Tai-Yuan Su1, Shu-Wen Chang2. 1. Department of Electrical Engineering, Yuan-Ze University, Chung-Li, Taiwan. Electronic address: tysu@saturn.yzu.edu.tw. 2. Department of Ophthalmology, Far Eastern Memorial Hospital, Banchiao District, New Taipei City, Taiwan; Department of Ophthalmology, National Taiwan University Hospital, Taipei, Taiwan. Electronic address: swchang2007@ntu.edu.tw.
Abstract
PURPOSE: To construct normalized ocular surface temperature (NOST) models for different tear film characteristics and evaluate its potential in dry-eyes screening. METHODS: We included 227 right eyes of 227 patients. Tear film characteristics were categorized into 4 types according to fluorescein tear film breakup time (FTBUT) and Schirmer test results, namely type 1: FTBUT >5 s and Schirmer >5 mm; type 2: FTBUT ≤5 s and Schirmer >5 mm; type 3: FTBUT ≤5 s and Schirmer ≤5 mm; and type4: FTBUT >5 s and Schirmer ≤5 mm. Ocular surface temperature was measured by a video-thermographer. Mean temperatures of the central cornea were calculated from the videos of each frame during the 4-s blink interval. We first constructed individual NOST model for every tear characteristic. Participants were included for further analysis when their OSDI was ≥23, FTBUT ≤5 s, and Schirmer >5 mm. They were subdivided into short-BUT and short BUT with SPK subgroups according to the absence or presence of corneal fluorescein-stain. The NOST models of the normal, short-BUT and short BUT with SPK groups were separately constructed and the potential of screening analyzed via ROC curves. RESULTS: Each tear film type had a different NOST model. At 3 s after blinking, the order of NOST was type 4 >type 1 >type 3 >type 2. In dry-eye screening, the NOST was normal > short-BUT > short BUT with SPK. The NOST displayed a sensitivity 0.87, specificity 0.80, and AUC 0.88 for diagnosing short BUT with SPK. CONCLUSION: NOST models are useful in differentiating tear film characteristics and screening dry-eyes. It alleviates the discomfort and inconvenience encountered during conventional dry-eye diagnosis.
PURPOSE: To construct normalized ocular surface temperature (NOST) models for different tear film characteristics and evaluate its potential in dry-eyes screening. METHODS: We included 227 right eyes of 227 patients. Tear film characteristics were categorized into 4 types according to fluorescein tear film breakup time (FTBUT) and Schirmer test results, namely type 1: FTBUT >5 s and Schirmer >5 mm; type 2: FTBUT ≤5 s and Schirmer >5 mm; type 3: FTBUT ≤5 s and Schirmer ≤5 mm; and type4: FTBUT >5 s and Schirmer ≤5 mm. Ocular surface temperature was measured by a video-thermographer. Mean temperatures of the central cornea were calculated from the videos of each frame during the 4-s blink interval. We first constructed individual NOST model for every tear characteristic. Participants were included for further analysis when their OSDI was ≥23, FTBUT ≤5 s, and Schirmer >5 mm. They were subdivided into short-BUT and short BUT with SPK subgroups according to the absence or presence of corneal fluorescein-stain. The NOST models of the normal, short-BUT and short BUT with SPK groups were separately constructed and the potential of screening analyzed via ROC curves. RESULTS: Each tear film type had a different NOST model. At 3 s after blinking, the order of NOST was type 4 >type 1 >type 3 >type 2. In dry-eye screening, the NOST was normal > short-BUT > short BUT with SPK. The NOST displayed a sensitivity 0.87, specificity 0.80, and AUC 0.88 for diagnosing short BUT with SPK. CONCLUSION: NOST models are useful in differentiating tear film characteristics and screening dry-eyes. It alleviates the discomfort and inconvenience encountered during conventional dry-eye diagnosis.