We sought to construct and evaluate a deep learning (DL) model to diagnose early glaucoma from spectral-domain optical coherence tomography (OCT) images.
Artificial intelligence diagnostic tool ...development, evaluation, and comparison.
This multi-institution study included pretraining data of 4316 OCT images (RS3000) from 1371 eyes with open angle glaucoma (OAG) regardless of the stage of glaucoma and 193 normal eyes. Training data included OCT-1000/2000 images from 94 eyes of 94 patients with early OAG (mean deviation > −5.0 dB) and 84 eyes of 84 normal subjects. Testing data included OCT-1000/2000 from 114 eyes of 114 patients with early OAG (mean deviation > −5.0 dB) and 82 eyes of 82 normal subjects. A DL (convolutional neural network) classifier was trained using a pretraining dataset, followed by a second round of training using an independent training dataset. The DL model input features were the 8 × 8 grid macular retinal nerve fiber layer thickness and ganglion cell complex layer thickness from spectral-domain OCT. Diagnostic accuracy was investigated in the testing dataset. For comparison, diagnostic accuracy was also evaluated using the random forests and support vector machine models. The primary outcome measure was the area under the receiver operating characteristic curve (AROC).
The AROC with the DL model was 93.7%. The AROC significantly decreased to between 76.6% and 78.8% without the pretraining process. Significantly smaller AROCs were obtained with random forests and support vector machine models (82.0% and 67.4%, respectively).
A DL model for glaucoma using spectral-domain OCT offers a substantive increase in diagnostic performance.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
We aimed to develop a model to predict visual field (VF) in the central 10 degrees in patients with glaucoma, by training a convolutional neural network (CNN) with optical coherence tomography (OCT) ...images and adjusting the values with Humphrey Field Analyzer (HFA) 24-2 test. The training dataset included 558 eyes from 312 glaucoma patients and 90 eyes from 46 normal subjects. The testing dataset included 105 eyes from 72 glaucoma patients. All eyes were analyzed by the HFA 10-2 test and OCT; eyes in the testing dataset were additionally analyzed by the HFA 24-2 test. During CNN model training, the total deviation (TD) values of the HFA 10-2 test point were predicted from the combined OCT-measured macular retinal layers' thicknesses. Then, the predicted TD values were corrected using the TD values of the innermost four points from the HFA 24-2 test. Mean absolute error derived from the CNN models ranged between 9.4 and 9.5 B. These values reduced to 5.5 dB on average, when the data were corrected using the HFA 24-2 test. In conclusion, HFA 10-2 test results can be predicted with a OCT images using a trained CNN model with adjustment using HFA 24-2 test.
Full text
Available for:
IZUM, KILJ, NUK, PILJ, PNG, SAZU, UL, UM, UPUK
Purpose
To report a new technique for fundus examination using a wide-angle viewing system combined with intraocular illumination without accessing the vitreous cavity.
Study design
Retrospective ...case series
Methods
Consecutive patients with atopic dermatitis-related cataracts who underwent standard cataract surgery and the novel fundus examination technique were included. After phacoemulsification, the anterior and posterior chambers were filled with ocular viscoelastic devices. A 27-gauge endo-illumination probe was inserted into the anterior chamber through a corneal incision made for cataract surgery. The fundus examination was performed with a wide-angle viewing system and scleral indentation. If any retinal breaks/detachments were detected, they were treated simultaneously. Finally, an intraocular lens was implanted.
Results
Ten patients (13 eyes) were included (mean age 26.8 years; 9 men). Retinal breaks were detected in 5 eyes (38%); 2 of the 5 had rhegmatogenous retinal detachment (RRD) (15%). Intraoperative cryopexy was performed for 3 eyes with retinal breaks, while 2 eyes with RRD underwent a scleral buckling procedure (SBP) during the same surgery. There were no intra- or postoperative complications, including posterior capsule damage. The average surgical time was 22 minutes for cases requiring only fundus examination and about 28 and for eyes with cryopexy and 80 minutes for SBP.
Conclusion
The described technique may reduce the disadvantages of creating scleral incisions and provide comparable visibility to inserting the illuminator into the vitreous cavity.
Full text
Available for:
EMUNI, FIS, FZAB, GEOZS, GIS, IJS, IMTLJ, KILJ, KISLJ, MFDPS, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, SBMB, SBNM, UKNU, UL, UM, UPUK, VKSCE, ZAGLJ
We investigated birefringence-derived artifacts that potentially mimic focal defects of the lamina cribrosa (focal LC defects) in optical coherence tomography (OCT) imaging of eyes with glaucoma. ...This study included 74 eyes of 48 patients with glaucoma. Five horizontal line B-scan images of the optic disc were obtained using commercial swept-source OCT. From a dataset of prototype swept-source polarization-diversity OCT, we calculated the following types of OCT images: polarization-dependent, polarization-dependent attenuation-coefficient, polarization-independent, and polarization-independent attenuation-coefficient. We assessed the commercial OCT images for the presence of birefringence-derived artifacts by comparison with the polarization-diversity OCT images. Commercial OCT showed suggestive findings of focal LC defects in 17 of 74 eyes. Reevaluation using polarization-independent OCT revealed that the focal LC defects in one of 17 eyes (5.9%) were actually birefringence-derived artifacts. This study demonstrated the existence of birefringence-derived artifacts mimicking focal LC defects in commercial OCT imaging and indicated that polarization-diversity OCT is an effective tool to evaluate the presence of these artifacts.
