MS MRI: How Scans Detect Lesions and Rule Out Mimics

MRI is the single most important tool for diagnosing and monitoring multiple sclerosis. No blood test or physical exam can confirm MS on its own, but MRI can reveal the characteristic lesions in the brain and spinal cord that define the disease, often before symptoms become obvious. Diagnosing MS requires showing that these lesions are scattered across different parts of the nervous system and that they developed at different times, and MRI is how clinicians prove both points. Yet the technology does far more than check a diagnostic box. Newer MRI techniques can now distinguish MS from conditions that mimic it, track silent disease activity between relapses, and even detect early signs of tissue repair.

What MRI Needs to Show for a Diagnosis

The current diagnostic framework for MS, rooted in the McDonald criteria, requires two things: evidence that lesions are spread across more than one region of the central nervous system (dissemination in space), and evidence that lesions formed at different points in time (dissemination in time). MRI can satisfy both requirements. Dissemination in space is typically established by finding lesions in at least two of four characteristic locations: near the brain’s fluid-filled ventricles, in the outer cortical or juxtacortical regions, in the brainstem or cerebellum, or in the spinal cord. Dissemination in time can be shown if a scan reveals both enhancing and non-enhancing lesions simultaneously, or if a follow-up scan shows a new lesion that was not on an earlier image. Alternatively, the presence of oligoclonal bands in spinal fluid can stand in for the time requirement on MRI.¹

In practice, this means a person’s very first MRI scan can sometimes be enough for a diagnosis if it shows both old and new-looking lesions in the right pattern. Other times, a second scan months later is needed to catch a new lesion forming. Either way, MRI is doing the heavy lifting. The clinical exam and symptom history help, but they confirm what MRI reveals rather than the other way around.

The Key Sequences and What They Reveal

Not all MRI images are created equal. Different pulse sequences highlight different aspects of MS pathology, and a standard MS protocol uses several of them together.

T2-weighted and FLAIR: These are the workhorses. Lesions appear as bright spots on T2-weighted images, and the FLAIR sequence suppresses signal from cerebrospinal fluid so that lesions near the ventricles are easier to see. FLAIR is particularly useful in the brain, where periventricular lesions are a hallmark of MS.²
T1-weighted with gadolinium contrast: When a gadolinium-based contrast agent is injected, active lesions light up because their blood-brain barrier is temporarily broken down. These enhancing lesions are considered markers of acute inflammation.³ Gadolinium-enhanced MRI is more sensitive than older imaging methods at detecting these transient barrier disruptions and can help link an active lesion to a new symptom.⁴
T1-weighted without contrast: Over time, some MS lesions become permanently dark on standard T1 images. These so-called “black holes” reflect severe tissue destruction and axonal loss, distinguishing old damage from more reversible inflammatory changes.⁵

Together, these sequences give clinicians a layered picture: where lesions are, whether any are currently inflamed, and how much permanent damage has accumulated. A follow-up scan months or years later uses the same sequences to look for new bright spots on FLAIR or new enhancing lesions, which would signal ongoing disease activity even if you feel fine.

Telling MS Apart From Its Mimics

White spots on a brain MRI are not unique to MS. Migraines, small-vessel disease, lupus, and other inflammatory conditions of the nervous system can all produce lesions that look similar on a standard scan. Two conditions in particular can be difficult to separate from MS: aquaporin-4 neuromyelitis optica spectrum disorder (AQP4-NMOSD) and MOG-antibody disease (MOGAD). Getting this distinction right matters because the treatments differ.

