteria in suspected multiple sclerosis. Neuroradiology 1999;41 (5) 347- 350
PubMedArticle
19.
Uitdehaag  BMKappos  LBauer  L  et al.  Discrepancies in the interpretation of clinical symptoms and signs in the diagnosis of multiple sclerosis: a proposal for standardization. Mult Scler 2005;11 (2) 227- 231
PubMedArticle
20.
Molyneux  PDTubridy  NParker  GJ  et al.  The effect of section thickness on MR lesion detection and quantification in multiple sclerosis. AJNR Am J Neuroradiol 1998;19 (9) 1715- 1720
PubMed
21.
Dolezal  ODwyer  MGHorakova  D  et al.  Detection of cortical lesions is dependent on choice of slice thickness in patients with multiple sclerosis. Int Rev Neurobiol 2007;79475- 489
PubMed
22.
Barkhof  FPolman  CHRadue  EW  et al.  Magnetic resonance imaging effects of interferon beta-1b in the BENEFIT study: integrated 2-year results. Arch Neurol 2007;64 (9) 1292- 1298
PubMedArticle
23.
Swanton  JKRovira  ATintore  M  et al.  MRI criteria for multiple sclerosis in patients presenting with clinically isolated syndromes: a multicentre retrospective study. Lancet Neurol 2007;6 (8) 677- 686
PubMedArticle
24.
Polman  CHReingold  SCEdan  G  et al.  Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald criteria.” Ann Neurol 2005;58 (6) 840- 846
PubMedArticle
25.
Bink  ASchmitt  MGaa  JMugler  JP  IIILanfermann  HZanella  FE Detection of lesions in multiple sclerosis by 2D FLAIR and single-slab 3D FLAIR sequences at 3.0 T: initial results. Eur Radiol 2006;16 (5) 1104- 1110
PubMedArticle
26.
Moraal  BRoosendaal  SDPouwels  PJ  et al.  Multi-contrast, isotropic, single-slab 3D MR imaging in multiple sclerosis. Eur Radiol 2008;18 (10) 2311- 2320
PubMedArticle
27.
Wattjes  MPLutterbey  GGHarzheim  M  et al.  Higher sensitivity in the detection of inflammatory brain lesions in patients with clinically isolated syndromes suggestive of multiple sclerosis using high field MRI: an intraindividual comparison of 1.5 T with 3.0 T. Eur Radiol 2006;16 (9) 2067- 2073
PubMedArticle
28.
Bedell  BJNarayana  PA Implementation and evaluation of a new pulse sequence for rapid acquisition of double inversion recovery images for simultaneous suppression of white matter and CSF. J Magn Reson Imaging 1998;8 (3) 544- 547
PubMedArticle
29.
Geurts  JJPouwels  PJUitdehaag  BMPolman  CHBarkhof  FCastelijns  JA Intracortical lesions in multiple sclerosis: improved detection with 3D double inversion-recovery MR imaging. Radiology 2005;236 (1) 254- 260
PubMedArticle
30.
Nelson  FPoonawalla  AHHou  PHuang  FWolinsky  JSNarayana  PA Improved identification of intracortical lesions in multiple sclerosis with phase-sensitive inversion recovery in combination with fast double inversion recovery MR imaging. AJNR Am J Neuroradiol 2007;28 (9) 1645- 1649
PubMedArticle
31.
Tan  ILvan Schijndel  RAFazekas  F  et al.  Image registration and subtraction to detect active T2 lesions in MS: an interobserver study. J Neurol 2002;249 (6) 767- 773
PubMedArticle
32.
Moraal  BMeier  DSPoppe  PA  et al.  Subtraction MR images in a multiple sclerosis multicenter clinical trial setting. Radiology 2009;250 (2) 506- 514
PubMedArticle
×