Multiple sclerosis MRI

Multiple sclerosis Microchapters
Home
Patient Information
Overview
Historical Perspective
Classification
Pathophysiology
Causes
Differentiating Multiple sclerosis from other Diseases
Epidemiology and Demographics
Risk Factors
Screening
Natural History, Complications and Prognosis
Diagnosis
Diagnostic Study of Choice
History and Symptoms
Physical Examination
Laboratory Findings
Electrocardiogram
X-ray
Echocardiography or Ultrasound
CT Scan
MRI
Other Imaging Findings
Other Diagnostic Studies
Treatment
Medical Therapy
Surgery
Alternative Therapies
Primary Prevention
Secondary Prevention
Tertiary Prevention
Cost-Effectiveness of Therapy
Future or Investigational Therapies
Case Studies
Case #1
Multiple sclerosis MRI On the Web
Most recent articles
cited articles
Review articles
CME Programs
Powerpoint slides
Images
American Roentgen Ray Society Images of Multiple sclerosis MRI All Images X-rays Echo & Ultrasound CT Images MRI
Ongoing Trials at Clinical Trials.gov
US National Guidelines Clearinghouse
NICE Guidance
FDA on Multiple sclerosis MRI
CDC on Multiple sclerosis MRI
Multiple sclerosis MRI in the news
Blogs on Multiple sclerosis MRI
Directions to Hospitals Treating Multiple sclerosis
Risk calculators and risk factors for Multiple sclerosis MRI

Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]; Associate Editor(s)-in-Chief: Fahimeh Shojaei, M.D. Julinka Auta Fernandes

Overview

On magnetic resonance imaging (MRI), multiple sclerosis (MS) is characterized by focal lesions with typical morphology and distribution in the brain, spinal cord, and optic nerve. Typical lesions are hyperintense on T2-weighted or fluid-attenuated inversion recovery (FLAIR) images and preferentially involve the periventricular, cortical/juxtacortical, infratentorial, spinal-cord, and optic-nerve regions. MRI can demonstrate dissemination in space and time, while susceptibility-sensitive MRI can demonstrate supportive features such as the central vein sign and paramagnetic rim lesions. MRI findings are not pathognomonic; they must be interpreted in the appropriate clinical setting and after exclusion of a better explanation.

MRI findings

Characteristic brain findings

T2-weighted and FLAIR lesions

• Typical MS lesions are focal areas of T2/FLAIR hyperintensity. They are usually round or ovoid, have relatively well-defined margins, and generally measure at least 3 mm along the longest axis. A lesion smaller than 3 mm may be real, particularly in the infratentorial compartment or spinal cord, but should be counted cautiously because small vessels, enlarged perivascular spaces, and artifacts can mimic lesions.
• Lesions may be sharply marginated or, during acute inflammation, larger and less sharply defined because of edema. Lesions can subsequently shrink as edema resolves. Size and border definition alone cannot reliably determine lesion age.
• Typical lesion topographies are:
  • Periventricular: a lesion directly contacts the ventricular surface without intervening normal white matter. Ovoid lesions oriented perpendicular to the lateral ventricles along medullary veins are termed Dawson fingers. Symmetric periventricular caps or smooth bands and lesions merely close to, but not touching, the ventricle should not be counted as typical periventricular MS lesions.
  • Cortical/juxtacortical: a cortical lesion involves the cortex; a juxtacortical lesion directly abuts the cortex without intervening normal white matter and may involve U-fibers. A subcortical or deep-white-matter lesion separated from the cortex by normal white matter is not juxtacortical. Cortical lesions are often underdetected on conventional MRI.
  • Infratentorial: lesions may occur in the brainstem, cerebellar peduncles, or cerebellum. Typical brainstem lesions are commonly focal and peripheral rather than symmetric and central.
  • Corpus callosum/callososeptal interface: callosal lesions, particularly at the callososeptal interface, strongly support a demyelinating pattern, although the corpus callosum is not a separate topography in the current dissemination-in-space count.
  • Deep white matter: deep-white-matter lesions can occur in MS but, by themselves, are less specific and do not form a separate diagnostic topography.
• Lesions meeting more than one anatomic description should not be double-counted when determining the number of involved topographies.

Brain MRI demonstrating multiple sclerosis lesions. Image by James Heilman, MD; Wikimedia Commons; licensed under CC BY-SA 4.0.

Axial double-inversion-recovery brain MRI. White circles identify two cortical lesions and one juxtacortical lesion. Image by Favaretto et al.; Wikimedia Commons; licensed under CC BY 4.0.

