Visionary AI: Decoding Systemic Vascular Health and Hypertensive Disorders in Pregnancy Through Retinal Imaging and...

Area of Science:

• Reproductive Medicine
• Medical Imaging
• Artificial Intelligence

Background:

• Hypertensive disorders of pregnancy (HDP) pose significant risks, including maternal mortality and long-term cardiovascular issues.
• Current screening methods for HDP are imprecise and challenging, necessitating earlier and more accurate detection.
• Early placental dysfunction and systemic endothelial disruption are linked to HDP, but underlying mechanisms require further elucidation.

Purpose of the Study:

• To develop and validate an AI-driven platform, Visionary AI, for early prediction of HDP using retinal imaging.
• To investigate the utility of ultra-widefield retinal imaging and vascular modeling for identifying HDP-specific microvascular signatures.
• To provide mechanistic insights into the pathophysiology of different HDP subtypes through retinal vascular analysis.

Main Methods:

• Visionary AI integrates 200° ultra-widefield retinal imaging with biologically grounded vascular modeling.
• The platform constructs interpretable, graph-based representations of the maternal retinal vasculature.
• Topological and geometric analyses are applied to identify condition-specific microvascular signatures in a prospective multiethnic cohort (n=1,267).

Main Results:

• Visionary AI demonstrated high predictive performance for preeclampsia (AUC=0.90), early-onset preeclampsia (AUC=0.93), and severe preeclampsia (AUC=0.89).
• The AI model also accurately predicted gestational hypertension (AUC=0.91) and chronic hypertension (AUC=0.90).
• Retinal vascular analyses revealed distinct remodeling patterns associated with different HDP subtypes, offering mechanistic insights.

Conclusions:

• The maternal retina serves as a minimally invasive biosensor for early systemic vascular health assessment.
• Visionary AI provides a clinically actionable, biologically grounded diagnostic framework for early HDP detection.
• This AI platform shows potential for broad global scalability in improving HDP screening and management.