Updates to the McDonald Diagnostic Criteria for Multiple Sclerosis

The 2024 updates to the McDonald diagnostic criteria for multiple sclerosis (MS), published in 2025, expand on the 2017 criteria and are designed to allow for earlier and more accurate diagnosis. This article provides an overview of these updates with an emphais on how neurologists can apply the McDonald criteria updates into clinical practice. 

Evolution of the Diagnostic Criteria for MS
Although there is no single diagnostic test for MS, several sets of criteria have been proposed over the years in an attempt to establish diagnostic thresholds. These frameworks reflected advances in clinical neurology, neuroimaging, and laboratory diagnostics.¹ Before these frameworks were established, early efforts such as Charcot’s triad (ie, the description of the triad of nystagmus, intention tremor, and scanning speech in 1868) provided the first clinical foundation for the diagnosis of MS.²

Schumacher Criteria (1965)
In 1965, a landmark publication by Schumacher et al³ proposed the first widely accepted set of clinical criteria for MS diagnosis. These criteria emphasized the concept of “relapse,” and diagnosis required careful history taking and neurologic examination. In addition, diagnosis required evidence of lesions that showed dissemination in space (DIS) and dissemination in time (DIT) within the central nervous system,³ although these concepts were not yet formalized through modern imaging. The Schumacher criteria laid the foundation for later refinements in the diagnosis of MS.

Poser Criteria (1983)
Two decades later, the Poser criteria introduced the use of paraclinical testing, including oligoclonal bands (OCBs) and visual evoked potentials, to support MS diagnosis.⁴ This marked a shift from relying solely on clinical presentation to integrating laboratory and electrophysiologic evidence. MS cases were classified into 4 categories, which reflected recognition of disease heterogeneity. 

McDonald Criteria (2001 and Subsequent Revisions)
The advent of MRI revolutionized MS diagnosis—a transition captured in the McDonald criteria. Introduced in 2001, the McDonald framework allowed for MRI-based demonstrations of DIS and DIT. Subsequent revisions (2005, 2010, and 2017) progressively broadened the sensitivity of MRI findings and reduced diagnostic delays. The 2017 update to the McDonald criteria also allowed OCB positivity to substitute for demonstration of DIT, increasing the role of laboratory support for MS diagnosis.

2024 Revisions of the McDonald Criteria
In 2024, a revision of the McDonald criteria was initiated to enhance diagnostic specificity and incorporate new scientific advances in the field.⁵ The changes aim to provide internationally applicable criteria for diagnosing MS throughout the lifespan.⁶ Additional measures were introduced to reduce misdiagnosis in certain vulnerable populations.⁷ The revision process was led by an international panel of 56 experts spanning diverse disciplines, including neurologists, radiologists, patient advocates, epidemiologists, and methodologists.^6,8 Proposed revisions were presented along with their supporting evidence, followed by group discussions and voting.^6,9 For a revision to be incorporated into the 2024 McDonald criteria, at least 90% of participants needed to vote, and 80% needed to be in agreement.^6,10

Key Updates to the McDonald Criteria

The Optic Nerve as a 5th Anatomic Location for DIS
About one-quarter of people with MS present with optic neuritis, and the majority will show clear involvement of the optic nerve over the disease course.¹¹ The optic nerve was incorporated into the 2024 McDonald criteria due to evidence which showed increased diagnostic sensitivity without decreasing the diagnostic specificity.^6,9 For detecting optic nerve involvement, a “classical” lesion must be found on an orbital MRI scan. Alternatively, the peripapillary retinal nerve fiber layer must show an intereye difference of ≥6 μm (or composite macular ganglion cell and inner plexiform layer intereye difference of ≥4 μm) on optical coherence tomography.^6,7 Visual evoked potential tests can also be used to assess demyelinating optic nerve injury when performed with appropriate technical knowledge; delayed latencies of 2.5 standard deviations above the mean for both peak P100 latency and interocular latency, with no better explanation, can be used to demonstrate unilateral optic nerve involvement.^6,7

