How Multiple Sclerosis Is Diagnosed: Key Tests

Multiple sclerosis is diagnosed through a combination of clinical evaluation, brain and spinal cord MRI, and sometimes spinal fluid analysis, all interpreted through a framework called the McDonald criteria. There is no single blood draw or scan that confirms MS on its own. Instead, doctors piece together evidence that inflammatory damage has occurred in at least two separate areas of the central nervous system at two different points in time. That framework sounds straightforward, but in practice the diagnostic path can stretch from a single clinic visit to months of follow-up testing, and the stakes of getting it right are high.

The McDonald Criteria and What They Require

The diagnostic rules neurologists follow today are called the McDonald criteria, first introduced in 2001 and revised several times since. The original criteria were a deliberate shift from older purely clinical standards toward incorporating MRI evidence, driven in part by the arrival of disease-modifying drugs that made early, accurate diagnosis medically urgent.

¹ Each revision has refined how lesions on MRI are counted and where they need to appear, but the core logic has stayed the same: you need to show dissemination in space (damage in more than one region of the central nervous system) and dissemination in time (damage that occurred on more than one occasion).²

Dissemination in space is typically satisfied by MRI lesions appearing in at least two of four classic locations: near the brain’s ventricles (periventricular), in the cortex or just beneath it (juxtacortical/cortical), in the brainstem or cerebellum (infratentorial), or in the spinal cord. The current criteria require lesions in at least two of these four regions, achieving about 87% sensitivity and 73% specificity when tested against the full diagnostic standard.³ Researchers are actively investigating whether adding a fifth region, such as the optic nerve, could improve detection. Early results suggest that including the optic nerve pushes sensitivity up to about 98% but drops specificity sharply, so future criteria revisions will need to balance those trade-offs.

Dissemination in time can be shown in a few ways. If a single MRI scan shows both gadolinium-enhancing lesions (which indicate fresh, active inflammation) alongside non-enhancing lesions (older damage), that counts. Alternatively, a follow-up MRI that shows new lesions compared to an earlier scan satisfies the requirement. The presence of oligoclonal bands in spinal fluid can also substitute for some of this time-based evidence under the 2017 revision of the criteria.⁴

MRI Scans and What Neurologists Look For

MRI is the backbone of MS diagnosis. It is far more sensitive than a neurological exam at detecting the white matter lesions MS causes, and it is the reason diagnoses can now happen earlier in the disease course than they could decades ago. A standard MS workup includes MRI of the brain, and often of the spinal cord as well, using sequences designed to highlight inflammatory plaques.

The two most informative types of findings are T2-hyperintense lesions, which appear as bright spots on certain scan sequences and represent areas where myelin has been damaged (whether recently or in the past), and gadolinium-enhancing lesions, which light up after a contrast agent is injected and indicate active, ongoing inflammation. In one study of patients during an MS relapse, about 62% had at least one gadolinium-enhancing brain lesion, and the most common location was subcortical.⁵ Having both enhancing and non-enhancing lesions on the same scan is diagnostically powerful because it simultaneously shows old and new damage.

The number and location of T2 lesions matters for prognosis as well as diagnosis. A higher lesion count at the first clinical event is one of the strongest predictors of whether someone will go on to develop full MS. Having more than ten T2 lesions at baseline roughly tripled the odds of conversion in one prospective study.⁶

The Lumbar Puncture and Oligoclonal Bands

A lumbar puncture (spinal tap) collects a small sample of cerebrospinal fluid (CSF) from the lower back. The main thing doctors look for in MS workups is oligoclonal bands, which are distinct bands of antibodies that appear in the spinal fluid but not in the blood. Their presence indicates that the immune system is active inside the central nervous system, and they are found in roughly two-thirds of people with MS.⁷

Oligoclonal bands serve two purposes in the diagnostic process. They help confirm that the central nervous system inflammation is consistent with MS rather than something else, and under the 2017 McDonald criteria, their presence can substitute for MRI evidence of dissemination in time. That means a person who has the right pattern of lesions in space, plus positive oligoclonal bands, can be diagnosed at their first clinical event without waiting for a follow-up scan. Oligoclonal band positivity was also one of the strongest independent predictors of conversion from a first demyelinating episode to clinically definite MS, with one study finding an odds ratio close to six.⁶