Full text
Available for:
IZUM, KILJ, NUK, PILJ, PNG, SAZU, UL, UM, UPUK
We evaluated the prevalence and characteristics of the defects of the lamina cribrosa (LC) in high myopia and glaucoma, and compared them with control eyes using swept-source optical coherence ...tomography (SS-OCT).
One hundred fifty-nine eyes of 108 participants were divided into four subgroups; high myopia with glaucoma (MG, 67 eyes of 46 subjects), glaucoma without high myopia (G, 22 eyes of 13 subjects), high myopia without glaucoma (M, 35 eyes of 29 subjects), and a control group with neither glaucoma nor high myopia (C, 35 eyes of 20 subjects). The LC defects were identified and located using a standardized protocol in serial horizontal OCT scans. The prevalence rates of the defects were compared among the groups. Demographic and ocular factors were compared between eyes with and without defects.
LC defects were observed in one eye (0.03%) in the C group, 8 eyes (22.9%) in the M group, 11 eyes (50%) in the G group, and 28 eyes (41.8%) in the MG group. The prevalence rates of the defects differed significantly among the groups (P = 0.0009). Most eyes with defects in the G and MG groups (79.5%) had damage in the corresponding visual hemifields. Other factors such as visual acuity, intraocular pressure, axial length, refractive error, disc ovality, or parapapillary atrophy area did not differ significantly between eyes with and without LC defects.
High myopia and glaucoma significantly increased the risk of LC damage. The LC damage in non-glaucomatous highly myopic eyes may at least partly explain the increased risk of developing glaucoma in myopic eyes.
Full text
Available for:
DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
To develop a large-scale real clinical database of glaucoma (Japanese Archive of Multicentral Databases in Glaucoma: JAMDIG) and to investigate the effect of treatment.
The study included a total of ...1348 eyes of 805 primary open-angle glaucoma patients with 10 visual fields (VFs) measured with 24-2 or 30-2 Humphrey Field Analyzer (HFA) and intraocular pressure (IOP) records in 10 institutes in Japan. Those with 10 reliable VFs were further identified (638 eyes of 417 patients). Mean total deviation (mTD) of the 52 test points in the 24-2 HFA VF was calculated, and the relationship between mTD progression rate and seven variables (age, mTD of baseline VF, average IOP, standard deviation (SD) of IOP, previous argon/selective laser trabeculoplasties (ALT/SLT), previous trabeculectomy, and previous trabeculotomy) was analyzed.
The mTD in the initial VF was -6.9 ± 6.2 dB and the mTD progression rate was -0.26 ± 0.46 dB/year. Mean IOP during the follow-up period was 13.5 ± 2.2 mm Hg. Age and SD of IOP were related to mTD progression rate. However, in eyes with average IOP below 15 and also 13 mm Hg, only age and baseline VF mTD were related to mTD progression rate.
Age and the degree of VF damage were related to future progression. Average IOP was not related to the progression rate; however, fluctuation of IOP was associated with faster progression, although this was not the case when average IOP was below 15 mm Hg.
To compare the rates of retinal nerve fiber layer (RNFL) loss in patients suspected of having glaucoma who developed visual field damage (VFD) with those who did not develop VFD and to determine ...whether the rate of RNFL loss can be used to predict the development of VFD.
Prospective, observational cohort study.
Glaucoma suspects, defined as having glaucomatous optic neuropathy or ocular hypertension (intraocular pressure, >21 mmHg) without repeatable VFD at baseline, from the Diagnostic Innovations in Glaucoma Study and the African Descent and Glaucoma Evaluation Study.
Global and quadrant RNFL thickness (RNFLT) were measured with the Spectralis spectral-domain optical coherence tomography (SD-OCT; Spectralis HRA+OCT Heidelberg Engineering, Heidelberg, Germany). Visual field damage was defined as having 3 consecutive abnormal visual fields. The rate of RNFL loss in eyes developing VFD was compared to eyes not developing VFD using multivariate linear mixed-effects models. A joint longitudinal survival model used the estimated RNFLT slope to predict the risk of developing VFD, while adjusting for potential confounding variables.
The rate of RNFL thinning and the probability of developing VFD.
Four hundred fifty-four eyes of 294 glaucoma suspects were included. The average number of SD-OCT examinations was 4.6 (range, 2-9), with median follow-up of 2.2 years (0.4-4.1 years). Forty eyes (8.8%) developed VFD. The estimated mean rate of global RNFL loss was significantly faster in eyes that developed VFD compared with eyes that did not develop VFD (-2.02 μm/year vs. -0.82 μm/year; P<0.001). The joint longitudinal survival model showed that each 1-μm/year faster rate of global RNFL loss corresponded to a 2.05-times higher risk of developing VFD (hazard ratio, 2.05; 95% confidence interval, 1.14-3.71; P = 0.017).
