Abstract Body: Background:
Hypertensive retinopathy (HR) serves as a non-invasive marker of microvascular damage caused by chronically elevated blood pressure, and is associated with an increased risk of stroke, cardiac events, and renal dysfunction. Despite its prognostic importance, HR is frequently underdiagnosed—especially in underserved populations with limited access to ophthalmologic care. The emerging field of oculomics leverages retinal imaging to assess systemic health, opening new avenues for artificial intelligence (AI)-driven screening and disease monitoring.

Objective:
We developed AHyReSolution, a multi-platform, privacy-preserving AI application for rapid detection, severity grading, and longitudinal monitoring of HR. Designed to be computationally efficient and universally accessible, AHyReSolution aims to advance equity in retinal and cardiovascular screening, particularly in resource-limited settings.

Methods:
AHyReSolution utilizes state-of-the-art ResNet architectures for accurate identification and classification of HR, optimized to function on even low-power mobile devices. Our models were trained and validated on a diverse set of expert-annotated fundus image datasets—including ODIR, HRSG, and JSIEC—to ensure global applicability and generalizability. Performance was benchmarked against other retinal pathologies to assess diagnostic specificity. The mobile interface delivers secure, real-time analysis and underwent independent usability testing with clinicians across multiple countries.

Results:
ResNet152 achieved >97% diagnostic accuracy and AUROC >0.99 on all external validation datasets, with image processing times of under one second. The platform showed consistent performance across varied image qualities, ethnic backgrounds, and device types. It supports four-level HR severity grading, highlights key retinal lesions, and provides explainable outputs to enhance clinical interpretability.

Conclusion:
AHyReSolution is a robust, scalable, and accessible AI tool for early detection and monitoring of hypertensive retinopathy. By facilitating timely intervention, it has the potential to reduce global disparities in the diagnosis and management of both retinal and cardiovascular diseases.