scholarly journals Deep learning models for screening of high myopia using optical coherence tomography

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Kyung Jun Choi ◽  
Jung Eun Choi ◽  
Hyeon Cheol Roh ◽  
Jun Soo Eun ◽  
Jong Min Kim ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Deep Learning ◽  
Receiver Operating Characteristic ◽  
Axial Length ◽  
High Myopia ◽  
Diagnostic Performance ◽  
Operating Characteristic ◽  
Optical Coherence ◽  
Column Model ◽  
Absolute Agreement

AbstractThis study aimed to validate and evaluate deep learning (DL) models for screening of high myopia using spectral-domain optical coherence tomography (OCT). This retrospective cross-sectional study included 690 eyes in 492 patients with OCT images and axial length measurement. Eyes were divided into three groups based on axial length: a “normal group,” a “high myopia group,” and an “other retinal disease” group. The researchers trained and validated three DL models to classify the three groups based on horizontal and vertical OCT images of the 600 eyes. For evaluation, OCT images of 90 eyes were used. Diagnostic agreements of human doctors and DL models were analyzed. The area under the receiver operating characteristic curve of the three DL models was evaluated. Absolute agreement of retina specialists was 99.11% (range: 97.78–100%). Absolute agreement of the DL models with multiple-column model was 100.0% (ResNet 50), 90.0% (Inception V3), and 72.22% (VGG 16). Areas under the receiver operating characteristic curves of the DL models with multiple-column model were 0.99 (ResNet 50), 0.97 (Inception V3), and 0.86 (VGG 16). The DL model based on ResNet 50 showed comparable diagnostic performance with retinal specialists. The DL model using OCT images demonstrated reliable diagnostic performance to identify high myopia.

scholarly journals Development and validation of a deep learning system to screen vision-threatening conditions in high myopia using optical coherence tomography images

2020 ◽  
pp. bjophthalmol-2020-317825
Author(s):  
Yonghao Li ◽  
Weibo Feng ◽  
Xiujuan Zhao ◽  
Bingqian Liu ◽  
Yan Zhang ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Deep Learning ◽  
High Myopia ◽  
Large Scale ◽  
Learning System ◽  
Youden Index ◽  
Optical Coherence ◽  
Test Dataset ◽  
Independent Test ◽  
Independent Test Dataset

Background/aimsTo apply deep learning technology to develop an artificial intelligence (AI) system that can identify vision-threatening conditions in high myopia patients based on optical coherence tomography (OCT) macular images.MethodsIn this cross-sectional, prospective study, a total of 5505 qualified OCT macular images obtained from 1048 high myopia patients admitted to Zhongshan Ophthalmic Centre (ZOC) from 2012 to 2017 were selected for the development of the AI system. The independent test dataset included 412 images obtained from 91 high myopia patients recruited at ZOC from January 2019 to May 2019. We adopted the InceptionResnetV2 architecture to train four independent convolutional neural network (CNN) models to identify the following four vision-threatening conditions in high myopia: retinoschisis, macular hole, retinal detachment and pathological myopic choroidal neovascularisation. Focal Loss was used to address class imbalance, and optimal operating thresholds were determined according to the Youden Index.ResultsIn the independent test dataset, the areas under the receiver operating characteristic curves were high for all conditions (0.961 to 0.999). Our AI system achieved sensitivities equal to or even better than those of retina specialists as well as high specificities (greater than 90%). Moreover, our AI system provided a transparent and interpretable diagnosis with heatmaps.ConclusionsWe used OCT macular images for the development of CNN models to identify vision-threatening conditions in high myopia patients. Our models achieved reliable sensitivities and high specificities, comparable to those of retina specialists and may be applied for large-scale high myopia screening and patient follow-up.

scholarly journals The Association between Macular Thickness and Axial Length in Myopic Eyes

2019 ◽  
Vol 2019 ◽  
pp. 1-7
Author(s):  
Yeon Woong Chung ◽  
Moon Young Choi ◽  
Jung-sub Kim ◽  
Jin-woo Kwon
Keyword(s):  
Optical Coherence Tomography ◽  
Axial Length ◽  
High Myopia ◽  
Polynomial Regression ◽  
Macular Thickness ◽  
Optical Coherence ◽  
Regression Analyses ◽  
Foveal Thickness ◽  
P Values ◽  
The Relationship

