Artificial intelligence capabilities have, recently, greatly improved. In the past few years, one of the deep learning algorithms, the recurrent neural network (RNN), has shown an outstanding ability in sequence labeling and prediction tasks for sequential data. We built a reliable visual field prediction algorithm using RNN and evaluated its performance in comparison with the conventional pointwise ordinary linear regression (OLR) method. A total of 1,408 eyes were used as a training dataset and another dataset, comprising 281 eyes, was used as a test dataset. Five consecutive visual field tests were provided to the constructed RNN as input and a 6 th visual field test was compared with the output of the RNN. The performance of the RNN was compared with that of OLR by predicting the 6 th visual field in the test dataset. The overall prediction performance of RNN was significantly better than OLR. The pointwise prediction error of the RNN was significantly smaller than that of the OLR in most areas known to be vulnerable to glaucomatous damage. The RNN was also more robust and reliable regarding worsening in the visual field examination. In clinical practice, the RNN model can therefore assist in decision-making for further treatment of glaucoma.

We developed a deep learning architecture based on Inception V3 to predict visual field using optical coherence tomography (OCT) imaging and evaluated its performance. Two OCT images, macular ganglion cell-inner plexiform layer (mGCIPL) and peripapillary retinal nerve fibre layer (pRNFL) thicknesses, were acquired and combined. A convolutional neural network architecture was constructed to predict visual field using this combined OCT image. The root mean square error (RMSE) between the actual and predicted visual fields was calculated to evaluate the performance. Globally (the entire visual field area), the RMSE for all patients was 4.79 ± 2.56 dB, with 3.27 dB and 5.27 dB for the normal and glaucoma groups, respectively. The RMSE of the macular region (4.40 dB) was higher than that of the peripheral region (4.29 dB) for all subjects. In normal subjects, the RMSE of the macular region (2.45 dB) was significantly lower than that of the peripheral region (3.11 dB), whereas in glaucoma subjects, the RMSE was higher (5.62 dB versus 5.03 dB, respectively). The deep learning method effectively predicted the visual field 24-2 using the combined OCT image. This method may help clinicians determine visual fields, particularly for patients who are unable to undergo a physical visual field exam.

The Spectralis SD-OCT showed excellent reproducibility in BMO-MRW measurements in both normal and glaucoma subjects. The measurements variability was worse in more advanced glaucoma.

PurposeWe evaluate the relationship between Bruch's membrane opening minimum rim width (BMO-MRW) and peripapillary retinal nerve fiber layer thickness (pRNFLT) and develop a new parameter combining BMO-MRW and pRNFLT using a neural network to maximize their compensatory values.MethodsA total of 402 subjects were divided into two groups: 273 (validation group) and 129 (neural net training) subjects. Linear quadratic and broken-stick regression models were used to explore the relationship between BMO-MRW and pRNFLT. A multilayer neural network was used to create a combined parameter, and diagnostic performances were compared using area under the receiver operating characteristic curves (AUROCs).ResultsRegression analyses between BMO-MRW and pRNFLT revealed that the broken-stick model afforded the best fit. Globally, the tipping point was a BMO-MRW of 226.5 μm. BMO-MRW and pRNFLT were correlated significantly with visual field. When differentiating normal from glaucoma subjects, the neural network exhibited the largest AUROC. When differentiating normal from early glaucoma subjects, the overall diagnostic performance decreased, but the neural network still exhibited the largest AUROC.ConclusionsThe optimal relationship between BMO-MRW and pRNFLT was revealed using the broken-stick model. Considerable BMO-MRW thinning preceded pRNFLT thinning. The neural network significantly improved diagnostic power by combining BMO-MRW and pRNFLT.Translational RelevanceA combined index featuring BMO-MRW and pRNFLT data can aid clinical decision-making, particularly when individual parameters yield confusing results. Our neural network effectively combines information from separate parameters.