Researchers have found that a combination of imaging features can reliably tell these conditions apart. When comparing relapsing MS to AQP4-NMOSD, the most useful single feature was the proportion of lesions showing a central vein sign, a tiny vein running through the middle of the lesion. In MS, about 84% of lesions had this feature compared with about 33% in AQP4-NMOSD. Combining the central vein sign with cortical lesion counts and optic nerve measurements reached 95% accuracy in distinguishing the two. For separating MS from MOGAD, the total number of white matter lesions was the strongest differentiator, with MS patients averaging around 39 lesions compared to roughly one in MOGAD patients.⁶

The Central Vein Sign and Iron Rim Lesions

The central vein sign has attracted considerable attention as a diagnostic biomarker because MS lesions tend to form around small veins in a way that other white matter lesions do not. On susceptibility-weighted or T2*-weighted images, you can sometimes see a thin dark line running through the center of a lesion, representing that vein. A large multicenter study of over 4,400 lesions in 606 people found that using a 35% threshold (meaning at least 35% of a person’s lesions show the sign) yielded about 68% sensitivity and 83% specificity for distinguishing MS from non-MS conditions. A stricter rule requiring at least three lesions with the sign pushed specificity to 89%.⁷ A separate study found even higher performance, reporting about 86% sensitivity and 100% specificity at a 45% cutoff, though that was in a smaller cohort.⁸

A related finding involves paramagnetic rim lesions, sometimes called iron rim lesions. These are lesions with a visible rim on susceptibility-sensitive MRI sequences, caused by iron-laden immune cells trapped at the lesion edge. They correspond to chronically active lesions where smoldering inflammation continues even after the initial attack has subsided.⁹ These lesions have emerged as potential biomarkers of disease severity and may help predict who is at risk for faster disability progression.¹⁰ Neither the central vein sign nor paramagnetic rim lesions are yet required by the McDonald criteria, but they are increasingly used in clinical practice to resolve ambiguous cases.

Cortical Lesions and Why They Are Hard to Find

Standard MRI sequences are good at finding lesions in the brain’s white matter but much worse at detecting those in the cortex, the brain’s outer gray matter layer. That is a problem because cortical pathology is a major feature of MS, especially in progressive forms, and cortical lesion burden correlates with cognitive decline and physical disability.

A specialized sequence called double inversion recovery (DIR) suppresses signals from both white matter and cerebrospinal fluid, making cortical lesions stand out more clearly. Consensus guidelines define cortical lesions on DIR as focal bright abnormalities compared with surrounding normal gray matter.¹¹ Newer, faster DIR techniques have improved detection further, catching significantly more juxtacortical lesions than conventional DIR in head-to-head comparisons.¹² Adding a T2 preparation pulse to DIR has been shown to reveal cortical lesions that conventional DIR misses entirely; in one study, over 40% of cortical lesions were visible only on the T2-prepared version.¹³

Ultra-high-field scanners operating at 7 Tesla take this even further. At that strength, subpial lesions, which sit just beneath the brain’s surface membrane and are notoriously invisible at standard field strengths, become detectable. In one 7T study, 86% of MS patients had at least one subpial lesion, and these lesions were distributed widely across the frontal, temporal, and parietal cortex.¹⁴ Surface-based analysis at 7T has confirmed that these subpial signal changes are widespread, reflecting the diffuse cortical pathology long described in autopsy studies but until recently invisible in living patients.¹⁵

Spinal Cord Imaging

Brain MRI gets most of the attention, but spinal cord lesions are common in MS and contribute independently to disability, especially problems with walking, coordination, and bladder control. Spinal cord imaging is technically harder than brain imaging because the cord is small, surrounded by moving structures (heartbeat, breathing, swallowing), and susceptible to imaging artifacts.

Research has shown that spinal cord lesion load and spinal cord atrophy are each independently linked to disability and are not well correlated with each other, meaning a person can have significant cord shrinkage without many visible lesions or vice versa.¹⁶ In one study, cervical cord lesion load and cord cross-sectional area were both independently associated with disability scores, along with overall brain atrophy. The statistical model incorporating all of these explained a substantial portion of the variation in disability.¹⁷ This is why thorough MS imaging protocols include the cervical and often the thoracic cord, not just the brain.

