ABSTRACT
PURPOSE: The aim of this study was to determine the feasibility of using readily accessible technology, anterior segment optical coherence tomography (AS-OCT), to detect and grade meibomian gland dropout and examine its interchangeability with the Oculus Keratograph 5M (K5M). METHODS: A total of 30 participants (30 eyes) with a median age of 21 (range = 19-28 years) were recruited. Meibography was performed using two commercially available imaging devices to look at the structure of the meibomian glands and grade them subjectively in real time, and image analysis was used to quantify meibomian gland loss objectively. Gland loss as imaged by the two techniques was graded using the meiboscore grading schema. Test-retest reliability was determined with intraclass correlation coefficients (ICCs). Weighted kappa was used to evaluate agreement between the two imaging devices and four methods of image analysis. Spearman and Pearson correlation coefficients were used to determine the association of structural measurements between each of the techniques. The agreement between the two imaging techniques was determined with the Bland-Altman analysis. RESULTS: Reliability of subjective grading was strong for AS-OCT (ICC: 0.92, 95% CI: 0.83-0.96, p < 0.001) and K5M (ICC: 0.96, 95% CI: 0.96-0.91, p = 0.001). Image analysis with ImageJ reliability was strong between the imaging devices (ICC: 0.84, 95% CI: 0.55-0.94, p < 0.001). Agreement between each subjective technique was fair, κ = 0.45 (95% CI: 0.17-0.73, p < 0.001) and a positive Spearman correlation was also observed (r = 0.52, p < 0.001). There was no significant difference between the mean meibomian gland loss measured with ImageJ between AS-OCT and K5M (0.92 ± 6.28, p = 0.26). The 95% limits of agreement were -12.45% to +14.04%. CONCLUSION: These findings suggest subjective real-time grading of meibomian gland loss could be performed using readily available AS-OCT technology and that this method was interchangeable with the K5M.
