Retinal hyper-reflective foci link retinal and cortical pathology in paediatric multiple sclerosis

Pediatric-onset multiple sclerosis is associated with high inflammatory activity early in the disease course, yet sensitive biomarkers of early disease pathology are limited. Hyper-reflective foci on optical coherence tomography have been proposed as markers of inflammation in adult multiple sclerosis, but their relevance in pediatric populations remains unclear. We aimed to identify and quantify retinal layer-specific hyper-reflective foci burden in children with pediatric-onset multiple sclerosis and evaluate associations with MRI markers of disease severity. In this cross-sectional study, 53 children with pediatric-onset multiple sclerosis and 36 age- and sex-matched healthy controls underwent spectral-domain optical coherence tomography and MRI, including three-dimensional T1-weighted and fluid-attenuated inversion recovery sequences, near disease onset. Patients with a history of optic neuritis were excluded. Hyper-reflective foci were quantified within the ganglion cell-inner plexiform layer and inner nuclear layer as counts and as a normalized hyper-reflective foci index (count divided by retinal layer volume). Brain MRIs underwent parcellation and multiple sclerosis lesion segmentation. Generalized estimating equation models accounted for inter-eye correlation and adjusted for demographic and retinal structural covariates. Linear regression assessed associations between hyper-reflective foci burden and magnetic resonance imaging measures, including thalamic volume, cortical volume, and white matter lesion volume. Hyper-reflective foci counts and indexes were significantly increased in pediatric-onset multiple sclerosis compared with controls across both retinal layers. Higher ganglion cell-inner plexiform layer hyper-reflective foci index correlated with greater white matter lesion volume (r = 0.38, P = 0.006) and lower thalamic volume (r = -0.35, P = 0.012), but not cortical volume. In adjusted models, multiple sclerosis was independently associated with higher ganglion cell-inner plexiform and inner nuclear layer hyper-reflective foci counts and indexes (both P < 0.001). Inner nuclear layer hyper-reflective foci were negatively associated with macular retinal nerve fiber layer thickness and volume (β = -0.18, P < 0.01; β = -0.006, P < 0.01). Higher ganglion cell-inner plexiform layer hyper-reflective foci index remained associated with lower thalamic volume (β = -0.321, P = 0.024), greater lesion volume (β = 5.253, P = 0.011), and lower cortical volume (β = -6.403, P = 0.016). Children with multiple sclerosis demonstrate increased retinal hyper-reflective foci burden early in disease, in the absence of optic neuritis. The observed relationships between ganglion cell-inner plexiform layer hyper-reflective foci burden, thalamic and cortical atrophy, and white matter lesion volume suggest that hyper-reflective foci capture aspects of both inflammatory and neurodegenerative disease activity. Hyper-reflective foci represent a promising non-invasive biomarker of disease severity in pediatric-onset multiple sclerosis.