Fast Progressors in Glaucoma: Prevalence Based on Global and Central Visual Field Loss.

PURPOSE: To determine the prevalence of fast global and central visual field (VF) progression in individuals with glaucoma under routine care. DESIGN: Observational study. PARTICIPANTS: Six hundred ninety-three eyes of 461 individuals with glaucoma followed up over a median of 4.5 years. METHODS: This study included (1) patients at a private ophthalmology clinic in Melbourne, Australia, and (2) individuals in 2 prospective longitudinal observational studies across 3 sites in the United States. All individuals had a diagnosis of glaucoma and were under routine care, and had performed 5 or more reliable 24-2 VF tests over a 1- to 5-year period. Ordinary least squares regression analyses were used to calculate the rate of global mean deviation (MD) change over time and the rate of the mean total deviation values of the 12 test locations within the central 10° region (MTD10) for each eye. MAIN OUTCOME MEASURES: Prevalence of progression based on the rate of MD and the MTD10 change across various fixed cutoffs and cutoffs based on the estimated normal distribution (from the positive slopes). RESULTS: Based on the MD and the MTD10, 12.5% and 11.7% of the eyes, respectively, exhibited a rate of change that was less than -1.0 dB/year (being a rate that typically is defined as "fast progression" for MD values), and 29.0% of the eyes showed a change of less than -0.5 dB/year on MTD10. Furthermore, 12.7% and 9.1% of the eyes exhibited a rate of change that exceeded the 1% cutoff of the estimated normal distribution MD and the MTD10 values, respectively. CONCLUSIONS: This study found that approximately 1 in 8 eyes with glaucoma receiving routine care showed fast progression based on global MD values (< -1.0 dB/year) and that nearly 1 in 3 eyes showed a < -0.5 dB/year decline centrally. These findings highlight the clinical importance of assessing progressive central VF loss and reinforce the need for new therapies to prevent functional disability in a notable proportion of individuals who continue to exhibit fast progression. FINANCIAL DISCLOSURE(S): Proprietary or commercial disclosure may be found after the references.

Retinal electrophysiologic response to IOP elevation in living human eyes.

PURPOSE: The relationships between intraocular pressure (IOP), ocular perfusion pressure (OPP), retinal perfusion, and retinal electrophysiologic responses have been explored experimentally across several animal models. These studies have demonstrated that elevated IOP reduces OPP, and when this reduction in OPP exceeds the autoregulatory capacity of the retina vasculature, retinal perfusion and electrophysiologic responses are reduced. This study aimed to evaluate these interactions for the first time in the living human eye. METHODS: Five eyes from three research-consented brain-dead organ donors underwent optical coherence tomography with angiographic (OCT/A; Spectralis, Heidelberg Engineering) and electroretinographic (ERG, Diagnosys LLC) measurements while IOP was manometrically-elevated stepwise to pressures of 10, 30 and 50 mmHg. Systemic blood pressure (BP) was monitored continuously during testing. Correlation analysis was applied to assess association between ERG and OPP changes. In a single eye, prolonged IOP elevation was induced with viscoelastic injection and serial ERG measurements were obtained. RESULTS: Reductions in inner retinal function defined by photopic ERG were observed with elevation in IOP and concomitant reduction in OPP. Reductions, especially in b-wave, and photopic negative response (PhNR) amplitudes and implicit times were significantly correlated with elevation in IOP and reduction in OPP. There were more appreciable changes in perfusion and functional responses in eyes tested while systemic blood pressure was lower. With prolonged IOP elevation, selective loss of the PhNR response was observed. CONCLUSIONS: In the living human eye, retinal perfusion and inner retinal function are acutely impacted by elevation of IOP, and this impact is related to systemic BP and OPP. This novel approach provides a viable model to study the autoregulatory responses to IOP elevation in the living human eye.

Impact of Smoking on Visual Field Progression in a Long-term Clinical Follow-up.

