Genomic and Epigenomic Factors Associated with Therapeutic Response in Diabetic Retinopathy: From Pharmacogenomic Evidence to Precision Ophthalmology

Diabetic retinopathy (DR) is a clinically heterogeneous complication of diabetes, and substantial variability in therapeutic response remains a challenge in the management of diabetic macular oedema (DME) and other vision-threatening manifestations of DR. Although clinical and systemic factors improve treatment outcomes, genomic and epigenomic variation may provide additional evidence for individualised therapeutic strategies. This review critically examines current evidence linking genomic and epigenomic factors to therapeutic response in DR, with particular emphasis on pharmacogenomic findings related to anti-vascular endothelial growth factor (anti-VEGF) therapy. Evidence strength varies substantially across domains. The strongest direct pharmacogenomic evidence concerns exploratory genome-wide associations with anti-VEGF treatment outcomes in DME, whereas evidence for corticosteroid and laser response remains insufficient. Importantly, most reported associations remain preliminary, with limited replication, heterogeneous phenotypes and treatment regimens, and insufficient functional or prospective validation. Epigenomic mechanisms, including DNA methylation and non-coding RNAs, may contribute to treatment-response heterogeneity. However, direct relevance to therapeutic outcomes remains less well established than their role in DR pathogenesis. The review further considers methodological, translational, ethical, regulatory, and economic barriers to clinical implementation. It discusses integrating genomic and epigenomic information with retinal imaging, clinical variables, and artificial intelligence-based approaches. Overall, genomic and epigenomic markers represent a promising but currently exploratory component of precision ophthalmology. Progress toward clinical application will require reproducible associations, standardised response phenotypes, functional validation, multimodal predictive models, and prospective demonstration of clinical utility.