Effect of ethnic diversity on the saccadic reaction time among healthy Indian and Dutch adults.

Eye movement perimetry (EMP) expresses the decline in visual field (VF) responsiveness based on the deviation in saccadic reaction times (SRTs) from their expected age-similar responses (normative database). Since ethnic dissimilarities tend to affect saccade parameters, we evaluated the effect of such a factor on SRT and its interaction with age, stimulus eccentricity, and intensity. 149 healthy adults, spread into five age groups, drawn from Indian and Dutch ethnicities underwent a customized EMP protocol integrated with a saccade task from which the SRTs to 'seen' visual stimuli were computed. The EMP test had a total of 54 coordinates (five stimulus eccentricities) tested using Goldmann size III visual stimuli presented at four stimulus intensity (SI) levels against a constant background. Considering SRT as a dependent variable, a Generalized Linear Mixed Model analysis was conducted that revealed a statistically significant (p < 0.001) influence of ethnicity and interaction between the tested factors (ethnicity × age × stimulus eccentricity × intensity). However, during the post hoc analysis, out of the 100 possible pair-wise comparisons, only 6% (minor proportion) of the estimates showed statistical significance. Hence, the ethnic-specific differences need not be accounted for while implementing EMP in a diverse set of populations instead a collective database might serve the purpose.

AIROGS: Artificial Intelligence for Robust Glaucoma Screening Challenge.

The early detection of glaucoma is essential in preventing visual impairment. Artificial intelligence (AI) can be used to analyze color fundus photographs (CFPs) in a cost-effective manner, making glaucoma screening more accessible. While AI models for glaucoma screening from CFPs have shown promising results in laboratory settings, their performance decreases significantly in real-world scenarios due to the presence of out-of-distribution and low-quality images. To address this issue, we propose the Artificial Intelligence for Robust Glaucoma Screening (AIROGS) challenge. This challenge includes a large dataset of around 113,000 images from about 60,000 patients and 500 different screening centers, and encourages the development of algorithms that are robust to ungradable and unexpected input data. We evaluated solutions from 14 teams in this paper and found that the best teams performed similarly to a set of 20 expert ophthalmologists and optometrists. The highest-scoring team achieved an area under the receiver operating characteristic curve of 0.99 (95% CI: 0.98-0.99) for detecting ungradable images on-the-fly. Additionally, many of the algorithms showed robust performance when tested on three other publicly available datasets. These results demonstrate the feasibility of robust AI-enabled glaucoma screening.

Frequency of Visual Fields Needed to Detect Glaucoma Progression: A Computer Simulation Using Linear Mixed Effects Model.

PRCIS: Irregular visual field test frequency at relatively short intervals initially and longer intervals later on in the disease provided acceptable results in detecting glaucoma progression. PURPOSE: It is challenging to maintain a balance between the frequency of visual field testing and the long-term costs that may result from insufficient treatment of glaucoma patients. This study aims to simulate real-world circumstances of visual field data to determine the optimum follow-up scheme for the timely detection of glaucoma progression using a linear mixed effects model (LMM). MATERIALS AND METHODS: An LMM with random intercept and slope was used to simulate the series of mean deviation sensitivities over time. A cohort study including 277 glaucoma eyes that were followed for 9.0±1.2 years was used to derive residuals. Data were generated from patients with early-stage glaucoma having various regular and irregular follow-up scenarios and different rates of visual field loss. For each condition, 10,000 series of eyes were simulated, and one confirmatory test was conducted to identify progression. RESULTS: By doing one confirmatory test, the percentage of incorrect progression detection decreased considerably. The time to detect progression was shorter for eyes with an evenly spaced 4-monthly schedule, particularly in the first 2 years. From then onward, results from twice-a-year testing were similar to results from examinations scheduled 3 times per year. CONCLUSIONS: Irregular visual field test frequency at relatively short intervals initially and longer intervals later on in the disease provided acceptable results in detecting glaucoma progression. This approach could be considered for improving glaucoma monitoring. Moreover, simulating data using LMM may provide a better estimate of the disease progression time.

Characteristics of a Large, Labeled Data Set for the Training of Artificial Intelligence for Glaucoma Screening with Fundus Photographs.

Purpose: Significant visual impairment due to glaucoma is largely caused by the disease being detected too late. Objective: To build a labeled data set for training artificial intelligence (AI) algorithms for glaucoma screening by fundus photography, to assess the accuracy of the graders, and to characterize the features of all eyes with referable glaucoma (RG). Design: Cross-sectional study. Subjects: Color fundus photographs (CFPs) of 113 893 eyes of 60 357 individuals were obtained from EyePACS, California, United States, from a population screening program for diabetic retinopathy. Methods: Carefully selected graders (ophthalmologists and optometrists) graded the images. To qualify, they had to pass the European Optic Disc Assessment Trial optic disc assessment with ≥ 85% accuracy and 92% specificity. Of 90 candidates, 30 passed. Each image of the EyePACS set was then scored by varying random pairs of graders as "RG," "no referable glaucoma (NRG)," or "ungradable (UG)." In case of disagreement, a glaucoma specialist made the final grading. Referable glaucoma was scored if visual field damage was expected. In case of RG, graders were instructed to mark up to 10 relevant glaucomatous features. Main Outcome Measures: Qualitative features in eyes with RG. Results: The performance of each grader was monitored; if the sensitivity and specificity dropped below 80% and 95%, respectively (the final grade served as reference), they exited the study and their gradings were redone by other graders. In all, 20 graders qualified; their mean sensitivity and specificity (standard deviation [SD]) were 85.6% (5.7) and 96.1% (2.8), respectively. The 2 graders agreed in 92.45% of the images (Gwet's AC2, expressing the inter-rater reliability, was 0.917). Of all gradings, the sensitivity and specificity (95% confidence interval) were 86.0 (85.2-86.7)% and 96.4 (96.3-96.5)%, respectively. Of all gradable eyes (n = 111 183; 97.62%) the prevalence of RG was 4.38%. The most common features of RG were the appearance of the neuroretinal rim (NRR) inferiorly and superiorly. Conclusions: A large data set of CFPs was put together of sufficient quality to develop AI screening solutions for glaucoma. The most common features of RG were the appearance of the NRR inferiorly and superiorly. Disc hemorrhages were a rare feature of RG. Financial Disclosures: Proprietary or commercial disclosure may be found after the references.

Magnetic resonance imaging reveals possible cause of diplopia after Baerveldt glaucoma implantation.

PURPOSE: To assess if ocular motility impairment, and the ensuing diplopia, after Baerveldt Glaucoma device (BGI) implantation, is related to the presence of a large fluid reservoir (bleb), using Magnetic Resonance Imaging (MRI). METHODS: In a masked observational study (CCMO-registry number: NL65633.058.18), the eyes of 30 glaucoma patients with (n = 12) or without diplopia (n = 18) who had previously undergone BGI implantation were scanned with a 7 Tesla MRI-scanner. The substructures of the BGI-complex, including both blebs and plate, were segmented in 3D. Primary outcomes were a comparison of volume and height of the BGI-complex between patients with and without diplopia. Comparisons were performed by using an unpaired t-test, Fisher's Exact or Mann-Whitney test. Correlations were determined by using Spearman correlation. RESULTS: The median volume and height of the BGI-complex was significantly higher in patients with compared to patients without diplopia (p = 0.007 and p = 0.025, respectively). Six patients had an excessively large total bleb volume (median of 1736.5mm3, interquartile range 1486.3-1933.9mm3), four of whom experienced diplopia (33% of the diplopia patients). Fibrotic strands through the BGI plate, intended to limit the height of the bleb, could be visualized but were not related to diplopia (75% versus 88%; p = 0.28). CONCLUSIONS: With MRI, we show that in a significant number of diplopia cases a large bleb is present in the orbit. Given the large volume of these blebs, they are a likely explanation of the development of diplopia in at least some of the patients with diplopia after BGI implantation. Additionally, the MR-images confirm the presence of fibrotic strands. As these strands are also visible in patients with a large bleb, they are apparently not sufficient to restrict the bleb height.

