This research advances medical imaging AI by introducing context-aware vision-language foundation models specifically designed for ocular disease screening.

• Leverages both visual features and clinical context through innovative prompting techniques
• Outperforms previous approaches in diabetic retinopathy detection
• Demonstrates significant improvements in handling domain shifts and data variability
• Requires minimal fine-tuning while maintaining high clinical accuracy

This breakthrough matters for healthcare by making AI-assisted retinal screening more reliable, adaptable, and clinically viable—potentially expanding access to early diagnosis in resource-limited settings.

Context-Aware Vision Language Foundation Models for Ocular Disease Screening in Retinal Images