You are here

Article Text

Article menu

Download PDFPDF

Multiple sclerosis
Research paper
A longitudinal study of cortical grey matter lesion subtypes in relapse-onset multiple sclerosis
  1. Varun Sethi1,
  2. Tarek Yousry1,2,
  3. Nils Muhlert3,
  4. Daniel J Tozer4,
  5. Daniel Altmann1,5,
  6. Maria Ron1,
  7. Claudia Wheeler-Kingshott1,
  8. David H Miller1,6,
  9. Declan T Chard1,6
  1. 1NMR Research Unit, Department of Neuroinflammation, Queen Square MS Centre, UCL Institute of Neurology, London, UK
  2. 2Lysholm Department of Neuroradiology, National Hospital for Neurology and Neurosurgery, London, UK
  3. 3School of Psychology and Cardiff University Brain Research Imaging Centre, Cardiff University, Cardiff, UK
  4. 4Department of Clinical Neurosciences, University of Cambridge, Cambridge, UK
  5. 5Medical Statistics Department, London School of Hygiene & Tropical Medicine, London, UK
  6. 6National Institute for Health Research (NIHR) University College London Hospitals (UCLH) Biomedical Research Centre, London, UK
  1. Correspondence to Dr Declan T Chard, Department of Neuroinflammation, Queen Square Multiple Sclerosis Centre, UCL Institute of Neurology, University College London, Queen Square, London, WC1N 3BG, UK; d.chard{at}ucl.ac.uk

Abstract

Background Cortical grey matter (GM) lesions are common in multiple sclerosis (MS), but little is known about their temporal evolution. We investigated this in people with relapsing–remitting (RR) and secondary progressive (SP) MS.

Methods 27 people with RRMS, and 22 with SPMS were included in this study. Phase-sensitive inversion recovery scans were acquired on 2 occasions. Cortical GM lesions were classified as intracortical (IC, only involving GM) and leucocortical (LC, mixed GM–white matter (WM)); WM lesions touching the cortex as juxtacortical (JC). On follow-up scans, new IC, LC and JC lesions were identified, and any change in classification of lesions previously observed was noted. WM lesion counts in the whole brain were assessed on PD/T2-weighted scans.

Results Over a mean (SD) of 21.0 (5.8) months, the number of new IC lesions per person per year was greater in SPMS (1.6 (1.9)) than RRMS (0.8 (1.9)) (Mann-Whitney p=0.039). All new LC lesions arose from previously seen IC lesions (SPMS 1.4 (1.8) per person per year, and RRMS 1.1 (1.0)), and none arose de novo, or from previously seen JC lesions. Changes in cortical GM (either new IC or IC converting to LC) lesion counts did not correlate with the changes in WM lesion counts.

Conclusions New cortical GM lesions rarely arise from the WM and the rate of new IC lesion formation is not closely linked with WM lesion accrual. IC lesion formation appears to be more common in SPMS than RRMS.

https://doi.org/10.1136/jnnp-2015-311102

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Histopathological1–4 and MRI studies5 ,6 have established that cortical grey matter (GM) lesions can be extensive and are clinically relevant in multiple sclerosis (MS). Cortical GM lesions are seen in all MS subtypes,1 ,2 but are particularly apparent in people with progressive MS.1 ,7 Cortical lesions play a significant role in the accumulation of irreversible disability.5

    Little is known about the development and evolution of cortical GM lesions.8–10 It is uncertain whether lesions confined to cortical GM (intracortical (IC) lesions) accrue at the same rate as leucocortical (LC) lesions (those involving both GM and white matter (WM)) or WM lesions, or whether the rate of cortical GM lesion formation is similar throughout the course of relapsing–remitting (RR) and secondary progressive (SP) MS. It is also not known if LC lesions form de novo, or through the extension of IC lesions into WM, or juxtacortical (JC) WM lesions in to GM. Cortical GM demyelination has been linked with meningeal inflammation, providing a plausible mechanism for cortical GM lesion formation to occur independently of WM demyelination.11 and so it cannot be assumed that an effect MS disease modifying treatments have on WM lesion accrual will be similar for GM lesions.

