Grading and baseline characteristics of meibomian glands in meibography images and their clinical associations in the Dry Eye Assessment and Management (DREAM) study

doi:10.1016/j.jtos.2019.04.003

The Ocular Surface

Volume 17, Issue 3, July 2019, Pages 491-501

Meeting Presentation: Presentation at ARVO May 2019

https://doi.org/10.1016/j.jtos.2019.04.003 Get rights and content

Abstract

Purpose

To describe and evaluate a comprehensive grading system for meibomian gland (MG) digital infrared images developed for the Dry Eye Assessment and Management (DREAM) Study.

Methods

Cross-sectional study. Reading Center (RC) certified readers independently evaluated MG features of both lids from meibography images of dry eye disease subjects. Dropout areas were measured using planimetry software. Inter-reader and grade-regrade agreement and comparison of meiboscale scores (Meiboscale^©; Pult) from clinical centers to RC percent dropout and of MG features with clinical parameters were evaluated.

Results

Among 551 eyes of 277 patients at baseline, 62 (11%) upper lid and 5 (1%) lower lid images were missing. Lid eversion was poor in 63 (13%) of upper lids compared to 15 (3%) of lower lids. Intraclass correlation for inter-reader and grade-regrade agreement was moderate to substantial for most MG features. MG features were more frequent in the upper lid (p < 0.001), except for dropout glands, gaps, fluffy gland areas and dropout areas. Clinic meiboscale score was associated with RC percent dropout (p < 0.001), a clinic score of 0% having a mean RC score of 19%, and a clinic score of >75% having a mean RC score of 66%. MG plugging was associated with ghost glands (p = 0.009), dropout glands (p < 0.001) and a composite severity score (p = 0.02); turbid and absent secretions were associated with ghost glands (p = 0.046).

Conclusion

RC readers identified MG features with good reproducibility. Upper lids had more MG features. RC dropout areas correlated well with clinic meiboscale scores. Ghost glands were associated with paste like and absent meibomian secretions.

Introduction

Meibomian gland dysfunction (MGD) is a leading cause of dry eye disease (DED) [1]. Clinically, fluorescein tear break up time, lid margin irregularity, vascular engorgement, glandular orifice obstruction, anterior or posterior displacement of the mucocutaneous junction, and the quality of expressed sebum have been used for assessing MGD related to DED [2,3]. However, there is no standardized, universal grading system that is in use to evaluate the features of meibomian glands (MG). The introduction of non-invasive, infrared photography of MGs has been a major step towards allowing assessment of two-dimensional details of the silhouette of the glands.

A variety of scoring systems have been used to quantify the degree of MG dropout in the upper and lower lids and to correlate the gland loss to clinical parameters [[4], [5], [6]].These scoring systems generally are intended for clinical use while examining patients. Further, only a small amount of literature describes the association of MG features seen on meibography with clinical parameters. Individual investigations and reviews have clearly spelt out the need for more exhaustive research to improve the correlation of ocular imaging with clinical findings in DED. There is consensus that the combination of both morphological and functional evaluation would be essential to providing further insights into the pathophysiology of DED [8].

The Dry Eye Assessment and Management (DREAM) Study [7]was a randomized clinical trial of omega-3 fatty acid supplementation for the treatment of moderate to severe DED. Images of MGs were obtained using infrared photography by centers that had the Oculus Keratograph^® 5M (OCULUS Optikgeräte, Wetzlar, Germany). The purpose of this paper is to introduce a comprehensive grading system for meibography images, assess the reproducibility of grading by certified readers, and evaluate the association of the gradings with clinical signs of MGD among participants of the DREAM Study.

Section snippets

Study population

From October 2014 through July 2016, 535 subjects from 27 clinical centers in the United States completed a screening and eligibility confirmation visits and were enrolled into the study. A detailed description of the DREAM study design has previously been described [9]. Briefly, subjects needed to be ≥ 18 years with ocular symptoms related to DED for at least 6 months with the use of or a desire to use artificial tears. The patient's average score from the two visits on the OSDI needed to be

Results

Among the 292 patients enrolled through centers with a keratograph, 277 (98%) patients had at least 1 image of an eyelid submitted to the reading center from the eligibility confirmation visit. Reasons for patients not having any images include machine malfunction and human error. Among the 551 eyes of 277 patients, 62 (11%) upper lid images and 5 (1%) lower lid images were missing. Among available lid images, the quality of lid eversion (good, fair or poor) when it was performed was similar

Discussion

Non-contact infrared photography facilitates assessment of morphological features of the MGs in DED and in other ocular and systemic conditions [19]. However, there is not an accepted standard classification of the MG features or method for identifying areas of MG dropout. We have compiled a comprehensive collection of morphological features from several previous studies and added a few more to facilitate investigation of possible associations with the clinical signs and symptoms of DED.

