Grey matter MRI tells neuromyelitis optica apart from MS

Multiple sclerosis and neuromyelitis optica (NMO) can look alike early on, but they need different treatment — and some MS therapies make NMO worse. Anything that sharpens the distinction from imaging people already have is clinically valuable.

This Neurology study, first-authored by Dr Arman Eshaghi, studied 90 participants in Tehran (25 MS, 30 NMO, 35 healthy controls) and a further 54 in Padua (24 MS, 20 NMO, 10 controls). Volume, thickness and surface area of 50 cortical grey matter regions plus deep grey matter nuclear volumes were derived from routine T1 and T2/FLAIR scans, and three random-forest classifiers were trained: MS versus NMO, MS versus controls, and NMO versus controls. Clinical diagnosis was the gold standard.

The classifier separated MS from NMO with 74% accuracy (sensitivity 77%, specificity 72%; p < 0.01), with MS showing greater atrophy particularly in the deep grey matter. Combining thalamic volume — the single most discriminating measure — with white matter lesion volume raised accuracy to 80%. Distinguishing each disease from healthy controls was easier still (92% and 88%). That this held across two centres, on measures extracted automatically from clinical-grade scans, is what makes it translatable rather than a single-site curiosity.

Read the paper: Gray matter MRI differentiates neuromyelitis optica from multiple sclerosis using random forest.