Not everyone with MS tests positive for oligoclonal bands, and that gap has pushed researchers to explore alternatives. Kappa free light chains, a byproduct of antibody production, can be measured in spinal fluid and may catch cases that traditional oligoclonal band testing misses. One study found that a kappa-index above a certain threshold was more sensitive than oligoclonal bands in distinguishing MS from controls, and it also predicted whether patients with a first episode would go on to develop MS.⁸ Because the kappa free light chain test can be automated rather than requiring the labor-intensive process of reading gel patterns by hand, it may eventually become a standard complement or replacement for traditional oligoclonal band testing.⁹

Evoked Potentials

Evoked potential tests measure how quickly electrical signals travel along nerve pathways by presenting a stimulus (a flashing checkerboard pattern, a clicking sound, or a mild electrical pulse) and recording the brain’s response with electrodes on the scalp. The most commonly used type in MS workups is the visual evoked potential (VEP), which checks the optic nerve pathway. If demyelination has slowed nerve conduction, the brain’s response arrives later than normal.

In one study, over half of MS patients showed delayed responses on visual evoked potential testing, even among some who had no obvious visual symptoms.¹⁰ This ability to detect subclinical damage makes evoked potentials useful as an additional line of evidence when MRI and spinal fluid results are borderline. Abnormal evoked potentials were among the strongest independent predictors of conversion from a first episode to definite MS in one prospective cohort, with an odds ratio above fourteen.⁶

Optical Coherence Tomography

Optical coherence tomography (OCT) is a quick, painless eye scan that measures the thickness of the retinal nerve fiber layer at the back of the eye. Because the optic nerve is part of the central nervous system and is commonly affected by MS, thinning of this layer can reflect past damage from optic neuritis or even subclinical injury that the patient never noticed.¹¹

OCT is not part of the formal McDonald diagnostic criteria, but it is increasingly used in MS clinics for two reasons. First, it provides objective, reproducible evidence of optic nerve involvement that can support a diagnosis when other tests are equivocal. Second, and perhaps more valuable in the long run, retinal measurements track disease progression. A meta-analysis found that people with MS whose retinal nerve fiber layer was thinner than about 88 micrometers at baseline had roughly double the risk of future disability progression. Faster-than-average thinning over time also predicted worsening.¹²

Advanced MRI Markers on the Horizon

Standard MRI identifies white matter lesions but cannot tell you much about what kind of lesion you are looking at. Two newer imaging markers are changing that, especially when it comes to telling MS apart from other conditions that produce similar-looking spots on a brain scan.

The central vein sign refers to a tiny vein visible running through the center of a white matter lesion. MS lesions form around small veins due to how the immune attack begins, so this sign is far more common in MS than in other conditions. In MS patients, a median of about 88% of lesions show a central vein, compared to roughly 14% in inflammatory vascular diseases. When researchers applied a threshold of 50% of lesions having this sign, it separated MS from inflammatory vasculopathies with perfect accuracy in one study.¹³ A recent meta-analysis found that a somewhat lower threshold, around 37.5%, achieved about 97% sensitivity and 90% specificity for diagnosing MS.¹⁴ The central vein sign also helps distinguish MS from a closely related condition called aquaporin-4 neuromyelitis optica spectrum disorder (NMOSD), where the sign is present in only about a third of lesions.¹⁵

Paramagnetic rim lesions are another emerging marker. These are chronic lesions with an iron-rich border of activated immune cells that appears as a dark rim on susceptibility-weighted MRI sequences. They indicate smoldering, ongoing inflammation at the lesion edge, and they seem to be highly specific to MS. An international multicenter study found paramagnetic rim lesions in about 52% of people with MS but only 7% of those with other neurological conditions, giving them roughly 93% specificity.¹⁶ A separate study estimated their prevalence at about 50% in people with established MS diagnoses and confirmed specificity near 90%.¹⁷ Neither the central vein sign nor paramagnetic rim lesions are yet required by the McDonald criteria, but both are expected to play a growing role in future revisions.