Changes to the Use of DIS and DIT
The 2024 McDonald criteria updates fundamentally changed the DIS determination by increasing the possible anatomic locations of MS lesions from 4 to 5.^7,8 The DIS criteria are now met when typical lesions are found in 2 of 5 regions (optic nerve, cortical/juxtacortical, periventricular, infratentorial, and spinal cord). The new criteria also no longer require a mandatory demonstration of DIT, which was previously required to distinguish MS from monophasic inflammatory syndromes; the extent of spatial involvement is enough to establish a diagnosis in certain cases.^7,8 The 2024 McDonald criteria also built on the 2017 criteria with additional paraclinical diagnostic tests, such as cerebrospinal fluid (CSF) markers, to substitute for demonstration of DIT.⁸

Paraclinical Tests
Central Vein Sign. The central vein sign (CVS) appears as a line or dot (depending on the MRI plane) that courses centrally through a lesion on susceptibility-weighted sequences.^6,12 The CVS is the radiologic correlate to histologic perivenular inflammation in an MS plaque, and its identification on an MRI scan can help differentiate MS lesions from those resulting from other conditions, such as vascular disease or migraine.⁶ The CVS can be applied for MS diagnosis if at least 6 lesions have CVS positivity; if <10 lesions are present, CVS-positive lesions must outnumber CVS-negative lesions.⁶ This method is called “Select-6.” CVS positivity is not necessary to make a diagnosis of MS, but may aid the diagnosis in certain cases. The CVS was found in 73% of lesions in people with MS, showing 95% sensitivity and 92% specificity.⁸ This results in diagnostic accuracy with 90% sensitivity and 89% specificity when using Select-6 as a diagnostic tool.⁸ Examples of the CVS are shown in Figure 1.

Figure 1. Examples of classic MRI findings in multiple sclerosis. Central vein sign on fluid-attenuated inversion recovery MRI^19,20 (A, B). Paramagnetic rim lesion on phase susceptibility MRI; reprinted from Hemond et al. Paramagnetic rim lesions are highly specific for multiple sclerosis in real-world data. Brain Commun. 2025;7(3), by permission of Oxford University Press (C).²¹ Left optic nerve signal hyperintensity on T2-weighted MRI sequence²² (D). Right cerebellar peduncle demyelinating lesion²² (E). Periventricular Dawson fingers on sagittal fluid-attenuated inversion recovery MRI sequence²² (F). Juxtacortical demyelinating lesions²² (G, H).

Paramagnetic Rim Lesions. Paramagnetic rim lesions (PRLs), which represent chronically active MS lesions, are characterized by an inflammatory rim of iron-laden microglia surrounding an inactive core. PRLs are specific MS markers and are not often found in MS radiologic mimics.⁶ PRLs are associated with high specificity (≥95%) for MS across numerous studies.^6-8 PRLs also yield prognostic significance, with studies showing that the presence of ≥4 PRLs is associated with greater neurologic disability.^6-8 PRLs appear as a hypointense (dark) rim on MRI susceptibility sequences, surrounding a correlating hyperintense lesion on T2/fluid-attenuated inversion recovery sequences. An example of a PRL is shown in Figure 1.

Kappa Free Light Chain. The 2024 McDonald criteria include the kappa free light chain (KFLC) as part of the biomarker-driven diagnostic approach to MS.⁷ CSF KFLC positivity is now considered an interchangeable biomarker with CSF OCB positivity.⁶ KFLC testing is more accessible and reliable than OCB testing, as it uses automated, cost-effective platforms⁷ with faster turnaround times and rater-independent results.^6,13 Numerous studies have also demonstrated that KFLC levels have diagnostic accuracy comparable to OCB tests. In a multicenter validation study, intrathecal KFLC synthesis showed diagnostic sensitivity of 95% in people with MS and 82% in people with clinically isolated syndrome (CIS), with specificity of 95%, demonstrating a valid alternative (or supplement) to OCB tests.⁸

