Retinal curvature as a regional morphologic phenotype of the posterior pole in high myopia: A two-center OCTA study

To evaluate retinal curvature across macular regions in eyes with high myopia and to examine its relationship with retinal thickness and optical coherence tomography angiography-derived retinal flow-density metrics within a unified spatial framework.

In this dual-center observational study, participants underwent swept-source optical coherence tomography angiography imaging using a 24 × 20 mm macular scan. Retinal curvature (RC) was derived from Bruch's membrane-based surface reconstruction and analyzed across six concentric macular rings. Corresponding regional measurements of retinal thickness (RT) and optical coherence tomography angiography (OCTA)-derived retinal flow-density metric (RF) were obtained. Eye-level comparisons between high myopia and non-high myopia were performed using generalized estimating equation models adjusted for age, sex, and study center. Additional models including axial length were performed as sensitivity analyses. Interocular analyses were conducted using Spearman correlation based on left-right differences.

A total of 288 eyes from 144 participants were included. Compared with non-high myopia (non-HM) eyes, high myopia (HM) eyes showed lower retinal curvature (RC) in the outer macular region, with significant differences at Rings 5 and 6; the difference at Ring 6 was β = -0.3636 (95% CI, -0.4722 to -0.2550; q < 0.001), whereas Rings 1 and 3 showed no significant differences. RT differences were limited and did not remain significant after false discovery rate (FDR) correction, including Ring 3 (β = -63.8362; 95% CI, -125.5506 to -2.1218; q = 0.128). RF metrics did not show consistent differences across rings after FDR correction. In eye-level analyses, RC was associated with RT at Ring 1 (β = -18.2924; 95% CI, -31.1265 to -5.4582; p = 0.005), but not at Rings 3 or 6. Interocular analysis showed a positive correlation between ΔRC and ΔRT at Ring 6 (ρ = 0.382; 95% CI, 0.225 to 0.523; q < 0.001), whereas ΔRC-ΔRF correlations were not significant.

Using a ring-based macular framework, retinal curvature differences between high myopia and non-HM eyes were observed primarily in outer macular regions. Corresponding variations in retinal thickness were limited to selected regions, and no consistent patterns were observed for OCTA-derived retinal flow-density metrics. These findings characterize regional posterior pole morphology in high myopia within a unified spatial framework. Because this framework can be implemented using OCT/OCTA imaging, device-derived RC may serve as a candidate structural descriptor of regional posterior pole morphology. However, its prognostic value and clinical utility require validation in longitudinal studies with clinically meaningful outcomes.