A novel computer aided quantification method of focal arteriolar narrowing using colour retinal image

Area of Science:

• Ophthalmology
• Cardiology
• Medical Imaging

Background:

• Focal arteriolar narrowing (FAN) is a key indicator of cardiovascular disease risk.
• Accurate quantification of FAN is crucial for early detection and management.
• Existing methods for FAN assessment may lack robustness and reproducibility.

Purpose of the Study:

• To introduce a novel, robust method for quantifying focal arteriolar narrowing (FAN) in the human retina.
• To establish a reliable measurement for assessing FAN severity and its correlation with cardiovascular health.
• To improve the reproducibility of FAN measurements compared to expert grading.

Main Methods:

• A graph-based optimization method incorporating intensity, gradient, and spatial priors for arteriolar boundary mapping.
• Analysis of arteriolar widths to quantify the severity of focal arteriolar narrowing (FAN).
• Validation on a dataset of 116 retinal arteriolar segments graded by two expert ophthalmologists.

Main Results:

• The proposed method's FAN measurement score (percent narrowing, PN) demonstrated strong correlation with expert grading (Spearman's ρ = 0.82–0.84, p<0.0001).
• The method exhibited superior reproducibility (Spearman's ρ = 0.92, p<0.0001) compared to inter-grader variability.
• Quantitative measurements provide a reliable link between FAN and systemic/ocular diseases.

Conclusions:

• The novel method offers accurate and reproducible quantification of retinal focal arteriolar narrowing.
• This tool can aid in the early detection of hypertension, stroke, and other cardiovascular conditions.
• Improved FAN measurement facilitates a better understanding of its association with systemic and eye diseases.