Conclusions

We have catalogued various morphological features among the MGs present in moderate to severe dry eye disease patients that will allow further investigation into their associations with demographic, clinical and laboratory tests common to DED. We have shown good agreement among readers in identifying different morphological features as well as measuring the percentage areas of MG dropout. There is good correlation between the clinic meiboscale scores and the RC drop out percentages. More MG

Conflicts of interest

Dr. Asbell:consultant for Sun Pharma, Dompe, Novaliq, Senju, Santen, Shire, Alcon, Kala, CLAO, Allakon, Medscape and Regeneron. Dr Bunya: Grant recipient from Bausch &Lomb/Immco Diagnostics and consultant for Celularity. Dr Massaro-giordano is consultant for GSK, Celularity and PRN. Ebenezer Daniel, Maureen Maguire, Maxwell Pistilli, Giacomina Massaro-giordano, Eli Smith and Pooja Kadakia have no conflicts of interest.

Financial support

Grants U10EY022881, U10EY022879 and R01 EY026972 from the National Eye Institute; Supplemental funding from the Office of Dietary Supplements, National Institutes of Health; Prevention of blindness.

References (42)

J.S. Wolffsohn et al.
TFOS DEWS II diagnostic methodology report
Ocul Surf
(2017)
S. McGinnigle et al.
Evaluation of dry eye
Surv Ophthalmol
(2012)
R. Arita et al.
Noncontact infrared meibography to document age-related changes of the meibomian glands in a normal population
Ophthalmology
(2008)
R. Arita et al.
New insights into the morphology and function of meibomian glands
Exp Eye Res
(2017)
P.A. Asbell et al.
Dry eye assessment and management (DREAM©) study research group. Dry eye assessment and management (DREAM©) study: study design and baseline characteristics
Contemp Clin Trials
(2018)
Y. Zhao et al.
The significance of meibomian gland changes in asymptomatic children
Ocul Surf
(2018)
H. Pult et al.
Non-contact meibography: keep it simple but effective
Contact Lens Anterior Eye
(2012)
A.J. Bron et al.
The contribution of meibomian disease to dry eye
Ocul Surf
(2004)
J.S. Kim et al.
Exploring the Asian ethnic predisposition to dry eye disease in a pediatric population
Ocul Surf
(2019)
I.A. Butovich
Meibomian glands, meibum, and meibogenesis
Exp Eye Res
(2017)

K.K. Nichols et al.

The international workshop on meibomian gland dysfunction: executive summary

Investig Ophthalmol Vis Sci

(2011)

Dry Eye Assessment and Management Study Research Group et al.

n-3 fatty acid supplementation for the treatment of dry eye disease

N Engl J Med

(2018)

D. Finis et al.

Evaluation of meibomian gland dysfunction and local distribution of meibomian gland atrophy by non-contact infrared meibography

Curr Eye Res

(2015)

T.C.Y. Chan et al.

Advances in dry eye imaging: the present and beyond

Br J Ophthalmol

(2018)

J.B. Robin et al.

Meibomian gland evaluation in patients with extended wear soft contact lens deposits

CLAO J

(1986)

R. Arita et al.

Proposed diagnostic criteria for obstructive meibomian gland dysfunction

Ophthalmology

(2009)

H. Pult et al.

Relation between upper and lower lids' meibomian gland morphology, tear film, and dry eye

Optom Vis Sci

(2012)

Y.W. Koh et al.

Detection of meibomian glands and classification of meibography images

J Biomed Opt

(2012)

R. Arita et al.

Objective image analysis of the meibomian gland area

Br J Ophthalmol

(2014)

J.R. Landis et al.

The measurement of observer agreement for categorical data

Biometrics

(1977)

K.Y. Liang et al.

Regression analysis for correlated data

Annu Rev Public Health

(1993)

Cited by (41)