Ruling Out Conditions That Look Like MS

A significant part of diagnosing MS is making sure it is not something else. White matter lesions on an MRI are common and nonspecific. Migraine, small-vessel vascular disease, infections, lupus, sarcoidosis, and a long list of other conditions can produce brain lesions that look similar on standard imaging.¹⁸

Two conditions deserve special attention because they overlap heavily with MS and have their own targeted treatments. Aquaporin-4 NMOSD is identified by blood antibodies against aquaporin-4, a water-channel protein. In one study, about 76% of NMOSD patients tested positive for this antibody, compared to roughly 12% of MS patients. Meanwhile, oligoclonal bands were positive in about 65% of the MS group but only 13% of the NMOSD group. That divergent pattern makes the combination of antibody and spinal fluid testing a powerful way to separate the two.⁷

MOG antibody-associated disease (MOGAD) is a more recently recognized condition. It involves antibodies against a different myelin protein and tends to cause optic neuritis and spinal cord inflammation that can mimic MS. Unlike MS, MOGAD patients rarely show oligoclonal bands in their spinal fluid.¹⁹ Proposed diagnostic criteria for MOGAD center on the detection of MOG-IgG antibodies in blood as a core requirement.²⁰ Testing for both aquaporin-4 and MOG antibodies has become a routine part of the MS workup, especially when clinical features are atypical.

How Often MS Is Misdiagnosed

MS misdiagnosis is not rare. A review of over 1,200 patients referred to two tertiary MS centers found that about 4.4% had been incorrectly diagnosed with MS.²¹ The most common reasons were nonspecific neurological symptoms paired with incidental MRI abnormalities, undifferentiated white matter disease, NMOSD, and MOGAD. Migraine accounted for about 7% of the misdiagnosed cases, which makes sense given that migraine frequently causes white matter lesions that can superficially resemble MS plaques.

One particularly tricky mimic is CADASIL, a genetic small-vessel disease that causes progressive white matter changes. A study of CADASIL patients found that about 10% had been misdiagnosed with MS. The median time to correction was over fourteen months, and nearly half of the misdiagnosed patients received unnecessary MS therapies during that period, with some experiencing adverse events from those treatments.²² These cases highlight why the McDonald criteria insist on a specific lesion pattern and why additional testing (spinal fluid, antibodies, sometimes genetic testing) is so important when the clinical picture is ambiguous.

When Lesions Come Before Symptoms

Sometimes MS-like lesions show up on a brain MRI done for an unrelated reason, such as a scan after a head injury or for chronic headaches, in someone who has never had neurological symptoms. This is called radiologically isolated syndrome (RIS). Data from large international cohorts indicate that about half of people with RIS develop clinical MS symptoms within ten years, suggesting RIS may represent a presymptomatic stage of the disease.²³

Not everyone with RIS progresses. The risk is highest in people under 35, those with spinal cord lesions, those with gadolinium-enhancing lesions, and those with positive oligoclonal bands. A study of over 350 RIS patients found that having two of three key risk factors (age under 37, spinal cord lesions, and enhancing lesions) was associated with about a 28% chance of a clinical event within two years. Having all three pushed that estimate above 90%.²⁴ Revised diagnostic criteria for RIS now incorporate these risk factors, allowing the diagnosis to be made even when fewer classic lesion-location criteria are met, as long as the right combination of risk factors is present.²⁵

Diagnosing MS in Children and Adolescents

MS in children poses extra diagnostic challenges. The same McDonald criteria used for adults can be applied from age twelve onward, and a prospective national cohort study found that the 2017 criteria performed well across the age span when applied at the time of a child’s first attack.²⁶ Below age twelve, however, the criteria are less reliable. The positive predictive value dropped to about 32% in children under eleven in one analysis, compared to around 76% in adolescents twelve and older.²⁷ This is partly because acute disseminated encephalomyelitis (ADEM), a post-infectious inflammatory condition that can look very similar on MRI, is much more common in younger children. Neurologists diagnosing young children often rely more heavily on the clinical picture and follow-up imaging to distinguish MS from ADEM.

Blood Biomarkers and Where the Field Is Headed

One of the biggest limitations of current MS diagnostics is that MRI and lumbar puncture, while effective, are expensive, uncomfortable, or both. There is strong interest in finding reliable blood-based markers. Neurofilament light chain (NfL), a protein released from damaged nerve fibers, is the furthest along. A systematic review and meta-analysis concluded that blood NfL levels can help distinguish MS from controls, differentiate between MS subtypes, and predict future worsening.²⁸ NfL is not specific to MS, though. It rises after any kind of nerve damage, so an elevated level alone does not point to MS rather than another neurological condition. Its primary value is in monitoring disease activity and treatment response rather than making a first-time diagnosis.

