Integrated oculomics and lipidomics characterise cross-sectional retinal–lipidomic associations in a generally healthy biobank population

Cardiovascular disease remains the leading global cause of mortality, yet current risk stratification methods may fail to capture early, subclinical vascular changes. Previous studies have rarely integrated retinal microvascular characteristics with comprehensive serum lipidomic profiles as potential indicators of cardiovascular health. In this study, we applied an imaging–omics framework that combines deep learning-derived retinal microvascular traits with serum lipidomic data to explore retinal-lipidomic association patterns using higher-dimensional serum lipidomic profiling alongside quantitative retinal vascular phenotypes. To our knowledge, this represents one of the first large-scale, covariate-adjusted correlation analyses of retinal microvascular phenotypes and serum lipidomics in a generally healthy biobank population. Retinal phenotypes were quantified using automated image analysis tools, while serum lipid profiling was performed using ultra-high-performance liquid chromatography coupled with electrospray ionisation high-resolution mass spectrometry. Modest but statistically significant age- and sex-adjusted partial correlations were observed, particularly between average artery width, vessel density, and lipid subclasses such as triacylglycerols, diacylglycerols, and ceramides. These relationships are biologically plausible in the context of microvascular–metabolic remodelling but should be interpreted as exploratory association signals rather than evidence of causality or clinical prediction. By linking quantitative retinal vascular phenotypes to annotated serum lipidomic features, this study characterises retinal–lipidomic association patterns relevant to microvascular–metabolic variation. This integration provides an imaging–omics workflow for identifying candidate association signals that connect automated retinal vascular phenotypes with circulating molecular profiles. Future longitudinal and outcome-linked studies will be important to assess the temporal stability and clinical relevance of these association patterns.