Diagnosis of multiple sclerosis: what is changing?

KEYWORDS:

• Dissemination in space
• dissemination in time
• MRI
• oligoclonal bands
• multiple sclerosis

1. Introduction

The conceptual framework of the diagnostic criteria for multiple sclerosis (MS) was originally conceived to generate a common definition for research purposes [1]. Whereas its basic concepts of dissemination in space (DIS) and time (DIT) in the absence of a better explanation for the symptoms are still valid, what is changing is the incorporation of evidence based, objective data to support the diagnosis, allowing the use of the diagnostic criteria in the clinical practice [2]. Nonetheless, such objective data is not without controversies, as we will outline below.

2. Number or periventricular lesions to demonstrate DIS

The 2016 MAGNIMS criteria suggested, based on an expert consensus, to increase the minimum number of periventricular lesions to 3 to assess DIS [3]. A single periventricular lesion was deemed insufficiently specific to confirm its demyelinating and inflammatory origin, as incidental periventricular lesions can be detected in older individuals and in those with cardiovascular risk factors. Later studies comparing different periventricular lesion number thresholds showed conflicting results [4–6]. The International Panel on Diagnosis of Multiple Sclerosis considered said results did not justify the added complexity of requiring different numbers of lesions in different topographies, and maintained the cutoff of 1 lesion [7]. Nevertheless, in older individuals, establishing DIS with ≥3 periventricular lesions can increase the diagnostic specificity [5]. Therefore, it could be considered advisable to increase the minimum number of periventricular lesions to 3 in older patients. We think this would add unnecessary complexity to the diagnostic criteria. Besides, the diagnosis of MS does not rely solely on periventricular lesions, but on the simultaneous presence of lesions in at least one more topography to establish DIS, which in turn still requires DIT [4–6] or +OB [8]. If reasonable doubts persist, it is advisable to wait for the occurrence of a more typical attack or lesion on MRI than to risk misdiagnosis.

3. The anterior visual system

The optic nerve was included in the 2016 MAGNIMS criteria as a fifth topography to determine DIS, based on an expert consensus [3]. This proposal stated clinical documentation of optic nerve atrophy or objective demonstration of optic nerve involvement via visual evoked potentials (VEP), MRI or optical coherence tomography (OCT) could support DIS and, in patients without concurrent visual symptoms, DIT. We recognize optic nerve involvement as an important feature of MS and consider its inclusion in the diagnostic criteria an unmet need, but more data are required to determine how to do it best. The vast differential diagnosis of subacute visual loss makes clinical documentation alone unreliable [9]. Therefore, we should focus our efforts on validating objective measures of optic nerve involvement that can be deemed typical of MS, alone or in combination with other radiological findings and/or OB to avoid an increase in false positives. To date, efforts to assess MRI or VEP as objective measures of optic nerve involvement have yielded very small improvements, if any, in the criteria´s diagnostic performance [6]. Additionally, we consider that symptomatic and asymptomatic optic nerves should be assessed separately at the present stage, although it is logical to make no distinction as stated in the 2017 criteria. In this regard, a study showed the criteria´s diagnostic performance increased after including the symptomatic optic nerve in DIS, but not in patients with clinically isolated syndromes (CIS) and asymptomatic optic nerve involvement (non-ON CIS) detected clinically or with VEP [10]. This could be partly explained by the low proportion of optic nerve involvement in non-ON CIS, raising the concern of whether asymptomatic cases could be of value to the diagnostic criteria. In any respect, standardization of normative data for MRI of the optic nerve, VEP, and OCT should be fully addressed before considering the inclusion of the optic nerve in the diagnostic criteria.

