Objective:

To explore how deep learning models can enhance the prediction of glaucoma progression and improve surgical intervention timing.

Approach:

Key Findings:

• Deep learning models can predict future Humphrey visual fields from a single baseline visual field.
• Fusing multimodal data (disc photos and OCT) significantly enhances prediction accuracy compared to single modality use.
• The model provides insights into which modality contributed to the prediction, improving interpretability.

Interpretation:

AI has the potential to transform glaucoma management by enabling proactive rather than reactive surgical decisions through improved prediction of disease progression.

Limitations:

• Model performance decreases with longer time intervals between baseline and future predictions.
• Requires large, harmonized longitudinal multimodal datasets for effective training.

Conclusion:

AI-driven prediction models can lead to earlier interventions and better outcomes for glaucoma patients, emphasizing the need for comprehensive data collection.

Attribution Notice

This content is an AI-generated, fully rewritten summary based on a published scholarly article. It does not reproduce the original text and is not a substitute for the original publication. Readers are encouraged to consult the source for full context, data, and methodology.