Gadolinium enhancement

• Active inflammatory lesions may enhance after gadolinium because of blood–brain barrier disruption. Enhancement is usually nodular, homogeneous, or open/incomplete ring-like; complete ring enhancement can occur but is less typical.
• Enhancement is generally transient and commonly lasts for several weeks. Persistent enhancement, especially when accompanied by progressive enlargement, marked mass effect, necrosis, or disproportionate edema, is atypical and should prompt consideration of neoplasm, infection, inflammatory vasculopathy, sarcoidosis, or another diagnosis.
• Simultaneous typical enhancing and nonenhancing lesions on the same examination demonstrate MRI dissemination in time. The absence of enhancement does not exclude recent inflammatory activity, and enhancement alone does not establish MS.

T1-weighted findings

• Many T2 lesions are isointense on noncontrast T1-weighted images. T1-hypointense lesions are commonly called black holes.
• A transient T1-hypointense lesion may become isointense as edema resolves and tissue partially repairs. A persistent T1 black hole is associated with more severe myelin and axonal loss, but it is neither specific for MS nor independently diagnostic.

Tumefactive demyelinating lesions

• Tumefactive demyelinating lesions are usually larger than 2 cm and may have edema or mass effect. Findings favoring demyelination include an open-ring enhancement pattern, relatively limited mass effect for lesion size, peripheral diffusion restriction at the active edge rather than central abscess-like restriction, and the presence of other typical MS lesions.
• A tumefactive lesion is not diagnostic of MS and requires careful exclusion of neoplasm, abscess, infarction, and other inflammatory disorders, particularly when it enlarges or enhances persistently.

Brain atrophy

• MS may be associated with global and regional brain-volume loss, including cortical and deep-gray-matter atrophy; thalamic atrophy is particularly common and correlates at group level with disease burden and disability.
• Atrophy is not specific for MS and is not part of the MRI dissemination-in-space or dissemination-in-time criteria. Apparent volume change can be affected by normal aging, hydration, inflammation, corticosteroids, scanner differences, and analysis software; visual or automated atrophy measurements should therefore not be interpreted as stand-alone proof of MS or interval inflammatory activity.

Characteristic spinal-cord findings

• Typical MS cord lesions are focal T2-hyperintense lesions, commonly involving the cervical cord but potentially occurring at any level. They are usually:
  • short, most often less than two vertebral-body segments in craniocaudal length;
  • peripheral, frequently dorsal or lateral, on axial images;
  • smaller than one-half of the cord cross-sectional area; and
  • cigar-shaped on sagittal images and wedge-shaped on axial images.
• An acute cord lesion can enhance and may produce mild focal swelling. A chronic lesion may become sharply marginated and may be associated with focal cord atrophy. Diffuse cord signal abnormality or atrophy without a definite focal lesion should not be counted as a spinal-cord lesion for dissemination in space.
• Findings atypical for conventional MS include a longitudinally extensive lesion spanning three or more vertebral-body segments, central gray-matter-predominant involvement, involvement of more than half of the cord cross-sectional area, marked swelling, cavitation, prominent conus involvement, or a lesion restricted to a level of severe mechanical compression. These patterns should prompt consideration of neuromyelitis optica spectrum disorder, myelin oligodendrocyte glycoprotein antibody-associated disease, vascular, infectious, metabolic, neoplastic, sarcoid, or compressive myelopathy.

Contrast-enhanced cervical spinal-cord MRI showing multiple sclerosis involvement. Image from the National Institutes of Health; Wikimedia Commons; United States public domain.

Characteristic optic-nerve findings

• On dedicated optic-nerve MRI, acute MS-associated optic neuritis commonly appears as unilateral optic-nerve T2 hyperintensity and/or enhancement. A typical lesion is short and most often affects the intraorbital segment. Mild nerve swelling may be present.
• Chronic injury may result in optic-nerve atrophy and persistent T2 signal abnormality; these findings are not specific for MS.
• Bilateral simultaneous involvement, a lesion extending through more than half the nerve, predominant posterior optic-nerve or chiasmal involvement, marked swelling, perineural enhancement, optic-tract involvement, or a long longitudinally extensive lesion is atypical for MS-associated optic neuritis and should prompt consideration of MOG-antibody-associated disease, neuromyelitis optica spectrum disorder, sarcoidosis, infection, infiltration, or ischemic optic neuropathy.

Susceptibility-sensitive MRI findings

Central vein sign

• The central vein sign (CVS) is a small vein running centrally through a white-matter lesion on susceptibility-sensitive MRI. In an eligible lesion, the vein appears as a thin hypointense line or dot, generally less than 2 mm in apparent diameter, lies centrally within the lesion, and is visible in at least two perpendicular planes, with a line visible in at least one plane.
• Lesions assessed for CVS should generally be discrete, at least 3 mm, and technically evaluable. Confluent lesions, lesions distorted by artifact, and lesions containing several veins without an identifiable central vein should be excluded rather than scored negative.
• Under the 2024 revised diagnostic framework, a positive CVS assessment can be determined by:
  • the select-6 rule when more than nine eligible lesions are present: six CVS-positive lesions; or
  • the majority rule when fewer than ten eligible lesions are present: more than 50% of eligible lesions are CVS-positive.
• CVS increases the specificity of a typical lesion pattern but is not unique to MS and is not a stand-alone diagnosis. A scan on which CVS cannot be evaluated reliably should be described as technically unevaluable rather than CVS-negative.

