Deep Learning Estimation of 10-2 Visual Field Map Based on Macular Optical Coherence Tomography Angiography Measurements.

PURPOSE: To develop deep learning (DL) models estimating the central visual field (VF) from optical coherence tomography angiography (OCTA) vessel density (VD) measurements. DESIGN: Development and validation of a deep learning model. METHODS: A total of 1051 10-2 VF OCTA pairs from healthy, glaucoma suspects, and glaucoma eyes were included. DL models were trained on en face macula VD images from OCTA to estimate 10-2 mean deviation (MD), pattern standard deviation (PSD), 68 total deviation (TD) and pattern deviation (PD) values and compared with a linear regression (LR) model with the same input. Accuracy of the models was evaluated by calculating the average mean absolute error (MAE) and the R2 (squared Pearson correlation coefficients) of the estimated and actual VF values. RESULTS: DL models predicting 10-2 MD achieved R2 of 0.85 (95% confidence interval [CI], 74-0.92) for 10-2 MD and MAEs of 1.76 dB (95% CI, 1.39-2.17 dB) for MD. This was significantly better than mean linear estimates for 10-2 MD. The DL model outperformed the LR model for the estimation of pointwise TD values with an average MAE of 2.48 dB (95% CI, 1.99-3.02) and R2 of 0.69 (95% CI, 0.57-0.76) over all test points. The DL model outperformed the LR model for the estimation of all sectors. CONCLUSIONS: DL models enable the estimation of VF loss from OCTA images with high accuracy. Applying DL to the OCTA images may enhance clinical decision making. It also may improve individualized patient care and risk stratification of patients who are at risk for central VF damage.

Impact of smoking on choroidal microvasculature dropout in glaucoma: a cross-sectional study.

OBJECTIVE: To investigate the effect of smoking on choroidal microvasculature dropout (MvD) in glaucoma. DESIGN: Cross-sectional study. SETTING: Tertiary glaucoma centre. PARTICIPANTS: 223 eyes of 163 patients with primary open-angle glaucoma who had undergone imaging with optical coherence tomography angiography and completed a questionnaire on smoking from the Diagnostic Innovations in Glaucoma Study. PRIMARY OUTCOME MEASURES: Linear mixed-effects models were used to determine the effect of each parameter on MvD area and angular circumference. The sensitivity analysis was performed by categorising the glaucoma severity determined by visual field mean deviation (MD). RESULTS: MvD was found in 37 (51.4%) eyes with smoking history and in 67 (44.4%) eyes with non-smokers (p=0.389). Larger MvD area and wider angular circumference were found in smokers compared with non-smokers (p=0.068 and p=0.046, respectively). In a multivariable model, smoking intensity was significantly associated with MvD area (0.30(95% CI 0.01 to 0.60) each 0.01 mm2 per 10 pack-years; p=0.044). In eyes with moderate-severe glaucoma (MD <-6), smoking intensity was associated with larger MvD area (0.47 (95% CI 0.11 to 0.83) each 0.01 mm2 per 10 pack-years; p=0.011), whereas no significant association was found in early glaucoma (MD ≥-6) (-0.08 (95% CI -0.26 to 0.11), p=0.401). CONCLUSIONS: Smoking intensity was associated with larger choroidal MvD area in eyes with glaucoma, especially in patients with more severe disease. TRIAL REGISTRATION NUMBER: NCT00221897.

Intraocular pressure increases the rate of macular vessel density loss in glaucoma.

BACKGROUND/AIMS: To evaluate the relationship over time between intraocular pressure (IOP) and the rate of macula whole image vessel density (wiVD) loss and whole image ganglion cell complex (wiGCC) thinning in glaucoma METHODS: From 62 patients in the Diagnostic Innovations in Glaucoma Study, 59 Primary open-angle glaucoma and 27 glaucoma suspect eyes with mean follow-up of 3.2 years were followed. Optical coherence tomography angiography (OCT-A)-based vessel density and OCT-based structural thickness of the same 6×6 mm GCC scan slab were evaluated. Univariable and multivariable linear mixed models were performed for all eyes and also a subset of them in which peak IOP <18 mm Hg to investigate the effect of IOP parameters on the rate of wiVD and wiGCC change. RESULTS: The mean baseline visual field mean deviation (95% CI) was -3.3 dB (-4.4 to -2.1). Higher mean IOP (-0.07%/year per 1 mm Hg (-0.14 to -0.01), p=0.033), peak IOP (-0.07%/year per 1 mm Hg (-0.13 to -0.02), p=0.004) and IOP fluctuation (IOP SD) (-0.17%/year per 1 mm Hg (-0.32 to 0.02), p=0.026) were associated with faster macular vessel density loss. Faster wiGCC thinning was associated with higher mean IOP (-0.05 µm/year per 1 mm Hg (-0.10 to -0.01), p=0.015), peak IOP (-0.05 µm/year per 1 mm Hg (-0.08 to -0.02), p=0.003) and IOP fluctuation (-0.12 µm/year per 1 mm Hg (-0.22 to -0.01), p=0.032). In eyes with peak <18 mm Hg, faster wiVD progression was associated with higher mean IOP (p=0.042). Faster wiGCC progression was associated with higher mean IOP in these eyes (p=0.025). CONCLUSION: IOP metrics were associated with faster rates of overall macular microvascular loss and also in the eyes with peak IOP <18 mm Hg. Future studies are needed to examine whether additional IOP lowering reduces the rate of microvascular loss in patients with glaucoma. TRIAL REGISTRATION NUMBER: NCT00221897.