Quantification of parapapillary atrophy and optic disc

Area of Science:

• Ophthalmology
• Medical Imaging
• Computer Vision

Background:

• Parapapillary atrophy (PPA) and optic disc (OD) are key indicators in retinal health.
• Accurate quantification of these regions is crucial for diagnosing and monitoring eye conditions.
• Existing methods for PPA and OD measurement can be time-consuming and subjective.

Purpose of the Study:

• To develop and validate a computer-aided tool for automatic detection and quantification of PPA and OD.
• To assess the accuracy and reliability of the proposed algorithm in measuring these retinal structures.
• To provide a novel software solution for objective PPA and OD size estimation from fundus images.

Main Methods:

• The optic disc (OD) region was segmented using a Chan-Vese model with shape constraints.
• Parapapillary atrophy (PPA) was estimated using a modified Chan-Vese approach after OD removal.
• Boundary refinement involved thresholding, scanning filters, and region-growing; dual color channels minimized artifact interference.

Main Results:

• The algorithm was tested on 94 images from 66 subjects with PPA.
• Mean accuracy for PPA and OD size estimation was 93.8% and 94.0%, respectively.
• High correlation (0.98) was observed between automated and ophthalmologist-defined measurements for both PPA and OD.

Conclusions:

• This software provides the first automated method for quantifying PPA size in retinal fundus images.
• The algorithm reliably detects and quantifies PPA and OD with high accuracy (>93%).
• Additional measurements like OD diameter may enhance its utility in eye screening programs.