PURPOSE: To investigate the effect of smoking on rates of progressive visual field (VF) damage over time in glaucoma. DESIGN: Retrospective cohort study. PARTICIPANTS: Five hundred eleven eyes of 354 patients with glaucoma followed up from multicenter glaucoma registries. METHODS: In this longitudinal study, 354 patients with primary open-angle glaucoma with a minimum of 3 years of follow-up and 5 VF tests were enrolled from the Diagnostic Innovations in Glaucoma Study and the African Descent and Glaucoma Evaluation Study. Univariate and multivariate linear mixed models were used to investigate the effects of smoking on rates of 24-2 VF mean deviation loss. Visual field progression was defined using pointwise linear and significant negative VF mean deviation loss. Logistic regression was used to identify baseline factors and whether different levels of smoking intensity were associated with VF progression. Kaplan-Meier survival analysis and the log-rank test were used to compare the cumulative risk ratio of progression between smoker and never smoker groups. MAIN OUTCOME MEASURES: Visual field progression. RESULTS: Five hundred eleven eyes of 354 patients were included over the median follow-up of 12.5 years. Median baseline age was 64.8 years. Of the 354 patients, 124 (35%) were Black, and 149 (42.1%) and 168 (59.8%) had reported a history of smoking or alcohol consumption, respectively. In a multivariate model, higher smoking intensity was associated with faster VF loss (coefficient, -0.05 decibels (dB)/year per 10 pack-years; 95% confidence interval [CI], -0.08 to -0.01 dB/year per 10 pack-years; P = 0.010). Developing VF progression in eyes of heavy smokers (≥ 20 pack-years) was 2.2 times more than in eyes of patients without smoking history (odds ratio, 2.21; 95% CI, 1.02-4.76; P = 0.044). Statistically significant differences were found between heavy smokers (≥ 20 pack-years) and never smokers by Kaplan-Meier analysis (P = 0.011, log-rank test). CONCLUSIONS: Heavy smokers are more likely to sustain VF loss in eyes with glaucoma. The prospective longitudinal design of this study supports the hypothesis that levels of smoking may be a significant predictor for glaucoma progression. Additionally, this information can be used for clinically relevant tobacco prevention and intervention messages.

Deep Learning Estimation of 10-2 and 24-2 Visual Field Metrics Based on Thickness Maps from Macula OCT.

PURPOSE: To develop deep learning (DL) systems estimating visual function from macula-centered spectral-domain (SD) OCT images. DESIGN: Evaluation of a diagnostic technology. PARTICIPANTS: A total of 2408 10-2 visual field (VF) SD OCT pairs and 2999 24-2 VF SD OCT pairs collected from 645 healthy and glaucoma subjects (1222 eyes). METHODS: Deep learning models were trained on thickness maps from Spectralis macula SD OCT to estimate 10-2 and 24-2 VF mean deviation (MD) and pattern standard deviation (PSD). Individual and combined DL models were trained using thickness data from 6 layers (retinal nerve fiber layer [RNFL], ganglion cell layer [GCL], inner plexiform layer [IPL], ganglion cell-IPL [GCIPL], ganglion cell complex [GCC] and retina). Linear regression of mean layer thicknesses were used for comparison. MAIN OUTCOME MEASURES: Deep learning models were evaluated using R2 and mean absolute error (MAE) compared with 10-2 and 24-2 VF measurements. RESULTS: Combined DL models estimating 10-2 achieved R2 of 0.82 (95% confidence interval [CI], 0.68-0.89) for MD and 0.69 (95% CI, 0.55-0.81) for PSD and MAEs of 1.9 dB (95% CI, 1.6-2.4 dB) for MD and 1.5 dB (95% CI, 1.2-1.9 dB) for PSD. This was significantly better than mean thickness estimates for 10-2 MD (0.61 [95% CI, 0.47-0.71] and 3.0 dB [95% CI, 2.5-3.5 dB]) and 10-2 PSD (0.46 [95% CI, 0.31-0.60] and 2.3 dB [95% CI, 1.8-2.7 dB]). Combined DL models estimating 24-2 achieved R2 of 0.79 (95% CI, 0.72-0.84) for MD and 0.68 (95% CI, 0.53-0.79) for PSD and MAEs of 2.1 dB (95% CI, 1.8-2.5 dB) for MD and 1.5 dB (95% CI, 1.3-1.9 dB) for PSD. This was significantly better than mean thickness estimates for 24-2 MD (0.41 [95% CI, 0.26-0.57] and 3.4 dB [95% CI, 2.7-4.5 dB]) and 24-2 PSD (0.38 [95% CI, 0.20-0.57] and 2.4 dB [95% CI, 2.0-2.8 dB]). The GCIPL (R2 = 0.79) and GCC (R2 = 0.75) had the highest performance estimating 10-2 and 24-2 MD, respectively. CONCLUSIONS: Deep learning models improved estimates of functional loss from SD OCT imaging. Accurate estimates can help clinicians to individualize VF testing to patients.

Strain by virtual extensometers and video-imaging optical coherence tomography as a repeatable metric for IOP-Induced optic nerve head deformations.

