IMPORTANCE A deep learning system (DLS) that could automatically detect glaucomatous optic neuropathy (GON) with high sensitivity and specificity could expedite screening for GON.OBJECTIVE To establish a DLS for detection of GON using retinal fundus images and glaucoma diagnosis with convoluted neural networks (GD-CNN) that has the ability to be generalized across populations. DESIGN, SETTING, AND PARTICIPANTSIn this cross-sectional study, a DLS for the classification of GON was developed for automated classification of GON using retinal fundus images obtained from the Chinese Glaucoma Study Alliance, the Handan Eye Study, and online databases. The researchers selected 241 032 images were selected as the training data set. The images were entered into the databases on June 9, 2009, obtained on July 11, 2018, and analyses were performed on December 15, 2018. The generalization of the DLS was tested in several validation data sets, which allowed assessment of the DLS in a clinical setting without exclusions, testing against variable image quality based on fundus photographs obtained from websites, evaluation in a population-based study that reflects a natural distribution of patients with glaucoma within the cohort and an additive data set that has a diverse ethnic distribution. An online learning system was established to transfer the trained and validated DLS to generalize the results with fundus images from new sources. To better understand the DLS decision-making process, a prediction visualization test was performed that identified regions of the fundus images utilized by the DLS for diagnosis. EXPOSURES Use of a deep learning system. MAIN OUTCOMES AND MEASURES Area under the receiver operating characteristics curve (AUC), sensitivity and specificity for DLS with reference to professional graders. RESULTS From a total of 274 413 fundus images initially obtained from CGSA, 269 601 images passed initial image quality review and were graded for GON. A total of 241 032 images (definite GON 29 865 [12.4%], probable GON 11 046 [4.6%], unlikely GON 200 121 [83%]) from 68 013 patients were selected using random sampling to train the GD-CNN model. Validation and evaluation of the GD-CNN model was assessed using the remaining 28 569 images from CGSA. The AUC of the GD-CNN model in primary local validation data sets was 0.996 (95% CI, 0.995-0.998), with sensitivity of 96.2% and specificity of 97.7%. The most common reason for both false-negative and false-positive grading by GD-CNN (51 of 119 [46.3%] and 191 of 588 [32.3%]) and manual grading (50 of 113 [44.2%] and 183 of 538 [34.0%]) was pathologic or high myopia.CONCLUSIONS AND RELEVANCE Application of GD-CNN to fundus images from different settings and varying image quality demonstrated a high sensitivity, specificity, and generalizability for detecting GON. These findings suggest that automated DLS could enhance current screening programs in a cost-effective and time-efficient manner.

Exfoliation syndrome (XFS) is the commonest known risk factor for secondary glaucoma and a significant cause of blindness worldwide. Variants in two genes, LOXL1 and CACNA1A have been previously associated with XFS. To further elucidate the genetic basis of XFS, we collected a global sample of XFS cases to refine the association at LOXL1, which previously showed inconsistent results between populations, and to identify new variants associated with XFS. We identified a rare, protective allele at LOXL1 (p.407Phe, OR = 25, P =2.9 × 10−14) through deep resequencing of XFS cases and controls from 9 countries. This variant results in increased cellular adhesion strength compared to the wild-type (p.407Tyr) allele. A genome-wide association study (GWAS) of XFS cases and controls from 24 countries followed by replication in 18 countries identified seven genome-wide significant loci (P < 5 × 10−8). Index variants at the new loci map to chromosomes 13q12 (POMP), 11q23.3 (TMEM136), 6p21 (AGPAT1), 3p24 (RBMS3) and 5q23 (near SEMA6A). These findings provide biological insights into the pathology of XFS, and highlight a potential role for naturally occurring rare LOXL1 variants in disease biology.

