Pediatrics
Pediatric MS Mimics: Which Demyelination Is It?
Fellowship-level neuroradiology teaching by E. Brooke Schrickel, MD. Open this topic in the interactive reading room →
Core clinical idea
The job isn't 'is there demyelination' but which one — MS, NMOSD, MOGAD, and ADEM diverge on antibody testing, spine protocol, and immunotherapy. Match the lesion's shape, location, and company to the right serology, confirm with the cord pattern, and deliberately interrogate the area postrema and optic-nerve segment before signing. Encephalopathy gates ADEM; discrete ovoid perpendicular periventricular lesions and black holes pull toward MS.
Bottom line
Don't say 'demyelination' — match lesion shape, distribution, and cord pattern to MS / MOGAD / NMOSD / ADEM, and check area postrema and the optic-nerve segment.
Core workstation questions
- Are the lesions discrete/ovoid/perpendicular (MS), fluffy/bilateral/BG-thalamic with peduncle and pontine tegmentum involvement (MOGAD), aquaporin-4-distributed (NMOSD), or large/bilateral with encephalopathy (ADEM)?
- Are there black holes, which favor MS?
- Is there encephalopathy, the defining gate for ADEM?
- Did I image the whole spine — short-segment dorsal (MS) vs LETM/central H-sign (NMOSD/MOGAD)?
- Did I look specifically at the area postrema and name the involved optic-nerve segment?
- Is gadolinium actually needed on this follow-up study?
What changes reporting / management
- Describe morphology and distribution and name the favored entity rather than writing 'demyelinating lesions' — it dictates the antibody panel and therapy path.
- Do not downgrade ADEM because lesions don't enhance or restrict; vasogenic, non/faintly enhancing, non-restricting lesions with encephalopathy are the expected look.
- Always include the whole spine in suspected NMOSD/MOGAD and report segment length and central-vs-peripheral pattern, because the cord finding can flip the diagnosis.
- Interrogate the area postrema in a child with intractable hiccups/nausea/vomiting (NMOSD) and report the optic-nerve segment (anterior+perineural MOGAD, mid/discrete MS, posterior+chiasm NMOSD).
- Give gadolinium at diagnosis to mark activity, but routine follow-up usually does not need it — new T2/FLAIR lesions track activity.
Practical traps
- Missing the tiny area postrema NMOSD lesion that explains the clinical syndrome because it wasn't looked for on purpose.
- Excluding ADEM because lesions lack enhancement or diffusion restriction — that is the expected appearance.
- Signing pediatric demyelination without spine imaging, losing the LETM/H-sign vs short-segment discriminator.
Teaching pearls
- Discrete ovoid perpendicular periventricular lesions + black holes = MS.
- Fluffy, bilateral, basal ganglia/thalamic with cerebellar peduncle and pontine tegmentum involvement = think MOGAD.
- Aquaporin-4 distribution and a tiny area postrema lesion = NMOSD — look for it on purpose.
- Large bilateral vasogenic lesions + encephalopathy = ADEM; absent enhancement/restriction doesn't exclude it.
- Always image the spine: LETM/H-sign = NMOSD/MOGAD; short-segment dorsal = MS.
- Optic-nerve segment is a clue: anterior+perineural (MOGAD), mid/discrete (MS), posterior+chiasm (NMOSD).
- Give gadolinium at diagnosis, then mostly skip it on follow-up.
Source lectures
- Pediatric MS and Mimics
Educational material for radiology residents and neuroradiology fellows. Nothing here drives individual patient care, and it contains no patient data.