Radiologically or Clinically Isolated Syndrome: What Now? 
The 2024 McDonald criteria enable individuals who were previously diagnosed with CIS to receive an MS diagnosis more quickly, as they no longer need to wait for further MS activity to occur (ie, DIT) if the proper conditions are met.^7,8 In some cases, a diagnosis of CIS may be updated to MS on the basis of criteria adjustments alone.⁸

The 2024 McDonald criteria also represent a conceptual shift in how radiologically isolated syndrome (RIS) is approached by reclassifying some cases of RIS from a purely radiologic designation to a true MS diagnosis grounded in biologic and imaging evidence.⁸ Long-term observational studies show that ~19% of individuals with RIS develop clinical MS within 2 years, 35% within 5 years, and 51% within 10 years.⁸ Previously, people with RIS had to wait for symptoms to appear to receive an MS diagnosis, but now neurologists can diagnose MS if certain criteria are met, saving individuals from missed treatment time.⁸

Diagnosis of Relapsing or Progressive MS Within a Joint Framework
The 2024 McDonald criteria mandate a single, unified diagnostic framework for relapsing and progressive MS.⁶ This change recognizes that the various forms of MS share similar mechanisms and constitute a continuum, with contributions from concurrent pathophysiologic processes that vary across individuals and over time.^6,8 The revised criteria help facilitate the diagnosis of MS in individuals with a variety of presentations, recognizing all of them as manifestations of the same underlying disease processes.⁸

Diagnosis in People Aged ≥50 Years or With Comorbidities
Recognizing that vascular risk factors and comorbidities increase the risk of MS misdiagnosis, especially in older individuals, the 2024 McDonald criteria focused on increasing diagnostic specificity in those populations.⁸ The CVS is particularly helpful in this setting, as it can differentiate MS lesions from vascular lesions.⁸ In individuals aged ≥50 years or with certain comorbidities (eg, vascular, migraine), the following additional features are recommended to confirm an MS diagnosis: a classical spinal cord lesion, positive CSF test (OCB or KFLC), or Select-6 positivity.⁸

A New Diagnostic Algorithm for MS

Diagnosis of Exclusion
The 2017 McDonald criteria emphasized the need for “no better explanation” to account for clinical and radiologic findings, and this fundamental requirement has not changed. The 2024 McDonald criteria continue to emphasize the importance of considering differential diagnoses, including reevaluating the diagnosis periodically over the disease course, even after a diagnosis of MS has been made.⁶

The Increased Role of MRI
The 2017 McDonald criteria were not validated for use in individuals who were asymptomatic or had nonspecific or prodromal presentations.¹⁴ Therefore, application of the criteria was limited to people with a classical demyelinating event. With the 2024 McDonald criteria updates, there is decreased emphasis on clinical presentation and increased emphasis on the use of MRI scans. Identifying classical lesions on MRI scans “is now the cornerstone of diagnosis.”⁶ MS can contribute to the development and accumulation of nonspecific lesions—particularly in the subcortical white matter—however, these cannot count toward fulfillment of DIS criteria. Examples of classical lesions that can satisfy DIS criteria are shown in Figure 1.

Beyond expanding the definition of DIS, the 2024 McDonald criteria also increase the role of MRI scans by incorporating the CVS and PRLs as MS markers. However, supporting evidence for the CVS is strongest in periventricular or subcortical lesions, and PRLs have been described primarily in periventricular lesions. Data on PRLs in other locations are limited, and there is little evidence regarding the proportion of lesions with the CVS in other locations.⁶

The Clinical Presentation or the MRI Results: Which Matters More?
MS evaluation in a real-world clinical setting typically starts with the possibility of MS being suggested on a radiology report, perhaps even in the absence of clinical suspicion. The misdiagnosis rate of MS is estimated at ~20%,¹⁵ with one of the most common pitfalls stemming from misinterpretation of lesions on brain MRI scans.¹⁶ Nevertheless, the panel proposed 2 algorithms: 1 for individuals with classical presentations of MS and 1 for individuals who are asymptomatic or with nonspecific presentations. We present a “reverse” algorithm (Figure 2), which starts with the MRI and then considers special features and an MS-like presentation—an approach that may more closely resemble a real-world encounter. 

