Spatial Decomposition of Longitudinal RNFL Maps Reveals Distinct Modes of Glaucomatous Progression With...

Area of Science:

• Ophthalmology
• Neuroscience
• Genetics

Background:

• Glaucoma is a leading cause of irreversible blindness.
• Conventional analysis of retinal nerve fiber layer (RNFL) thickness changes averages progression, potentially masking distinct patterns.
• Understanding these patterns is crucial for accurate diagnosis and treatment.

Purpose of the Study:

• To identify distinct spatial modes of glaucomatous progression in RNFL thickness maps.
• To validate these modes using structure-function mapping and genetic association.
• To determine if spatial decomposition improves prediction over conventional averaging.

Main Methods:

• Pixel-wise RNFL rates of change were calculated from longitudinal OCT scans of 15,242 eyes.
• Non-negative matrix factorization decomposed RNFL maps into spatial progression components.
• Components were validated for structure-function concordance, visual field classification, and genetic association at POAG loci.

Main Results:

• Six distinct anatomical progression patterns were identified, including diffuse, focal, and arcuate defects.
• Pattern-based models significantly improved prediction of glaucomatous progression compared to global RNFL rates.
• Spatial pattern weights showed stronger genetic associations at POAG loci, suggesting reduced phenotypic heterogeneity.

Conclusions:

• Glaucomatous structural progression occurs via distinct spatial modes.
• These modes possess independent structure-function and genetic signatures.
• Spatial decomposition of RNFL maps provides a more nuanced understanding of glaucoma progression than conventional averaging.