AI driven quantitative analysis of meibomian glands in children and adolescents: a benchmark dataset study

To analyze relations between upper lid (UL) and lower lid (LL) meibomian gland (MG) morphology and tear film and the MG criteria ability to predict dry eye. MG, lipid layer, and non-invasive break-up time (NIBUT) were evaluated of the OD of 20 randomly selected subjects (female = 10; median age = 44.5 years, interquartiles = 39.5 to 55 years). Subjects were grouped into nine Ocular Surface Disease Index (OSDI)- and 11 OSDI+ by the OSDI. Non-contact infrared meibography and image analysis were performed to evaluate MG loss, MG thickness, and MG bent angle. MG loss (Pearson correlation; r = 0.647, p = 0.003) and MG bent angle (r = 0.489, p = 0.027) were significantly correlated between lids, but MG thickness was not (r = -0.059, p = 0.413). MG loss was significantly (t-test; p = 0.048) less in the UL (median = 26.9%; LL = 32.3%), thicker in the LL (p < 0.001) and were more bent in the LL (p = 0.001). MG loss was significantly correlated to lipid-layer thickness (r < -0.597, p < 0.003) and NIBUT (r < -0.453, p < 0.030), whereas MG thickness and bent angle of the UL only were related to NIBUT (r < -0.563, p < 0.018). Combining MG loss of both lids (linear regression analysis) resulted in the best predictive ability of OSDI± (area under the receiver operative characteristic curve = 0.929, p = 0.001). MG scores between lids were significantly different but correlated. MG loss was significantly correlated to tear film characteristics including lipid layer thickness and stability. MG thickness and bent angle of the UL were related to NIBUT. The combination of both lids' MG loss showed best predictive ability of dry eye.

Effect of meibomian gland morphology on functionality with applied treatment

Meibomian Gland Morphology in Japanese Infants, Children, and Adults Observed Using a Mobile Pen-shaped Infrared Meibography Device

To determine the impact of age, sex, ethnicity, and contact lens wear on the detailed morphology of the Meibomian glands as quantified by a deep learning segmentation model. A large dataset of meibography images (n = 2233) from 560 subjects was compiled and input to a supervised machine learning model to quantify gland length, width, tortuosity, contrast, atrophy, density, and number of glands. These morphology outcomes were modeled as functions of age, sex, ethnicity, and contact lens wear parameters. Age was significantly associated with shorter glands, more atrophy, and lower gland density (all p < 0.001). No Meibomian gland morphological characteristics were related to sex. Asian subjects exhibited the longest and densest glands, and black subjects exhibited the most gland atrophy. Although contact lens wearers overall had significantly longer glands (∼4%-5%; p < 0.001) than non-wearers, no other contact lens wear parameter was significantly related to any Meibomian gland morphological feature. What constitutes "normal-looking" Meibomian glands in a meibography image depends on the age and ethnicity of the patient. There appear to be no significant female/male Meibomian gland morphological differences. Meibomian gland morphology is robust to contact lens wear based on a large-sample analysis of younger, successful contact lens wearers. Now that the impact of these external factors has been established, work is ongoing to determine exactly what alterations in Meibomian gland morphology contribute to downstream pathology. This study used artificial intelligence to provide clinicians with novel insights into normal versus abnormal Meibomian gland morphological features in their patients.

