Purpose: This study investigated the efficacy of automated ocular thermography metrics for the screening of dry eye disease (DED).
Methods: This was a prospective study that enrolled 20 participants with DED, sex- and age-matched to 20 non-DED controls. Ocular Surface Disease Index (OSDI), Dry Eye Questionnaire-5 (DEQ5), noninvasive tear-break-up time (NITBUT), tear meniscus height (TMH), meibomian gland dysfunction (MGD) score, and corneal staining were measured in a screening visit. The DED group was defined as: OSDI score of ≥13 or DEQ-5 score of ≥6, and DED signs in at least one eye (corneal/conjunctival/lid margin staining, NITBUT <5 seconds, tear film osmolarity ≥308 miliosmoles [mOsm]/L). Thermography recording of the ocular surface (natural blinking over a period of 30 seconds) was conducted the next day, and the thermal cooling rate and thermal interblink interval (IBI) were derived.
Results: Thermal IBI was significantly shorter in the DED group compared to the non-DED group (P = 0.034). The thermal cooling rate was significantly faster in the DED group (P = 0.047). Thermal IBI significantly correlated with DEQ5 (r = -0.37, P = 0.025) and OSDI (r = -0.37, P = 0.026). The thermal cooling rate significantly correlated with DEQ5 (r = -0.39, P = 0.022) and OSDI (r = -0.36, P = 0.036). The best discrimination was achieved by combining the thermal cooling rate and TMH, with an area under the curve (AUC) = 0.80 (sensitivity = 0.87 and specificity = 0.63).
Conclusions: The thermal IBI and thermal cooling rate were significant predictors of DED, suggesting the utility of ocular thermography for DED screening.
Translational Relevance: Automated ocular thermography may help to assess ocular dryness in a noninvasive, quantifiable, and real-time manner.