A Deep Learning Approach for Meibomian Gland Appearance Evaluation
2023, Ophthalmology Science
To develop and evaluate a deep learning algorithm for Meibomian gland characteristics calculation.
Evaluation of diagnostic technology.
A total of 1616 meibography images of both the upper (697) and lower (919) eyelids from a total of 282 individuals.
Images were collected using the LipiView II device. All the provided data were split into 3 sets: the training, validation, and test sets. Data partitions used proportions of 70/10/20% and included data from 2 optometry settings. Each set was separately partitioned with these proportions, resulting in a balanced distribution of data from both settings. The images were divided based on patient identifiers, such that all images collected for one participant could end up only in one set. The labeled images were used to train a deep learning model, which was subsequently used for Meibomian gland segmentation. The model was then applied to calculate individual Meibomian gland metrics. Interreader agreement and agreement between manual and automated methods for Meibomian gland segmentation were also carried out to assess the accuracy of the automated approach.
Meibomian gland metrics, including length ratio, area, tortuosity, intensity, and width, were measured. Additionally, the performance of the automated algorithms was evaluated using the aggregated Jaccard index.
The proposed semantic segmentation–based approach achieved average aggregated Jaccard index of mean 0.4718 (95% confidence interval [CI], 0.4680–0.4771) for the ‘gland’ class and a mean of 0.8470 (95% CI, 0.8432–0.8508) for the ‘eyelid’ class. The result for object detection–based approach was a mean of 0.4476 (95% CI, 0.4426–0.4533). Both artificial intelligence–based algorithms underestimated area, length ratio, tortuosity, width_mean, width_median, width_10th, and width_90th. Meibomian gland intensity was overestimated by both algorithms compared with the manual approach. The object detection–based algorithm seems to be as reliable as the manual approach only for Meibomian gland width_10th calculation.
The proposed approach can successfully segment Meibomian glands; however, to overcome problems with gland overlap and lack of image sharpness, the proposed method requires further development. The study presents another approach to utilizing automated, artificial intelligence–based methods in Meibomian gland health assessment that may assist clinicians in the diagnosis, treatment, and management of Meibomian gland dysfunction.
The authors have no proprietary or commercial interest in any materials discussed in this article.
Latest developments in meibography: A review
2022, Ocular Surface
Citation Excerpt :
Proposed metrics and their potential clinical relevance are summarised in Table 2. An objective method that would allow detection of subtle changes in Meibomian gland morphology is currently a field of interest for many researchers [42,81,86,90–96]. Koh et al. [90] were the first to develop an algorithm that could detect Meibomian gland characteristics with minimal observer input.
Meibography is a visualisation technique that has been used for over 40 years. There have been significant improvements in image quality, examination technique and image interpretation over this period. Although meibography has received sporadic reviews in the past, an updated review is timely due to the rapid recent rise of relevant technology and advances in both image processing and artificial intelligence. The primary aim of this paper is to review recent research into Meibomian gland imaging and update the community about the most relevant technologies and approaches used in the field.
Morphological characteristics of Meibomian Glands and their Influence on Dry Eye disease in contact lens wearers
2022, Ocular Surface
Citation Excerpt :
The Dry Eye Assessment and Management (DREAM) Study has since introduced a long list of new MG morphology metrics, which include the terms ‘distorted’, ‘tortuous’, ‘hooked’, ‘abnormal gap’, ‘overlapping’, ‘fluffy areas’, ‘tadpoling’, ‘thinned’, ‘thickened’, ‘ghost’, ‘no extension to lid margin’, ‘shortened’, and ‘dropout (atrophy)’ [27]. Although there is some literature suggesting greater amounts of MG atrophy and MG tortuosity are associated with worse DED states, most of the DREAM Study morphology metrics have yet to be fully vetted as DED diagnostic tools [27]. This information is needed not only because it would be of value to know their clinical utility, but also because eliminating some of the metrics would allow clinicians to streamline their MG evaluations.
Meibomian glands (MG) are now easily imaged via clinical meibography machines. The purpose of this work was to explore the utility of the known MG morphology metrics for predicting dry eye disease (DED) in contact lens (CL) wearers.
Successful and previous CL wearers were recruited. DED was diagnosed if the participant's worst eye had a reduced tear meniscus height (TMH) of <0.2 mm or non-invasive tear break-up time (NITBUT) of <10 s and a Standardized Patient Evaluation of Eye Dryness (SPEED) score >5.0. Meibography was performed and images were subjectively graded by two examiners for the following MG characteristics: distorted, tortuous, hooked, abnormal gap, overlapping, fluffy areas, tadpoling, thinned, thickened, ghost, no extension to lid margin, shortened and dropout (atrophy). DED diagnostic ability of each metric was determined with receiver operating characteristic (ROC) analysis.
A total of 112 participants were recruited, with 18.8% having DED and 60.7% being female. The only MG morphology metrics that were marginally predictive of DED were thickened upper eyelid MGs (p = 0.046), thickened mean upper plus lower eyelid MGs (p = 0.007), and atrophy of upper eyelid MGs (p = 0.043); however, none of these metrics reached a meaningful area under the curve in ROC analysis (all <0.70).
While abnormal MG morphology is likely suggestive of DED in CL wearers, none of the MG morphology metrics evaluated alone in this study had clinically meaningful predictive value for detecting DED in this group of current and previous CL wearers.
Quantitative analysis of morphological and functional features in Meibography for Meibomian Gland Dysfunction: Diagnosis and Grading
2021, eClinicalMedicine
Citation Excerpt :
The GA was calculated as the whole MGs area ratio by eliminating the gaps between glands, which can reflect additional information of gland thickening, thinning and abnormal gaps compared to MGs drop out quantified as meibograde [22,34]. The thinned glands and large gaps had been observed and determined to represent incomplete gland atrophy in the DREAM study [35]. While MGs drop out was reported to represent the final stage of MGD [22], our study demonstrated that a decreased GA was independently associated with the presence of MGD from level 2 to level 5, revealed a thinned glands and large gaps were also involved in the progression of the disease.
To explore the performance of quantitative morphological and functional analysis in meibography images by an automatic meibomian glands (MGs) analyser in diagnosis and grading Meibomian Gland Dysfunction (MGD).
A cross-sectional study collected 256 subjects with symptoms related to dry eye and 56 healthy volunteers who underwent complete ocular surface examination was conducted between January 1, 2019, and December 31, 2020. The 256 symptomatic subjects were classified into MGD group (n = 195) and symptomatic non-MGD group (n = 61). An automatic MGs analyser was used to obtained multi-parametric measurements in meibography images including the MGs area ratio (GA), MGs diameter deformation index (DI), MGs tortuosity index (TI), and MGs signal index (SI). Adjusted odds ratios (ORs) of the multi-parametric measurements of MGs for MGD, and the area under the receiver operating characteristic (AUC-ROC) curves of multi-parametric measurements for MGD diagnosing and grading were conducted.
When consider age, sex, ocular surface condition together, the estimated ORs for DI was 1.62 (95% CI, 1.29-2.56), low-level SI was 24.34 (95% CI, 2.73-217.3), TI was 0.76(95% CI, 0.54-0.90), and GA was 0.86 (95% CI, 0.74-0.92) for MGD. The combination of DI-TI-GA-SI showed an AUC = 0.82 (P < 0.001) for discriminating MGD from symptomatic subjects. The DI had a higher AUC in identifying early-stage MGD (grade 1-2), while TI and GA had higher AUCs in moderate and advanced stages (grade 3-5). Merging DI-TI-GA showed the highest AUCs in distinguish MGD severities.
The MGs area ratio, diameter deformation, tortuosity and signal intensity could be considered promising biomarkers for MGD diagnosis and objective grading.
This work was supported by the Key-Area Research and Development Program of Guangdong Province (No. 2019B010152001), the National Natural Science Foundation of China under Grant (81901788) and Guangzhou Science and Technology Program (202002030412).
A Machine Learning Approach to Predicting Dry Eye-Related Signs, Symptoms and Diagnoses from Meibography Images
2024, SSRN
Factors affecting long-term changes of meibomian gland in MGD patients
2024, Graefe's Archive for Clinical and Experimental Ophthalmology