DenseUNets with feedback non-local attention for the segmentation of specular microscopy images of the corneal endothelium with guttae.

Corneal guttae, which are the abnormal growth of extracellular matrix in the corneal endothelium, are observed in specular images as black droplets that occlude the endothelial cells. To estimate the corneal parameters (endothelial cell density [ECD], coefficient of variation [CV], and hexagonality [HEX]), we propose a new deep learning method that includes a novel attention mechanism (named fNLA), which helps to infer the cell edges in the occluded areas. The approach first derives the cell edges, then infers the well-detected cells, and finally employs a postprocessing method to fix mistakes. This results in a binary segmentation from which the corneal parameters are estimated. We analyzed 1203 images (500 contained guttae) obtained with a Topcon SP-1P microscope. To generate the ground truth, we performed manual segmentation in all images. Several networks were evaluated (UNet, ResUNeXt, DenseUNets, UNet++, etc.) and we found that DenseUNets with fNLA provided the lowest error: a mean absolute error of 23.16 [cells/mm[Formula: see text]] in ECD, 1.28 [%] in CV, and 3.13 [%] in HEX. Compared with Topcon's built-in software, our error was 3-6 times smaller. Overall, our approach handled notably well the cells affected by guttae, detecting cell edges partially occluded by small guttae and discarding large areas covered by extensive guttae.

Pointwise and Region-Wise Course of Visual Field Loss in Patients With Glaucoma.

Purpose: Accurate assessment of visual field (VF) trend may help clinicians devise the optimum treatment regimen. This study was conducted to investigate the behavior of VF sequences using pointwise and region-wise linear, exponential, and sigmoid regression models. Materials and Methods: In a retrospective cohort study, 277 eyes of 139 patients with glaucoma who had been followed for at least 7 years were investigated. Linear, exponential, and sigmoid regression models were fitted for each VF test location and Glaucoma Hemifield Test (GHT) region to model the trend of VF loss. The model with the lowest root mean square error (RMSE) was selected as the best fit. Results: The mean age (standard deviation [SD]) of the patients was 59.9 years (9.8) with a mean follow-up time of 9.3 (0.7) years. The exponential regression had the best fit based on pointwise and region-wise approaches in 39.3% and 38.1% of eyes, respectively. The results showed a better performance based on sigmoid regression in patients with initial VF sensitivity threshold greater than 22 dB (71.6% in pointwise and 62.2% in region-wise approaches). The overall RMSE of the region-wise regression model was lower than the overall RMSE of the pointwise model. Conclusions: In the current study, nonlinear regression models showed a better fit compared to the linear regression models in tracking VF loss behavior. Moreover, findings suggest region-wise analysis may provide a more appropriate approach for assessing VF deterioration. Translational Relevance: Findings may confirm a nonlinear progression of VF deterioration in patients with glaucoma.

Contrast-to-Noise Ratios to Evaluate the Detection of Glaucomatous Progression in the Superior and Inferior Hemiretina.

Purpose: To determine the sensitivity of optical coherence tomography (OCT) and standard automated perimetry (SAP) for detecting glaucomatous progression in the superior and inferior hemiretina. Methods: We calculated contrast-to-noise ratios (CNRs) for OCT retinal nerve fiber layer (RNFL) thickness of hemiretinas and for SAP mean total deviation (MTD) of the corresponding hemifields from longitudinal data (205 eyes, 125 participants). The glaucoma stage for each hemiretina was based on the corresponding hemifield's MTD. Contrast was defined as the difference of the parameter between two consecutive glaucoma stages, whereas noise was the measurement variability of the parameter in those stages. The higher the CNR of a parameter, the more sensitive it is to detecting progression in the transition between successive stages. Results: There were no statistically significant differences for the RNFL CNR and MTD CNR between superior and inferior hemiretinas. As the glaucoma stage of the opposite hemiretina worsened, the MTD CNR in the transition from moderate to advanced glaucoma significantly increased. The RNFL CNR in the transition from mild to moderate glaucoma significantly decreased in case of advanced glaucoma in the opposite hemiretina. Conclusions: Similar to full retinas, detecting conversion to glaucoma in hemiretinas is more sensitive with OCT than SAP, whereas with more advanced disease, SAP is more sensitive for detecting progression. More importantly, the sensitivity for detecting progression in one hemiretina with either technique depends on the glaucoma severity in the opposite hemiretina. Translational Relevance: Monitoring glaucomatous progression with either OCT or SAP partly depends on the glaucoma severity in the opposite hemiretina.

