Document resource
Background/aims To evaluate the respective roles of systemic vascular vulnerability and local vascular anatomy in the onset of branch retinal vein occlusion (BRVO) by analysing the retinal age gap and arteriovenous crossing patterns.Methods This observational study enrolled 202 patients with unilateral BRVO and 100 age-matched, sex-matched and axial length-matched healthy controls. A deep learning model was applied to colour fundus images to estimate the retinal age; the retinal age gap (predicted retinal age minus chronological age) in the fellow eye was used as a surrogate residual marker of systemic vascular vulnerability. At the occlusion site, arteriovenous crossing was classified (arterial vs venous overcrossing) using optical coherence tomography angiography (OCTA).Results The mean retinal age gap was 1.8±6.5 years and 0.2±6.0 years in the BRVO and control groups, respectively. Retinal age gap was significantly larger in BRVO eyes than in control eyes (p=0.046). A larger retinal age gap in the fellow eye correlated with younger age at BRVO onset (ρ= –0.350, p<0.001). Among 156 BRVO eyes with OCTA data, venous overcrossing at the occlusion site was more frequent in younger-onset major BRVO (p=0.031), whereas in macular BRVO, onset age did not differ by crossing pattern (p=0.734).Conclusion Both systemic vascular vulnerability, reflected by a larger retinal age gap and local anatomical configuration, represented by venous overcrossing, may jointly contribute to BRVO susceptibility, particularly in younger patients. These findings should be interpreted with caution because the retinal age gap is a prediction, using a deep learning model, rather than a direct biological measure.