Radiological predictors of visual outcome in myelin oligodendrocyte glycoprotein-related optic neuritis.

Abstract

Objective: This study aimed to determine whether magnetic resonance imaging (MRI) biomarkers are associated with visual prognosis in myelin oligodendrocyte protein (MOG)-associated optic neuritis (MOG-ON).

Design: Cross-sectional analysis.

Subjects: Patients meeting 2023 international diagnostic criteria for MOG antibody-associated disease who were seen for first episodes of MOG-ON at three tertiary neuro-ophthalmology practices between January 2017 and July 2023 were enrolled. Patients who received less than 3 months of neuro-ophthalmic follow-up and did not demonstrate visual recovery (visual acuity [VA] ≥20/20 and visual field mean deviation [VFMD] >-5.0 dB) during this time were excluded.

Methods: Patients received contrast-enhanced, fat-suppressed MRI of the brain and orbits within one month of symptom onset.

Main outcome measures: The associations between radiological biomarkers and poor VA outcome (<20/40), incomplete VA recovery (<20/20), and poor VFMD outcome (VFMD <-5.0 dB) were assessed using multivariable logistic regression adjusting for time from symptom onset to treatment and nadir VA or VFMD. Radiological biomarkers included length of optic nerve enhancement (below vs. above 25%, 50%, and 75%); degree of orbital, canalicular, and intracranial or chiasmal optic nerve enhancement (mild vs. moderate-severe compared to the lacrimal gland); and absence vs. presence of optic nerve sheath enhancement on baseline T1-weighted MRI.

Results: A total of 129 eyes of 92 patients (median [IQR] age 37.0 [20.8-51.3], 65.2% female) were included. Poor VA outcome was seen in 6.2% of cases, incomplete VA recovery in 19.4%, and poor VFMD outcome in 16.9%. Compared to eyes with moderate-severe enhancement, eyes with mild orbital optic nerve enhancement were more likely to have poor VA outcome (OR 8.57; 95% CI [1.85, 51.14], P=0.009), incomplete VA recovery (OR 7.31, 95% CI [2.42, 25.47], P=0.001), and poor VFMD outcome (adjusting for time to treatment: OR 6.81, 95% CI [1.85, 28.98], P=0.005; adjusting for nadir VFMD: OR 11.65, 95% CI [1.60, 240.09], P=0.04). Lack of optic nerve sheath enhancement was additionally associated with incomplete VA recovery (OR 3.86, 95% CI [1.19, 12.85], P=0.02) compared to the presence of enhancement. These associations remained consistent in subgroup logistic regression analysis of MRIs performed before initiation of treatment but were not seen in pairwise analysis of MRIs performed after treatment.

Conclusions: In eyes with first MOG-ON episodes, milder enhancement in the orbital optic nerve is associated with poorer VA and VF recovery. Prospective and mechanistic studies are needed to confirm the prognostic utility of MRI in MOG-ON.