MRI of the relevant domain should be performed to confirm whether clinical symptoms represent an attack of NMOSD: Commentary

Neuromyelitis optica spectrum disorder (NMOSD) attacks are often severe and leave little doubt about the diagnosis. However, diagnosis may be challenging for one or more reasons:

1. Preexisting neurological deficits, which may create ‘ceiling’ effects that complicate clinical interpretation or measurement.

2. Some NMOSD attacks have no objective signs, for example, area postrema syndrome characterized by intractable nausea and vomiting; early myelitis presenting with radicular pain.

3. Patients in clinical trials may present very early before clinical signs are well developed.

4. ‘Pseudo attacks’ may occur due to other pathologies (e.g. complicating CNS infection and injuries) or due to systemic illnesses that aggravate preexisting disability.

The consequences of misdiagnosis are important for clinical practice and especially for clinical trials, the outcomes of which hinge generally on the accurate diagnosis of a single event for each participant; typically, a single attack leads to withdrawal from the randomized portion of a clinical trial and either exit from the trial or transition to an open label extension.

Contentti and Wingerchuk argue that magnetic resonance imaging (MRI) (detection of new T2 lesions or gadolinium enhancement) can enhance both sensitivity and specificity of diagnosis, particularly in situations noted above. Marignier and Cree agree that MRI may be helpful when diagnostic uncertainty exists but contend that MRI is unnecessary when clinical signs are clear and there are no confounding issues such as significant prior deficits. Furthermore, MRI is not entirely sensitive or specific for diagnosis of relapse. The two groups take different views on the feasibility of requiring MRI for adjudication of relapses. Marignier and Cree argue that requiring MRI would be impossible in many countries and may not be available in a timely fashion for decisions about treatment. Contentti and Wingerchuk cite a survey suggesting that most neurologists confirm relapses with MRI and contend that rapid MRI is immediately available in 59% and within 72 h in 96%.

Both sides appear to agree that MRI may be helpful in the face of diagnostic uncertainty and might agree that MRIs should be performed to enhance the reliability of attack adjudication in clinical trials. The recent NMOmentum study of inebilizumab that routinely required MRIs as part of the event evaluation process found that agreement between investigators and adjudicators was 87%, compared to 21% in the eculizumab and 50%–75% in the two trials of satralizumab where MRIs were not routinely performed ¹ ; routine MRI is likely an important, but not the sole factor accounting for these findings. However, Marignier and Cree correctly point out that interpretation of MRIs, while usually straightforward, may be difficult when there is scant gadolinium enhancement or preexisting T2 signal abnormality, particularly for the optic nerve. Silent MRI changes in asymptomatic individuals, a helpful way to assess disease activity for MS, occur rarely in NMOSD, and when they do, may herald an incipient attack. ²

MRI is likely a helpful adjunct to diagnosis of acute relapses of NMOSD, but may be unnecessary in all individuals, especially outside of a clinical trial, may not be feasible in some medical settings, and may be difficult to interpret in some instances.