Development and validation of a multi-omics model for early glaucoma diagnosis and prediction in high-risk populations

To develop and validate a multi-omics integration-based predictive model for early glaucoma diagnosis in high-risk populations, thereby providing a quantitative tool for early clinical screening and intervention. A total of 342 consecutive participants with high-risk factors for glaucoma were enrolled and randomly allocated into a training set and a validation set at a 7:3 ratio. Clinical, imaging, and multi-omics indicators—including intraocular pressure (IOP), vertical cup-to-disc ratio (vCDR), retinal nerve fiber layer thickness (RNFLT), ganglion cell complex thickness (GCCT), latent transforming growth factor beta binding protein 2 (LTBP2), and D-glucose-6-phosphate—were collected. Univariate analysis and multivariate logistic regression was applied to identify independent influencing factors (vCDR was entered into the model per 0.1-unit increment). Three machine learning models—Random Forest (RF), Support Vector Machine, and logistic regression—were then constructed. Model performance was evaluated using the area under curve (AUC) the receiver operating characteristic curve, calibration curves, and decision curves. Furthermore, SHAP analysis was applied to interpret feature contributions. Univariate analysis revealed significant differences between the early glaucoma and non-glaucoma groups for all six indicators ( P < 0.05). And all six factors as independent predictors of early glaucoma in high-risk individuals ( P < 0.05). RF demonstrated the best predictive performance, achieving an AUC of 0.863 (95% confidence interval [CI]: 0.803–0.923) in the training set and 0.785 (95% CI 0.670–0.900) in the validation set. SHAP analysis ranked feature contributions in descending order as: RNFLT, vCDR, IOP, D-glucose-6-phosphate, GCCT, and LTBP2. Calibration curves showed good agreement between predicted and observed probabilities. Additionally, decision curves confirmed a high net clinical benefit across a wide range of threshold probabilities. A multi-omics integration-based predictive model for early glaucoma diagnosis was successfully constructed and validated, which serves as an objective, quantitative auxiliary tool for identifying early glaucoma in high-risk populations.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2026-11-20
DOI
https://doi.org/10.1038/s41598-026-80752-2
Primary Topic
Glaucoma and retinal disorders
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Development and validation of a multi-omics model for early glaucoma diagnosis and prediction in high-risk populations

Mingyuan Yang, Chaofeng Yuan, XueTao Huang, Jia Meng
Scientific Reports
Glaucoma and retinal disorders
article

Development and validation of a multi-omics model for early glaucoma diagnosis and prediction in high-risk populations

Mingyuan Yang, Chaofeng Yuan, XueTao Huang, Jia Meng
article en

Abstract

To develop and validate a multi-omics integration-based predictive model for early glaucoma diagnosis in high-risk populations, thereby providing a quantitative tool for early clinical screening and intervention. A total of 342 consecutive participants with high-risk factors for glaucoma were enrolled and randomly allocated into a training set and a validation set at a 7:3 ratio. Clinical, imaging, and multi-omics indicators—including intraocular pressure (IOP), vertical cup-to-disc ratio (vCDR), retinal nerve fiber layer thickness (RNFLT), ganglion cell complex thickness (GCCT), latent transforming growth factor beta binding protein 2 (LTBP2), and D-glucose-6-phosphate—were collected. Univariate analysis and multivariate logistic regression was applied to identify independent influencing factors (vCDR was entered into the model per 0.1-unit increment). Three machine learning models—Random Forest (RF), Support Vector Machine, and logistic regression—were then constructed. Model performance was evaluated using the area under curve (AUC) the receiver operating characteristic curve, calibration curves, and decision curves. Furthermore, SHAP analysis was applied to interpret feature contributions. Univariate analysis revealed significant differences between the early glaucoma and non-glaucoma groups for all six indicators ( P < 0.05). And all six factors as independent predictors of early glaucoma in high-risk individuals ( P < 0.05). RF demonstrated the best predictive performance, achieving an AUC of 0.863 (95% confidence interval [CI]: 0.803–0.923) in the training set and 0.785 (95% CI 0.670–0.900) in the validation set. SHAP analysis ranked feature contributions in descending order as: RNFLT, vCDR, IOP, D-glucose-6-phosphate, GCCT, and LTBP2. Calibration curves showed good agreement between predicted and observed probabilities. Additionally, decision curves confirmed a high net clinical benefit across a wide range of threshold probabilities. A multi-omics integration-based predictive model for early glaucoma diagnosis was successfully constructed and validated, which serves as an objective, quantitative auxiliary tool for identifying early glaucoma in high-risk populations.

Scientific Reports
First Affiliated Hospital of Zhengzhou University (CN)
Good health and well-being
Glaucoma and retinal disorders
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon