Meibum Lipidomic Analysis in Evaporative Dry Eye Subjects.

Meibomian Glands (MG) are sebaceous glands responsible for the production of meibum, the main component of the Tear Film Lipid Layer (TFLL). The TFLL facilitates the spread of the tear film over the ocular surface, provides stability and reduces tear evaporation. Alterations in meibum composition lead to different ocular alterations like Meibomian Gland Dysfunction (MGD) and subsequent Evaporative Dry Eye (EDE). The aim of the present study was to investigate the composition and abundance of meibum lipids and their relationship with eyelid margin abnormalities, lipid layer patterns and MG status. The study utilizes a lipidomic approach to identify and quantify lipids in meibum samples using an Elute UHPLC system. This system considered all four dimensions (mass/charge, retention time, ion mobility and intensity) to provide the accurate identification of lipid species. Samples were categorized as healthy or low/no signs of alteration (group 1) or severe signs of alteration or EDE/MGD (group 2). The current investigation found differences in Variable Importance in Projection lipid abundance between both groups for the MGD signs studied. Changes in meibum composition occur and are related to higher scores in eyelid margin hyperaemia, eyelid margin irregularity, MG orifice plugging, MG loss and lipid layer pattern.

A Comprehensive Study on Tear Meniscus Height Inter-Eye Differences in Aqueous Deficient Dry Eye Diagnosis.

(1) Background: Dry eye disease (DED) is a chronic ocular surface condition that requires precise diagnostic tools. The present study aimed to investigate the diagnostic potential of the absolute inter-eye difference (|OD-OS|) in tear meniscus height (TMH) for the detection of the presence of aqueous deficient dry eye (ADDE). (2) Methods: A sample of 260 participants with dry eye complaints underwent ocular surface examinations thorough diagnostic assessments based on the Tear Film and Ocular Surface Society guidelines (TFOS DEWS II). Participants were subsequently categorized as No ADDE and ADDE based on TMH. Statistical analyses to determine the optimal TMH|OD-OS| cut-off value in a randomly selected study group (200 participants) were performed, while a separate validation analysis of the cut-off value obtained in a random cross-validation group (60 participants) was also performed. (3) Results: The significant diagnostic capability of TMH|OD-OS| (area under the curve = 0.719 ± 0.036, p < 0.001) was found. The identified cut-off value of 0.033 mm demonstrated reliable specificity (77.6%) and moderate sensitivity (59.1%). Cross-validation confirmed the cut-off value's association with the TFOS DEWS II diagnostic criterion (Cramer's V = 0.354, p = 0.006). (4) Conclusions: The present study provides evidence for the diagnostic potential of TMH|OD-OS| in identifying ADDE. The identified cut-off value enhances the specificity and offers moderate sensitivity, providing an objective tool for clinical decision making.

Capability of the Inter-Eye Differences in Osmolarity, Break-Up Time and Corneal Staining on the Diagnostic of Dry Eye.

PURPOSE: Inter-eye variability is a recognized characteristic of Dry Eye Disease (DED) and has been proposed as a diagnostic indicator in clinical practice. This study aimed to assess the diagnostic potential of the absolute difference between eyes in three key diagnostic tests recommended by the Tear Film and Ocular Surface Society Dry Eye Workshop II (TFOS DEWS II) Diagnostic Methodology report: tear film osmolarity, Fluorescein Break-Up Time (FBUT), and ocular surface staining. METHODS: A total of 180 participants were included in a cross-sectional study. Before a dry eye examination, participants completed an online self-administered OSDI questionnaire. The TFOS DEWS II diagnostic criteria for DED assessment were used: along with OSDI, osmolarity, FBUT and ocular surface staining were measured in all participants in both eyes following standardized methodology. Based on signs and symptoms, participants were diagnosed as having No DED or DED. After diagnosis, the parameters were computed as the right and left eyes' absolute inter-eye difference (|OD-OS|). RESULTS: Receiver Operating Characteristics analyses for computed parameters were used based on the previous diagnosis. ROC analyses showed that Osmolarity|OD-OS| have a diagnostic capability to differentiate between No DED and DED participants with a cut-off value of 9.5 mOsm/L (AUC = 0.745 ± 0.052, p < 0.003), whereas FBUT|OD-OS| and Corneal Stainning|OD-OS| have not (AUC, both p ≥ 0.160). CONCLUSION: The present study found that the Osmolarity|OD-OS| parameter could be used as a diagnostic indicator for DED assessment, while the FBUT|OD-OS| and the Corneal Staining|OD-OS| parameters do not have this capability.

Meibomian gland secretion quality association with ocular parameters in university students during COVID- 19 restrictions.

PURPOSE: To determine if the Meibomian Gland (MG) secretion quality is associated with symptoms of ocular discomfort, hours of Video Display Terminals (VDT) use, eyelid margin abnormalities, conjunctival hyperemia, and Meibomian Gland Loss Area (MGLA) in a sample of university students. METHODS: An online survey that included an Ocular Surface Disease Index (OSDI) questionnaire and an extra question about hours of VDT use recruited an initial sample of 183 participants. Only 120 participants that fulfilled the inclusion criteria were scheduled for a battery of ocular surface and MG specific exam. The tests include: 1) meibometry, 2) slit lamp exploration of eyelid margin abnormalities (irregularity, hyperemia and MG orifices plugging), MG secretion quality and conjunctival hyperemia, and 3) Meibography. RESULTS: Significant positive correlations between the MG secretion quality and eyelid margin hyperemia, MG orifices plugging, MGLA, nasal conjunctival hyperemia, and temporal conjunctival hyperemia (Spearman Rho; all r>0.186, p<0.042) were found. Multivariate regression found association between OSDI with hours of VDT use (B=0.316, p=0.007), and eyelid hyperemia (B=0.434, p≤ 0.001). A statistical association between MG secretion quality and eyelid margin hyperemia, MG orifices plugging, MGLA and conjunctival hyperemia (Fisher's exact; all p<0.039) were found. Multivariate regression found association between MG secretion quality with MG orifices plugging (B=0.295, p=0.004) and meibometry (B=-0.001, p=0.029). CONCLUSION: Participants with higher values in MG secretion quality have higher values in eyelid margin hyperemia, MG plugging, MGLA, and conjunctival hyperemia. No direct relationship between MG secretion quality and hours of VDT use or OSDI were found.

Diurnal variations of tear film osmolarity on the ocular surface.

The measurement of tear film osmolarity has been suggested as a gold standard in the diagnosis of dry eye. Many tear film physiological variables oscillate during the day. This review summarises current clinical knowledge regarding diurnal osmolarity variation in the tear film. A critical analysis is presented in respect of of sample size and characteristics, differences in the diurnal osmolarity variation on healthy versus altered tear film conditions or environment, and time of day and number of measurements undertaken. A comparison of 21 studies was made for studies in which one of the main objectives was to analyse the variance of tear film osmolarity at different time-points in a day on human cohorts. Tear film osmolarity appeard to be somewhat influenced by the time of day in healthy subjects and patients with ocular surface disease, or altered by environmental conditions. Both healthy and non-physiological tear film cohorts showed variations in results depending on the study: no variations during the day or statistically different values at some point in the day. These differences could be in the middle of the day or between the beginning and the end of the day, with higher values in the morning than in the afternoon, or even the opposite situation. The possibility of diurnal variations in tear film osmolarity should be considered by the clinician since the time of day when the tear film measurements are made can be critical in making the right diagnosis. Future studies in the diurnal variation field may have to use a well-established range of measurement time-points and a larger group of healthy subjects and and subjects who have a tear film altered by pathological or environmental conditions.

Analysis of the Differences in Ocular Surface Damage and Inflammatory Signs between Healthy and Evaporative Dry Eye Participants.

