Computer-aided diagnostic system for hypertensive retinopathy: A review

Area of Science:

• Ophthalmology and Medical Imaging
• Biomedical Engineering
• Cardiovascular Health

Background:

• Hypertensive Retinopathy (HR) is a serious eye condition linked to prolonged high blood pressure.
• Early detection is crucial as HR indicates broader health risks like stroke and kidney disease.
• Retinal examination can reveal microcirculation changes before major clinical symptoms appear.

Purpose of the Study:

• To comprehensively review automated methods for Hypertensive Retinopathy (HR) detection and grading.
• To analyze computer-aided diagnosis (CAD) techniques for Artery-Vein (A/V) classification and Arteriovenous Ratio (AVR) computation.
• To identify challenges and propose future research directions in automated HR diagnosis.

Main Methods:

• Systematic literature review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol.
• Analysis of automated tasks: A/V classification, AVR computation, HR detection (binary classification), and HR severity grading.
• Evaluation of datasets, methodologies, performance metrics, and computational platforms used in reviewed studies.

Main Results:

• Summaries and critical analysis of existing methods for automated HR diagnosis tasks.
• Comparison of classifier details, methodologies, and performance across reviewed studies.
• Identification of common research issues, including data availability and grading accuracy.

Conclusions:

• Automated CAD systems offer efficient and accurate early detection of Hypertensive Retinopathy (HR).
• Significant challenges remain in data availability and precise HR severity grading.
• Future research should focus on addressing these challenges to improve clinical applicability.