# Artificial Intelligence for Diagnostic Guidance in Ocular Surface Disorders

## TL;DR

This paper reviews how artificial intelligence is being used to help diagnose eye surface diseases, showing promise but needing more validation.

## Contribution

The paper provides a comprehensive review of AI applications in ocular surface diagnostics, highlighting current limitations and future directions.

## Key findings

- AI models for ocular surface disorders match or exceed ophthalmologists' performance in diagnostic accuracy.
- Current AI studies often lack external validation and use small, device-specific datasets.
- Future research should focus on large datasets, standardization, and real-world AI-human collaboration.

## Abstract

Artificial intelligence (AI) has been explored as a promising diagnostic aid for ocular surface diseases (OSDs). The spectrum of OSD ranges from highly prevalent benign conditions such as dry eye disease (DED) to rare but potentially dangerous disorders, including ocular surface squamous neoplasia (OSSN) and conjunctival melanoma. This review provides an overview of current applications of AI across the major categories of ocular surface pathology and specifically highlights anterior segment imaging modalities, including slit-lamp examination, optical coherence tomography (OCT), and in vivo confocal microscopy (IVCM). Meibography, tear film dynamics, biochemical profiling, and other DED-related measures are also examined. Across these domains, reported AI model performance matches or exceeds that of ophthalmologists, offering consistent, reproducible, and accurate approaches for guiding diagnosis. However, studies with limited external or prospective validation, variable labeling strategies, and small, device-specific datasets predominate in the current literature, thereby limiting generalizability. Large multicenter datasets, standardized diagnostic frameworks, multimodal integration, and prospective trials that assess human–AI cooperation in practical settings should be an emphasis in future research. By filling these gaps, AI systems could advance from experimental tools to clinically reliable applications that improve access and diagnostic accuracy in the care of ocular surface disease and tumors.