OBJECTIVE: Distinguish between EDE severity levels by analysing the MGLA, conjunctival hyperemia and corneal staining. METHODS: One hundred participants were recruited based on OSDI, TO, TFBUT, TMH, and LLP to be categorised as healthy (Group 1) or EDE (Group 2). Group 2 was divided into Group 2A (mild symptoms), 2B (moderate), and 2C (severe). MGLA, conjunctival hyperemia, and corneal staining were measured. RESULTS:  Positive correlation between MGLA, conjunctival hyperemia, and corneal staining were found (all r ≥ 0.221, p ≤ 0.027). Significant differences were found: MGLA between Group 1 vs. 2C and 2C vs. 2A or 2B; conjunctival hyperemia between Group 1 vs. 2A, 2B or 2C; corneal staining between Group 1 vs. 2B or 2C and 2A vs. 2B or 2C (all p ≤ 0.049). CONCLUSION: Severe EDE participants have higher MGLA, conjunctival hyperemia, and corneal staining values than healthy, mild, or moderate EDE participants.

A cross-sectional study of non-modifiable and modifiable risk factors of dry eye disease states.

PURPOSE: The present study aimed to determine the relationship of non-modifiable (rheumatoid arthritis, thyroid diseases, and arterial hypertension) and modifiable risk factors (diuretics, antidepressants, or anxiolytics tranquilizers) with the different Dry Eye Disease (DED) diagnostics in a sample adjusted by antihistamines intake. METHODS: A total of 400 participants were included in a cross-sectional study. Before a dry eye examination, participants completed an online self-administered OSDI questionnaire with six additional questions about possible DED risk factors. The Tear Film and Ocular Surface Society Dry Eye Workshop II (TFOS DEWS-II) diagnostic criteria of DED was used. Based on signs and/or symptoms, participants were divided into 4 groups: No DED, Pre-clinical DED, Predisposition to DED and DED. Since the symptom scores would have been altered by the use of antihistamines, the analysis of each outcome was adjusted for this factor, where those participants were assumed to be symptomatic. RESULTS: Multivariable logistic regression found thyroid disease as a possible risk factor for DED (OR 4.53, 95 % CI 1.04-19.73; Fisher's exact, p = 0.044; Cramers V = 0.140, p = 0.024). No association was found between the studied parameters and Pre-clinical DED (Fisher's exact, all p ≥ 0.398; Cramers V, all p ≥ 0.242) or Predisposition to DED (Fisher's exact, all p ≥ 0.065; Cramers V, all p ≥ 0.031). CONCLUSION: Participants with thyroid disease were more likely to develop DED, therefore, thyroid disease could be a risk factor for DED.

Categorization of the Aqueous Deficient Dry Eye by a Cut-Off Criterion of TMH Measured with Tearscope.

A decrease of the Tear Meniscus Height (TMH) has been proposed as a useful indicator for Aqueous Deficient Dry Eye (ADDE) categorization. The present study aimed to calculate a TMH cut-off criterion for the categorization or severity assessment of ADDE with the Tearscope. 200 participants with a previous Dry Eye Disease (DED) diagnosis according to TFOS DEWS-II criteria were recruited. TMH by slit-lamp illumination and Lipid Layer Pattern (LLP) with Tearscope were assessed to categorise the participants into the ADDE or the Evaporative Dry Eye (EDE) group. The ADDE group was also subdivided into Mild-moderate ADDE and Moderate-severe ADDE based on TMH with slit-lamp. Additionally, the TMH was measured by Tearscope (TMH-Tc). Receiver Operating Characteristics showed that the TMH-Tc have a diagnostic capability to differentiate between ADDE and EDE participants, and between Mild-moderate or Moderate-severe ADDE, with a cut-off value of 0.159 mm (AUC = 0.843 ± 0.035, p < 0.001; sensitivity: 86.4%; specificity: 75.4%) and 0.105 mm (AUC = 0.953 ± 0.025, p < 0.001; sensitivity: 98.1%; specificity: 80.0%), respectively. The present study proposed a cut-off criterion to differentiate between ADDE and EDE participants, or between ADDE severities through TMH assessed by Tearscope.