PURPOSE: To determine if in vivo strain response of the Optic Nerve Head (ONH) to IOP elevation visualized using Optical Coherence Tomography (OCT) video imaging and quantified using novel virtual extensometers was able to be provided repeatable measurements of tissue specific deformations. METHODS: The ONHs of 5 eyes from 5 non-human primates (NHPs) were imaged by Spectralis OCT. A vertical and a horizontal B-scan of the ONH were continuously recorded for 60 s at 6 Hz (video imaging mode) during IOP elevation from 10 to 30 mmHg. Imaging was repeated over three imaging sessions. The 2D normal strain was computed by template-matching digital image correlation using virtual extensometers. ANOVA F-test (F) was used to compare inter-eye, inter-session, and inter-tissue variability for the prelaminar, Bruch's membrane opening (BMO), lamina cribrosa (LC) and choroidal regions (against variance the error term). F-test of the ratio between inter-eye to inter-session variability was used to test for strain repeatability across imaging sessions (FIS). RESULTS: Variability of strain across imaging session (F = 0.7263, p = 0.4855) and scan orientation was not significant (F = 1.053, p = 0.3066). Inter session variability of strain was significantly lower than inter-eye variability (FIS = 22.63, p = 0.0428) and inter-tissue variability (FIS = 99.33 p = 0.00998). After IOP elevation, strain was highest in the choroid (-18.11%, p < 0.001), followed by prelaminar tissue (-11.0%, p < 0.001), LC (-3.79%, p < 0.001), and relative change in BMO diameter (-0.57%, p = 0.704). CONCLUSIONS: Virtual extensometers applied to video-OCT were sensitive to the eye-specific and tissue-specific mechanical response of the ONH to IOP and were repeatable across imaging sessions.

Histologic validation of optical coherence tomography-based three-dimensional morphometric measurements of the human optic nerve head: Methodology and preliminary results.

PURPOSE: To compare the three-dimensional (3D) morphology of the deep load-bearing structures of the human optic nerve head (ONH) as revealed in vivo by spectral domain optical coherence tomography (SDOCT) with ex vivo quantitative 3D histology. METHODS: SDOCT imaging of the ONH was performed in six eyes from three brain-dead organ donors on life-support equipment awaiting organ procurement (in vivo conditions). Following organ procurement (ex vivo conditions), the eyes were enucleated and underwent a pars plana vitrectomy followed by pressurization to physiologic IOP and immersion fixation. Ex vivo ONH morphology was obtained from high-fidelity episcopic fluorescent 3D reconstruction. Morphologic parameters of the observed ONH canal geometry and peripapillary choroid, as well as the shape, visibility and depth of the lamina cribrosa were compared between ex vivo and in vivo measurements using custom software to align, scale, and manually delineate the different regions of the ONH. RESULTS: There was significant correspondence between in vivo and ex vivo measurements of the depth and shape of the lamina cribrosa, along with the size and shape of Bruch's membrane opening (BMO) and anterior scleral canal opening (ASCO). Weaker correspondence was observed for choroidal thickness; as expected, a thinner choroid was seen ex vivo due to loss of blood volume upon enucleation (-79.9%, p < 0.001). In addition, the lamina was shallower (-32.3%, p = 0.0019) and BMO was smaller ex vivo (-3.38%, p = 0.026), suggesting post mortem shrinkage of the fixed tissue. On average, while highly variable, only 31% of the anterior laminar surface was visible in vivo with SDOCT (p < 0.001). CONCLUSIONS: Morphologic parameters by SDOCT imaging of the deep ONH showed promising correspondence to histology metrics. Small but significant shrinkage artifact, along with large effects of exsanguination of the choroid, was seen in the ex vivo reconstructions of fixed tissues that may impact the quantification of ex vivo histoarchitecture, and this should be considered when developing models and biomarkers based on ex vivo imaging of fixed tissue. Lack of visibly of most of the lamina surface in SDOCT images is an important limitation to metrics and biomarkers based on in vivo images of the ONH deep tissues.

Deep Learning Approaches Predict Glaucomatous Visual Field Damage from OCT Optic Nerve Head En Face Images and Retinal Nerve Fiber Layer Thickness Maps.