4. To include or not to include oligoclonal bands?

After new evidence demonstrated that the presence of OB together with DIS could be used as an alternative criterion for diagnosing MS in the absence of DIT [8], this proposal was included in the 2017 McDonald criteria [7] but not without controversy. Some authors rightly point out that OB are not specific for MS and that the implications of this proposal rely on the analytical accuracy of the test [11,12]. Others support this proposal as it increases the proportion of patients diagnosed with MS at the time of the CIS, which in turn can have important therapeutic and prognostic consequences [13]. Yet other authors consider the role of OB insufficient: they advocate for their use in all patients as proof of the immune and inflammatory nature of the lesions to exclude alternative diagnoses [14]. Indeed, the combination of ≥2 T2 typical MS lesions plus OB as an alternative definition of DIS in previous versions of the diagnostic criteria was intended to avoid misdiagnosing nonspecific vascular lesions as inflammatory [15]. When considering that the increased proportion of diagnosed patients is due mostly to +OB [13], clinicians should have a low threshold to undertake additional diagnostic tests if deemed necessary. This is particularly important in populations with a lower prevalence of MS given the evidence informing the changes in the 2017 diagnostic criteria originates from studies conducted in Europe and North America. In this sense, we consider relevant to validate the 2017 criteria in other regions/countries as well as in different age groups, such as children ≤10 years of age, adolescents or patients with late-onset MS. For example, a study assessing the criteria in Korean patients demonstrated similar diagnostic properties as a Dutch study [16,17]. Another study in children aged 0.46–17.87 years showed the 2017 criteria performed well across the studied age span, with presence of OB improving the diagnostic accuracy [18]. These results suggest that if the McDonald criteria are applied to patients presenting with typical symptoms suggestive of MS, they should be valid in any population, always keeping in mind that the diagnosis of MS should be done only when other diseases have been ruled out. Finally, we should not forget that an important factor in interpreting OB results is the use of appropriate analytical techniques, as has been already recommended elsewhere [19].

5. Are the 2017 McDonald criteria less specific than the 2010 revisions?

Several studies compared the diagnostic performance of the 2010 and 2017 criteria in CIS. Whereas the sensitivity of the 2017 iteration increased, its specificity decreased compared to the 2010 revisions [17,20]. We have several points to consider when interpreting these findings: 1. OB were the largest contributors to the increase in the proportion of patients diagnosed at baseline. 2. The outcome was clinically definite MS (CDMS). 3. The median follow-up was 2–3 years. 4. Although specificity was lower in the 2017 criteria compared to the 2010 revisions, the risk of developing MS was at least double in patients with OB after correcting for confounding factors. Thus, what we might be observing is the Will Rogers phenomenon [21]. There is a subset of patients with DIS and OB with no concomitant DIT that may have a less inflammatory disease at onset and thus may require a longer time for a second demyelinating attack to occur. We have certainly seen this in our CIS inception cohort in which patients with a typical CIS but no lesions on baseline MRI present the second attack more than 10 years after the first episode [22]. CDMS has been considered the gold standard to evaluate the diagnostic performance of the different versions of the McDonald criteria. Nevertheless, we are now diagnosing patients before they present the second attack. Therefore, we should also take other factors contributing to the diagnosis of MS into account, such as the demonstration of new lesions on follow-up MRIs. In our experience, a proportion of CIS patients with 0 lesions at baseline has new lesions observed on follow-up imaging studies before the second attack occurs. Indeed, MRI is a more sensitive tool to assess disease activity than clinical data [23], and the specificity of the McDonald criteria considering +OB was high when assessing new lesions with or without a second attack as the outcome [8,24].

6. Other considerations

Probably the greatest change in the diagnosis of MS occurred when the 2001 McDonald criteria substituted the Poser criteria. Conversely, changes in the diagnostic performance of the different iterations of the McDonald criteria are minimal, although their application has been simplified. Interestingly, the McDonald criteria were initially deemed too stringent but later versions are considered perhaps too liberal. How do we reach the middle ground? Should we now focus on increasing their specificity, although the criteria should only be applied when there is a high index of suspicion of MS? Could additional tests that are less rater-dependent (kappa free light chains) [25] or that appear to be more specific of MS (central vein sign) [7] be used in all CIS patients as part of a new version of the criteria, or only in cases with diagnostic uncertainties? This is a matter of debate that should be addressed in ongoing/future studies.

