We aimed to establish an artificial intelligence (AI) system based on deep learning and transfer learning for meibomian gland (MG) segmentation and evaluate the efficacy of MG density in the diagnosis of MG dysfunction (MGD). First, 85 eyes of 85 subjects were enrolled for AI system-based evaluation effectiveness testing. Then, from 2420 randomly selected subjects, 4006 meibography images (1620 upper eyelids and 2386 lower eyelids) graded by three experts according to the meiboscore were analyzed for MG density using the AI system. The updated AI system achieved 92% accuracy (intersection over union, IoU) and 100% repeatability in MG segmentation after 4 h of training. The processing time for each meibography was 100 ms. We discovered a significant and linear correlation between MG density and ocular surface disease index questionnaire (OSDI), tear break-up time (TBUT), lid margin score, meiboscore, and meibum expressibility score (all p < 0.05). The area under the curve (AUC) was 0.900 for MG density in the total eyelids. The sensitivity and specificity were 88% and 81%, respectively, at a cutoff value of 0.275. MG density is an effective index for MGD, particularly supported by the AI system, which could replace the meiboscore, significantly improve the accuracy of meibography analysis, reduce the analysis time and doctors’ workload, and improve the diagnostic efficiency.

Purpose: This study aimed to establish an image-based classification that can reveal the clinical characteristics of patients with dry eye using unsupervised learning methods. Methods: In this study, we analyzed 82,236 meibography images from 20,559 subjects. Using the SimCLR neural network, the images were categorized. Data for each patient were averaged and subjected to mini-batch k-means clustering, and validated through consensus clustering. Statistical metrics determined optimal category numbers. Using a UNet model, images were segmented to identify meibomian gland (MG) areas. Clinical features were assessed, including tear breakup time (BUT), tear meniscus height (TMH), and gland atrophy. A thorough ocular surface evaluation was conducted on 280 cooperative patients. Results: SimCLR neural network achieved clustering patients with dry eye into six image-based subtypes. Patients in different subtypes harbored significantly different noninvasive BUT, significantly correlated with TMH. Subtypes 1 and 5 had the most severe MG atrophy. Subtype 2 had the highest corneal fluorescent staining (CFS). Subtype 4 had the lowest TMH, whereas subtype 5 had the highest. Subtypes 3 and 6 had the largest MG areas, and the upper MG areas of a person's bilateral eyes were highly correlated. Image-based subtypes are related to meibum quality, CFS, and morphological characteristics of MG. Conclusions: In this study, we developed an unsupervised neural network model to cluster patients with dry eye into image-based subtypes using meibography images. We annotated these subtypes with functional and morphological clinical characteristics.

The purpose of this study was to compare the differences between uneven meibomian gland (MG) atrophy with and without Demodex infestation based on the index of uneven atrophy score (UAS). In this retrospective cohort study, 158 subjects were recruited, including 66 subjects in the Demodex-positive MGD group, 49 subjects in the Demodex-negative MGD group, and 43 subjects as normal control. No significant difference was verified in OSDI, TMH, TBUT, CFS, lid margin score, and meibograde (all p > 0.05) between the Demodex-positive MGD group and the Demodex-negative MGD group. The UAS index of the upper eyelid or both eyelids was significantly higher in the Demodex-positive group in comparison with the normal control group and Demodex-negative group and the difference was statistically significant between the three groups. The UAS was significantly positive correlation with OSDI (r = 0.209, p < 0.05), lid margin score (r = 0.287, p < 0.001), and meibograde (r = 0.356, p < 0.001), which has a significant negative correlation with TBUT (r = −0.248, p < 0.05). Thus, Demodex infestation can cause uneven MG atrophy and we propose a novel index of UAS, which is used to evaluate uneven atrophy of MGs and as a morphological index of Demodex infestation.

Purpose: To evaluate the ability of swept-source optical coherence tomography angiography (SS-OCTA) to assess lid margin vascularity. Methods: This prospective, cross-sectional trial enrolled 125 participants, including 15 control subjects and 110 meibomian gland dysfunction (MGD) patients. Lid margin blood flow density (LMBFD) was obtained using SS-OCTA. LMBFD was assessed for repeatability in 54 of 125 participants and for reproducibility in 23 of 125 participants. The efficacy of LMBFD was validated in the 125 participants, who were divided into mild (n = 46), moderate (n = 42), and severe groups (n = 37) according to the lid margin vascularity severity shown in the slit-lamp photographs. Correlations between LMBFD and MG-related parameters, such as ocular surface disease index (OSDI), fluorescein tear break-up time (FTBUT), cornea fluorescein staining (CFS), lid margin score (LMS), and meibomian gland expressibility (ME), were analyzed in all 125 participants. Results: Repeatability and reproducibility coefficients were satisfactorily high in the scan mode with a scan area of 6 mm × 6 mm (intraclass correlation coefficient [ICC] repeatability = 0.905; ICC reproducibility = 0.986) and a scan area of 9 mm × 9 mm (ICC repeatability = 0.888; ICC reproducibility = 0.988). The LMBFD gradually increased in the mild, moderate, and severe groups (P < 0.001). LMBFD was significant correlated with OSDI (r = 0.290, P = 0.001), FTBUT (r = −0.195, P = 0.030), CFS (r = 0.352, P < 0.001), ME (r = 0.191, P = 0.033), and LMS (r = 0.370, P < 0.001). Conclusions: LMBFD may be a noninvasive, repeatable, reproducible, and efficient index for the quantitative evaluation of eyelid margin vascularity in the future. Translational Relevance: We demonstrated that SS-OCTA has the potential to evaluate the eyelid margin vascularity in MGD patients and guide future treatment strategies in clinics.