To evaluate the diagnostic utility of meibomian gland uneven atrophy (UA) in identifying Demodex infestation among patients with meibomian gland dysfunction (MGD) and to compare its diagnostic performance with that of cylindrical dandruff (CD).

Patients with MGD were prospectively enrolled and stratified into four groups (n=25/group) based on the presence/absence of CD and UA. Comparative analyses and multivariate adjustments were performed using Demodex detection confirmed by in vivo confocal microscopy and light microscopy. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis.

A total of 100 patients were recruited. The median age was 31.0y (range: 21.0–45.0y; Q1, Q3: 27.0–37.0y), including 29 males and 71 females. Demodex counts per eyelash and positivity rates were highest in the CD⁺UA⁺ group, followed by CD⁺UA⁻, and CD⁻UA⁺, and then CD⁻UA⁻. UA was associated with a 5.521-fold increased likelihood of Demodex positivity [95% confidence interval (CI): 2.098–14.531, P=0.001]; CD conferred a 6.460-fold increased risk (95%CI: 2.454–17.002, P<0.001). Incorporating UA status significantly improved model fit (Δχ²=14.096, P<0.001). The areas under the ROC curve (AUC) for CD (in UA^–), UA (in CD^–), and CD+UA (CD^–UA^– and CD⁺UA⁺) were 0.776, 0.762, and 0.865, respectively. CD and UA demonstrated comparable sensitivity (CD: 87.5%, UA: 86.7%) and specificity (CD: 65.6%, UA: 65.7%), which improved to 90.9% and 82.1%, respectively, when both were combined.

UA is a significant predictor of Demodex infestation, with diagnostic performance comparable to CD. Combined assessment of CD and UA offers high sensitivity and good specificity, providing a novel and practical tool and substantially reducing misdiagnosis.

To explore whether unilateral herpes simplex keratitis (HSK) can cause morphological changes of bilateral meibomian glands (MGs) based on artificial intelligence (AI) analytical system.

In the retrospective study, 29 patients with unilateral HSK and 29 participants matched in terms of age and sex were included as control group. Meibographic images of the upper eyelid using Keratograph 5M and assessed ocular surface parameters including tear meniscus height and tear break-up time. MG density and vagueness values were automatically analyzed and calculated using an AI analytical system. We compared the differences between the affected and the contralateral unaffected eyes in HSK patients, and the normal control eyes. We employed either the paired t-test or the Wilcoxon signed-rank test to compare significant difference between the affected and unaffected eyes in HSK patients or between the HSK group and control group.

The MG density was 0.19±0.09 in the HSK-affected eye and 0.18±0.07 in contralateral unaffected eye, which had no significant difference (P=0.616). The MG density between the affected eye with HSK and the normal control group was statistically significant (P=0.028). There was a significant difference in MG density between the contralateral unaffected eye and the normal control group (P=0.012). However, no significant difference in vagueness value was observed between the eye with HSK and the control group or between HSK eye and contralateral eye.

The MG density between the HSK-affected eye and the contralateral unaffected eye don’t significantly differ, whereas there is a significant decrease in the HSK group compared to that of the normal participants. Unilateral HSV keratitis may suffer from bilateral changes of MG morphology indicating bilateral dry eye. Therefore, the fellow eye of patients with unilateral HSK should be considered a potential case of MG dysfunction, necessitating early treatment for bilateral dry eye in the clinic.