PURPOSE: To develop and evaluate a deep learning system for differentiating between eyes with and without glaucomatous visual field damage (GVFD) and predicting the severity of GFVD from spectral domain OCT (SD OCT) optic nerve head images. DESIGN: Evaluation of a diagnostic technology. PARTICIPANTS: A total of 9765 visual field (VF) SD OCT pairs collected from 1194 participants with and without GVFD (1909 eyes). METHODS: Deep learning models were trained to use SD OCT retinal nerve fiber layer (RNFL) thickness maps, RNFL en face images, and confocal scanning laser ophthalmoscopy (CSLO) images to identify eyes with GVFD and predict quantitative VF mean deviation (MD), pattern standard deviation (PSD), and mean VF sectoral pattern deviation (PD) from SD OCT data. MAIN OUTCOME MEASURES: Deep learning models were compared with mean RNFL thickness for identifying GVFD using area under the curve (AUC), sensitivity, and specificity. For predicting MD, PSD, and mean sectoral PD, models were evaluated using R2 and mean absolute error (MAE). RESULTS: In the independent test dataset, the deep learning models based on RNFL en face images achieved an AUC of 0.88 for identifying eyes with GVFD and 0.82 for detecting mild GVFD significantly (P < 0.001) better than using mean RNFL thickness measurements (AUC = 0.82 and 0.73, respectively). Deep learning models outperformed standard RNFL thickness measurements in predicting all quantitative VF metrics. In predicting MD, deep learning models based on RNFL en face images achieved an R2 of 0.70 and MAE of 2.5 decibels (dB) compared with 0.45 and 3.7 dB for RNFL thickness measurements. In predicting mean VF sectoral PD, deep learning models achieved high accuracy in the inferior nasal (R2 = 0.60) and superior nasal (R2 = 0.67) sectors, moderate accuracy in inferior (R2 = 0.26) and superior (R2 = 0.35) sectors, and lower accuracy in the central (R2 = 0.15) and temporal (R2 = 0.12) sectors. CONCLUSIONS: Deep learning models had high accuracy in identifying eyes with GFVD and predicting the severity of functional loss from SD OCT images. Accurately predicting the severity of GFVD from SD OCT imaging can help clinicians more effectively individualize the frequency of VF testing to the individual patient.

Experimental myopia increases and scleral crosslinking using genipin inhibits cyclic softening in the tree shrew sclera.

PURPOSE: Myopia progression is thought to involve biomechanical weakening of the sclera, which leads to irreversible deformations and axial elongation of the eye. Scleral crosslinking has been proposed as a potential treatment option for myopia control by strengthening the mechanically weakened sclera. The biomechanical mechanism by which the sclera weakens during myopia and strengthens after crosslinking is not fully understood. Here, we assess the effect of lens-induced myopia and exogenous crosslinking using genipin on the inelastic mechanical properties of the tree shrew sclera measured by cyclic tensile tests. METHODS: Cyclic tensile tests were performed on 2-mm wide scleral strips at physiological loading conditions (50 cycles, 0-3.3 g, 30 s cycle-1 ). Two scleral strips were obtained from each eye of juvenile tree shrews exposed to two different visual conditions: normal and 4 days of monocular -5 D lens wear to accelerate scleral remodelling and induce myopia. Scleral strips were mechanically tested at three alternative conditions: immediately after enucleation; after incubation in phosphate buffered saline (PBS) for 24 h at 37°C; and after incubation for 24 h in PBS supplemented with genipin at a low cytotoxicity concentration (0.25 mm). Cyclic softening was defined as the incremental strain increase from one cycle to the next. RESULTS: -5D lens treatment significantly increased the cyclic softening response of the sclera when compared to contralateral control eyes (0.10% ± 0.029%, mean ± standard error, P = 0.037). Exogenous crosslinking of the lens treated sclera significantly decreased the cyclic softening response (-0.12% ± 0.014%, P = 2.2 × 10-5 ). Contrary to all other groups, the genipin-cross-linked tissue did not exhibit cyclic softening significantly different from zero within the 50-cycle test. CONCLUSIONS: Results indicated that cyclic tensile loading leads to an inelastic, cyclic softening of the juvenile tree shrew sclera. The softening rate increased during lens-induced myopia and was diminished after genipin crosslinking. This finding suggests that axial elongation in myopia may involve a biomechanical weakening mechanism that increased the cyclic softening response of the sclera, which was inhibited by scleral crosslinking using genipin.

Diagnostic Accuracy of Optic Nerve Head and Macula OCT Parameters for Detecting Glaucoma in Eyes With and Without High Axial Myopia.

PURPOSE: To characterize structural differences and assess the diagnostic accuracy of optic nerve head (ONH) and macula optical coherence tomography (OCT) parameters to detect glaucoma in eyes with and without high axial myopia. DESIGN: Cross-sectional study. METHODS: Three hundred sixty-eight glaucoma and 411 healthy eyes with no axial myopia, 393 glaucoma and 271 healthy eyes with mild axial myopia and 124 glaucoma and 85 healthy eyes with high axial myopia were included. Global and sectoral peripapillary retinal nerve fiber layer thickness (pRNFLT), Bruch's membrane opening minimum rim width (BMO-MRW), ganglion cell inner plexiform layer thickness (GCIPLT), and macula RNFLT (mRNFLT) were compared and the diagnostic accuracy for glaucoma detection was evaluated using the adjusted area under the receiver operating characteristic curve (AUC). RESULTS: Diagnostic accuracy for ONH and macula parameters to detect glaucoma was generally high and differed by myopia group. For ONH parameters the diagnostic accuracy was highest for global (AUC = 0.95) and inferotemporal (AUC = 0.91) pRNFLT for high myopes and global BMO-MRW for nonmyopes (AUC = 1.0) and mild myopes (AUC = 0.97). For macula parameters, the diagnostic accuracy was higher in high myopes with 6 of the 11 GCIPLT global/sectors having adjusted AUCs > 0.90 compared to nonhigh myopes with no AUCs > 0.90. In all myopia groups, mRNFLT had lower AUCs than GCIPLT. CONCLUSIONS: The diagnostic accuracy for pRNFL and GCIPL was high for high axial myopic eyes and shows promise for glaucoma detection in high myopes. Further analysis is needed to determine whether the high diagnostic accuracy can be confirmed in other populations.

