Ai-Driven Analysis Of Ocular Surface & Tear Biomarkers.

To develop and validate an AI-integrated biomarker scoring system as a surrogate for ocular surface health and inflammation, comparing its performance against standard dry eye diagnostics.

Prospective cross-sectional study conducted at Narayana Nethralaya, Bangalore, India, in collaboration with the Imaging, Biomechanics and Mathematical Modelling Solutions (IBMS) Lab, GROW Research Lab, between January 2024 and January 2026. Patients were recruited from the dry eye outpatient clinic.

2000 patients (1115 subclinical, 885 dry eye) underwent Schirmer’s, TBUT, Meibography, epithelial mapping and confocal microscopy for dendritic cell metrics. Tear biomarkers (via BioM Pathfinder) were integrated using AI to generate a composite score (0–1): healthy (0–0.3), subclinical (0.3–0.7), and dry eye (0.7–1), correlated with HOAs and ocular surface changes.

Patients with AI scores above 0.7 showed clear signs of dry eye—higher levels of inflammatory biomarkers, more dendritic cells, uneven epithelium and increased HOAs, all statistically significant (p<0.05). The AI system was able to accurately group patients into healthy, subclinical, and dry eye categories.

Patients with AI scores above 0.7 showed clear signs of dry eye—higher levels of inflammatory biomarkers, more dendritic cells, uneven epithelium and increased HOAs, all statistically significant (p<0.05). The AI system was able to accurately group patients into healthy, subclinical, and dry eye categories.