Paramagnetic rim lesions

• A paramagnetic rim lesion (PRL) is a chronic active white-matter lesion with a T2-hyperintense core and a discrete paramagnetic rim on susceptibility-sensitive MRI at a field strength of at least 1.5 T. For a definite PRL, the rim should:
  • follow at least two-thirds of the lesion's outer white-matter edge on the slice of greatest visibility;
  • be visible on at least two consecutive two-dimensional slices or in two orthogonal planes of a three-dimensional acquisition;
  • be distinct from a central or peripheral vein and from susceptibility artifact; and
  • surround a nonenhancing lesion.
• Use particular caution with lesions smaller than 3 mm, an apparent rim thicker than approximately 2 mm, lesions near air–tissue interfaces, calcification, dipole artifact, and lesions touching cortex or ventricle where the complete white-matter margin cannot be assessed.
• If contemporaneous postcontrast images are unavailable, a candidate lesion should have been present for at least 3 months, ideally 6 months, before it is called a definite nonenhancing PRL; otherwise it should be considered possible and confirmed on a later comparable examination.
• One or more definite PRLs can strongly support MS in the appropriate diagnostic setting, but PRLs are not stand-alone proof of MS.

MRI dissemination in space and time

The 2024 revisions of the McDonald criteria were published in 2025. The following describes only the MRI elements; MRI findings must be used within the complete diagnostic criteria.

• Dissemination in space (DIS): typical lesions in at least two of five MRI topographies:

1. periventricular;
2. cortical/juxtacortical;
3. infratentorial;
4. spinal cord; and
5. optic nerve.

• Dissemination in time (DIT): either of the following MRI findings:
  • simultaneous typical gadolinium-enhancing and nonenhancing lesions on one examination; or
  • a new or unequivocally enlarging T2 lesion and/or a new enhancing lesion compared with an earlier technically comparable MRI.
• Typical lesions in at least four of the five topographies can satisfy the imaging component of a specified 2024 diagnostic pathway without separate MRI evidence of dissemination in time.
• CVS and PRLs can add specificity and contribute to specified pathways in the revised criteria; neither finding should be treated as independent proof of MS.
• In a progressive-onset presentation with at least 12 months of progression, two or more typical spinal-cord lesions can satisfy the relevant MRI requirement within the complete revised algorithm.
• In a radiologically isolated presentation, typical lesion morphology and topography, evidence of new MRI activity, and a positive CVS assessment can contribute to the revised diagnostic pathway. Incidental nonspecific white-matter hyperintensities should not be relabeled as demyelination solely because of their number.
• In people aged 50 years or older, or in people with headache disorders or vascular risk factors, cerebral white-matter lesions have lower specificity. A typical spinal-cord lesion or a positive CVS assessment increases MRI specificity in these settings.

Atypical MRI findings and important mimics

• Findings favoring small-vessel ischemic disease or migraine rather than MS include small rounded lesions confined to the deep or subcortical white matter, lesions that spare the U-fibers and do not touch the ventricles, lacunes, prominent perivascular spaces, and numerous microbleeds. Age and vascular risk must be considered.
• Symmetric confluent white-matter disease, prominent anterior temporal-pole or external-capsule involvement, basal-ganglia lesions, cortical infarcts, hemorrhage, or extensive diffusion restriction should prompt consideration of hereditary or acquired vascular, metabolic, toxic, mitochondrial, infectious, or leukodystrophic disorders.
• Long bilateral optic-nerve lesions, chiasmal involvement, perineural enhancement, or longitudinally extensive central cord lesions are red flags for neuromyelitis optica spectrum disorder, MOG-antibody-associated disease, sarcoidosis, infection, or another inflammatory disorder.
• Leptomeningeal or ependymal enhancement, persistent lesion enhancement, abscess-like central diffusion restriction, substantial mass effect, necrosis, cavitation, or progressive lesion enlargement is atypical and requires an alternative diagnostic evaluation.
• In children, large bilateral poorly marginated lesions, widespread deep-gray-matter involvement, and longitudinally extensive cord lesions should raise consideration of acute disseminated encephalomyelitis or MOG-antibody-associated disease; the presence of typical MS morphology and topography remains essential.
• A normal MRI reduces but does not completely exclude early MS, especially when the relevant brain, spinal-cord, or optic-nerve region has not been adequately assessed. Conversely, a high lesion count does not establish MS when lesion morphology, distribution, or the clinical presentation is atypical.
• Automated lesion detection, segmentation, subtraction, or classification can support comparison, but false-positive and false-negative results occur. The source images require expert review before a lesion is classified as new, enlarging, enhancing, CVS-positive, or a PRL.