Approved Research

Ischemic and Hemorrhagic Stroke Prediction Using Retinal Biomarkers and Medical Records

NEC Corporation Of America

Lay summary

Stroke is a major global health issue, impacting 15 million people annually, with 10 million resulting in death or long-term disability. Prompt administration of treatments is crucial for effective outcomes, as their effectiveness diminishes over time. Consequently, patients must be quickly transported to a stroke center for a CT or MRI scan to confirm the diagnosis and identify the specific type of stroke, which requires different types of treatments. While CT and MRI are powerful tools for stroke type detection, they are high cost and invasive.

The blood vessels in the retina, which can be seen through basic retina photography, have a similar origin and structure as the blood vessels in the brain. Damage to the blood vessels in the retina can indicate damage to the brain. This means that retinal images offer a convenient way to assess cardiovascular health and can serve as a non-invasive method for evaluating the risk of stroke.

The aim of this study is to develop prediction models for ischemic and hemorrhagic stroke based on retinal characteristics and fundamental medical record features. Using these models, patients and medical care centers could get an advanced warning and prepare properly for potential stroke events. Proper preparation and awareness of stroke events can save people lives and prevent disability. In addition, our models could provide a simple, non-invasive, and cost-effective tool for stroke type classification, which can save time and money.