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

Die Hu; Rui-Yang Hu; Yan-Yan Zhang; Yu-Yue Cheng; Fang-Liang Yu; Xin Huang

doi:10.2147/OPTH.S527088

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

Clin Ophthalmol. 2026 Jul 3:20:527088. doi: 10.2147/OPTH.S527088. eCollection 2026.

Authors

Die Hu^#^{1

2

3

4

5}, Rui-Yang Hu^#⁵, Yan-Yan Zhang^#⁶, Yu-Yue Cheng^{1

2

3

4

5}, Fang-Liang Yu^{1

2

3

4

5}, Xin Huang^{1

2

3

4

5}

Affiliations

¹ The Affiliated Eye Hospital, Jiangxi Medical College, Nanchang University, Nanchang, Jiangxi, People's Republic of China.
² Jiangxi Province Key Laboratory of Ophthalmology and Vision Sciences, Nanchang, Jiangxi, People's Republic of China.
³ Jiangxi Clinical Research Center for Ophthalmic Disease, Nanchang, Jiangxi, People's Republic of China.
⁴ Jiangxi Provincial Key Laboratory of Vitreoretinal Diseases for Health, Nanchang, Jiangxi, People's Republic of China.
⁵ School of Ophthalmology and Optometry, Jiangxi Medical College, Nanchang University, Nanchang, Jiangxi, People's Republic of China.
⁶ Department of Ophthalmology, Jiangxi Provincial People's Hospital, the First Affiliated Hospital of Nanchang Medical College, Nanchang, Jiangxi, People's Republic of China.

^# Contributed equally.

Abstract

Purpose: To investigate large-scale structural brain network reorganization in primary angle-closure glaucoma (PACG) using regional radiomics similarity networks (R2SNs) and to characterize their associated molecular and neurobiological substrates.

Design: Case-control study.

Methods: Structural magnetic resonance imaging data were acquired from 44 patients with PACG and 44 age- and sex-matched healthy controls. Individualized R2SNs were constructed to identify PACG-related structural network alterations. Partial least squares regression was used to link R2SN alterations with brain-wide transcriptomic profiles, followed by enrichment, cell-type, neurochemical, and epicenter analyses. Supervised machine learning was used to evaluate the discriminative value of R2SN-derived features.

Results: Compared with controls, patients with PACG demonstrated widespread R2SN alterations extending beyond the visual pathway, involving frontal, temporal, limbic, and subcortical regions. Network-level analyses revealed disrupted structural covariance across intrinsic functional systems, particularly within limbic, default mode, attentional, and frontoparietal control networks. Imaging-transcriptomic analyses showed spatial coupling between PACG-related network alterations and gene expression profiles enriched in synaptic, cytoskeletal, and immune-related processes. Epicenter analysis identified highly connected hub regions within default mode and attentional systems. Machine learning classifiers based on R2SN features achieved robust discrimination between groups.

Conclusion: PACG is associated with widespread structural brain network reorganization beyond the visual system.R2SN-derived network features may provide sensitive imaging markers of central structural reorganization and offer a multiscale framework for understanding the neural mechanisms of PACG.

Keywords: Imaging transcriptomics; machine learning; neurochemical mapping; primary angle-closure glaucoma; regional radiomics similarity network; structural brain network.