A Comprehensive Review Comparing Artificial Intelligence and Clinical Diagnostic Approaches for Dry Eye Disease

Area of Science:

• Ophthalmology
• Artificial Intelligence
• Medical Diagnostics

Background:

• Dry eye disease (DED) affects a significant portion of the population, necessitating accurate and efficient diagnostic methods.
• Current clinical assessments for DED can be subjective and time-consuming.
• The integration of artificial intelligence (AI) offers potential advancements in objective and automated DED diagnosis.

Purpose of the Study:

• To systematically review and synthesize studies comparing AI-based diagnostic models with traditional clinical tests for DED.
• To evaluate the performance of AI models across various imaging modalities used in ophthalmology.
• To identify limitations and propose recommendations for future AI research in DED diagnosis.

Main Methods:

• Systematic literature search adhering to PRISMA guidelines across four major databases (Google Scholar, PubMed, ScienceDirect, Cochrane Library).
• Inclusion of 30 peer-reviewed articles published between 2020 and 2025, focusing on AI in DED diagnosis.
• Analysis of studies based on AI models, imaging modalities, diagnostic performance, and reported limitations.

Main Results:

• Deep learning models across diverse imaging modalities (e.g., videokeratography, OCT, meibography) demonstrated high diagnostic accuracy for DED, ranging from 82% to 99%.
• AI models frequently matched or surpassed the performance of established clinical assessments.
• Limited external validation and lack of focus on inter-expert variability were noted as common limitations.

Conclusions:

• AI holds significant potential to enhance the accuracy and efficiency of dry eye disease diagnosis in ophthalmology.
• Addressing gaps in data diversity, external validation, and reproducibility is crucial for the clinical implementation of AI tools.
• Future research should focus on robust validation and standardization to facilitate AI integration into routine eye care practices.