Brain and spinal cord lesion criteria distinguishes AQP4-positive neuromyelitis optica and MOG-positive disease from multiple sclerosis

Objective: Test the ability of a brain and spinal cord MRI criteria to differentiate neuromyelitis optica spectrum disorders and MOG-disease from MS. MRI criteria was further tested in patients with CIS and pediatric MS. Background: MOG-disease and neuromyelitis optica spectrum disorders can present clinical and radiological features strikingly similar to those of MS. Previously, diagnostic criteria based on brain MRI have been proposed to distinguish between these demyelinating diseases (Matthews-Jurynczik criteria), but spinal cord imaging and its relevance in CIS have not been evaluated. Simple brain and spinal cord MRI criteria may help separate these three inflammatory CNS diseases both in adults and children, aiding in early diagnostic decision-making, such as need for antibody testing. Design/methods: We included 150 participants (23 with aquaporin-4-positive neuromyelitis optica spectrum disorder, 14 with MOG-disease, 20 with aquaporin-4-negative neuromyelitis optica spectrum disorder, 48 with adult-onset relapsing remitting MS, 24 with pediatric-onset MS and 21 with clinically isolated syndrome). Brain and spinal cord MRI scans were anonymised and scored by 2 separate raters, based on two sets of criteria: one previously described by Matthews and colleagues (including presence of at least one lesion adjacent to the body of lateral ventricle and in the inferior temporal lobe, or presence of subcortical U-fiber lesion or a Dawson's finger-type lesion), and an extended version including spinal cord features (non-longitudinally extensive cervical lesion). Results: Extended MRI brain and spinal cord lesion criteria were able to separate adult-onset relapsing remitting MS with 100% sensitivity and 87% specificity from aquaporin-4-positive neuromyelitis optica spectrum disorder; and with 100% sensitivity and 79% specificity from MOG-disease. Additionally, brain and spinal cord criteria showed 100% sensitivity and specificity in patients presenting optic neuritis. Brain and spinal cord criteria were less sensitive in patients with CIS and in pediatric MS patients. Conclusions: Our data suggest radiological criteria can be useful to separate MS from MOG- and aquaporin-4-positive neuromyelitis optica spectrum disorders, in particular in patients with optic neuritis. Further work is needed to support their use in CIS.