View all citing articles on Scopus

¹: Credit roster for the DREAM Research Group may be found in the Appendix.

View full text

Original ResearchGrading and baseline characteristics of meibomian glands in meibography images and their clinical associations in the Dry Eye Assessment and Management (DREAM) study

Abstract

Purpose

Methods

Results

Conclusion

Introduction

Section snippets

Study population

Results

Discussion

Conclusions

Conflicts of interest

Financial support

Ocul Surf

Surv Ophthalmol

Ophthalmology

Exp Eye Res

Contemp Clin Trials

Ocul Surf

Contact Lens Anterior Eye

Ocul Surf

Ocul Surf

Exp Eye Res

The international workshop on meibomian gland dysfunction: executive summary

Investig Ophthalmol Vis Sci

n-3 fatty acid supplementation for the treatment of dry eye disease

N Engl J Med

Evaluation of meibomian gland dysfunction and local distribution of meibomian gland atrophy by non-contact infrared meibography

Curr Eye Res

Advances in dry eye imaging: the present and beyond

Br J Ophthalmol

Meibomian gland evaluation in patients with extended wear soft contact lens deposits

CLAO J

Proposed diagnostic criteria for obstructive meibomian gland dysfunction

Ophthalmology

Relation between upper and lower lids' meibomian gland morphology, tear film, and dry eye

Optom Vis Sci

Detection of meibomian glands and classification of meibography images

J Biomed Opt

Objective image analysis of the meibomian gland area

Br J Ophthalmol

The measurement of observer agreement for categorical data

Biometrics

Regression analysis for correlated data

Annu Rev Public Health

Original Research
Grading and baseline characteristics of meibomian glands in meibography images and their clinical associations in the Dry Eye Assessment and Management (DREAM) study