Circulating and extracellular vesicle-associated non-coding RNAs in diabetic retinopathy: a stage-aware framework for network-informed biomarker panels

Diabetic retinopathy (DR) is a progressive microvascular complication of diabetes in which prognosis and management change substantially across non-proliferative and proliferative stages. Current imaging accurately detects established structural and vascular abnormalities but provides limited insight into early molecular changes or the risk of clinically important transitions within non-proliferative diabetic retinopathy (NPDR). Circulating non-coding RNAs (ncRNAs), including microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs), have emerged as candidate biomarkers because they regulate pathways involved in endothelial dysfunction, inflammation, oxidative stress, and angiogenesis and can be detected in blood and extracellular vesicle (EV)-associated fractions. In this structured narrative review, we synthesize circulating and EV-associated ncRNA evidence within a stage-aware framework aligned with three clinical use cases: early screening, risk stratification within NPDR, and monitoring at the severe NPDR-early proliferative diabetic retinopathy boundary. miRNAs currently provide the most mature evidence for stage-associated expression patterns and, in selected studies, stage-discriminatory performance. By contrast, circulating lncRNA and circRNA evidence remains limited and is predominantly cross-sectional and advanced-stage associated, supporting their use mainly as complementary, network-informing components. EV association may improve RNA stability and enrich disease-relevant signals, but it does not establish retinal origin, and differences in biofluid source, EV isolation and characterization, normalization, and systemic confounders limit cross-study comparability. Importantly, direct evidence for clinically relevant adjacent-stage discrimination and longitudinal prediction remains scarce. We therefore propose that future biomarker development should prioritize stage-aware, network-informed multi-ncRNA panels, standardized pre-analytical and EV workflows, adjustment for major clinical confounders, and longitudinal imaging-integrated validation to determine whether these biomarkers add value beyond conventional retinal staging.