The long-term postoperative effect of the Baerveldt glaucoma drainage device and of a trabeculectomy on the corneal endothelium.

PURPOSE: To determine whether the postoperative corneal endothelial cell density (ECD) differs between glaucoma patients who underwent Baerveldt implant (BGI) surgery and patients who underwent a trabeculectomy (TE) over 5 years ago. METHODS: Cross-sectional, observational study including 34 patients who underwent TE and 36 patients who underwent BGI surgery 5-11 years ago, as part of a randomized clinical trial. None of the patients had a history of intraocular surgery prior to their glaucoma surgery. Central and peripheral ECD was measured by using a non-contact specular microscope. RESULTS: Central and peripheral ECD in the TE group was 2285 ± 371 cells/mm2 (mean ± SD) and 2463 ± 476 cells/mm2 , respectively. Central and peripheral ECD in the BGI group was 1813 ± 745 cells/mm2 and 1876 ± 764 cells/mm2 , respectively. The central and peripheral ECD was statistically significantly higher in the TE group than in the BGI group (p = 0.001 for both). Additional intraocular surgical interventions were more prevalent in the BGI group (23) than in the TE group (5) (p < 0.001). In a subanalysis, without eyes that had undergone additional surgical interventions, only the peripheral ECD was statistically significantly higher in the TE group compared with the BGI group (p = 0.011). For the BGI group, a longer postoperative period resulted in a lower central ECD (r = -0.614, p = 0.004). CONCLUSION: Long-term ECD in eyes that underwent a BGI was considerably lower compared with eyes that underwent a TE, mainly in the peripheral cornea. This suggests that BGI causes a larger decrease of ECD than TE. Additionally, the decrease after BGI appears to continue for a longer period than after TE.

Slowed Saccadic Reaction Times in Seemingly Normal Parts of Glaucomatous Visual Fields.

Purpose: In eye movement perimetry, peripheral stimuli are confirmed by goal-directed eye movements toward the stimulus. The saccadic reaction time (SRT) is regarded as an index of visual field responsiveness, whereas in standard automated perimetry (SAP), the visual field sensitivity is tested. We investigated the relation between visual field sensitivity and responsiveness in corresponding locations of the visual field in healthy controls and in patients with mild, moderate and advanced glaucoma. Materials and Methods: Thirty-four healthy control subjects and 42 glaucoma patients underwent a 54-point protocol in eye movement perimetry (EMP) and a 24-2 SITA standard protocol in a Humphrey Field Analyzer. The visual field points were stratified by total deviation sensitivity loss in SAP into 6 strata. A generalized linear mixed model was applied to determine the influence of the various factors. Results: The generalized linear mixed model showed that the mean SRT increased with increasing glaucoma severity, from 479 ms in the control eyes to 678 ms in the eyes of patients with advanced glaucoma (p < 0.001). Mean SRTs significantly increased with increasing SAP sensitivity loss. Even at the locations where no sensitivity loss was detected by SAP (total deviation values greater or equal than 0 dB), we found lengthened SRTs in mild, moderate and advanced glaucoma compared to healthy controls (p < 0.05) and in moderate and advanced glaucoma compared to mild glaucoma (p < 0.05). At locations with total deviation values between 0 and -3 dB, -3 and -6 dB and -6 and -12 dB, we found similar differences. Conclusions: The lengthened SRT in areas with normal retinal sensitivities in glaucomatous eyes, i.e., planning and execution of saccades to specific locations, precede altered sensory perception as assessed with SAP. Better understanding of altered sensory processing in glaucoma might allow earlier diagnosis of emerging glaucoma.

Ocular motility changes and diplopia in sutured versus unsutured implantation of the Baerveldt glaucoma device.

PURPOSE: To investigate the effect of two surgical techniques in primary Baerveldt glaucoma implant (BGI) surgery, that is the sutured technique and the unsutured (free) plate technique, on the ocular motility and prevalence of diplopia. We hypothesize that the free plate technique results in a lower diplopia prevalence. METHODS: We performed a prospective study of patients who underwent BGI surgery with the free plate technique and compared them with patients from a previous study who had undergone BGI surgery with the sutured technique. Their ductions, ocular alignment and fusion range and the prevalence of diplopia were measured before surgery and at 3 months, 6 months and 1 year postoperatively. RESULTS: We analysed 57 free plate and 51 sutured plate patients. One year postoperatively, we found no statistically significant difference in the prevalence of diplopia between the two techniques. All duction changes between baseline and 1-year follow-up were restrictions and occurred statistically significantly more frequently in the free plate than in the sutured plate group (p = 0.03; 60% versus 34%). About the ocular alignment, in the horizontal direction, a change in exodirection was more common in both groups, while in the vertical direction, a hyperdeviation of the operated eye was more common. The vertical ocular alignment change was smaller in the free plate group than in the sutured plate group (p = 0.04 at near and p = 0.02 at distance). CONCLUSIONS: One year postoperatively, the prevalence of diplopia was not significantly different between patients with the sutured plate and patients with the free plate technique. Both surgical techniques induce diplopia and changes in ocular motility and/or in ocular alignment.

Differences in clinical presentation of primary open-angle glaucoma between African and European populations.

PURPOSE: Primary open-angle glaucoma (POAG) has been reported to occur more frequently in Africans, and to follow a more severe course compared to Europeans. We aimed to describe characteristics of POAG presentation and treatment across three ethnic groups from Africa and one from Europe. METHODS: We ascertained 151 POAG patients from South African Coloured (SAC) and 94 South African Black (SAB) ethnicity from a university hospital in South Africa. In Tanzania, 310 patients were recruited from a university hospital and a referral hospital. In the Netherlands, 241 patients of European ancestry were included. All patients were over 35 years old and had undergone an extensive ophthalmic examination. Patients were diagnosed according to the ISGEO criteria. A biogeographic ancestry analysis was performed to estimate the proportion of genetic African ancestry (GAA). RESULTS: The biogeographic ancestry analysis showed that the median proportion of GAA was 97.6% in Tanzanian, 100% in SAB, 34.2% in SAC and 1.5% in Dutch participants. Clinical characteristics at presentation for Tanzanians, SAB, SAC and Dutch participants, respectively: mean age: 63, 57, 66, 70 years (p < 0.001); visual acuity in the worse eye: 1.78, 1.78, 0.3, 0.3 LogMAR (p < 0.001); maximum intraocular pressure of both eyes: 36, 34, 29, 29 mmHg (panova  < 0.001); maximum vertical cup to disc ratio (VCDR) of both eyes: 0.90, 0.90, 0.84, 0.83 (p < 0.001); mean central corneal thickness: 506, 487, 511, 528 µm (p < 0.001). Fourteen percent of Tanzanian patients presented with blindness (<3/60 Snellen) in the better eye in contrast to only 1% in the Dutch. CONCLUSION: In this multi-ethnic comparative study, Sub-Saharan Africans present at a younger age with lower visual acuity, higher IOP, larger VCDR, than SAC and Dutch participants. This indicates the more progressive and destructive course in Sub-Saharan Africans.

[Artificial intelligence for eye care]. / Artificiële intelligentie in de oogzorg.