Validation of a New Software Application for Tear Breakup Measurement.

SIGNIFICANCE: Tear film instability has been recognized as one key mechanism on dry eye disease. There is a need for new instruments, methods, or protocols to improve the repeatability and reproducibility of the tear film stability to facilitate its clinical evaluation use as a rapid and reliable primary diagnostic test. PURPOSE: The present pilot study aimed to validate a method to automatically measure the tear film breakup time (BUT). METHODS: A total of 264 videos of BUT were randomly selected among the clinical history of subjects attending the optometry clinic. Videos were stored in a backup server through a secure file transfer protocol and analyzed by three different examiners: two masked observers (subjective evaluation) and a third investigator using the automatic software application (objective evaluation). Subjective evaluation was conducted only once on an online software designed for this protocol where videos were presented in random masked order. Automatic evaluation based on color and texture analysis was performed by (1) automatic localization of sequences of interest in the video, (2) extraction of the region of interest within each frame, and (3) automatic BUT measurement from evolution curves in the region of interest as time elapsed from the beginning of the sequence of interest until the curve exceeds a threshold. RESULTS: Substantial correlation was observed among the examiners (intraclass correlation coefficient, 0.752). There was a statistical difference between observer 1 and 2 evaluations (t test, P < .001), whereas data provided by the software showed no significant differences from those of the observers (t test, P ≥ .26). Similar results to the whole data set analysis were obtained when the sample was reassessed only considering mean BUT values ≤15 seconds. CONCLUSIONS: The present pilot study showed acceptable clinical results for the software application designed to objectively measure the BUT.

Inter-week variation of meibometry and tear break-up time in healthy subjects.

CLINICAL RELEVANCE: Variation with time in the tear film parameters should be considered by the clinician since the time when measurements are made can influence proper diagnosis. BACKGROUND: A hallmark of dry eye is an unstable tear film associated with variability in objective daily measures. The purpose of the present study was to evaluate the inter-week repeatability of meibometry, break-up time (BUT) and maximum blink interval (MBI) in healthy subjects. METHODS: Forty healthy subjects were recruited for the study. Meibometry, BUT and MBI were performed twice in two sessions, one-week apart. Meibum from the lower eyelid was collected and quantified with the Meibometer MB550. Five meibomian curves were generated by tape and each peak value was averaged as a mean value per session. BUT and MBI were determined three times after fluorescein instillation and recorded by a DV-3 camera attached to the slit lamp. BUT and MBI videos were analysed by a second masked observer. Both BUT and MBI values were calculated by averaging the two most similar measurements of the three evaluated. RESULTS: No statistical difference between inter-week sessions was obtained for the meibometry (p > 0.340; Wilcoxon test), BUT (p > 0.326; Wilcoxon test) and MBI values (p ≥ 0.248; Wilcoxon test). Inter-week differences were low for BUT and MBI when time intervals were no longer than 15 seconds (both p ≥ 0.586; Wilcoxon test). A correlation between BUT and MBI was found (r ≥ 0.668, p < 0.001; Spearman Rho), while a no statistical correlation was obtained between meibometry results and BUT or MBI (all p ≥ 0.194; Spearman Rho). CONCLUSION: Meibum secretion and the tear film stability present good intersession repeatability, and are stable along with sessions in healthy patients.

Osmolality and pH of commercially available contact lens care solutions and eye drops.

PURPOSE: The physical properties of contact lens care solutions and Eye Drops (ED) may affect initial comfort and dry eye symptomatology in contact lens wearers, although these properties are not always provided by manufacturers. The present study aimed to measure and compare the osmolality and pH of commercially available contact lens care solutions and ED. METHODS: Forty-four solutions were tested (17 lens care solutions and 27 ED) and classified by the presence and/or combination of the viscosity/lubrication-enhancing ingredients. Solution osmolality was obtained with the Fiske 110 osmometer and pH was measured with a micro-pH 2000. Each measurement was taken ten times, following the manufacturer's instructions, while controlling for room temperature and humidity. Differences between the values of the physical properties of the solutions were analysed by type and viscosity/lubrication-enhancing agent subclassification. RESULTS: Osmolality ranged from 192.6 ± 2.17 to 364.6 ± 2.88 mOsm/Kg, while pH ranged from 6.35 [6.35-6.26] to 7.99 [7.99-8.00]. A significant difference in the osmolality and pH values of contact lens care solutions and ED was found when classified by type or viscosity/lubrication-enhancing agent (ANOVA and Kruskal-Wallis respectively, both p < 0.001). CONCLUSIONS: The physical properties of some contact lens care solutions and ED are not readily available. The osmolality and pH values of various commercially available lens care solutions and ED vary significantly both by type and viscosity/lubrication-enhancing ingredients.

Inter-eye osmolarity differences in patients with symptomatic and non-symptomatic dry eyes / Diferenças de osmolaridade entre olhos de pacientes com olhos secos sintomáticos e não sintomáticos

ABSTRACT Purpose: To analyze whether inter-eye osmo larity differences were related to dry eye symptomatology. Methods: A total of 135 participants were randomly recruited from those who visited in the Optometry Clinic of the Optometry Faculty (Universidade de Santiago de Compostela). In a single scheduled session after the recruitment, Ocular Surface Disease Index was filled out following the standard instructions and TearLab measurements were made in both the participants' eyes (10-15 min lapse). Osmolarity values were compared between the right and left eyes and the absolute inter-ocular difference (-OD-OS-) correlated with the Ocular Surface Disease Index score for the whole sample. Based on the Ocular Surface Disease Index score, the sample was divided into four symptomatic subgroups, and differences in the -OD-OS- values were calculated. Results: The whole sample showed a statistically significant inter-eye osmolarity difference (p=0.025; -OD-OS- = 9.2 ± 9.3 mOsm/l) and the correlation between Ocular Surface Disease Index and -OD-OS- (r=0.369; p<0.001). A statistically significant difference was found in the -OD-OS- value between symptomatic subgroups (Kruskal-Wallis, p=0.003). Mann-Whitney U test showed a significant difference between asymptomatic vs. moderate (p=0.006) vs. severe symp tomatic patients (p=0.001) and between mild vs. severe symptomatic patients (p=0.045), whereas no difference on -OD-OS- was found between participants with contiguous symptomatic subgroups (all p³0.174). Conclusion: Tear film inter-eye osmolarity differences are significantly higher in severe dry eye disease symptoms.

RESUMO Objetivo: Analisar se as diferenças entre osmolaridade entre os olhos foram relacionadas à sintomatologia do olho seco. Métodos: Um total de 135 participantes foram recrutados aleatoriamente entre os indivíduos da Clínica de Optometria da Faculdade de Optometria (Universidade de Santiago de Compostela). Em uma única sessão agendada após o recrutamento, o Índice de Doenças da Superfície Ocular foi preenchido seguindo as instruções padrão e as mensurações do TearLab foram feitas em ambos os olhos dos participantes (lapso de 10 a 15 min). Os valores de osmolaridade foram com parados entre os olhos direito e o esquerdo e a diferença absoluta ocular (-OD-OS-) correlacionada com a pontuação do Índice de Doença da Superfície Ocular para toda a amostra. Com base na pontuação do Índice de Doença da Superfície Ocular, a amostra foi dividida em quatro subgrupos sintomáticos, e as diferenças nos -OD-OS- os valores foram calcula dos. Resultados: A amostra total mostrou uma diferença de osmolarida de entre os olhos estatisticamente significativa (p=0,025; -OD-OS- = 9,2 ± 9,3 mOsm/l) e a correlação entre o Índice de Doença da Superfície Ocular e -OD-OS- (r=0,369; p<0,001). Diferença estatisticamente significativa foi encontrada no valor -OD-OS- entre os subgrupos sintomáticos (Kruskal-Wallis, p=0,003). O teste U de Mann-Whitney mostrou uma diferença significativa entre pacientes assintomáticos versus moderados (p=0,006) versus sintomáticos graves (p=0,001) e entre pacientes sinto máticos leves e graves (p=0,045), enquanto que nenhuma di ferença de -OD-OS- foi encontrada entre os participantes de subgrupos sintomáticos contíguos (todos p³0,174). Conclusão: As diferenças entre osmolaridade inter-ocular do filme lacrimal são significativamente maiores nos sintomas graves da doença do olho seco.

Inter-eye osmolarity differences in patients with symptomatic and non-symptomatic dry eyes.

PURPOSE: To analyze whether inter-eye osmo larity differences were related to dry eye symptomatology. METHODS: A total of 135 participants were randomly recruited from those who visited in the Optometry Clinic of the Optometry Faculty (Universidade de Santiago de Compostela). In a single scheduled session after the recruitment, Ocular Surface Disease Index was filled out following the standard instructions and TearLab measurements were made in both the participants' eyes (10-15 min lapse). Osmolarity values were compared between the right and left eyes and the absolute inter-ocular difference (|OD-OS|) correlated with the Ocular Surface Disease Index score for the whole sample. Based on the Ocular Surface Disease Index score, the sample was divided into four symptomatic subgroups, and differences in the |OD-OS| values were calculated. RESULTS: The whole sample showed a statistically significant inter-eye osmolarity difference (p=0.025; |OD-OS| = 9.2 ± 9.3 mOsm/l) and the correlation between Ocular Surface Disease Index and |OD-OS| (r=0.369; p<0.001). A statistically significant difference was found in the |OD-OS| value between symptomatic subgroups (Kruskal-Wallis, p=0.003). Mann-Whitney U test showed a significant difference between asymptomatic vs. moderate (p=0.006) vs. severe symp tomatic patients (p=0.001) and between mild vs. severe symptomatic patients (p=0.045), whereas no difference on |OD-OS| was found between participants with contiguous symptomatic subgroups (all p³0.174). CONCLUSION: Tear film inter-eye osmolarity differences are significantly higher in severe dry eye disease symptoms.

Reproducibility in measuring tear samples using a freezing point depression osmometer.

BACKGROUND: Hyperosmolarity of tear fluid has been recognised as a common feature of all types of dry eye disease. This study was designed to assess the inter-session reproducibility of a freezing point depression osmometer (Fiske 110) as the most common and precise way of measuring osmolality, by using two different volumes of tear samples on healthy subjects, and to determine the possible applications of this device in tear film research and clinical practice. METHODS: Measurements were made by using the Fiske 110 osmometer under two different tear sample volumes (4 µl and 2 µl). In both cases, samples were diluted in purified water to obtain the 20 µl required by the device to perform the measurement (1:4 and 1:9 dilutions, respectively). Inter-session reproducibility was determined in two groups of 40 healthy subjects, in two sessions, one week apart. In each group, one of the two different tear sample volumes was used to determine the reproducibility of each technique. RESULTS: No significant differences were detected between the measurements obtained in the two sessions using the 4 µl (paired t-test, p = 0.772; mean difference ± SD = -0.85 ± 18.77 mOsm/L; 95 per cent limits of agreement [LoAs] = -37.64/+35.94) or the 2 µl volume sample (paired t-test, p = 0.054; mean difference ± SD = 9.27 ± 29.44 mOsm/L; 95 per cent LoAs = -48.43/+66.97). CONCLUSIONS: Whereas both techniques show an acceptable inter-session reproducibility, the bias range with the present protocol was higher using the 2 µl tear sample volume than the 4 µl one. Therefore, it seems that the diluted 4 µl sample was the only dilution that could be acceptable for use in routine clinical practice for tear film analysis.

Clinical relationship of meibometry with ocular symptoms and tear film stability.

PURPOSE: To evaluate the relationship between meibometry with both ocular symptoms and tear film stability by: (1) to find out whether meibometry is able to differentiate between dry eye symptomatic and asymptomatic subjects classified by standardized dry eye questionnaires (OSDI and McMonnies), and (2) to assess the clinical relationship between meibometry with both tear break-up time (BUT) and maximum blink interval (MBI). METHODS: 140 Patients were recruited for the study. Using Meibometer MB550, five curves were generated for each patient. Subjects performed OSDI and McMonnies questionnaires and were stratified following a two- and a three-subgroup stratification for each questionnaire. BUT/MBI were repeated three times (by video recordings), and they were determined by counting their frames. RESULTS: Subjects grouped by OSDI showed a trend to present lower meibometry values as the OSDI score were higher (ANOVA, p≤0.044). For McMonnies questionnaire this was only true for the two-subgroup stratification (ANOVA, p=0.04), but not for three-subgroup stratification (one-way ANOVA, p=0.30). On the other hand, meibometry values showed a statistical correlation with both BUT (r=0.305, p<0.001) and MBI (r=0.265, p<0.001). When the sample was divided in three groups regarding BUT value (≤5s, between 5 and 10s and≥10s), significant differences of meibometry values were found between BUT subgroups (p=0.008). CONCLUSION: Meibometer MB550 can discriminate asymptomatic from dry eye symptomatic patients. Furthermore, there is a relationship between meibometry and the tear film stability.

On the development of conjunctival hyperemia computer-assisted diagnosis tools: Influence of feature selection and class imbalance in automatic gradings.

OBJECTIVE: The sudden increase of blood flow in the bulbar conjunctiva, known as hyperemia, is associated to a red hue of variable intensity. Experts measure hyperemia using levels in a grading scale, a procedure that is subjective, non-repeatable and time consuming, thus creating a need for its automatisation. However, the task is far from straightforward due to data issues such as class imbalance or correlated features. In this paper, we study the specific features of hyperemia and propose various approaches to address these problems in the context of an automatic framework for hyperemia grading. METHODOLOGY: Oversampling, undersampling and SMOTE approaches were applied in order to tackle the problem of class imbalance. 25 features were computed for each image and regression methods were then used to transform them into a value on the grading scale. The values and relationships among features and experts' values were analysed, and five feature selection techniques were subsequently studied. RESULTS: The lowest mean square error (MSE) for the regression systems trained with individual features is below 0.1 for both scales. Multi-layer perceptron (MLP) obtains the best values, but is less consistent than the random forest (RF) method. When all features are combined, the best results for both scales are achieved with MLP. Correlation based feature selection (CFS) and M5 provide the best results, MSE=0.108 and MSE=0.061 respectively. Finally, the class imbalance problem is minimised with the SMOTE approach for both scales (MSE<0.006). CONCLUSIONS: Machine learning methods are able to perform an objective assessment of hyperemia grading, removing both intra- and inter-expert subjectivity while providing a gain in computation time. SMOTE and oversampling approaches minimise the class imbalance problem, while feature selection reduces the number of features from 25 to 3-5 without worsening the MSE. As the differences between the system and a human expert are similar to the differences between experts, we can therefore conclude that the system behaves like an expert.

iDEAS: A web-based system for dry eye assessment.

BACKGROUND AND OBJECTIVES: Dry eye disease is a public health problem, whose multifactorial etiology challenges clinicians and researchers making necessary the collaboration between different experts and centers. The evaluation of the interference patterns observed in the tear film lipid layer is a common clinical test used for dry eye diagnosis. However, it is a time-consuming task with a high degree of intra- as well as inter-observer variability, which makes the use of a computer-based analysis system highly desirable. This work introduces iDEAS (Dry Eye Assessment System), a web-based application to support dry eye diagnosis. METHODS: iDEAS provides a framework for eye care experts to collaboratively work using image-based services in a distributed environment. It is composed of three main components: the web client for user interaction, the web application server for request processing, and the service module for image analysis. Specifically, this manuscript presents two automatic services: tear film classification, which classifies an image into one interference pattern; and tear film map, which illustrates the distribution of the patterns over the entire tear film. RESULTS: iDEAS has been evaluated by specialists from different institutions to test its performance. Both services have been evaluated in terms of a set of performance metrics using the annotations of different experts. Note that the processing time of both services has been also measured for efficiency purposes. CONCLUSIONS: iDEAS is a web-based application which provides a fast, reliable environment for dry eye assessment. The system allows practitioners to share images, clinical information and automatic assessments between remote computers. Additionally, it save time for experts, diminish the inter-expert variability and can be used in both clinical and research settings.

Diurnal variations in tear film break-up time determined in healthy subjects by software-assisted interpretation of tear film video recordings.

BACKGROUND: This study was designed to examine diurnal variations in tear film break-up time (BUT) and maximum blink interval (MBI) and to assess two different ways of calculating these variables on video recordings of the BUT test interpreted with the help of especially designed software. The repeatability of interpreting BUT video recordings was also addressed. METHODS: Twenty-six healthy young adults were enrolled after ruling out dry eye according to a battery of tests (ocular surface disease index, McMonnies questionnaire, Schirmer test, phenol red test and corneal staining). BUT and maximum blink interval were determined on video-recordings of the BUT test conducted over a day in four sessions (9.30 am, 12.30 pm, 3.30 pm and 6.30 pm). In each session, the test was repeated three times to give three videos in which three BUT and MBI values were obtained by a masked observer. BUT and MBI were determined by averaging the three measurements and by averaging only the two closest measurements. Finally, two further experienced observers re-examined the videos to assess the repeatability of the BUT measurements made. RESULTS: No diurnal variation in BUT was observed regardless of whether three or two video measurements were averaged. Significant correlation was detected between BUT and MBI. Inter-observer repeatability was better when BUT times were no longer than 15 seconds. CONCLUSIONS: Tear film BUT was not influenced by the time of day and moderate to strong correlation with MBI was observed in all four sessions. The software-assisted method proved useful and identified the need to clarify the BUT end-point and to limit the test to 15 seconds to improve observer repeatability.

Interobserver variability of an open-source software for tear meniscus height measurement.

PURPOSE: Different values of the lower tear meniscus height (TMH) can be obtained depending on the method and technique of measurement employed. The aim of this study was to analyse the interobserver variability of a method for measuring TMH by using an open source software. MATERIAL AND METHODS: On a group of 176 subjects, two videos of the central lower tear meniscus, first under slit-lamp illumination and ten minutes later under Tearscope illumination, were generated by a digital camera attached to a slit-lamp. Images were extracted from each video by a masked observer. Two further observers measured in a masked and randomized order the TMH in each illumination method by using an open source software based on Java (NIH ImageJ). TMH was measured from the lower lid to the upper limit of the tear meniscus for both slit-lamp (TMH-SL) and Tearscope (TMH-Tc) illumination methods. Subsequently, in different order, observers assigned a four-grading and a healthy/abnormal subjective classification to each central meniscus. RESULTS: No significant differences were found between the TMH measurements obtained by both investigators in slit-lamp or Tearscope image datasets (t-test; both p≥0.136). When comparing TMH measurements stratified by grade, only interobserver significant differences were observed for grades 3 and 4 with silt-lamp (t-test; both p≤0.009). Significant differences on TMH results between subjective subgroups were observed for both illumination techniques (ANOVA, all p≤0.045). CONCLUSION: This study showed a useful tool to objectively measure TMH by photography.