Brain Shrinkage as a Measure of Progression

Everyone’s brain shrinks a little with age, but in MS the rate is faster, and it starts earlier. Brain atrophy can be measured on serial MRI scans and has become one of the more reliable predictors of future disability, both physical and cognitive.¹⁸ Interestingly, gray matter atrophy appears to be a better marker of progression than white matter atrophy. In one long-term study, patients whose disability worsened over time had significantly faster gray matter loss, while white matter atrophy rates were similar regardless of whether someone’s disability progressed.¹⁹

The challenge is that atrophy changes are small year to year and can be influenced by hydration, medication effects (some MS drugs cause temporary brain volume changes unrelated to real tissue loss), and normal aging. For this reason, atrophy measurements are more useful across longer time intervals and are primarily a research and clinical-trial endpoint for now rather than something that guides individual treatment decisions in routine practice.

When the Scan Doesn’t Match the Symptoms

One of the longstanding puzzles in MS is the so-called clinico-radiological paradox: sometimes a person’s brain MRI looks alarming but they function well, and sometimes someone with few visible lesions has severe disability. This disconnect has real consequences because if MRI findings do not reliably track symptoms, using scans alone to judge treatment effectiveness becomes tricky.²⁰

More recent research suggests this paradox may partly be a myth created by the limitations of conventional imaging. Standard T2 lesion counts ignore lesion location, cortical involvement, spinal cord damage, diffuse tissue changes outside visible lesions, and atrophy, all of which contribute to disability. When researchers account for where lesions sit rather than just how many there are, the mismatch between imaging and symptoms shrinks considerably.²¹ Techniques like diffusion tensor imaging, which detects subtle damage in tissue that looks normal on standard scans, also help close the gap. Diffusion tensor imaging has identified significantly altered tissue properties in normal-appearing white matter of MS patients compared with healthy controls.²² In other words, a lot of damage is hiding in tissue that conventional MRI calls “normal.”

Gadolinium Contrast and When It Is Needed

Gadolinium-based contrast agents are injected intravenously to highlight areas where the blood-brain barrier has broken down, revealing active inflammation. They have been a cornerstone of MS imaging for decades, but there is growing awareness that repeated gadolinium exposure leaves trace deposits in certain brain structures. Whether these deposits cause harm is still debated, but the concern has prompted a shift in how often contrast is used.²³

Updated guidelines from the French Observatory of MS, which have influenced practice in many centers, recommend gadolinium at the time of diagnosis, when starting or switching a disease-modifying therapy, at a six-month re-baseline scan, and when previous scans are unavailable for comparison. During a relapse or when another condition like progressive multifocal leukoencephalopathy is suspected, gadolinium may also be used. For routine follow-up scans where a good baseline exists, contrast can often be skipped as long as the protocol includes high-quality 3D FLAIR sequences.²⁴ If your neurologist has been ordering contrast-free follow-up scans, this shift is likely why.

Myelin Imaging and Watching Repair in Real Time

Standard MRI can tell you that a lesion exists, but it cannot easily tell you whether the myelin coating around nerve fibers has been destroyed or is starting to grow back. Myelin water imaging, a technique that separates the MRI signal from water trapped between myelin layers, can do exactly that. In newly forming lesions, myelin water imaging has shown that individual lesions go through repeated cycles of demyelination and remyelination in their early weeks and months, a finding that was previously only available from autopsy data.²⁵ Quantitative magnetization transfer, a related method, has been able to track partial myelin recovery in enhancing lesions, with inflammation resolving quickly and remyelination following more slowly.²⁶

This kind of imaging matters most for the development of therapies that promote myelin repair. If a drug is designed to boost remyelination, you need a way to measure whether it is working in living people, not just in lab dishes. Myelin water fraction imaging has the potential to serve that role as a treatment biomarker.²⁷ These techniques are still primarily research tools, but they are moving steadily toward clinical use as scanner software catches up.