Technological developments in ophthalmic imaging and artificial intelligence (AI) create new possibilities for diagnostics in eye care. AI has already been applied in ophthalmic diabetes care. AI-systems currently detect diabetic retinopathy in general practice with a high sensitivity and specificity. AI-systems for the screening, monitoring and treatment of age-related macular degeneration and glaucoma are promising and are still being developed. AI-algorithms, however, only perform tasks for which they have been specifically trained and highly depend on the data and reference-standard that were used to train the system in identifying a certain abnormality or disease. How the data and the gold standard were established and determined, influences the performance of the algorithm. Furthermore, interpretability of deep learning algorithms is still an ongoing issue. By highlighting on images the areas that were critical for the decision of the algorithm, users can gain more insight into how algorithms come to a particular result.

Deep Learning for Assessing the Corneal Endothelium from Specular Microscopy Images up to 1 Year after Ultrathin-DSAEK Surgery.

Purpose: To present a fully automatic method to estimate the corneal endothelium parameters from specular microscopy images and to use it to study a one-year follow-up after ultrathin Descemet stripping automated endothelial keratoplasty. Methods: We analyzed 383 post ultrathin Descemet stripping automated endothelial keratoplasty images from 41 eyes acquired with a Topcon SP-1P specular microscope at 1, 3, 6, and 12 months after surgery. The estimated parameters were endothelial cell density (ECD), coefficient of variation (CV), and hexagonality (HEX). Manual segmentation was performed in all images. Results: Our method provided an estimate for ECD, CV, and HEX in 98.4% of the images, whereas Topcon's software had a success rate of 71.5% for ECD/CV and 30.5% for HEX. For the images with estimates, the percentage error in our method was 2.5% for ECD, 5.7% for CV, and 5.7% for HEX, whereas Topcon's software provided an error of 7.5% for ECD, 17.5% for CV, and 18.3% for HEX. Our method was significantly better than Topcon's (P < 0.0001) and was not statistically significantly different from the manual assessments (P > 0.05). At month 12, the subjects presented an average ECD = 1377 ± 483 [cells/mm2], CV = 26.1 ± 5.7 [%], and HEX = 58.1 ± 7.1 [%]. Conclusions: The proposed method obtains reliable and accurate estimations even in challenging specular images of pathologic corneas. Translational Relevance: CV and HEX, not currently used in the clinic owing to a lack of reliability in automatic methods, are useful biomarkers to analyze the postoperative healing process. Our accurate estimations allow now for their clinical use.

A very short version of the Visual Function Questionnaire (VFQ-3oo7) for use as a routinely applied Patient-Reported Outcome Measure.

BACKGROUND: Patient-reported outcome measures (PROMs) are valuable supplements in regular care to facilitate routine monitoring of quality of life from the patient's perspective. The 25-item National Eye Institute Visual Function Questionnaire (NEI-VFQ-25) is a widely used PROM in ophthalmology. However, the NEI-VFQ-25 is too time-consuming and cumbersome for routine evaluations in regular care. The aim of this study is to construct a 7-item questionnaire of which only 3 items are presented to the patient, by means of routing. This VFQ 3 out of 7 (VFQ-3oo7) should have a minimal loss of information compared with the NEI-VFQ-25. METHODS: An historical database including 3293 administrations of the NEI-VFQ-25 was constructed involving patients with retinal detachment, cataract, corneal diseases, glaucoma, macular degeneration, uveal melanoma and a normal population sample. The data were subjected to Rasch analyses, in particular a generalized partial credit model. Items were sorted on the latent trait and divided into seven categories. From each category, the item with the highest discriminative value was selected. Through routing, only three out of the seven remaining questions are used, where the answers navigate patients to a fitting trait level. RESULTS: A one-dimensional structure was considered fitting. The VFQ-3oo7 showed a small loss of information compared with the total score of the NEI-VFQ-25: correlation 0.927 and a relative precision of 0.868. CONCLUSION: The very short, but valid, VFQ-3oo7 can be applied to evaluate the patient's perceived vision-related health status in routine evaluations of treatments in regular care, with a small burden for patients.

Evaluation of an AI system for the automated detection of glaucoma from stereoscopic optic disc photographs: the European Optic Disc Assessment Study.