Purpose. To investigate the relationship between macular thickness and axial length (AL) in myopic eyes. Methods. We included 441 myopic eyes in this study and measured macular thickness at the fovea and in other macular regions, using optical coherence tomography. We got thickness difference indices (TDIs) which by definition are the values of thickness difference obtained by subtracting the foveal thickness from that of each macula sector to evaluate macular contour. We then analyzed the relationships between AL and foveal thickness and AL and the TDIs of each macular sector. Results. In polynomial regression analyses, foveal thickness slope was relatively flat up to an AL of 25.5 mm and began to rise from 25.5–26.0 mm. The TDIs were also relatively flat up to AL of 25.5mm and started to show steepened negative slopes from around AL of 25.5 mm. When grouping myopia participants as high myopia or non-high myopia based on AL of 25.5mm, all macular indices of the high myopia group showed significant correlation with AL (all p values <0.01). But all indices of non-high myopia group had no significant correlation with AL. Conclusions. Average macular thickness profiles showed that appreciable changes started at an AL of 25.5mm.

Towards ‘automated gonioscopy’: a deep learning algorithm for 360° angle assessment by swept-source optical coherence tomography

2021 ◽  
pp. bjophthalmol-2020-318275
Author(s):  
Natalia Porporato ◽  
Tin A Tun ◽  
Mani Baskaran ◽  
Damon W K Wong ◽  
Rahat Husain ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Deep Learning ◽  
Diagnostic Performance ◽  
Learning Algorithm ◽  
Previous History ◽  
Angle Closure ◽  
Index Test ◽  
Optical Coherence ◽  
Cross Sectional ◽  
Swept Source

AimsTo validate a deep learning (DL) algorithm (DLA) for 360° angle assessment on swept-source optical coherence tomography (SS-OCT) (CASIA SS-1000, Tomey Corporation, Nagoya, Japan).MethodsThis was a reliability analysis from a cross-sectional study. An independent test set of 39 936 SS-OCT scans from 312 phakic subjects (128 SS-OCT meridional scans per eye) was analysed. Participants above 50 years with no previous history of intraocular surgery were consecutively recruited from glaucoma clinics. Indentation gonioscopy and dark room SS-OCT were performed. Gonioscopic angle closure was defined as non-visibility of the posterior trabecular meshwork in ≥180° of the angle. For each subject, all images were analysed by a DL-based network based on the VGG-16 architecture, for gonioscopic angle-closure detection. Area under receiver operating characteristic curves (AUCs) and other diagnostic performance indicators were calculated for the DLA (index test) against gonioscopy (reference standard).ResultsApproximately 80% of the participants were Chinese, and more than half were women (57.4%). The prevalence of gonioscopic angle closure in this hospital-based sample was 20.2%. After analysing a total of 39 936 SS-OCT scans, the AUC of the DLA was 0.85 (95% CI:0.80 to 0.90, with sensitivity of 83% and a specificity of 87%) to classify gonioscopic angle closure with the optimal cut-off value of >35% of circumferential angle closure.ConclusionsThe DLA exhibited good diagnostic performance for detection of gonioscopic angle closure on 360° SS-OCT scans in a glaucoma clinic setting. Such an algorithm, independent of the identification of the scleral spur, may be the foundation for a non-contact, fast and reproducible ‘automated gonioscopy’ in future.

scholarly journals Solar retinopathy: a new setting of red, green, and blue channels

2020 ◽  
pp. 112067212091485
Author(s):  
Davide Borroni ◽  
Renars Erts ◽  
Neeru A Vallabh ◽  
Chiara Bonzano ◽  
Svetlana Sepetiene ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Visual Acuity ◽  
Receiver Operating Characteristic Curve ◽  
Receiver Operating Characteristic ◽  
Operating Characteristic ◽  
Characteristic Curve ◽  
Spectral Domain ◽  
Optical Coherence ◽  
Operating Characteristic Curve ◽  
Receiver Operating