Meibography is a non-contact imaging technique used by ophthalmologists and eye care practitioners to acquire information on the characteristics of meibomian glands. One of its most important applications is to assist in the evaluation and diagnosis of meibomian gland dysfunction (MGD). As the artificial qualitative analysis of meibography images can lead to low repeatability and efficiency, automated and quantitative evaluation would greatly benefit the image analysis process. Moreover, since the morphology and function of meibomian glands varies at different stages of MGD, multiparametric analysis offering more comprehensive information could help in discovering subtle changes to glands during MGD progression. Therefore, an automated and multiparametric objective analysis of meibography images is urgently needed. An algorithm was developed to perform multiparametric analysis of meibography images with fully automatic and repeatable segmentation based on image contrast enhancement and noise reduction. The full architecture can be divided into three steps: (I) segmentation of the tarsal conjunctiva area as the region of interest (ROI); (II) segmentation and identification of glands within the ROI; and (III) quantitative multiparametric analysis including a newly defined gland diameter deformation index (DI), gland tortuosity index (TI), and gland signal index (SI). To evaluate the performance of this automated algorithm, the similarity index (k) and the segmentation error including the false-positive rate (rP ) and the false-negative rate (rN ) were calculated between the manually defined ground truth and the automatic segmentations of both the ROI and meibomian glands of 15 typical meibography images. The results of the performance evaluation between the manually defined ground truth and automatic segmentations were as follows: for ROI segmentation, the similarity index (k)=0.94±0.02, the false-positive rate (rP )=6.02%±2.41%, and the false-negative rate (rN )=6.43%±1.98%; for meibomian gland segmentation, the similarity index (k)=0.87±0.01, the false-positive rate (rP )=4.35%±1.50%, and the-false negative rate (rN )=18.61%±1.54%. The algorithm was successfully applied to process typical meibography images acquired from subjects of different meibomian gland health statuses, by providing the gland area ratio (GA), the gland length (L), gland width (D), gland diameter deformation index (DI), gland tortuosity index (TI), and gland signal index (SI). A fully automated algorithm was developed which demonstrated high similarity with moderate segmentation errors for meibography image segmentation compared with the manual approach, offering multiple parameters to quantify the morphology and function of meibomian glands for the objective evaluation of meibography images.

Changes in the tear film and meibomian gland morphology between preclinical dry eye and normal subjects represented by ocular surface disease index scores

Is contact lens discomfort related to meibomian gland morphology?

To study changes in the lid margin and meibomian glands and their association with aging, sex, and tear function. We examined 354 eyes in 177 subjects (76 men and 101 women; 21-93 years; mean age, 63.0 +/- 14.3 years) with no ocular symptoms or ocular surface disorders. Anatomic changes in the lid margin were studied using slit-lamp biomicroscopy. Meibomian gland function and morphology were evaluated on the basis of meibum expression and meibography, respectively. Tear function and ocular surface epithelium were assessed with the Schirmer test, by tear film break-up time, and with a fluorescein staining test. Eyes with abnormal lid margin anatomy, hyposecretion of meibum, and meibomian gland dropout were seen in 26 (7.3%), 46 (12.4%), and 68 eyes (18.6%), respectively, with a significant association between each finding and aging (P = <0.0001, 0.0498, and <0.0001, respectively). In patients < or =69 years of age, no significant association was found between meibomian gland-related findings and sex. However, a high incidence of abnormal lid margin and gland dropout was noted in men > or =70 years of age compared with women. No significant association was found between changes in the lid margin and meibomian glands and tear function in patients > or =40 years of age. Among symptom-free subjects, we found that changes in the lid margin and meibomian glands were closely related to aging. Among elderly subjects, changes in the anatomic lid margin and meibomian gland morphology were observed more frequently in men than in women. Tear function showed no association with either changes in the lid margin or function of the meibomian glands.

To compare, in vivo, differences in meibomian gland morphology between children and adolescents. Sixty-nine patients were included in this study and divided into two groups: children (n = 31; age range: 3 to 11 years) and adolescents (n = 39; age range: 12 to 18 years). Images of meibomian glands were obtained by infrared meibography and analyzed using ImageJ software (developed by the National Institutes of Health; available at http://rsb.info.nih.gov/ij/download.html). Meibomian gland loss, the number of meibomian gland ducts, the relative width of the meibomian gland ducts, and the percent area of the meibomian gland acini were compared between the two groups. Meibomian gland loss was found in both groups, but the meiboscore was not significantly different between the two groups (0.35 ± 0.6 vs 0.41 ± 0.8, t = -0.314, P > .05). The number of meibomian gland ducts (25.85 ± 3.25 vs 23.23 ± 3.06, t = -3.437, P < .05), relative width of the meibomian gland ducts (69.62% ± 5% vs 66.1% ± 7%, t = -2.454, P < .05), and percent area of the meibomian gland acini (57.7% ± 4% vs 55.5% ± 4%, t = 2.571, P < .05) in the upper eyelid were significantly greater in adolescents than in children. However, no significant differences were found in the lower eyelid between the two groups. Meibography is useful for the assessment of ocular surface conditions in children and adolescents. Meibomian gland loss occurs in both children and adolescents. The meibomian glands of the upper eyelid exhibit more morphological changes in adolescents than in children. [J Pediatr Ophthalmol Strabismus. 2017;54(2):78-83.].