OBJECTIVES: To evaluate the performance of a deep learning based Artificial Intelligence (AI) software for detection of glaucoma from stereoscopic optic disc photographs, and to compare this performance to the performance of a large cohort of ophthalmologists and optometrists. METHODS: A retrospective study evaluating the diagnostic performance of an AI software (Pegasus v1.0, Visulytix Ltd., London UK) and comparing it with that of 243 European ophthalmologists and 208 British optometrists, as determined in previous studies, for the detection of glaucomatous optic neuropathy from 94 scanned stereoscopic photographic slides scanned into digital format. RESULTS: Pegasus was able to detect glaucomatous optic neuropathy with an accuracy of 83.4% (95% CI: 77.5-89.2). This is comparable to an average ophthalmologist accuracy of 80.5% (95% CI: 67.2-93.8) and average optometrist accuracy of 80% (95% CI: 67-88) on the same images. In addition, the AI system had an intra-observer agreement (Cohen's Kappa, κ) of 0.74 (95% CI: 0.63-0.85), compared with 0.70 (range: -0.13-1.00; 95% CI: 0.67-0.73) and 0.71 (range: 0.08-1.00) for ophthalmologists and optometrists, respectively. There was no statistically significant difference between the performance of the deep learning system and ophthalmologists or optometrists. CONCLUSION: The AI system obtained a diagnostic performance and repeatability comparable to that of the ophthalmologists and optometrists. We conclude that deep learning based AI systems, such as Pegasus, demonstrate significant promise in the assisted detection of glaucomatous optic neuropathy.

Contrast-to-Noise Ratios for Assessing the Detection of Progression in the Various Stages of Glaucoma.

PURPOSE: We determine the contrast-to-noise ratios (CNRs) of structural and functional measurements to assess their sensitivity to detect progression in the various stages of glaucoma. METHODS: We calculated the CNRs for the mean peripapillary retinal nerve fiber layer (RNFL) thickness measured by spectral domain optical coherence tomography, and the mean deviation (MD) and visual field index (VFI) determined by standard automated perimetry for the transitions between five stages. Longitudinal data from healthy and glaucomatous eyes from a prospective study were used. Contrast was defined as the change in the mean value of the parameter between two successive stages. Noise was defined as the variability of the parameter and calculated from the residuals of linear regression on the data from five subsequent visits per eye. RESULTS: We studied 205 eyes from 125 participants (46% men, 54% women). CNRs for different parameters varied considerably across the range of disease severity (0.8-12.2). The RNFL thickness had a higher CNR in the transition from normal to mild glaucoma (12.2) compared to the CNRs of the functional measures (MD 4.1, VFI 4.5). The CNRs for the functional measures were higher in the transition from moderate to advanced (MD 5.2, VFI 5.8) and advanced to severe glaucoma (MD 7.2, VFI 5.8) compared to the RNFL thickness (CNR 0.8 and 3.2, respectively). CONCLUSIONS: The RNFL thickness is more sensitive for detecting glaucomatous progression at the onset of glaucoma compared to the functional measures, while the latter are more sensitive for detecting progression in the later stages of glaucoma. TRANSLATIONAL RELEVANCE: The CNR method can be used to determine which measurement is most sensitive for detecting progression in glaucoma, differentiated for the severity of the disease. Furthermore, it creates a basic toolset for determining the most sensitive measurement in detecting progression not only in glaucoma, but other (ophthalmic) diseases as well.

Fully convolutional architecture vs sliding-window CNN for corneal endothelium cell segmentation.