    Investigating the evolution of cortical GM lesions in histopathological studies is not possible, as serial tissue sampling cannot be undertaken. Investigating lesion evolution in vivo with MRI has been hampered by limited detection of GM lesions using conventional scanning methods. However, with the development of MRI techniques that improve the detection of GM lesions, such as double inversion recovery (DIR)12 ,13 and recent phase-sensitive inversion recovery (PSIR),14 ,15 it is now possible to see GM lesions in nearly everyone with MS. Compared with DIR, PSIR appears to increase GM lesion detection 2–3 times and, importantly, allows cortical GM lesions to be more robustly distinguished from JC WM lesions, and subclassified into IC and LC lesions.15 ,16 Using a high-resolution PSIR sequence, we studied the formation and evolution of cortical GM lesions on a lesion-by-lesion basis, and examined their relationship with WM lesion accrual.

    Methods

    Sixty-five people took part in this study; all gave written informed consents. This study had approval from our local institutional ethics committee.

    Data from three people with RRMS had to be excluded due to MRI motion artefacts; of the remaining 62 people, 27 had RRMS, 22 SPMS and 13 were healthy controls (table 1). For the MS groups, a detailed clinical history was obtained and neurological examination undertaken. Disease duration, number of relapses since the baseline visit, and change in disease modifying therapy were all noted. Expanded disability status scale (EDSS) scores were estimated.17

    View this table:
    Table 1

    Participant demographics

    PSIR (0.5×0.5×2 mm) and PD/T2-weighted (1×1×3 mm) scans were acquired (as per our previous work15) on a 3 T Philips Achieva system, on two occasions at least 1 year apart. Follow-up images were registered to baseline scans using NiftyReg (http://sourceforge.net/projects/niftyreg)18 with an affine transformation. On baseline scans, cortical GM lesions were classified as IC or LC; WM lesions touching but not entering the cortex were classified as JC lesions. As in our previous work,15 a cortical GM lesion was defined as a focal hypointensity relative to the surrounding normal cortex that involved the cortex in part or whole. When small or ill-defined on a single slice, a lesion was only counted if it was present on at least one other adjacent slice. If a lesion was only seen to involve cortical GM it was classified as being IC, and if it also involved WM it was classified as LC. Total cortical GM lesion counts were calculated as the sum of IC and LC lesion counts. On follow-up scans, new cortical GM lesions were counted, and on a lesion-by-lesion basis any change in classification of lesions previously identified in or adjacent to the cortex (eg, from IC or JC to LC) was noted. The intraoperator intraclass correlation coefficient, using this PSIR sequence and these criteria, was 0.995 for total cortical GM lesion counts (IC and LC), and for IC and LC lesions separately, 0.990 and 0.984, respectively.15 The PD/T2-weighted scans were used to count new WM lesions appearing over the follow-up interval. All lesion marking was undertaken by VS under the supervision of an experienced neuroradiologist (TY) using guidelines previously reported.15 Marking of all scans was blind to clinical details.

    Statistical analyses were performed using SPSS V.21 (SPSS, Chicago, Illinois, USA). Differences in demographic and clinical variables between RR and SPMS were examined using independent sample t tests. Since lesion numbers tend not to be normally distributed, changes in lesion numbers were compared between patient groups using the Mann-Whitney test. Spearman rank correlation was used to investigate associations between GM and WM lesion accrual. The threshold for statistical significance was p≤0.05.

    Results

    The demographic and clinical details of the participants are shown in table 1. As expected, people with SPMS were older (mean 53.4 (standard deviation (SD) 7.3) compared with 41.7 (10.9) years; p<0.001) and had longer disease durations than RRMS (25.4 (10.0) compared with 13.1 (9.5) years; p<0.001). The mean (SD) follow-up period for the RR group was 22.2 (6.3) months, and for SP 19.5 (4.8) months, which was not significantly different. The median EDSS for people with RRMS was 1.0 at baseline (range 1.0–6.0) and 1.5 (range 1.0–6.0) at follow-up, and for SPMS was 6.5 (range 4.0–8.5) at both time points. At baseline, 23 RRMS and 9 of the SPMS group were on a disease modifying treatment, and during follow-up, disease modifying therapy was started in one person and changed in five people with RRMS.