To evaluate the effect of long-term contact lens (CL) wear on the morphology of meibomian glands (MGs) using meiboscore and digital analysis. Retrospective study. Sixty right eyes of sixty patients were involved in this study, and the data were analyzed retrospectively. According to the duration of CL wear, all patients were divided into three groups, nonwear group (n=21), short-term group (duration of CL wear ≤3 years, n=19) and long-term group (duration of CL wear>3 years, n=20). Digital images of MGs obtained by meibography were analyzed using Image J software, providing the area percentage of MGs loss. The meiboscores were also examined, and the data were analyzed using one-way ANOVA and Kruskal-Wallis test. Ten out of 21 nonwearers were scored 0 point, and 11 were scored 1 point in the upper lid meiboscores, while 7 were scored 0 point, 9 were scored 1 point, and 5 were scored 2 points in the total meiboscores. Seven out of 19 short-term wearers were scored 0 point, 10 were scored 1 point, and 2 were scored 2 points in the upper lid meiboscores, while 5 were scored 0 point, 6 were scored 1 point, 6 were scored 2 points, and 2 were scored 3 points in the total meiboscores. Four out of 20 long-term wearers were scored 0 point, 7 were scored 1 point, and 9 were scored 2 points in the upper lid meiboscores, while 3 were scored 0 point, 4 were scored 1 point, 4 were scored 2 points, 4 were scored 3 points, 4 were scored 4 points, and 1 was scored 5 points in the total meiboscores. The meiboscores of the upper eyelid and total meiboscores among the three groups were significantly different (Hc=9.967, P=0.007; Hc=9.725, P=0.008). The meiboscores of the upper eyelid and total meiboscores were significantly higher in the long-term group compared to the nonwear group (Z=102.500, P=0.003, Z=100.500, P=0.003) and the short-term group (Z=120.500, P=0.050, Z=117.500, P=0.041). No significant difference was found between the short-term group and the nonwear group. The median of the MGs loss area percentage in the upper eyelid of the nonwear, short-term and long-term groups was 9.2%, 13.3% and 16.7%, respectively. The median of the total MGs loss area percentage in the nonwear, short-term and long-term groups were 6.6%, 8.8% and 13.0%, respectively. The above medians were significantly different among the three groups (Hc=6.390, P=0.041; Hc=7.019, P=0.030). They were significantly larger in the long-term wearers than the nonwearers (Z=120.500, P=0.019, Z=120.500, P=0.009). No significant difference was found between the short-term group and the nonwear group, or between the short-term group and the long-term group. No significant differences in the meiboscores or MGs loss area percentage in the lower eyelid were noticed among the three groups. The area under the curve of total area percentage of MGs loss in receiver operating characteristic analysis was 0.981 (P<0.001). Long-term (more than 3 years) CL wear can cause MGs loss. Digital analysis is helpful in the morphologic evaluation of MGs. (Chin J Ophthalmol, 2016, 52: 604-609).

To determine the morphology and volume of Meibomian glands (MG) of dogs with microCT before and after partial tarsal plate excision (PTPE), cryotherapy, and laser therapy. MicroCT scans were made of 12 upper lids (ULs) and lower lids (LLs) of 12 dogs. After undergoing PTPE, 10 ULs and LLs were scanned again, and one UL and one LL was scanned after laser therapy and one UL and one LL after cryotherapy. The length of the area containing MGs did not change pre- and post-PTPE, and cryo- or laser therapy. The mean number of MGs in the ULs and LLs was 30.50 and 29.42, respectively, and did not change during the procedures. The average length of one individual MG was 2.60 mm. The mean volume of MGs in the 10 ULs and LLs pre-PTPE was 21.45 and 17.2 mm3 , respectively, and 12.84 and 11.25 mm3 in the UL and LL after PTPE, respectively. The mean volume of MGs decreased from 29.78 mm3 precryotherapy to 28.91 mm3 post-treatment and in the lower eyelid from 22.87 to 22.4 mm3 after cryotherapy. The mean volume of MGs in the UL and LL before laser therapy was 8.95 and 6.78 mm3 , respectively, and after 9.25 and 6.38 mm3 , respectively. MicroCT is a valuable tool to determine the morphology and the volume of MGs and to demonstrate changes that occur after PTPE, laser-, and cryotherapy. There is no need for additional preparation, such as staining, of the specimen prior to scanning.