Concerning the 2017 revisions to the McDonald criteria specifically, they review several critical definitions and areas for further study. Unfortunately, the 2017 revisions do not explain very well whether the diagnosis of MS can still be done based purely on objective clinical evidence, as stated in previous versions [26]. We believe this issue should be clarified in future versions of the McDonald criteria. In this day and age, with neuromyelitis optica spectrum disorder, MOG-associated disease, and probably other yet to define inflammatory diseases with symptoms similar to those of MS, we advice against it. We should always complement clinical findings (or lack thereof) with the necessary paraclinical studies. This is also an important issue if the breadth of the MS disease spectrum in the diagnostic criteria will be expanded to include radiologically isolated syndromes and progressive solitary sclerosis [7].

7. Conclusions

The diagnosis of MS still requires clinical judgment, but it is increasingly aided by MRI and laboratory assessments. As long as controversies on the application of these paraclinical studies remain, it is important to remember that the criteria were developed to confirm the diagnosis of MS and not to distinguish it from other diseases. But as new data emerges, imaging technology improves, and other biomarkers are discovered or validated, the criteria will require periodic reexaminations with the intention of improving their accuracy.

In the end, in a sense we seem to be going back to the Poser criteria. More specifically, to an evidence-based version of the Poser criteria. As the diagnosis and differential diagnosis of MS are evolving rapidly, education is key. Depending on the setting, MS specialists should play an important role informing other neurologists and general practitioners about how to apply the McDonald criteria integrating clinical, radiological, and biological information to make an accurate diagnosis.

8. Expert opinion and five year view

What is changing in the diagnosis of MS? The way objective data from observational and multicenter studies are generated to support the diagnosis. Yet, such studies have limitations and can be contested. These limitations underscore the importance of 1. Periodically reexamining the criteria as new evidence emerges, and 2. Applying the McDonald criteria only if there is a high index of suspicion of MS. Finally, as MS can now be diagnosed before the second attack, we should reevaluate our methods to measure the diagnostic performance of present and future versions of the McDonald criteria.

It is possible that enough evidence will be available to determine whether the anterior visual pathway can be considered to establish DIS. While concerns may arise regarding the presence of OB in diseases other than MS, assessing their value with potentially highly specific tests such as the central vein sign could add a higher degree of certainty to the diagnosis of MS.

Article highlights

• The diagnosis of MS is still clinical, but is increasingly aided by MRI and laboratory assessments

• Whereas the cut-off of 1 periventricular lesion is maintained in the 2017 McDonald criteria, special care should be taken in older patients to identify red flags and rule out other possible diagnoses

• The inclusion of the anterior visual pathway in the diagnostic criteria is an unmet need, but objective measures of optic nerve involvement and evidence of their relation to other lesion topographies are lacking

• The 2017 revisions reincorporated OB into the diagnostic scheme, which together with DIS can be used as an alternative criterion for MS in the absence of DIT

• The usefulness of the 2017 criteria is being validated in other populations and age groups

• We should consider new MRI findings besides CDMS when evaluating the diagnostic performance of current and future diagnostic criteria

• It is yet to be determined if the central vein sign or the kappa free light chains will increase the diagnostic performance of the McDonald criteria

• In the future, the breadth of the MS disease spectrum could be expanded to include radiologically isolated syndromes and progressive solitary sclerosis

• As new data emerges the diagnostic criteria require periodic re-examinations

• With more patients diagnosed at the time of the CIS, it is important to remember the diagnostic criteria were developed to confirm the diagnosis of MS when there is a high index of suspicion and alternative diagnoses have been ruled out

• MS specialists should play an active role in informing other medical practitioners about the correct application of the diagnostic criteria for MS.

Declaration of interest

G Arrambide compensation for consulting services from Sanofi, Merck and Roche, research support from Novartis, travel expenses for scientific meetings from Novartis, Roche and Stendhal, and speaking honoraria from Sanofi as well as personal fees from Excemed, outside the submitted work. M Tintore reports compensation for consulting services and speaking honoraria from Bayer Schering Pharma, Merck-Serono, Biogen-Idec, Teva Pharmaceuticals, Sanofi-Aventis, Novartis, Almirall, Genzyme, and Roche as well as personal fees from Excemed, outside the submitted work. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or conflict with the subject matter or materials discussed in this manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded