ESCRS - PO0797 - Ocular Surface Disease: Correlational Analysis Of Signs And Symptoms Using An Artificial Intelligence Based Non Invasive Imaging Methodology

Ocular Surface Disease: Correlational Analysis Of Signs And Symptoms Using An Artificial Intelligence Based Non Invasive Imaging Methodology

Published 2023 - 41st Congress of the ESCRS

Reference: PO0797 | Type: Free paper | DOI: 10.82333/b926-0b75

Authors: Rushad Shroff* 1 , Mathew Kurian 2 , Bharatkumar Hegde 3 , Apoorva Agrawal 4

1Cataract,Shroff Eye Centre,New Delhii,India, 2Cataract,Chaitanya Eye Hospital,Cochin,India, 3Autoyos,Bangalore,India, 4Cataract,Shroff Eye Centre,New Delhi,India

Purpose

Dry eye disease (DED) is characterised by loss of tear film homeostasis and a lack of association between signs and symptoms.  This study used a novel non-invasive artificial intelligence imaging methodology to create a reproducible diagnostic test to bridge the knowledge gap between signs and symptoms in dry eye disease (DED).

Setting

Multicentre study at 2 tertiary eye care organisations in North and South India

Methods

This was a prospective single armed cross sectional study of new patients and those with prior dry eye symptoms, signs or on treatment referred for dry eye evaluation. Patients with history of ocular surgery, laser treatment, contact lens use and topical or systemic medication or diseases known to be associated with dry eye disease were also included. The sample size was 284 subjects.

The OSDI questionnaire was administered by a trained interviewer. The assessment of the variability in ocular surface temperature during the inter-blink interval using the Ocular Surface Imager was done by an independent investigator blind to the questionnaire outcome.  

Results

Based on the OSDI questionnaire the subjects enrolled for the study could be identified as normal or having dry eye disease. 

The ocular surface imaging showed more non-uniform and quicker variation of relative temperature profiles across the surface during the inter-blink interval  among the dry eye patients. The analysis and classification of the data was done through a set of classical image processing techniques and artificial intelligence based approaches

Sensitivity (90%) and Specificity (98%), Correlation coefficient and Area under the curve were better than current available tests.

Conclusions

Variability in ocular surface temperature could signal a paradigm shift in understanding DED.  Ocular surface temperature profile is observed to be in correlation with tear film stability. This has important therapeutic implications in the management of symptoms in DED including use of eye drops kept in the fridge door and Bandage contat lenses placed in the fridge in eyes post refractive surgery