Prospective Evaluation of the 2022 Optic Neuritis Criteria When Optic Neuritis Is Suspected

This multicenter prospective observational study evaluates the 2022 International Criteria for Optic Neuritis, alone and with an antibody-stratified adjunct, in patients initially suspected of having optic neuritis. Structured prospective data collection began on January 6, 2025. Each participant contributes one clinical episode. The cohort is not restricted to participants ultimately diagnosed with optic neuritis. Diagnostic testing and treatment are determined by the treating clinicians as part of routine care.

Baseline symptoms and ophthalmic findings are recorded using a structured electronic case report form. Available visual field, optical coherence tomography, magnetic resonance imaging, and laboratory findings are collected. Baseline and follow-up optical coherence tomography are encouraged when clinically feasible but are not required by study participation. Available follow-up findings can support reference-standard adjudication. Absence of follow-up optical coherence tomography alone does not exclude a participant from the primary analysis.

Both index strategies use baseline clinical data and eligible paraclinical information obtained within 3 months after symptom onset for the current episode, based on examination or sample collection dates. Later information will not revise either index classification. The 2022 criteria assign definite optic neuritis, possible optic neuritis, or not optic neuritis using prospectively recorded clinical findings, including actual color vision assessments rather than assumed abnormalities.

The antibody-stratified adjunct described in Table 2 of Liang and colleagues (2026) is applied only to participants with aquaporin-4 immunoglobulin G positivity or clear-positive myelin oligodendrocyte glycoprotein immunoglobulin G results in the index window. Under that framework, acute or subacute visual loss with at least one objective finding consistent with optic neuropathy supports definite optic neuritis; without such an objective finding, the adjunct category is possible optic neuritis. The final index classification is the higher-certainty category assigned by the 2022 criteria or the adjunct, using the hierarchy definite optic neuritis, possible optic neuritis, then not optic neuritis. Participants without either qualifying positive antibody result retain their original 2022 criteria classification. Unavailable antibody results are not treated as negative.

Participants are followed for up to 6 months after enrollment whenever feasible. Reference-standard adjudication uses baseline information and available clinical course, repeat examinations, investigations, treatment response, and alternative diagnoses within this period. The target is whether the current episode represents optic neuritis, not merely whether an antibody-associated disease is present.

For each participant, two experts independently assess the case in two stages. In Stage 1, current and historical aquaporin-4 immunoglobulin G and myelin oligodendrocyte glycoprotein immunoglobulin G results, including directly revealing diagnostic labels, are withheld. Stage 1 judgments are recorded and locked before antibody disclosure. In Stage 2, the same experts reassess the case after receiving available antibody results and relevant assay information. Both stages use otherwise identical case information and the same information cutoff, no later than 6 months after enrollment. At each stage, cases are classified as optic neuritis, non-optic neuritis, or indeterminate. A third expert resolves disagreements under the corresponding information restrictions; Stage 1 disagreements are resolved before disclosure.

At both stages, experts are masked to the index classifications. Personnel assigning or verifying index classifications are masked to the expert reference diagnoses. Masking antibody results from experts does not remove those results from either index strategy. Index classifications and reference diagnoses are recorded separately and finalized before comparative analysis. Stage 2 full-information adjudication is the primary reference standard. Stage 1 provides a supplementary reference for sensitivity analyses and assessment of diagnostic changes after antibody disclosure. Initial agreement between experts is summarized at each stage using percentage agreement and Cohen's kappa.

The primary analysis treats definite optic neuritis as index-positive and not optic neuritis as index-negative; possible optic neuritis is excluded from this conditional binary analysis and reported separately. A secondary low-threshold analysis treats definite or possible optic neuritis as positive. Indeterminate reference diagnoses are excluded from the corresponding binary accuracy calculations but retained in reported category distributions. Definite-classification yield is also compared within the same reference-positive participants, retaining possible classifications as not definite. Newly upgraded cases are described according to their reference diagnoses. First-stage indeterminacy does not exclude a participant from second-stage analyses. Detailed outcome definitions are provided in the outcome measures section.

At initial public registration, recruitment was ongoing, and database lock, formal reference-standard adjudication, final index classification, and the primary diagnostic accuracy analysis had not occurred. The antibody-masked stage is an amendment-added assessment; full-information expert adjudication remains the primary reference standard.