Association of Long-Term Intraocular Pressure Variability and Rate of Ganglion Complex Thinning in Patients With Glaucoma.

PURPOSE: To evaluate the association of mean intraocular pressure (IOP) and IOP variability (IOP fluctuation [SD of IOP] and the IOP range) with the rate of ganglion cell complex (GCC) layer thinning over time in patients with glaucoma. DESIGN: Prospective cohort study. METHODS: Participants with at least 4 visits and 2 years of follow-up of optical coherence tomography tests were included. A linear mixed-effect model was used to investigate the association of IOP parameters with the rates of GCC thinning. Subgroup analyses were conducted for eyes with early (MD ≥ -6 dB), and moderate to advanced stage (MD < -6 dB) at baseline. RESULTS: The cohort consisted of 369 eyes of 249 glaucoma patients (282 early glaucoma and 87 moderate to advanced glaucoma) with mean (standard deviation [SD]) age of 68.2 (10.7) years over 5.1 years of follow-up. The mean rate of GCC change was -0.59 (95% confidence interval [CI], -0.67 to -0.52) µm per year. In multivariable models, faster annual rate of GCC thinning was associated with a higher IOP fluctuation (-0.17 [95% CI, -0.23 to -0.11] µm per 1-mmHg higher, P < .001) or higher IOP range (-0.07 [95% CI, -0.09 to -0.05] µm per 1-mmHg higher, P < .001) after adjustment for mean IOP and other confounding factors. Similar results were found for early and moderate to advanced stages of glaucoma. CONCLUSIONS: IOP variability showed an independent association with macular change in patients with glaucoma regardless of severity at baseline, even after adjustment for mean IOP, supporting its potential value as a therapeutic target for clinical decision-making.

Central visual field damage in glaucoma eyes with choroidal microvasculature dropout with and without high axial myopia.

PURPOSE: To characterise the relationship between a deep-layer microvasculature dropout (MvD) and central visual field (VF) damage in primary open-angle glaucoma (POAG) patients with and without high axial myopia. DESIGN: Cross-sectional study. METHODS: Seventy-one eyes (49 patients) with high axial myopia and POAG and 125 non-highly myopic POAG eyes (97 patients) were enrolled. Presence, area and angular circumference of juxtapapillary MvD were evaluated on optical coherence tomography angiography B-scans and en-face choroidal images. RESULTS: Juxtapapillary MvD was detected more often in the highly myopic POAG eyes (43 eyes, 86%) than in the non-highly myopic eyes (73 eyes, 61.9%; p=0.002). In eyes with MvD, MvD area and angular circumference (95% CI) were significantly larger in the highly myopic eyes compared with the non-highly myopic eyes (area: (0.69 (0.40, 0.98) mm2 vs 0.31 (0.19, 0.42) mm2, p=0.011) and (angular circumference: 84.3 (62.9, 105.8) vs 74.5 (58.3, 90.9) degrees, p<0.001), respectively. 24-2 VF mean deviation (MD) was significantly worse in eyes with MvD compared with eyes without MvD in both groups (p<0.001). After adjusting for 24-2 MD VF, central VF defects were more frequently found in eyes with MvD compared with eyes without MvD (82.7% vs 60.9%, p<0.001). In multivariable analysis, higher intraocular pressure, worse 24-2 VF MD, longer axial length and greater MvD area and angular circumference were associated with worse 10-2 VF MD. CONCLUSIONS: MvD was more prevalent and larger in POAG eyes with high myopia than in non-highly myopic POAG eyes. In both groups, eyes with MvD showed worse glaucoma severity and more central VF defects.

Deep Learning Identifies High-Quality Fundus Photographs and Increases Accuracy in Automated Primary Open Angle Glaucoma Detection.

