An automated optical coherence tomography to finite element analysis pipeline reveals key morphological determinants of optic nerve head biomechanics in glaucoma

自动化光学相干断层扫描到有限元分析流程揭示了青光眼视神经乳头生物力学的关键形态学决定因素

阅读:1

Abstract

BACKGROUND: Glaucoma, the leading cause of irreversible blindness, is closely linked to optic nerve head (ONH) damage, particularly within the lamina cribrosa (LC). Prior biomechanical studies have either relied on manual OCT‑based modelling, which is accurate but labour-intensive, or on idealised "standard" eye models, which allow high‑volume analysis but lack anatomical realism. To bridge this gap, we developed an automated pipeline that converts routine OCT into patient‑specific finite element (FE) models and quantifies biomechanical and morphological determinants. METHODS: We analysed OCT volumes from 154 healthy and 170 glaucomatous eyes using automated segmentation of the retina, choroid, sclera, and lamina cribrosa, followed by patient-specific ONH reconstruction and finite-element simulation of LC strain under an intraocular pressure of 15 mmHg. We quantified eight morphological parameters. A Random Forest regression combined with SHAP analysis was used to determine morphological predictors of LC strain. RESULTS: LC depth showed the strongest linear association with LC strain (r = -0.31). Other parameters with smaller associations included BMO radius (r = 0.17), pre-lamina volume (r =  -0.14) and mean choroidal thickness (r = -0.12). After adjusting for age and gender, glaucomatous eyes exhibited significantly lower LC strain than healthy eyes (coefficient = 0.0018, p = 0.011), suggesting potential tissue remodelling. SHAP ranked pre-lamina depth, LC curvature, BMO radius, and LC depth as the most influential predictors despite modest explained variance. CONCLUSIONS: This study presents a scalable, image-based biomechanical framework enables high-throughput, patient-specific assessment and offers new opportunities to identify morphological biomarkers for glaucoma risk stratification and disease monitoring.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。