If MRI scans are used as the starting point for MS diagnosis, the person to whom this algorithm is being applied must be more likely to have MS than any other explanation for their clinical and paraclinical findings. The differential diagnosis of MS is broad; for example, there are >60 known etiologies for optic neuropathy or neuritis.^17,18 MRI results are necessary for MS diagnosis, but the overall disease presentation matters more. 

The Reverse Algorithm
Clinical Scenario 1. An individual in their late 40s with a history of hypertension presents because of increased headaches. A brain MRI scan is obtained, and the radiology report describes a few scattered “periventricular” white matter lesions and suggests the possibility of MS, among other differential diagnoses, including migraine, Lyme disease, vasculitis, and chronic microvascular ischemic change. 

In this scenario, there is an overall lack of DIS, with only 1 anatomic location of the possible 5 being involved. Further clinical history must be suggestive of MS, with either a classical demyelinating event or 1 year of progressive neurologic decline, without any better explanation. If so, then MS may be diagnosed if DIT is also demonstrated, along with either 1 PRL or Select-6 positivity. 

For our purposes here, DIT can be demonstrated by a “classical” attack, identification of a new lesion on follow-up MRI scan, identification of simultaneous enhancing and nonenhancing lesions on the same MRI scan, or demonstration of either OCB (ie, ≥ 2 bands unmatched to the corresponding serum sample) or KFLC (eg, index ≥6.1) positivity in CSF. 

Clinical Scenario 2. An individual in their early 50s experienced a 1.5-year history of slowly progressive neurologic decline, characterized by bilateral leg stiffness, weakness, spasms, and bladder dysfunction, and now requires a cane to ambulate. Neurologic examination demonstrates clonus and pathologic reflexes in the lower extremities. Brain MRI results are normal for the individual’s age, but the spinal cord demonstrates 2 lesions that are dorsolaterally located in the cord, short-segmented, nonenhancing, nonexpansile, and not associated with disc compression. There is also mild diffuse brain and spinal cord atrophy. Considering the individual’s family history and results from electromyography/nerve conduction studies, genetic testing, and infectious and metabolic workup, the neurologist begins to strongly suspect MS.

In this scenario, the brain MRI results are normal, so Select-6 or PRL criteria are unlikely to be captured. However, there are 2 classical lesions in the spinal cord. The clinical course demonstrates at least 12 months of progressive neurologic decline typical for MS, with other differential diagnoses less likely. If DIT criteria can be satisfied in 1 of the 4 previously described ways, a diagnosis of MS can be made. 

A more common clinical scenario than either case example is an individual who presents with ≥2 anatomic locations observed on MRI scan, such as periventricular plus infratentorial, or optic nerve plus spinal cord. For patients who meet the criteria for DIS, a diagnosis can be established either by demonstrating DIT criteria (in 1 of the 4 ways previously described) or by fulfilling the Select-6 criteria—regardless of whether the patient is symptomatic or their clinical course. 

In the rare circumstance that an individual initially presents with 4 or 5 of the 5 anatomic locations involved, the clinician need not evaluate for DIT, repeat MRI, obtain a lumbar puncture, or assess for CVS or PRLs. In this setting, diagnosis requires only a classical clinical presentation historically (ie, 1 demyelinating event or ≥12 months of progression), with no better alternative explanation. All of these scenarios are illustrated in Figure 2.

Summary
The MS diagnostic criteria have undergone many revisions over the years in response to rapid advancements in the field of MS. By increasing the number of anatomic locations for DIS from 4 to 5, and by expanding the number of CSF biomarkers substituting for DIT, the hope is that more MS cases will be diagnosed and treated earlier, improving long-term prognosis. By incorporating highly specific MRI features for MS, including the CVS and PRLs, the diagnostic accuracy of MS can be improved, hopefully mitigating the appalling 20% misdiagnosis rate. The ultimate goal is correct and early diagnosis, rapid treatment initiation, and improved prognosis for people with MS.