    Baseline lesion characteristics

    Baseline and follow-up lesion counts are given in table 2. While the total number of cortical GM lesions (IC and LC combined) was higher in the SP compared with RRMS groups (46.6 (19.4) compared with 36.8 (14.8), respectively), this did not reach statistical significance. IC, LC and JC lesion counts (table 2) also did not differ significantly between the RR and SP groups.

    View this table:
    Table 2

    Number of cortical lesions seen at baseline and follow-up

    Lesion accrual

    At follow-up, the SPMS group had a statistically higher total number of cortical GM lesions when compared with the RRMS group (mean 49.0 (19.7) and 38.0 (15.7), respectively; Mann-Whitney p=0.043), but again, there was no statistically significant difference between the groups for IC, LC or JC lesion counts. All lesions seen at baseline remained visible at follow-up. The accrual of new IC lesions per person per year was greater in SPMS (1.6 (1.9)) than RRMS (0.8 (1.9)) (Mann-Whitney p=0.039). No de novo JC or LC lesions were seen, and new LC lesions were seen to evolve from IC lesions but not JC lesions (table 2). An example of an IC lesion converting to LC is shown in figure 1.

    Figure 1
    Figure 1

    An intracortical lesion evolving into a leucocortical lesion. Corresponding PSIR images at baseline (left) and follow-up (right) showing an intracortical lesion at baseline that extends into white matter on follow-up, so becoming a leucocortical lesion.

    The mean number of lesions converting from IC (at baseline) to LC lesions (at follow-up) per person per year was not significantly greater in SPMS than RRMS (1.4 (1.8) compared with 1.1 (1.0), respectively). No new cortical lesions were seen in controls.

    WM lesion accrual per year was not significantly different in RRMS and SPMS groups (mean (SD): 0.7 (1.3)) compared with 0.2 (0.4), respectively). No new WM lesions were seen in controls. In the entire MS group, there was no significant correlation between the number of new WM lesions and the number of lesions converting from IC to LC, or the number of new IC lesions. Similar results were seen in the RRMS and SPMS subgroups, and so are not shown.

    Discussion

    The most striking finding in our study was that new LC lesions (lesions that involve both cortical GM and adjacent (JC) WM) evolved from pre-existing IC lesions (which are confined to GM), and not from JC WM lesions or de novo. We also found significantly more new IC lesions in SPMS than RRMS and, consistent with previous work10 we found that changes in IC lesion numbers did not correlate with changes in WM lesion counts. The latter finding suggests that cortical GM lesion accrual is not closely linked with WM lesion formation.

    In a previous study using DIR scans to identify and count cortical GM lesions, Calabrese et al5 followed up a mixed cohort of people with MS over 6 years. When assessed at 3 years,10 the average rate of new cortical lesions was approximately 0.6 per person per year in RRMS (n=76) and approximately 0.7 in SPMS (n=31). At 5 years,5 the new cortical GM lesion accrual was 0.6 per person per year in RRMS (n=157) and SPMS (n=31), and the rate of new WM lesions was approximately 0.8 and 0.3 per person per year in RRMS and SPMS groups, respectively. At 6 years,10 in RRMS, the mean rate of cortical lesion accrual was 0.9 per person per year (n=95), but SPMS figures were not reported. Roosendaal et al9 also assessed cortical GM lesion formation in 13 people with RRMS (=9) and SPMS (n=4) over 3 years, but did not report on lesion accrual in these groups separately.

    In the present study, new cortical GM lesions formed at a rate of approximately 0.8 in RRMS and approximately 1.6 per person per year in SPMS, and WM lesions at a rate of approximately 0.7 and 0.2 per person per year, respectively. With the exception of new cortical GM lesion accrual in SPMS, our results are very similar to those of Calabrese et al. With regard to the rate of cortical GM lesion accumulation in SPMS, the discrepancy in results between studies may be due to chance alone. However, the SPMS group included in the present work also had a much longer mean disease duration (25.4 years compared with 11.6 years), and greater disability as measured by EDSS scores (median 6.5 compared with 4.5) than the previous study,5 and so these two studies may actually be looking different phases of progressive MS and it is possible that with longer disease duration there is an increase in formation of new IC lesions in SPMS. It is also possible that the difference reflects the use of different MRI sequences to detect cortical lesions in the studies (PSIR in ours and DIR in Calabrese et al).

    When we tracked individual lesions, we found that LC lesions did not occur de novo, but instead arose from previously seen IC lesions, and the rate of conversion from IC to LC lesions was matched by the rate of new IC lesion formation. As such, the apparently static total IC lesion counts (table 2) gives a false impression of stability, when actually new IC lesions did occur (more so in SPMS), and other IC lesions extended into WM so increasing LC counts. This has important implications when assessing treatment effects, as counting either total IC or JC lesions alone may overlook substantial increases in the number of LC lesions and significantly underestimate the rate of new IC lesion formation.

    The total number of cortical GM lesions was significantly greater in the SPMS compared with the RRMS group at follow-up. However, when we subdivided cortical GM lesions into IC and LC, while both lesion types were more numerous in SPMS compared with RRMS, and the difference was about twice as great for LC compared with IC lesions (table 2), the differences did not reach statistical significance. We think that this relates to the size and heterogeneity of the cohorts, which while large enough to demonstrate a difference in total cortical GM lesion counts were too small to distinguish subtypes of cortical GM lesions.

    Consistent with previous work,10 ,19 we did not find a significant correlation between the formation of new cortical GM lesions and new WM lesions. There are a couple of possible explanations for this. First, that cortical GM and WM lesion genesis have at least partially different underlying mechanisms. As noted above, meningeal inflammation has been seen in tissue samples from people with SPMS, and linked with both the extent of cortical GM lesions,20 and the magnitude of neuronal loss,21 providing a plausible mechanism for GM lesions to form independently of WM lesions. Second, that similar mechanisms underlie both WM and GM lesion formation but have their greatest effect on WM and GM at different times. WM lesion accrual occurs most rapidly earlier in the clinical course of MS.10 ,22 By contrast, the results of our study, and those from previous work,5 show that the rate of GM lesion formation is similar, if not higher, in SPMS compared with RRMS. With improved discrimination between IC and LC lesions using the PSIR sequence,16 our study also suggests that new IC lesion formation is probably higher in SPMS than RRMS.

    In recent work, Rinaldi et al23 looked at the effects of natalizumab, interferons and glatiramer acetate, on the rate of IC lesion formation in people with RRMS. Over the 2 years of their study, the average number of new IC lesions per person per year was approximately 1.5 in the untreated group (n=35), approximately 0.7 in those taking interferon or glatiramer acetate (n=50) and approximately 0.1 in those receiving natalizumab. This demonstrates a potentially useful role for cortical GM lesion measures in treatment trials. Our results suggest that not only are new IC lesion counts of interest, but the expansion of IC lesions into LC may also provide additional information about treatment efficacy.

    A remaining limitation of our in vivo study is the sensitivity of cortical lesion detection on MRI. While using PSIR, about three times more IC lesions are identified when compared with DIR,15 using DIR in vitro, <10% of histopathologically confirmed IC lesions are detected.4 It is therefore likely that many IC lesions are still not identified on the PSIR sequence. This is particularly the case for subpial lesions, which are the most common type of cortical GM lesion seen in histopathological studies,1 but are rarely seen on DIR or PSIR.12 ,15 While this does not negate the results of the present study, it would be of great interest to know if the new IC lesions seen in this study began in a subpial location. It may be possible to address using ultra-high field MRI. Recent work at 7T acquiring T2*-weighted scans has demonstrated that a substantial number of subpial lesions can be seen.24 In addition, we registered the follow-up to baseline scans, which allowed us to more readily match lesions seen at baseline with those seen at follow-up. However, this registration step will have subtly blurred the follow-up images, and so reduced the contrast between lesions and surrounding tissues, and between GM and WM. Given that the contrast between GM lesions and normal-appearing GM is lower than that between WM lesions and normal-appearing WM, and between GM and WM, slight blurring of the follow-up images will tend to make it relatively more difficult to detect IC lesions when compared with LC or JC lesions. As we only identified new IC, but no new LC or new JC lesions, we do not think this can be explained by registration-associated methodological bias, but we may have slightly underestimated the number of new IC lesions.

    In conclusion, we have found that lesions that involve cortical GM rarely arise from the WM; the rate of cortical lesion formation is not closely linked with WM lesion accrual; and that new IC lesions probably arise more often in SPMS than RRMS. These results suggest that measuring WM lesion accrual alone is not a sufficient marker of MS disease activity and progression, and should encourage further efforts to develop MRI techniques that improve the in vivo detection of cortical lesions in MS.

    Acknowledgments

    The authors thank the MS Society of Great Britain and Northern Ireland and the National Institute for Health Research University College London Hospitals Biomedical Research Centre for financial support, and the people who took part in the study.

    References

    1. ↵
      1. Bø L,
      2. Vedeler CA,
      3. Nyland HI, et al
      . Subpial demyelination in the cerebral cortex of multiple sclerosis patients. J Neuropathol Exp Neurol 2003;62:723–32.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Geurts JJG,
      2. Bø L,
      3. Pouwels PJW, et al
      . Cortical lesions in multiple sclerosis: combined postmortem MR imaging and histopathology. AJNR Am J Neuroradiol 2005;26:572–7.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Vercellino M,
      2. Plano F,
      3. Votta B, et al
      . Grey matter pathology in multiple sclerosis. J Neuropathol Exp Neurol 2005;64:1101–7. doi:10.1097/01.jnen.0000190067.20935.42
      OpenUrlCrossRefPubMed
    4. ↵
      1. Seewann A,
      2. Kooi EJ,
      3. Roosendaal SD, et al
      . Postmortem verification of MS cortical lesion detection with 3D DIR. Neurology 2012;78:302–8. doi:10.1212/WNL.0b013e31824528a0
      OpenUrlCrossRefPubMed
    5. ↵
      1. Calabrese M,
      2. Poretto V,
      3. Favaretto A, et al
      . Cortical lesion load associates with progression of disability in multiple sclerosis. Brain 2012;135:2952–61. doi:10.1093/brain/aws246
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Calabrese M,
      2. Favaretto A,
      3. Poretto V, et al
      . Low degree of cortical pathology is associated with benign course of multiple sclerosis. Mult Scler 2013;19:904–11. doi:10.1177/1352458512463767
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Kutzelnigg A,
      2. Lucchinetti CF,
      3. Stadelmann C, et al
      . Cortical demyelination and diffuse white matter injury in multiple sclerosis. Brain 2005;128:2705–12. doi:10.1093/brain/awh641
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Calabrese M,
      2. Filippi M,
      3. Rovaris M, et al
      . Morphology and evolution of cortical lesions in multiple sclerosis. A longitudinal MRI study. Neuroimage 2008;42:1324–8. doi:10.1016/j.neuroimage.2008.06.028
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Roosendaal S,
      2. Moraal B,
      3. Pouwels P, et al
      . Accumulation of cortical lesions in MS: relation with cognitive impairment. Mult Scler 2009;15:708–14. doi:10.1177/1352458509102907
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Calabrese M,
      2. Rocca MA,
      3. Atzori M, et al
      . A 3-year magnetic resonance imaging study of cortical lesions in relapse-onset multiple sclerosis. Ann Neurol 2010;67:376–83. doi:10.1002/ana.21906
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Magliozzi R,
      2. Howell O,
      3. Vora A, et al
      . Meningeal B-cell follicles in secondary progressive multiple sclerosis associate with early onset of disease and severe cortical pathology. Brain 2006;130:1089–104. doi:10.1093/brain/awm038
      OpenUrlCrossRef
    12. ↵
      1. Turetschek K,
      2. Wunderbaldinger P,
      3. Bankier AA, et al
      . Double inversion recovery imaging of the brain: initial experience and comparison with fluid attenuated inversion recovery imaging. Magn Reson Imaging 1998;16:127–35. doi:10.1016/S0730-725X(97)00254-3
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Geurts JJG,
      2. Pouwels PJW,
      3. Uitdehaag BMJ, et al
      . Intracortical lesions in multiple sclerosis: improved detection with 3D double inversion-recovery MR imaging. Radiology 2005;236:254–60. doi:10.1148/radiol.2361040450
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Nelson F,
      2. Poonawalla AH,
      3. Hou P, et al
      . 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;28:1645–9. doi:10.3174/ajnr.A0645
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Sethi V,
      2. Yousry TA,
      3. Muhlert N, et al
      . Improved detection of cortical MS lesions with phase-sensitive inversion recovery MRI. J Neurol Neurosurg Psychiatry 2012;83:877–82. doi:10.1136/jnnp-2012-303023
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Sethi V,
      2. Muhlert N,
      3. Ron M, et al
      . MS cortical lesions on DIR: not quite what they seem? PLoS ONE 2013;8:e78879. doi:10.1371/journal.pone.0078879
      OpenUrlCrossRefPubMed
    17. ↵
      1. Kurtzke JF
      . Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:1444–52. doi:10.1212/WNL.33.11.1444
      OpenUrlCrossRefPubMed
    18. ↵
      1. Modat M,
      2. Ridgway GR,
      3. Taylor ZA, et al
      . Fast free-form deformation using graphics processing units. Comput Methods Programs Biomed 2010;98:278–84. doi:10.1016/j.cmpb.2009.09.002
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Bø L,
      2. Geurts JJG,
      3. van der Valk P, et al
      . Lack of correlation between cortical demyelination and white matter pathologic changes in multiple sclerosis. Arch Neurol 2007;64:76–80. doi:10.1001/archneur.64.1.76
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Howell OW,
      2. Reeves CA,
      3. Nicholas R, et al
      . Meningeal inflammation is widespread and linked to cortical pathology in multiple sclerosis. Brain 2011;134(Pt 9): 2755–71. doi:10.1093/brain/awr182
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Magliozzi R,
      2. Howell OW,
      3. Reeves C, et al
      . A Gradient of neuronal loss and meningeal inflammation in multiple sclerosis. Ann Neurol 2010;68:477–93. doi:10.1002/ana.22230
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Fisher E,
      2. Lee JC,
      3. Nakamura K, et al
      . Gray matter atrophy in multiple sclerosis: a longitudinal study. Ann Neurol 2008;64:255–65. doi:10.1002/ana.21436
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Rinaldi F,
      2. Calabrese M,
      3. Seppi D, et al
      . Natalizumab strongly suppresses cortical pathology in relapsing-remitting multiple sclerosis. Mult Scler 2012;18:1760–7. doi:10.1177/1352458512447704
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Mainero C,
      2. Benner T,
      3. Radding A, et al
      . In vivo imaging of cortical pathology in multiple sclerosis using ultra-high field MRI. Neurology 2009;73:941–8. doi:10.1212/WNL.0b013e3181b64bf7
      OpenUrlCrossRefPubMed