Glial fibrillary acidic protein (GFAP), a marker of a type of brain cell called an astrocyte, is another emerging blood biomarker. In people with primary progressive MS, elevated GFAP levels were associated with nearly triple the risk of disability progression. That association was especially pronounced in patients whose NfL levels were not elevated, suggesting GFAP captures a different aspect of the disease process, possibly chronic astrocyte-driven injury rather than acute nerve damage.²⁹ A separate study confirmed that NfL, GFAP, and a third marker called CHI3L1 all independently predicted disability worsening in primary progressive MS.³⁰

None of these blood biomarkers have replaced MRI or spinal fluid testing in the diagnostic criteria, but they are increasingly used alongside them in clinical practice and research trials. As assay techniques become more standardized, NfL and GFAP could eventually offer a less invasive way to track whether someone’s disease is truly under control between MRI scans, which for many people with MS would be a meaningful improvement in how their care feels day to day.

1
PubMed Central. The evolution of disease criteria in multiple sclerosis: underlying motives and broader implications
2
PubMed. Diagnostic evolution in multiple sclerosis: A narrative review of the McDonald criteria from 2001 to 2024
3
PubMed Central. Improving criteria for dissemination in space in multiple sclerosis by including additional regions
4
PubMed Central. Investigating Whether Dissemination in Time Is Essential to Diagnose Relapsing Multiple Sclerosis
5
PubMed. Gadolinium-enhanced brain lesions in multiple sclerosis relapse
6
PubMed. Conversion of clinically isolated syndrome to multiple sclerosis: a prospective study
7
PubMed Central. Differential diagnosis between NMOSD and MS: a retrospective study based on clinical and imaging features
8
PubMed. Cerebrospinal fluid free light chains compared to oligoclonal bands as biomarkers in multiple sclerosis
9
PubMed Central. The Diagnostic Utility of Oligoclonal Bands in Multiple Sclerosis: A Time-Course Analysis
10
PubMed Central. Color vision versus pattern visual evoked potentials in the assessment of subclinical optic pathway involvement in multiple sclerosis
11
PubMed Central. Optical coherence tomography for retinal imaging in multiple sclerosis
12
PubMed Central. Retinal Optical Coherence Tomography Longitudinal Measures as Prognostic Biomarkers in Multiple Sclerosis: Systematic Review and Meta-Analysis
13
PubMed Central. Central vein sign differentiates Multiple Sclerosis from central nervous system inflammatory vasculopathies
14
PubMed. Accuracy of the central vein sign in diagnosis of multiple sclerosis: a systematic review and meta-analysis
15
PubMed Central. Differentiating Multiple Sclerosis From AQP4-Neuromyelitis Optica Spectrum Disorder and MOG-Antibody Disease With Imaging
16
PubMed Central. Paramagnetic Rim Lesions are Specific to Multiple Sclerosis: An International Multicenter 3T MRI Study
17
PubMed Central. Frequency and Diagnostic Implications of Paramagnetic Rim Lesions in People Presenting for Diagnosis to a Multiple Sclerosis Clinic
18
PubMed. What Can Mimic Multiple Sclerosis?
19
PubMed. Clinical and laboratory features distinguishing MOG antibody disease from multiple sclerosis and AQP4 antibody-positive neuromyelitis optica
20
The Lancet Neurology. Myelin oligodendrocyte glycoprotein antibody-associated disease: proposed diagnostic criteria
21
PubMed. Misdiagnosis of multiple sclerosis: Two tertiary centers experience in Saudi Arabia
22
PubMed Central. Multiple Sclerosis Misdiagnosis in Cerebral Autosomal Dominant Arteriopathy With Subcortical Infarcts and Leukoencephalopathy: A Genotype-Misphenotype Study
23
PubMed. Radiologically isolated syndrome
24
JAMA Network Open. Risk Factors and Time to Clinical Symptoms of Multiple Sclerosis Among Patients With Radiologically Isolated Syndrome
25
Brain. The radiologically isolated syndrome: revised diagnostic criteria
26
The Lancet Child & Adolescent Health. MRI and CSF measures in the diagnosis of paediatric multiple sclerosis: a prospective national cohort study
27
PubMed Central. Paediatric Multiple Sclerosis: Update on Diagnostic Criteria, Imaging, Histopathology and Treatment Choices
28
PLOS ONE. Neurofilament light chain in blood as a diagnostic and predictive biomarker for multiple sclerosis: A systematic review and meta-analysis
29
PubMed Central. Serum glial fibrillary acidic protein and disability progression in progressive multiple sclerosis
30
Journal of Neurology, Neurosurgery & Psychiatry. Serum biomarker levels predict disability progression in patients with primary progressive multiple sclerosis