A cluster and scotoma analysis based on empiric criteria

Area of Science:

• Ophthalmology
• Medical Imaging
• Computational Science

Background:

• Glaucoma diagnosis relies on detecting visual field defects.
• Subjective interpretation of visual field clusters can be inconsistent.
• Objective methods are needed to standardize glaucoma visual field analysis.

Purpose of the Study:

• To develop and evaluate a computer algorithm for analyzing glaucomatous visual field defects.
• To simulate expert interpreter's method of identifying visual field clusters or scotomas.
• To assess the algorithm's applicability across a broad spectrum of glaucoma damage.

Main Methods:

• Analysis of 30 randomly selected visual fields from 288 glaucomatous or suspect cases.
• Expert interpreter subjectively defined clusters based on adjacency criteria.
• Development of a computer algorithm to replicate the interpreter's clustering method.
• Comparison of algorithm output with expert interpretation.

Main Results:

• The computer algorithm, with standard parameter settings, produced a similar number of clusters as the expert interpreter.
• An alternative display mode grouped clusters using a Voronoi diagram.
• The algorithm demonstrated broad applicability due to the diverse range of visual field defects studied.

Conclusions:

• The developed computer algorithm effectively simulates expert evaluation of glaucomatous visual field damage.
• The algorithm's ability to identify clusters and scotomas shows promise for objective glaucoma assessment.
• Its broad applicability makes it a potentially valuable tool for managing diverse glaucoma cases.