A predictive model for triple negative optic neuritis patients: Identifying the subset of patients with favorable outcomes

Abstract

Background and objectives: A subset of optic neuritis (ON) patients remains seronegative for AQP4-IgG, MOG-IgG, and CSF oligoclonal bands (OCBs), complicating diagnosis and management. This study aimed to characterize these triple-negative ON (TNON) cases and compare them to MS, MOGAD, and NMOSD.

Methods: We retrospectively analyzed 103 patients with a first ON episode tested for AQP4-IgG, MOG-IgG, and OCBs, classifying them into MS (n=26), MOGAD (n=25), NMOSD (n=13), and TNON (n=39). TNON was further divided into evolving MS (evMS) and idiopathic TNON. A support vector machine (SVM) was used to classify TNON based on onset data.

Results: TNON patients showed less optic disc edema and radiological involvement than MOGAD and responded more poorly to steroids. However, TNON had lower relapse rates and less need for chronic treatment than MS or NMOSD. SVM analysis revealed that most evMS cases clustered with MS, while idiopathic TNON aligned more closely with MOGAD.

Conclusion: TNON is a heterogeneous entity encompassing both early MS and a distinct idiopathic subgroup. While clinical features alone could not differentiate these subgroups, our predictive model successfully stratified them at onset. Idiopathic TNON may reflect an undetected antibody-mediated process with a generally favorable prognosis.

Keywords: MOGAD; MS; NMOSD; ON; SVM model; TNON.