Suspected Myelitis on MRI: Narrowing the Differential Diagnosis and Deciding What to Check Next

Intramedullary T2 hyperintensity on spinal magnetic resonance imaging is not specific for myelitis; infarction, compression, venous congestion, tumor, and nutritional deficiency produce the same finding. Nearly half of the patients referred for transverse myelitis have a non-inflammatory cause, and compressive myelopathy and spinal arteriovenous fistula are often misdiagnosed as inflammatory or neoplastic disease, delaying treatment by months. Because the tests that establish a specific cause cannot be performed exhaustively and take time, the initial scan usually sets the direction of the work-up. This review shows radiologists reading a "suspected myelitis" scan what to consider first and what to ask the clinician to confirm or refute that suspicion. Two axes organize the assessment: the time from symptom onset to nadir, and the lesion distribution on sagittal, axial, and contrast-enhanced images. Eight imaging patterns defined by these axes are each linked to the first diagnosis to consider and to the single piece of information that changes the decision. A distinctive feature is the handling of numbers: published sensitivities and specificities in spinal cord disease mostly derive from small studies with restricted populations and comparators, and exceptions are common, so each number is presented with the population and comparator in which it was obtained and is used only within that range. The normal-appearing scan, discordance between antibody or rapid polymerase chain reaction results and imaging, and techniques that secure decision-changing images are also addressed. The report should state what the imaging supports, what cannot be excluded, and what to check next.