Purpose To investigate the function and morphology of meibomian glands (MG) in night shift medical staff (MS). Methods Sixty-two eyes of 31 patients in the MS group and 59 eyes of 31 patients in the control group were consecutively enrolled. All participants completed Ocular Surface Disease Index (OSDI) and Standard Patient Dry Eye Evaluation (SPEED) questionnaires for dry eye severity, as well as Schirmer I and tear break-up time (TBUT) tests. LipiView® II Ocular Surface Interferometer was used for lipid layer thickness (LLT), MG dropout, and partial blink (PB) rate tests. MG expression was measured with an MG evaluator. Results The OSDI score in the MS group was 22.39 ± 13.42, which was significantly higher than that in the control group (9.87 ± 6.64 Z = −3.997, P=0.001). The SPEED score in the MS group was 7.94 ± 3.81, which was significantly higher than in the control group (3.65 ± 2.11, Z = −4.766, P=0.001). There was no significant difference in Schirmer I test between the MS group and control group (Z = −1.346, P=0.178). TBUT in MS group was significantly shorter than that in the control group (Z = −5.201, P=0.001). The mean LLT of the MS group was 55.02 ± 21.17 nm significantly thinner than that of the control group 72.76 ± 21.62 nm (Z = −4.482, P=0.001). MG loss occurred in 45.16% of affected eyes in the MS group and 16.13% of affected eyes in the control group, and the difference was statistically significant (χ2 = 14.352, P=0.001). MG yielding liquid secretion and MG yielding secretion score were significantly lower in the MS group than in the control group (Z = −3.641, P=0.001; Z = −3.146, P=0.001, resp.). There was a negative correlation between mean LLT and SPEED score (Spearman r = −0.363, P=0.045). Conclusions Night shift MS had a higher incidence of MGD compared to day workers.

PurposeTo investigate changes in meibomian gland morphology and impact of electronic device usage time on meibomian glands in pediatric age group.MethodsIn this prospective study, 149 eyes of 149 children were enrolled. The participants also completed the Standard Patient Evaluation of Eye Dryness (SPEED) questionnaire and provided information regarding weekly hours spent in front of a digital screen. Meibography was performed in all subjects. Grading of images was evaluated using a previously validated 5-point meiboscale (0–4) for meibomian gland atrophy and a 3-point scale for meibomian gland tortuosity (0–2).ResultsOf the 149 enrolled children, 83 (55.7%) were female and 66 (44.3%) male. The mean age was 13.0 3.0 (range, 5–18) years. The mean loss of meibomian gland area was 20.80 9.32%. The mean meiboscore was 1.20 0.58 for gland atrophy and the mean tortuosity score was 0.99 0.62. The mean screen time was 29.32 16.18 hr/week. There was a weak and significantly positive correlation between loss of meibomian gland area and screen time (r = 0.210, P = 0.010). There was a weak and significantly positive correlation between meiboscore for gland atrophy and screen time (r = 0.188, P = 0.022). We found a weak but significantly positive correlation between meibomian gland tortuosity and screen time (r = 0.142, P = 0.033).ConclusionMeibomian gland morphology may show changes in pediatric age group and excessive screen time may be a factor triggering these changes in gland morphology.

The influence of overnight orthokeratology and soft contact lens on the meibomian gland evaluated using an artificial intelligence analytic system

To evaluate whether the anterior segment topographic measurements, meibomian gland (MG), and non-invasive tear film break-up time (NITFBUT) differ between healthy children and children with isolated growth hormone deficiency (GHD). A total of 74 eyes of 37 children with GHD and 84 eyes of 42 age- and sex-matched healthy children were included in the study. The spherical equivalence (SE), mean keratometry (Km), corneal thickness, corneal volume (CV), anterior chamber depth (ACD), topographic NITFBUT, qualitative and quantitative MG measurements, corneal endothelial cell density (CD), and proportion of hexagonal cells (HG) were analysed. The mean SE level of GHD group was similar between groups (p = 0.017). Back Km values were insignificantly steep in children with GHD (p = 0.004, with Bonferroni correction). Specular microscopy analysis was not different between groups. MG loss of GHD group were higher than control group (p < 0.001). The MG morphology analysis and distortion grade were not different between groups (p > 0.05). Our results showed that the growth hormone (GH) may have an important role on the anterior segment parameters; however, it is not clear that this misregulation leads to a clinical scenario in childhood. Future studies investigating GHD and/or GH therapy on the ocular surface system are required to clearly demonstrate basic mechanism of GH action.