What it is
Hypertensive disorders of pregnancy arise from early dysfunction of the placenta and of blood vessels throughout the body, yet current first-trimester screening for them remains imprecise, biologically opaque and logistically challenging. The authors present Visionary AI, an interpretable platform that combines images of the retina (the light-sensitive tissue at the back of the eye) with graph-based modeling of its blood vessels to predict these disorders before clinical onset. It converts retinal images into topological and geometric representations of the microvascular structure (how the smallest vessels connect and branch, and their shapes), so that risk is predicted from biologically grounded vascular features rather than from generic image embeddings or clinical variables. Across population-wide control settings in a prospective development cohort of 1,267 pregnancies, it predicted preeclampsia, one of these disorders, with an area under the curve (AUC, a measure of how well a test separates future cases from non-cases) of 0.91 and an average precision (the share of flagged pregnancies that turn out to be true cases, averaged across every level of sensitivity) of 0.81.
Why it matters
Current first-trimester screening for hypertensive disorders of pregnancy is imprecise, and this study presents the mother's retinal vessels as a minimally invasive way to predict them. The authors call Visionary AI a biologically interpretable framework for sorting pregnancies by risk: it works from features of the vessel network, and the features that recurred involved the vessels' topology, geometry, network complexity and hierarchical organization. A version optimized for stability transferred without retraining to an independent external validation cohort, where it reached an AUC of 0.81 and an average precision of 0.68 and outperformed clinical and deep learning benchmarks.
Underlined numbers link to their source. Every metric and quoted figure is listed under Sources and data below.
Filed underPregnancy and preeclampsia studies, Retinal Imaging and Analysis, Neurological Complications and Syndromes