BackgroundWe evaluated the relationships between corneal biomechanical properties and structural parameters in patients with newly diagnosed, untreated normal-tension glaucoma (NTG).MethodsAll subjects were evaluated using an Ocular Response Analyzer (ORA) measuring corneal hysteresis (CH) and the corneal resistance factor (CRF). Central corneal thickness (CCT), Goldmann applanation tonometric (GAT) data, axial length, and the spherical equivalent (SE), were also measured. Confocal scanning laser ophthalmoscopy was performed with the aid of a Heidelberg retina tomograph (HRT III). We sought correlations between HRT parameters and different variables including CCT, CH, and the CRF. Multiple linear regression analysis was performed to identify significant associations between corneal biomechanical properties and optic nerve head parameters.ResultsWe enrolled 95 eyes of 95 NTG patients and 93 eyes of 93 normal subjects. CH and the CRF were significantly lower in more advanced glaucomatous eyes (P = 0.001, P = 0.008, respectively). The rim area, rim volume, linear cup-to-disc ratio (LCDR), and mean retinal nerve fiber layer (RNFL) thickness were significantly worse in more advanced glaucomatous eyes (P < 0.001, P < 0.001, P < 0.001, and P = 0.001). CH was directly associated with rim area, rim volume, and mean RNFL thickness (P = 0.012, P = 0.028, and P = 0.043) and inversely associated with LCDR (P = 0.015), after adjusting for age, axial length, CCT, disc area, GAT data, and SE. However, in normal subjects, there were no significant associations between corneal biomechanical properties and HRT parameters.ConclusionsA lower CH is significantly associated with a smaller rim area and volume, a thinner RNFL, and a larger LCDR, independent of disc size, corneal thickness, intraocular pressure, and age.

To develop a deep learning model to estimate the visual field (VF) from spectral-domain optical coherence tomography (SD-OCT) and swept-source OCT (SS-OCT) and to compare the performance between them. Methods:Two deep learning models based on Inception-ResNet-v2 were trained to estimate 24-2 VF from SS-OCT and SD-OCT images. The estimation performance of the two models was evaluated by using the root mean square error between the actual and estimated VF. The performance was also compared among different glaucoma severities, Garway-Heath sectorizations, and central/peripheral regions. Results:The training dataset comprised images of 4391 eyes from 2350 subjects, and the test dataset was obtained from another 243 subjects (243 eyes). In all subjects, the global estimation errors were 5.29 ± 2.68 dB (SD-OCT) and 4.51 ± 2.54 dB (SS-OCT), and the estimation error of SS-OCT was significantly lower than that of SD-OCT (P < 0.001). In the analysis of sectors, SS-OCT showed better performance in all sectors except for the inferonasal sector in normal vision and early glaucoma. In advanced glaucoma, the estimation error of the central region was worsened in both OCTs, but SS-OCT was still significantly better in the peripheral region. Conclusions:Our deep learning model estimated the VF 24-2 better with a wide field image of SS-OCT than did with retinal nerve fiber layer and ganglion cell-inner plexiform layer images of SD-OCT.Translational Relevance: This deep learning method can help clinicians to determine the VF from OCT images. OCT manufacturers can equip this system to provide additional VF data.

BackgroundRetinal ganglion cell (RGC) death is a common cause of loss of vision during glaucoma. Pattern electroretinogram (PERG) is an objective measure of the central retinal function that correlates with macular GCL thickness. The aim of this study is to determine possible relationships between the N95 amplitude of pattern electroretinogram (PERGamp) and macular ganglion cell/inner plexiform layer thickness (GCIPLT).Methods and findingsThis was a retrospective and comparative study including 74 glaucoma patients (44 early stage and 30 advanced stage cases) and 66 normal control subjects. Macular GCIPLT was measured using Cirrus spectral domain-optical coherence tomography. Standard automated perimetry and pattern ERGs were used in all patient examinations. Three types of regression analysis (broken stick, linear regression, and quadratic regression) were used to evaluate possible relationships between PERGamp and GCIPLT. Correlations between visual field parameters and GCIPLT were evaluated according to glaucoma severity. The best fit model for the relationship between PERGamp and GCIPLT was the linear regression model (r2 = 0.22; P < 0.001). The best-fit model for the relationship between visual field parameters and GCIPLT was the broken stick model. During early glaucoma, macular GCIPLT was positively correlated with PERGamp, but not with visual field loss. In advanced glaucoma, macular GCIPLT was positively correlated with both PERGamp and visual field loss.ConclusionsPERGamp was significantly correlated with macular GCIPT in early glaucoma patients, while visual field performance showed no correlation with GCIPLT. PERGamp can therefore assist clinicians in making an early decision regarding the most suitable treatment plan, especially when GCIPLT is thinning with no change in visual field performance.

No abstract