Purpose: To develop and evaluate a deep learning (DL) model to assess fundus photograph quality, and quantitatively measure its impact on automated POAG detection in independent study populations. Methods: Image quality ground truth was determined by manual review of 2815 fundus photographs of healthy and POAG eyes from the Diagnostic Innovations in Glaucoma Study and African Descent and Glaucoma Evaluation Study (DIGS/ADAGES), as well as 11,350 from the Ocular Hypertension Treatment Study (OHTS). Human experts assessed a photograph as high quality if of sufficient quality to determine POAG status and poor quality if not. A DL quality model was trained on photographs from DIGS/ADAGES and tested on OHTS. The effect of DL quality assessment on DL POAG detection was measured using area under the receiver operating characteristic (AUROC). Results: The DL quality model yielded an AUROC of 0.97 for differentiating between high- and low-quality photographs; qualitative human review affirmed high model performance. Diagnostic accuracy of the DL POAG model was significantly greater (P < 0.001) in good (AUROC, 0.87; 95% CI, 0.80-0.92) compared with poor quality photographs (AUROC, 0.77; 95% CI, 0.67-0.88). Conclusions: The DL quality model was able to accurately assess fundus photograph quality. Using automated quality assessment to filter out low-quality photographs increased the accuracy of a DL POAG detection model. Translational Relevance: Incorporating DL quality assessment into automated review of fundus photographs can help to decrease the burden of manual review and improve accuracy for automated DL POAG detection.

Racial Differences in Diagnostic Accuracy of Retinal Nerve Fiber Layer Thickness in Primary Open-Angle Glaucoma.

Purpose: To evaluate the diagnostic accuracy of retinal nerve fiber layer thickness (RNFLT) by spectral-domain optical coherence tomography (OCT) in primary open-angle glaucoma (POAG) in eyes of African (AD) and European descent (ED). Design: Comparative diagnostic accuracy analysis by race. Participants: 379 healthy eyes (125 AD and 254 ED) and 442 glaucomatous eyes (226 AD and 216 ED) from the Diagnostic Innovations in Glaucoma Study and the African Descent and Glaucoma Evaluation Study. Methods: Spectralis (Heidelberg Engineering GmbH) and Cirrus (Carl Zeiss Meditec) OCT scans were taken within one year from each other. Main Outcome Measures: Diagnostic accuracy of RNFLT measurements. Results: Diagnostic accuracy for Spectralis-RNFLT was significantly lower in eyes of AD compared to those of ED (area under the receiver operating curve [AUROC]: 0.85 and 0.91, respectively, P=0.04). Results for Cirrus-RNFLT were similar but did not reach statistical significance (AUROC: 0.86 and 0.90 in AD and ED, respectively, P =0.33). Adjustments for age, central corneal thickness, axial length, disc area, visual field mean deviation, and intraocular pressure yielded similar results. Conclusions: OCT-RNFLT has lower diagnostic accuracy in eyes of AD compared to those of ED. This finding was generally robust across two OCT instruments and remained after adjustment for many potential confounders. Further studies are needed to explore the potential sources of this difference.

Proactive Decision Support for Glaucoma Treatment: Predicting Surgical Interventions with Clinically Available Data.

A longitudinal ophthalmic dataset was used to investigate multi-modal machine learning (ML) models incorporating patient demographics and history, clinical measurements, optical coherence tomography (OCT), and visual field (VF) testing in predicting glaucoma surgical interventions. The cohort included 369 patients who underwent glaucoma surgery and 592 patients who did not undergo surgery. The data types used for prediction included patient demographics, history of systemic conditions, medication history, ophthalmic measurements, 24-2 VF results, and thickness measurements from OCT imaging. The ML models were trained to predict surgical interventions and evaluated on independent data collected at a separate study site. The models were evaluated based on their ability to predict surgeries at varying lengths of time prior to surgical intervention. The highest performing predictions achieved an AUC of 0.93, 0.92, and 0.93 in predicting surgical intervention at 1 year, 2 years, and 3 years, respectively. The models were also able to achieve high sensitivity (0.89, 0.77, 0.86 at 1, 2, and 3 years, respectively) and specificity (0.85, 0.90, and 0.91 at 1, 2, and 3 years, respectively) at an 0.80 level of precision. The multi-modal models trained on a combination of data types predicted surgical interventions with high accuracy up to three years prior to surgery and could provide an important tool to predict the need for glaucoma intervention.

Multimodal Deep Learning Classifier for Primary Open Angle Glaucoma Diagnosis Using Wide-Field Optic Nerve Head Cube Scans in Eyes With and Without High Myopia.

