Abstract
OBJECTIVE: To investigate changes in peripapillary vessel density (pVD) and retinal nerve fiber layer thickness (pRNFL) in highly myopic glaucoma patients compared to non-highly myopic glaucoma patients and healthy controls, and to evaluate their diagnostic capabilities using optical coherence tomography angiography (OCTA), and to explore the relationship between these biomarkers and visual function. METHODS: A total of 382 eyes were recruited, including 101 highly myopic glaucoma eyes, 101 highly myopic control eyes, 90 non - highly myopic glaucoma eyes, and 90 non - highly myopic control eyes. Propensity score matching (PSM) was applied to balance age and axial length between groups. All subjects received comprehensive ophthalmic examinations. OCTA was used to measure pVD, spectral - domain OCT (SD - OCT) for pRNFL, and Humphrey 30-2 SITA standard visual field (VF) testing was performed. pVD and pRNFL were measured across eight peripapillary sectors. Pearson correlation and linear regression analyses were used to assess the relationships between pVD, pRNFL, and visual field mean sensitivity (VFMS). Receiver operating characteristic (ROC) curve analyses were carried out to evaluate the diagnostic performance. RESULTS: Both highly myopic glaucoma and non-highly myopic glaucoma groups exhibited significantly lower pVD, pRNFL, and VFMS compared to their respective controls (p < 0.001). In highly myopic glaucoma, average pVD was 37.66% versus 46.40% in controls (p < 0.001), and pRNFL was 71.13 μm versus 101.22 μm in controls (p < 0.001). pVD showed stronger correlations with VFMS than pRNFL in both glaucoma groups (highly myopic: r = 0.681 vs. r = 0.504; non-highly myopic: r = 0.749 vs. r = 0.722; p < 0.001). ROC analysis demonstrated that the pRNFL and pVD have comparable diagnostic abilities in the early-stage of glaucoma (p > 0.05). However, the pRNFL outperforms the pVD in average diagnostic ability (p < 0.05). Combining superior-temporal (ST) and inferior-temporal (IT) regions achieved the highest diagnostic accuracy (AUC(pVD): 0.905 and 0.965; AUC(pRNFL): 0.934 and 0.942) for both glaucoma groups. CONCLUSION: pVD and pRNFL are valuable diagnostic biomarkers for myopic glaucoma. pVD demonstrated a stronger correlation with visual function, making it a promising tool for early glaucoma diagnosis and monitoring in highly myopic patients. Integrating pVD with pRNFL enhances diagnostic precision, particularly in highly myopic patients.