Statistical atlas-based descriptor for an early detection of optic disc abnormalities

Area of Science:

• Ophthalmology and Medical Imaging
• Computational Anatomy
• Biomedical Engineering

Background:

• Optic disc (OD) appearance in fundus images is crucial for assessing retinal diseases like glaucoma.
• The cup-to-disc ratio (CDR) is a common glaucoma metric but only uses diameter-based relative sizes.
• Existing methods lack detailed geometric characterization of the OD and its internal vasculature.

Purpose of the Study:

• To develop and validate an atlas-based shape descriptor (ASD) for statistically characterizing optic disc (OD) geometric deformations.
• To assess the ASD's ability to differentiate between healthy and glaucomatous optic disc regions.
• To explore the utility of a deviation map for early detection of OD abnormalities.

Main Methods:

• Constructing a statistical atlas of the optic disc region using nonlinear registration and deformation field analysis.
• Defining the atlas-based shape descriptor (ASD) using statistical measures derived from the atlas.
• Analyzing the average model and principal modes of deformation in a healthy population.
• Comparing ASD components between healthy and pathological (glaucomatous) optic discs.

Main Results:

• The ASD statistically characterizes geometric deformations of the OD shape and internal blood vessel configuration.
• Significant differences were observed in ASD components between healthy and pathological optic discs.
• The ASD demonstrated effectiveness in distinguishing healthy from glaucomatous OD regions.
• A deviation map derived from the atlas can highlight abnormalities.

Conclusions:

• The atlas-based shape descriptor (ASD) offers a novel approach to quantify optic disc geometry beyond the CDR.
• ASD components show significant potential for differentiating healthy and glaucomatous optic discs.
• This method can assist clinicians in the early detection of glaucomatous changes in the optic disc.