Mapping Multiscale Brain Changes in Primary Angle-Closure Glaucoma Using Regional Radiomics Similarity Networks

Area of Science:

• Neuroimaging
• Ophthalmology
• Computational Neuroscience

Background:

• Primary angle-closure glaucoma (PACG) is a leading cause of irreversible blindness.
• The central nervous system's role in PACG pathogenesis is increasingly recognized.
• Understanding brain network alterations in PACG is crucial for elucidating disease mechanisms.

Purpose of the Study:

• To investigate large-scale structural brain network reorganization in PACG using regional radiomics similarity networks (R2SNs).
• To characterize the molecular and neurobiological substrates associated with these R2SN alterations.
• To evaluate the diagnostic potential of R2SN-derived features in PACG.

Main Methods:

• Structural MRI data from 44 PACG patients and 44 controls.
• Construction of individualized R2SNs to identify structural network alterations.
• Integration of R2SN data with transcriptomic profiles using partial least squares regression and machine learning.

Main Results:

• PACG patients exhibited widespread R2SN alterations beyond the visual pathway, affecting frontal, temporal, limbic, and subcortical regions.
• Disrupted structural covariance was observed across intrinsic functional systems, notably within limbic, default mode, and attentional networks.
• R2SN alterations spatially correlated with gene expression profiles related to synaptic, cytoskeletal, and immune processes, with machine learning achieving robust group discrimination.

Conclusions:

• PACG is associated with significant structural brain network reorganization extending beyond the visual system.
• R2SNs offer sensitive imaging markers for detecting central structural reorganization in PACG.
• This study provides a multiscale framework for understanding the neural mechanisms underlying PACG.