Fellow-eye retinal age gap and high injection burden over four years in neovascular age-related macular degeneration

The study evaluated whether fellow-eye retinal age gap, a deep learning-derived fundus metric, is associated with long-term anti-vascular endothelial growth factor injection burden in treatment-naive neovascular age-related macular degeneration. Retinal age gap has mostly been studied in relation to systemic or ocular disease risk; this work places it in the setting of long-term injection burden during anti-vascular endothelial growth factor therapy. Retrospective cohort design, 80 treatment-naive patients, follow-up of at least 4 years, with retinal age estimated from fellow-eye fundus photographs by the publicly available Japan Ocular Imaging Registry model.

Retinal age correlated with baseline best-corrected visual acuity and central retinal thickness, while raw retinal age gap correlated with central choroidal thickness. This indicates cross-sectional links between the deep learning metric and baseline ocular structural and functional parameters, providing context for the longitudinal analyses. Univariable analyses in a sample of 80 treatment-naive neovascular age-related macular degeneration patients.

In an exploratory threshold-based analysis, the ≥ 24-injection group had higher raw retinal age gap than the < 24-injection group, and higher raw retinal age gap was associated with requiring ≥ 24 injections. This is the first report of an association signal between fellow-eye retinal age gap and high injection burden within a four-year threshold framework. Between-group comparison P = 0.039; odds ratio 1.11 (95% CI, 1.01-1.22; P = 0.029), from an exploratory threshold-based analysis.

The threshold association was attenuated after age correction and was nonsignificant in alternative-threshold and count analyses. This shows the observed signal is sensitive to the analytic approach and is not yet a stable predictive marker. Association attenuated after age correction; not replicated at alternative thresholds; in the count analysis neither raw nor age-corrected retinal age gap was associated with total injection number.

The results apply to treatment-naive neovascular age-related macular degeneration patients receiving anti-vascular endothelial growth factor therapy and followed for at least 4 years, with retinal age estimated from fellow-eye fundus photographs by the publicly available Japan Ocular Imaging Registry model; their significance is to suggest fellow-eye retinal age gap as a fundus photograph-based candidate marker for further studies to test in larger samples and different models in relation to sustained high injection burden.

Readers should still note that the association was attenuated after age correction and was nonsignificant in alternative-threshold and count analyses, indicating sensitivity to the analytic approach; the sample of 80 patients is retrospective and retinal age depends on a specific publicly available model; whether a causal or predictive relationship exists between retinal age gap and injection burden, and how it performs in other populations and models, remain open questions for further study.