PRCIS: An optical coherence tomography (OCT)-based multimodal deep learning (DL) classification model, including texture information, is introduced that outperforms single-modal models and multimodal models without texture information for glaucoma diagnosis in eyes with and without high myopia. BACKGROUND/AIMS: To evaluate the diagnostic accuracy of a multimodal DL classifier using wide OCT optic nerve head cube scans in eyes with and without axial high myopia. MATERIALS AND METHODS: Three hundred seventy-one primary open angle glaucoma (POAG) eyes and 86 healthy eyes, all without axial high myopia [axial length (AL) ≤ 26 mm] and 92 POAG eyes and 44 healthy eyes, all with axial high myopia (AL > 26 mm) were included. The multimodal DL classifier combined features of 3 individual VGG-16 models: (1) texture-based en face image, (2) retinal nerve fiber layer (RNFL) thickness map image, and (3) confocal scanning laser ophthalmoscope (cSLO) image. Age, AL, and disc area adjusted area under the receiver operating curves were used to compare model accuracy. RESULTS: Adjusted area under the receiver operating curve for the multimodal DL model was 0.91 (95% CI = 0.87, 0.95). This value was significantly higher than the values of individual models [0.83 (0.79, 0.86) for texture-based en face image; 0.84 (0.81, 0.87) for RNFL thickness map; and 0.68 (0.61, 0.74) for cSLO image; all P ≤ 0.05]. Using only highly myopic eyes, the multimodal DL model showed significantly higher diagnostic accuracy [0.89 (0.86, 0.92)] compared with texture en face image [0.83 (0.78, 0.85)], RNFL [0.85 (0.81, 0.86)] and cSLO image models [0.69 (0.63, 0.76)] (all P ≤ 0.05). CONCLUSIONS: Combining OCT-based RNFL thickness maps with texture-based en face images showed a better ability to discriminate between healthy and POAG than thickness maps alone, particularly in high axial myopic eyes.

Displacement of the Lamina Cribrosa With Acute Intraocular Pressure Increase in Brain-Dead Organ Donors.

Purpose: To examine deformations of the optic nerve head (ONH) deep tissues in response to acute elevation of intraocular pressure (IOP). Methods: Research-consented brain-dead organ donors underwent imaging by spectral domain optical coherence tomography (OCT). OCT imaging was repeated while the eye was sequentially maintained at manometric pressures of 10, 30, and 50 mm Hg. Radial scans of the ONH were automatically segmented by deep learning and quantified in three dimensions by a custom algorithm. Change in lamina cribrosa (LC) depth and choroidal thickness was correlated with IOP and age by linear mixed-effect models. LC depth was computed against commonly utilized reference planes. Results: Twenty-six eyes from 20 brain-dead organ donors (age range, 22-62 years; median age, 43 years) were imaged and quantified. LC depth measured against a reference plane based on Bruch's membrane (BM), BM opening, and an anterior sclera canal opening plane showed both a reduction and an increase in LC depth with IOP elevation. LC depth universally increased in depth when measured against a sclera reference plane. Choroidal (-0.5222 µm/mm Hg, P < 0.001) and retinal nerve fiber layer thickness (-0.0717 µm/mm Hg, P < 0.001) significantly thinned with increasing IOP. The magnitude of LC depth change with IOP was significantly smaller with increasing age (P < 0.03 for all reference planes). Conclusions: LC depth changes with IOP reduce with age and are significantly affected by the reference plane of choice, which highlights a need for standardizing LC metrics to properly follow progressive remodeling of the loadbearing tissues of the ONH by OCT imaging and for the definition of a reference database.

Long-term variability of retinal nerve fibre layer thickness measurement in patients with glaucoma of African and European descents.

BACKGROUND: To examine long-term retinal nerve fibre layer thickness (RNFLT) variability and associated clinical factors in African (AD) and European descent (ED) individuals with glaucoma. METHODS: This retrospective cohort study included glaucoma eyes of AD and ED from Diagnostic Innovations in Glaucoma Study/The African Descent and Glaucoma Evaluation Study with ≥4 visits/2 years of follow-up. We calculated optic nerve head RNFLT variability per-examination/visit as the absolute error of its residuals across follow-up. Full, baseline and parsimonious linear-mixed models were fit to evaluate the effects of clinical factors (demographics and ocular characteristics, prior/intervening glaucoma surgeries and cataract extraction (CE), RNFLT thinning rate, scan quality, visit/testing frequency, etc) on RNFLT variability in both races. RESULTS: There were 376 and 625 eyes (226 and 349 participants) of AD and ED, and the mean (95% CI) RNFLT variability was 1.62 (1.52, 1.71) µm and 1.42 (1.34, 1.50) µm, respectively (p=0.002). AD and ED had some shared predictors of RNFLT variability, including intraocular pressure fluctuation and scan quality, although the effects varied (p<0.05). In both races, intervening CE was most strongly correlated with higher RNFLT variability (ß: 0.24-0.92, p<0.05). After excluding eyes with intervening CE, RNFLT variability was reduced and the small racial difference was no longer significant (AD: 1.40 (1.31, 1.48) µm vs ED: 1.34 (1.27, 1.40) µm; p=0.280). CONCLUSIONS: Although some predictors were identified, long-term RNFLT variability appeared small for both AD and ED eyes. Moreover, the racial difference did not remain once intervening CE, the strongest predictor of variability, was eliminated. Our findings inform on strategies to optimise structural assessment and suggest that, when accounting for relevant factors, RNFLT is reliable across races.

