Related Experiment Video
Updated: Jan 10, 2026

Full-Circle Cauterization of Limbal Vascular Plexus for Surgically Induced Glaucoma in Rodents
Published on: February 15, 2022
Multimodal AI-based modeling of glaucoma progression: a 3PM-guided approach integrating structural, functional, and
Natalia I Kurysheva1, Oxana Ye Rodionova2, Alexey L Pomerantsev2
1The Ophthalmological Center of the Federal Medical and Biological Agency of the Russian Federation, 15 Gamalei Street, Moscow, 123098 Russian Federation.
This study developed a personalized model to predict primary open-angle glaucoma (POAG) progression using multimodal biomarkers. This approach aids in stratifying patients for tailored treatments and preventing irreversible blindness.
Area of Science:
- Ophthalmology
- Biomedical Engineering
- Data Science
Background:
- Glaucoma is a leading cause of irreversible blindness globally.
- High inter-individual variability in glaucoma progression necessitates predictive, preventive, and personalized medicine (3PM) strategies.
- Effective patient stratification is crucial for protecting individuals from disease advancement.
Purpose of the Study:
- To develop and validate a personalized, multimodal predictive modeling framework for primary open-angle glaucoma (POAG).
- To integrate structural, functional, and vascular biomarkers for individualized risk stratification of POAG progression rates.
- To enhance patient management through precise prediction of disease trajectory.
Main Methods:
- Longitudinal monitoring (≥36 months) of POAG patients at various stages.
- Comprehensive multimodal data collection: optical coherence tomography (OCT), OCT angiography (OCT-A), automated perimetry, biomechanical assessments.
- Predictive modeling using Ranked Partial Least Squares Discriminant Analysis (Ranked PLS-DA) with Procrustes Cross-Validation.
Main Results:
- Developed models with up to 27 parameters for early and 20 for advanced POAG, achieving high prognostic accuracy (AUC up to 0.90).
- Identified key predictive biomarkers varying by disease stage: RNFL thickness, microvascular dropout, vascular density, and corneal hysteresis in early POAG; age, ganglion cell complex thickness, macular thickness, and perfusion parameters in advanced POAG.
- Demonstrated the framework's ability to classify slow, moderate, and rapid glaucoma progression rates.
Conclusions:
- The multimodal predictive modeling framework enables accurate risk stratification and personalized glaucoma management.
- Clinical application involves initial profiling, regular recalibration, and adaptive treatment strategies for improved visual outcomes.
- This 3PM approach optimizes resource utilization and enhances patient protection against disease progression.
Related Concept Videos
Glaucoma: Overview
Open Angle Glaucoma: Treatment
Drugs such as carbonic anhydrase inhibitors, α2- and...
Angle Closure Glaucoma: Treatment

