Non-Contrast Fat-Saturated FLAIR MRI: Utility for Detecting Posterior and Seronegative-Associated Lesions in Optic Neuritis

Abstract

Background/Objectives: To compare the lesion detection rates of fluid-attenuated inversion recovery with fat saturation (FLAIR-FS) versus T2-weighted short tau inversion recovery (T2WI-STIR) and contrast-enhanced T1-weighted imaging (CE-T1WI) in detecting lesions across antibody-defined subtypes (AQP4-ON, MOG-ON, seronegative ON) and anatomic segments of optic neuritis (ON). Methods: Forty-two patients (56 eyes) aged 5–57 years with acute/subacute optic neuritis underwent orbital MRI (T2WI-STIR, FLAIR-FS, and CE-T1WI). Two blinded neuroradiologists independently reviewed the images to assess lesion detectability and inter-observer reliability (kappa statistics). Generalized estimating equations (GEE) were used to account for within-patient correlation, with Bonferroni correction for pairwise comparisons. Results: FLAIR-FS demonstrated the highest detection rate (96.4%, 54/56 eyes), numerically higher than T2WI-STIR (80.4%, 45/56 eyes; corrected p = 0.051) and significantly higher than CE-T1WI (76.8%, 43/56 eyes; corrected p = 0.006). Inter-observer agreement was highest for FLAIR-FS (κ = 0.791), followed by CE-T1WI (κ = 0.729) and T2WI-STIR (κ = 0.681). In the posterior visual pathway, FLAIR-FS detected significantly more lesions than T2WI-STIR (corrected p = 0.024) and CE-T1WI (corrected p = 0.024). The Sequence × ON subtype interaction was not significant (p = 0.158); descriptively, FLAIR-FS showed the highest detection rate across all subtypes. Conclusions: FLAIR-FS demonstrated the highest lesion detection rate among the three sequences, particularly in the posterior visual pathway. FLAIR-FS should be regarded as a useful complementary non-contrast sequence, particularly when gadolinium cannot be administered.