How the Brain Compensates, and How MRI Shows It

Functional MRI, which measures brain activity by tracking blood flow, has revealed something encouraging about MS: the brain actively rewires itself to compensate for damage. Compared with healthy controls, MS patients tend to activate larger areas of the brain when performing the same motor, visual, or cognitive task, and sometimes recruit brain regions not normally involved.²⁸ This compensatory plasticity is most prominent in relapsing-remitting MS, though it has been observed even in people with substantial brain damage.²⁹

In one study of people with early, non-disabled MS, functional MRI during both rest and a hand-movement task showed that patients had higher baseline connectivity in their motor network and recruited additional prefrontal areas during movement. The degree of extra connectivity at rest directly correlated with the extra activation seen during the task, suggesting these two aspects of brain reorganization are linked.³⁰ Eventually, as damage accumulates beyond a threshold, this compensatory capacity may be exhausted, which could explain why some people with MS maintain function for years and then decline. Functional MRI research is helping to define where that threshold sits.

AI-Assisted Lesion Detection

Counting and measuring MS lesions by hand is tedious, time-consuming, and prone to disagreement between radiologists. Automated tools powered by deep learning are beginning to change that. One open-source tool, LST-AI, uses an ensemble of neural networks trained on nearly 500 MS cases and outperformed several existing segmentation methods, achieving Dice scores above 0.62 on public test data while older tools scored below 0.56. It also automatically labels lesions by location according to the McDonald criteria categories.³¹

Detecting new lesions between two scans taken months or years apart is an even harder task because the differences are subtle and the scans may not be perfectly aligned. A deep-learning pipeline designed specifically for this problem used transfer learning and synthetic data generation to train on the limited annotated longitudinal datasets available, achieving the best score in a MICCAI challenge dedicated to new MS lesion segmentation.³² Tools like these do not replace the radiologist, but they flag lesions that a tired human eye might miss and dramatically speed up the quantification that clinical trials depend on.

Pediatric MS Looks Different on MRI

Children and adolescents who develop MS tend to present with more dramatic MRI findings than adults. At the time of their first attack, pediatric MS patients show substantially more T2-bright lesions, more large and poorly defined lesions, and more gadolinium-enhancing lesions compared with adults at their initial presentation. However, follow-up scans show something the adult pattern does not: a significant resolution of that initial lesion burden over time. This is consistent with the idea that early lesions in children involve more reversible edema and inflammation, with less permanent axonal destruction, or that children’s brains are better at remyelination.³³

This has practical implications for radiologists and neurologists. A first MRI in a child that shows an overwhelming number of lesions can look alarming, but a substantial portion of those lesions may fade. Conversely, applying adult imaging benchmarks to pediatric scans can lead to misinterpretation in both directions.

Portable and Ultra-Low-Field MRI

Standard MS imaging is done at 1.5 or 3 Tesla, requiring large, expensive, climate-controlled scanners that are unavailable in much of the world. Portable ultra-low-field MRI devices, operating at around 0.064 Tesla, weigh a fraction of conventional scanners and can be wheeled to a bedside. These are not going to replace high-field MRI for detailed MS monitoring anytime soon, but early research has found that with adequate training, radiologists reading ultra-low-field brain scans achieved high sensitivity and positive predictive value for MS lesions.³⁴

The most promising application is triage and screening. In regions where the nearest 3T scanner is hundreds of miles away, a portable device could identify people who likely have MS and need to be referred for full diagnostic imaging. It could also serve patients with severe disability who have difficulty being transported to a radiology department. Traditional MRI studies have tended to exclude this population entirely, but portable scanners can image them at the bedside.³⁵ The image quality is crude by comparison with high-field machines, and the approach should be used conservatively, but it represents a meaningful step toward making MS imaging more accessible.