    Footnotes

    • Contributors TY, MR, DHM and DTC contributed to study concept and design. VS, NM and DTC contributed to acquisition of the data. VS, NM, DJT, DA, CW-K, DHM, DTC, TY and MR contributed to analysis and interpretation. TY, NM, DJT, DA, MR, CW-K, DHM and DTC contributed to critical revision of the manuscript for important intellectual content. DHM and DTC contributed to study supervision.

    • Competing interests VS received research support from Biogen Idec and Novartis. TY serves as editor for European Radiology Journal and has received honoraria (Board Membership) from UCB, Bristol-Myers Squibb, Biogen Idec, and grants (PI or Co-PI Coordinator) from NIHR CBRC, MRC, MS Society, PSP, Stroke, BHF, Wellcome Trust, GSK, Biogen Idec, Novartis; DJT previously had a portion of his salary paid by Novartis and Biogen Idec; CW-K is on the advisory board for BG12 (Biogen); DHM has received honoraria from Biogen Idec, Novartis, Mitsubishi Pharma Europe, and Bayer Schering, and research grant support for doing MRI analysis in multiple sclerosis trials sponsored by Biogen Idec, and Novartis; DTC has received honoraria (paid to his employer) from Bayer, Teva and the Serono Symposia International Foundation for faculty-led education work, and Teva for advisory board work; meeting expenses from Teva and Novartis; and holds stock in GlaxoSmithKline.

    • Ethics approval Local institutional ethics committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.