BACKGROUND: Corneal endothelium (CE) images provide valuable clinical information regarding the health state of the cornea. Computation of the clinical morphometric parameters requires the segmentation of endothelial cell images. Current techniques to image the endothelium in vivo deliver low quality images, which makes automatic segmentation a complicated task. Here, we present two convolutional neural networks (CNN) to segment CE images: a global fully convolutional approach based on U-net, and a local sliding-window network (SW-net). We propose to use probabilistic labels instead of binary, we evaluate a preprocessing method to enhance the contrast of images, and we introduce a postprocessing method based on Fourier analysis and watershed to convert the CNN output images into the final cell segmentation. Both methods are applied to 50 images acquired with an SP-1P Topcon specular microscope. Estimates are compared against a manual delineation made by a trained observer. RESULTS: U-net (AUC=0.9938) yields slightly sharper, clearer images than SW-net (AUC=0.9921). After postprocessing, U-net obtains a DICE=0.981 and a MHD=0.22 (modified Hausdorff distance), whereas SW-net yields a DICE=0.978 and a MHD=0.30. U-net generates a wrong cell segmentation in only 0.48% of the cells, versus 0.92% for the SW-net. U-net achieves statistically significant better precision and accuracy than both, Topcon and SW-net, for the estimates of three clinical parameters: cell density (ECD), polymegethism (CV), and pleomorphism (HEX). The mean relative error in U-net for the parameters is 0.4% in ECD, 2.8% in CV, and 1.3% in HEX. The computation time to segment an image and estimate the parameters is barely a few seconds. CONCLUSIONS: Both methods presented here provide a statistically significant improvement over the state of the art. U-net has reached the smallest error rate. We suggest a segmentation refinement based on our previous work to further improve the performance.

Convolutional neural network-based regression for biomarker estimation in corneal endothelium microscopy images.

The morphometric parameters of the corneal endothelium - cell density (ECD), cell size variation (CV), and hexagonality (HEX) - provide clinically relevant information about the cornea. To estimate these parameters, the endothelium is commonly imaged with a non-contact specular microscope and cell segmentation is performed to these images. In previous work, we have developed several methods that, combined, can perform an automated estimation of the parameters: the inference of the cell edges, the detection of the region of interest (ROI), a post-processing method that combines both images (edges and ROI), and a refinement method that removes false edges. In this work, we first explore the possibility of using a CNN-based regressor to directly infer the parameters from the edge images, simplifying the framework. We use a dataset of 738 images coming from a study related to the implantation of a Baerveldt glaucoma device and a standard clinical care regarding DSAEK corneal transplantation, both from the Rotterdam Eye Hospital and both containing images of unhealthy endotheliums. This large dataset allows us to build a large training set that makes this approach feasible. We achieved a mean absolute percentage error (MAPE) of 4.32% for ECD, 7.07% for CV, and 11.74% for HEX. These results, while promising, do not outperform our previous work. In a second experiment, we explore the use of the CNN-based regressor to improve the post-processing method of our previous approach in order to adapt it to the specifics of each image. Our results showed no clear benefit and proved that our previous post-processing is already highly reliable and robust.

Genome-wide association study of primary open-angle glaucoma in continental and admixed African populations.

Primary open angle glaucoma (POAG) is a complex disease with a major genetic contribution. Its prevalence varies greatly among ethnic groups, and is up to five times more frequent in black African populations compared to Europeans. So far, worldwide efforts to elucidate the genetic complexity of POAG in African populations has been limited. We conducted a genome-wide association study in 1113 POAG cases and 1826 controls from Tanzanian, South African and African American study samples. Apart from confirming evidence of association at TXNRD2 (rs16984299; OR[T] 1.20; P = 0.003), we found that a genetic risk score combining the effects of the 15 previously reported POAG loci was significantly associated with POAG in our samples (OR 1.56; 95% CI 1.26-1.93; P = 4.79 × 10-5). By genome-wide association testing we identified a novel candidate locus, rs141186647, harboring EXOC4 (OR[A] 0.48; P = 3.75 × 10-8), a gene transcribing a component of the exocyst complex involved in vesicle transport. The low frequency and high degree of genetic heterogeneity at this region hampered validation of this finding in predominantly West-African replication sets. Our results suggest that established genetic risk factors play a role in African POAG, however, they do not explain the higher disease load. The high heterogeneity within Africans remains a challenge to identify the genetic commonalities for POAG in this ethnicity, and demands studies of extremely large size.