1
PubMed Central. Investigating Whether Dissemination in Time Is Essential to Diagnose Relapsing Multiple Sclerosis ↩︎
2
PubMed Central. Brain MRI lesion load at 1.5T and 3T versus clinical status in multiple sclerosis ↩︎
3
PubMed Central. Characterizing contrast-enhancing and re-enhancing lesions in multiple sclerosis ↩︎
4
PubMed. Multiple sclerosis: gadolinium enhancement in MR imaging ↩︎
5
JAMA Neurology. Multiple Sclerosis Lesions and Irreversible Brain Tissue Damage: A Comparative Ultrahigh-Field Strength Magnetic Resonance Imaging Study ↩︎
6
PubMed Central. Differentiating Multiple Sclerosis From AQP4-Neuromyelitis Optica Spectrum Disorder and MOG-Antibody Disease With Imaging ↩︎
7
JAMA Neurology. Evaluation of the Central Vein Sign as a Diagnostic Imaging Biomarker in Multiple Sclerosis ↩︎
8
PubMed Central. Sensitivity and Specificity of Central Vein Sign as a Diagnostic Biomarker in Egyptian Patients with Multiple Sclerosis ↩︎
9
PubMed Central. Chronic active lesions in multiple sclerosis: classification, terminology, and clinical significance ↩︎
10
PubMed. Paramagnetic iron rim lesions as a prognostic factor for disability progression in multiple sclerosis: A systematic review of the literature ↩︎
11
PubMed. Consensus recommendations for MS cortical lesion scoring using double inversion recovery MRI ↩︎
12
PubMed. Improved detection of juxtacortical lesions using highly accelerated double inversion-recovery MRI in patients with multiple sclerosis ↩︎
13
PubMed Central. Improved detection of multiple sclerosis lesions with T2‐prepared double inversion recovery at 3T ↩︎
14
PubMed Central. Cortical lesion hotspots and association of subpial lesions with disability in multiple sclerosis ↩︎
15
PubMed Central. In vivo evidence of disseminated subpial T2* signal changes in multiple sclerosis at 7 T: a surface-based analysis ↩︎
16
PubMed Central. Multiple sclerosis lesions and atrophy in the spinal cord: Distribution across vertebral levels and correlation with disability ↩︎
17
PubMed. Cervical cord lesion load is associated with disability independently from atrophy in MS ↩︎
18
PubMed Central. Brain atrophy in multiple sclerosis: mechanisms, clinical relevance and treatment options ↩︎
19
PubMed Central. Gray Matter Atrophy Correlates With MS Disability Progression Measured with MSFC But Not EDSS ↩︎
20
PubMed Central. Resolving the clinico-radiological paradox in multiple sclerosis ↩︎
21
PubMed Central. The clinical-radiological paradox in multiple sclerosis: myth or truth? ↩︎
22
PubMed. Diffusion tensor imaging of lesions and normal-appearing white matter in multiple sclerosis ↩︎
23
PubMed Central. Gadolinium-Based Contrast Agent Use, Their Safety, and Practice Evolution ↩︎
24
Journal of Neuroradiology. New OFSEP recommendations for MRI assessment of multiple sclerosis patients: Special consideration for gadolinium deposition and frequent acquisitions ↩︎
25
PubMed Central. Exploring in vivo lesion myelination dynamics: Longitudinal Myelin Water Imaging in early Multiple Sclerosis ↩︎
26
PubMed. Quantitative magnetization transfer and myelin water imaging of the evolution of acute multiple sclerosis lesions ↩︎
27
PubMed Central. Myelin water imaging to detect demyelination and remyelination and its validation in pathology ↩︎
28
PubMed. Functional brain reorganization in multiple sclerosis: evidence from fMRI studies ↩︎
29
PubMed Central. Neural Plasticity in Multiple Sclerosis: The Functional and Molecular Background ↩︎
30
PubMed. Brain functional plasticity at rest and during action in multiple sclerosis patients ↩︎
31
PubMed Central. LST-AI: A deep learning ensemble for accurate MS lesion segmentation ↩︎
32
PubMed Central. Longitudinal detection of new MS lesions using deep learning ↩︎
33
JAMA Neurology. Difference in Disease Burden and Activity in Pediatric Patients on Brain Magnetic Resonance Imaging at Time of Multiple Sclerosis Onset vs Adults ↩︎
34
PubMed Central. High-Field-Blinded Assessment of Portable Ultra-Low-Field Brain MRI for Multiple Sclerosis ↩︎
35
PubMed Central. Ultra-low-field MRI for bedside imaging of severe multiple sclerosis ↩︎