Purpose: To determine if optical coherence tomography angiography (OCTA)-derived vessel density measurements can extend the available dynamic range for detecting glaucoma compared to spectral-domain optical coherence tomography (SDOCT)-derived thickness measurements.Design: Observational, cross-sectional study. Participants: A total of 509 eyes from 38 healthy participants, 63 glaucoma suspects and 193 glaucoma patients enrolled in the Diagnostic Innovations in Glaucoma Study. Methods: Relative vessel density and tissue thickness measurement floors of perifoveal superficial vessel density (pfVD), circumpapillary capillary density (cpCD), circumpapillary retinal nerve fiber (cpRNFL) thickness, ganglion cell complex (GCC) thickness and visual field mean deviation were investigated and compared with a previously reported linear change point model (CPM) and locally weighted scatterplot smoothing (LOWESS) curves. Main Outcome Measures: Estimated vessel density and tissue thickness measurement floors and corresponding dynamic ranges. Results: Visual field MD ranged from −30.1 dB to 2.8 dB. No measurement floor was found for pfVD which continued to decrease constantly until very advanced disease. A true floor (i.e. slope ~ 0 after observed CPM change point) was detected for cpRNFL thickness only. Post-CPM estimated floors were 49.5±2.6 μm for cpRNFL thickness, 70.7±1.0 μm for GCC thickness and 31.2± 1.1% for cpCD. pfVD reached the post-CPM estimated floor later in the disease (VF MD: −25.8±3.8 dB) than cpCD (VF MD: −19.3±2.4 dB), cpRNFL thickness (VF MD: −17±3.3 dB) and GCC thickness (VF MD: −13.9±1.8 dB) (p<0.

No abstract

Background: We describe corneal endothelial cell density and morphology in normal Iranian eyes and compare endothelial cell characteristics in the Iranian population with data available in the literature for American and Indian populations.

Optical coherence tomography angiography (OCTA) is a relatively new, noninvasive, dye-free imaging modality that provides a qualitative and quantitative assessment of the vasculature in the retina and optic nerve head. OCTA also enables visualization of the choriocapillaris, but only in areas of parapapillary atrophy. With OCTA, the movement of red blood cells is used as a contrast to delineate blood vessels from static tissues. The features seen with OCTA in eyes with glaucoma are reduction in the superficial vessel density in the peripapillary and macular areas, and complete loss of choriocapillaris in localized regions of parapapillary atrophy (called deep-layer microvascular dropout). These OCTA changes correlate well topographically with the functional changes seen on visual field examination and structural changes seen on optical coherence tomography (OCT) (ie, parapapillary retinal nerve fiber layer changes and inner retinal layer thickness changes at macula). The OCTA measurements also have acceptable test-retest variability and well differentiate glaucomatous from normal eyes. OCTA measurements can be affected by various subject-related, eye-related, and disease-related factors. Vessel density reduction on OCTA reaches a base level (floor) at a more advanced disease stage than the structural changes on OCT and therefore has the potential to monitor progression in eyes with advanced glaucomatous damage. OCTA also adds information about glaucoma patients at risk of faster progression. OCTA, therefore, complements visual field and OCT examinations to diagnose glaucoma, detect progression, and assess risk of progression.

Purpose: To characterize and compare the ganglion cell complex (GCC) thickness and macula vessel density in pre-perimetric and early primary open angle glaucoma (POAG) eyes. Design: Cross-sectional study. Methods: 57 healthy eyes, 68 pre-perimetric and 162 early POAG eyes enrolled in the Diagnostic Innovations in Glaucoma Study. Optical coherence tomography angiography (OCT-A) based superficial macula vessel density and OCT based GCC thickness were evaluated simultaneously. Percent loss from normal of GCC thickness and macula vessel density was compared. Area under the receiver operating characteristic curves was used to describe the diagnostic utility. Results: Both GCC thickness and vessel density were significantly lower in pre-perimetric and early POAG eyes compared to healthy eyes. Compared to the pre-perimetric POAG group, the early POAG group showed larger GCC thickness percent loss (whole image 4.72% vs. 9.86%; all P<0.01) but similar vessel density percent loss (whole image 4.97% vs. 6.93%; all P>0.05). In preperimetric POAG, GCC thickness and vessel density percent losses were similar (all P>0.1). In contrast, in early POAG, GCC thickness percent loss was larger than that of vessel density (all P≤ 0.001). To discriminate pre-perimetric or early glaucoma eyes from healthy eyes, GCC thickness and macula vessel density showed similar diagnostic accuracy (all P> 0.05).

Loss of OCT-A macula vessel density is associated with central 10-2 VF defects. Macula vessel density is a clinically relevant parameter that may enhance monitoring of glaucoma suspects and patients.