OCT prognostic biomarkers for progression to late age-related macular degeneration: A systematic review and meta-analysis

To evaluate which OCT prognostic biomarkers best predict the risk of progression from early/intermediate to late age-related macular degeneration (AMD). Amongst over 100 OCT prognostic biomarkers for AMD, it is unclear which are the most relevant for clinicians and researchers to focus on. This review evaluated which OCT biomarkers confer the greatest magnitude of prediction for progression to late AMD. Study protocol was registered on PROSPERO (CRD42023400166). PubMed and Embase were searched from inception to 2nd March 2023, and eligible studies assessed following the Grade of Recommendations, Assessment, Development and Evaluation (GRADE) approach. The primary outcome was any quantified risk of progression from treatment-naïve early/intermediate AMD to late AMD, including hazard ratios (HRs), odds ratios (ORs), and standardised mean differences (at baseline, between eyes with versus without progression), sub-grouped by each OCT biomarker. Further meta-analyses were sub-grouped by progression to geographic atrophy or neovascularisation. A total of 114 quantified OCT prognostic biomarkers were identified. With high GRADE certainty of evidence, the greatest magnitudes of prediction to late AMD belonged to: external limiting membrane abnormality (OR, 15.42 [7.63, 31.17]), ellipsoid zone abnormality (OR, 10.8 [4.58, 25.46]), interdigitation zone abnormality (OR, 7.68 [2.57, 23]), concurrent large drusen and reticular pseudodrusen (HR, 6.73 [1.35, 33.65], hypo-reflective drusen cores (HR, 2.48 [1.8, 3.4]; OR 1.85 [1.29, 2.66]), intra-retinal hyper-reflective foci (IHRF; HR, 2.16 [0.92, 5.07]; OR 5.08 [3.26, 7.92]), and large drusen (HR, 2.01 [1.35, 2.99]); OR, 1.98 [1.27, 3.08]). There was greater risk of geographic atrophy for IHRF and hypo-reflective drusen cores (P < 0.05), and neovascularisation for ellipsoid zone abnormality (P < 0.05). Other OCT biomarkers such as drusenoid pigment epithelium detachment, shallow irregular retinal pigment epithelium elevations, and nascent geographic atrophy, exhibited large magnitudes of risk but required further studies for validation. This review synthesises the six most relevant OCT prognostic biomarkers for AMD with greater predictive ability than large drusen alone, for clinicians and researchers to focus on. Further study is required to validate other biomarkers with less than high certainty of evidence, and assess how the co-presence of biomarkers may affect risks.