Detecting Glaucoma from Fundus Photographs Using Deep Learning without Convolutions: Transformer for Improved Generalization.

Purpose: To compare the diagnostic accuracy and explainability of a Vision Transformer deep learning technique, Data-efficient image Transformer (DeiT), and ResNet-50, trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS) to detect primary open-angle glaucoma (POAG) and identify the salient areas of the photographs most important for each model's decision-making process. Design: Evaluation of a diagnostic technology. Subjects Participants and Controls: Overall 66 715 photographs from 1636 OHTS participants and an additional 5 external datasets of 16 137 photographs of healthy and glaucoma eyes. Methods: Data-efficient image Transformer models were trained to detect 5 ground-truth OHTS POAG classifications: OHTS end point committee POAG determinations because of disc changes (model 1), visual field (VF) changes (model 2), or either disc or VF changes (model 3) and Reading Center determinations based on disc (model 4) and VFs (model 5). The best-performing DeiT models were compared with ResNet-50 models on OHTS and 5 external datasets. Main Outcome Measures: Diagnostic performance was compared using areas under the receiver operating characteristic curve (AUROC) and sensitivities at fixed specificities. The explainability of the DeiT and ResNet-50 models was compared by evaluating the attention maps derived directly from DeiT to 3 gradient-weighted class activation map strategies. Results: Compared with our best-performing ResNet-50 models, the DeiT models demonstrated similar performance on the OHTS test sets for all 5 ground-truth POAG labels; AUROC ranged from 0.82 (model 5) to 0.91 (model 1). Data-efficient image Transformer AUROC was consistently higher than ResNet-50 on the 5 external datasets. For example, AUROC for the main OHTS end point (model 3) was between 0.08 and 0.20 higher in the DeiT than ResNet-50 models. The saliency maps from the DeiT highlight localized areas of the neuroretinal rim, suggesting important rim features for classification. The same maps in the ResNet-50 models show a more diffuse, generalized distribution around the optic disc. Conclusions: Vision Transformers have the potential to improve generalizability and explainability in deep learning models, detecting eye disease and possibly other medical conditions that rely on imaging for clinical diagnosis and management.

Vision-Related Quality of Life Among Healthy, Preperimetric Glaucoma, and Perimetric Glaucoma Patients.

PURPOSE: To investigate the association of vision-related quality of life (VRQOL) with the central visual field and macular ganglion cell complex (GCC) thickness in healthy control participants, patients with preperimetric glaucoma, and patients with perimetric glaucoma. DESIGN: Retrospective cross-sectional study. METHODS: A total of 39 healthy, 34 preperimetric glaucoma, and 145 perimetric glaucoma patients completed the 25-item National Eye Institute Visual Function Questionnaire (NEI-VFQ). A linear mixed-effect models was used to investigate the association between the glaucoma stage as measured by binocular 10-2 visual field mean sensitivity (VFMS) and GCC thickness with the Rasch-calibrated NEI-VFQ score. RESULTS: A total of 436 eyes from 218 participants (mean age = 67.2 [95% CI = 65.1 to 69.2] years) were enrolled. VRQOL calculated by the NEI-VFQ Rasch-calibrated score was worst for patients with perimetric glaucoma (50.7 [95% CI = 47.2 to 54.2]), followed by patients with preperimetric glaucoma (41.2 [95% CI = 34.5 to 47.9]) and healthy controls (29.3 [95% CI = 24.0 to 34.7]. Worse VRQOL had a moderate association with a worse global binocular 10-2 VFMS (-3.4 [95% CI = -5.0 to -1.9] dB per 1 score; P < .001; adjusted R2 = 0.27), but not with a thinner global GCC in the better eye (-0.1 [95% CI = -0.2 to 0.1] µm per 1 score; P =.0485; adjusted R2 = 0.17). CONCLUSIONS: These findings suggest that patients with perimetric and preperimetric glaucoma have worse VRQOL than patients with healthy eyes. As compared to macular thickness measurements, the central visual field is more strongly associated with VRQOL and may better help to identify patients in need of intervention.