The rate of global RNFL loss was more than twice as fast in eyes that developed VFD compared with eyes that did not develop VFD. A joint longitudinal survival model showed that a 1-μm/year faster rate of RNFLT loss corresponded to a 2.05-times higher risk of developing VFD. These results suggest that measuring the rate of SD-OCT RNFL loss may be a useful tool to help identify patients who are at a high risk of developing visual field loss.
To characterize the corneal biomechanical properties of glaucoma eyes by comparing the dynamic Scheimpflug biomechanical parameters between untreated glaucoma and control eyes.
Cross-sectional ...observational data of dynamic Scheimpflug analyzer (Corvis ST) examinations were retrospectively collected from 35 eyes of 35 consecutive patients with untreated normal tension glaucoma and 35 eyes of 35 healthy patients matched on age and IOP. Ten biomechanical parameters were compared between the two groups using multivariable models adjusting for IOP, central corneal thickness, age, and axial length. The Benjamini-Hochberg method was used to correct for multiple comparison.
In multivariable models, glaucoma was associated with smaller applanation 1 time (P < 0.001, coefficient = -0.5865), applanation 2 time (P = 0.012, coefficient = -0.1702), radius (P = 0.006, coefficient = -0.5447), larger peak distance (P = 0.011, coefficient = 0.1023), deformation amplitude ratio at 1 mm (P < 0.001, coefficient = 0.072), and integrated radius (P < 0.001, coefficient = 1.094). These associations consistently indicate greater compliance of the cornea in glaucoma eyes.
Untreated normal tension glaucoma eyes were more compliant than healthy eyes. The greater compliance (smaller stiffness) of normal tension glaucoma eyes may increase the risk of optic nerve damage. These results suggest the relevance of measuring biomechanical properties of glaucoma eyes.
To test the detection of subclinical corneal ectasia using integrated Scheimpflug tomography and biomechanical assessment in a Japanese population.
This prospective, case-control study included 23 ...patients with very asymmetric ectasia (VAE) and 70 normal controls. Patients with VAE had defined clinical ectasia in one eye and a fellow eye with normal topography (VAE-NT). Objective topography for confirming normal topography in VAE-NT cases included having 0% similarity and 0% severity derived from Placido-disk based topography. Scheimpflug-based corneal tomography and corneal biomechanical assessment were performed. The Belin/Ambrósio Enhanced Ectasia Deviation index (BAD-D), Corvis Biomechanical Index (CBI), and Tomographic Biomechanical Index (TBI) were compared and their discriminating ability for detecting ectasia was assessed.
For differentiating normal and VAE-NT eyes, the areas under the receiver operating curve for the BAD-D, CBI, and TBI were 0.668, 0.660, and 0.751, respectively. The TBI cut-off of 0.259 provided 52.17% sensitivity and 88.57% specificity. Fourteen VAE-NT cases (60.9%) were abnormal in at least one of the criteria of the BAD-D > 1.60 (39.1%), CBI > 0.5 (26.1%), or TBI > 0.29 (43.5%). Conversely, nine VAE-NT cases (39.1%) exhibited normal values for the BAD-D, CBI, and TBI.
In the current study, 40% of VAE-NT eyes were classified as normal by the BAD-D, CBI, and TBI. Although some of these cases may truly represent unilateral ectasia, further advances are needed to enhance ectasia detection and characterize the susceptibility for ectasia progression. J Refract Surg. 2019;35(6):383-390..
To predict the visual field (VF) of glaucoma patients within the central 10° from optical coherence tomography (OCT) measurements using deep learning and tensor regression.
Cross-sectional study.
...Humphrey 10-2 VFs and OCT measurements were carried out in 505 eyes of 304 glaucoma patients and 86 eyes of 43 normal subjects. VF sensitivity at each test point was predicted from OCT-measured thicknesses of macular ganglion cell layer + inner plexiform layer, retinal nerve fiber layer, and outer segment + retinal pigment epithelium. Two convolutional neural network (CNN) models were generated: (1) CNN-PR, which simply connects the output of the CNN to each VF test point; and (2) CNN-TR, which connects the output of the CNN to each VF test point using tensor regression. Prediction performance was assessed using 5-fold cross-validation through the root mean squared error (RMSE). For comparison, RMSE values were also calculated using multiple linear regression (MLR) and support vector regression (SVR). In addition, the absolute prediction error for predicting mean sensitivity in the whole VF was analyzed.
RMSE with the CNN-TR model averaged 6.32 ± 3.76 (mean ± standard deviation) dB. Significantly (P < .05) larger RMSEs were obtained with other models: CNN-PR (6.76 ± 3.86 dB), SVR (7.18 ± 3.87 dB), and MLR (8.56 ± 3.69 dB). The absolute mean prediction error for the whole VF was 2.72 ± 2.60 dB with the CNN-TR model.
The Humphrey 10-2 VF can be predicted from OCT-measured retinal layer thicknesses using deep learning and tensor regression.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
PDF
