Optic nerve lesions on optic nerve 2D-STIR vs brain 3D-FLAIR in clinically isolated syndromes

Abstract Objectives The 2024 McDonald criteria for multiple sclerosis include the optic nerve as the fifth location for dissemination in space. While optic nerve 2D-STIR is the recommended diagnostic approach, brain 3D-FLAIR is more routinely acquired during initial assessments of clinically isolated syndrome (CIS). This study compares optic nerve lesion detection between the two imaging techniques in CIS and evaluates the structural and functional significance of the lesions. Materials and methods This retrospective study included 51 CIS patients from a prospective cohort at University College London. All underwent optic nerve 2D-STIR, brain 3D-FLAIR, and optical coherence tomography (OCT), as well as visual function assessments. Optic nerve lesions were identified, and lesion lengths were recorded using both imaging techniques. Diagnostic performance was evaluated using clinical optic neuritis diagnosis as the reference standard. Visual function and OCT (pRNFL, GCIPL, and GCL) were compared between 2D-STIR-positive and negative eyes. Results 2D-STIR and 3D-FLAIR both showed 100% sensitivity. Specificity and accuracy were 85.5% and 91.2% for 2D-STIR, and 88.7% and 93.1% for 3D-FLAIR, respectively. Lesion detection agreement was near-perfect (Cohen’s Kappa = 0.961, p < 0.001), and lesion length measurements were highly concordant (ICC = 0.886). In non-symptomatic eyes, 2D-STIR detected slightly more lesions (14.5% vs 11.3%). 2D-STIR-positive eyes showed significantly worse visual function and thinner retinal layers. Conclusion Brain 3D-FLAIR and optic nerve 2D-STIR show high agreement in detecting optic nerve lesions, with comparable diagnostic performance. 3D-FLAIR is a practical alternative for assessing the optic nerves in routine brain imaging, especially when dedicated optic nerve acquisitions are unavailable. Key Points Question Compare brain 3D-FLAIR with dedicated optic nerve 2D-STIR for detecting optic nerve lesions in CIS and determine their association with early clinical measures . Findings 2D-STIR and 3D-FLAIR showed excellent agreement and performance in detecting optic nerve lesions. Optic nerve lesions were associated with worse vision and retinal layer thinning . Clinical relevance 3D-FLAIR offers a feasible alternative for evaluating the optic nerves on routine brain MRI, particularly when dedicated optic nerve acquisitions are unavailable .