Purpose: To introduce a new color imaging technique using improved settings of red, green, and blue channels for improved delineation of retinal damage in patients with solar retinopathy. Method: A retrospective case series of patients with poor vision secondary to solar retinopathy were analyzed. All patients underwent visual acuity, refraction, and dilated fundus examination. A spectral domain–optical coherence tomography of the macula and color fundus imaging using optimized red, green, and blue color setting was performed. Patients were reviewed over a 6-month period. The data were analyzed for statistical significance using an independent t test and a receiver operating characteristic curve. Results: In total, 20 eyes of 10 patients were included between 2009 and 2017. The mean age was 24.9 ± 18.1 years. Best corrected visual acuity at first consultation was 0.78 ± 0.11 and after 6 months was 0.83 ± 0.09. Spectral domain–optical coherence tomography demonstrated retinal abnormalities at the myoid zone, ellipsoid zone, and the outer segment of photoreceptors. Receiver operating characteristic curve analysis showed an improving effect (area under the curve = 0.62; 95% confidence interval = 0.42–0.79). The color channels parameters, which improve visualization of the lesions were found to be 67-0.98-255 for the R-guided setting, 19-0.63-121 for the B-guided setting, and 7-1.00-129 for the G-guided setting. The ideal red, green, and blue setting was in 24-0.82-229. Conclusion: The use of a new setting of red, green, and blue channels could improve the diagnosis and monitoring of solar retinopathy, hence improving patient care.

Accuracy of intraocular lens calculation formulas using a swept-source optical coherence tomography biometer in high myopia

2021 ◽  
Author(s):  
Chi Zhang ◽  
Zi Ye ◽  
Wenqian Chen ◽  
Yi Gao ◽  
Tianju Ma ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Intraocular Lens ◽  
Axial Length ◽  
High Myopia ◽  
Absolute Error ◽  
Refractive Errors ◽  
Optical Coherence ◽  
Swept Source ◽  
The Mean ◽  
The Hill

Abstract Background: To compare the accuracy of intraocular lens (IOL) calculation formulas in cataract patients with high myopia using the measurements of a swept-source optical coherence tomography (SS-OCT) biometer, the IOLMaster700.Methods: Patients with axial length (AL) equal to or longer than 26.00mm undergoing uneventful cataract surgery were enrolled. Kane, Hill-RBF3.0, EVO, Barrett Universal II, Haigis, and SRK/T formulas were evaluated with the measurements taken by IOLMaster700. The manifest refraction was measured at one month postoperatively. After the mean refractive errors were zeroed out, the mean absolute error (MAE), the median absolute error (MedAE), the standard deviation of prediction error (SD), and the percentage of eyes with refractive errors within ±0.25, ±0.50, ±0.75, and ±1.00 diopter (D) were calculated. A subgroup analysis was based on the axial length.Results: 65 eyes of 65 patients were included. There were significant differences between the absolute refractive errors predicted by these formulas (P＜0.05), but no significant differences between the percentage of eyes within a certain range of refractive errors (P＞0.05). The Kane formula achieved the lowest MAE (0.323D), SD (0.402D), followed by EVO, Hill-RBF3.0, Barrett Universal II, Haigis, and SRK/T formulas. The Kane formula also had the highest percentage of eyes with refractive errors within ±0.50D (80.0%) and ±1.00D (98.5%). The Hill-RBF3.0 formula achieved the lowest MedAE (0.240D). In eyes with an AL ≥ 30.00mm, the Kane formula had the lowest MAE (0.358D).Conclusion: Newer formulas such as Kane, EVO, Hill-RBF3.0, and Barrett Universal II show the highest accuracy in refractive prediction in eyes with high myopia while using the measurements of IOLMaster700. In extremely myopic eyes with an AL ≥ 30.00mm, the Kane formula is the most accurate.

Correlation between Choroidal Thickness and Degree of Myopia Assessed with Enhanced Depth Imaging Optical Coherence Tomography

2017 ◽  
Vol 27 (5) ◽  
pp. 577-584 ◽  
Author(s):  
Amany A. El-Shazly ◽  
Yousra A. Farweez ◽  
Marwa E. ElSebaay ◽  
Walid M.A. El-Zawahry
Keyword(s):  
Optical Coherence Tomography ◽  
Axial Length ◽  
High Myopia ◽  
Choroidal Thickness ◽  
Significant Negative Correlation ◽  
Enhanced Depth Imaging ◽  
Temporal Region ◽  
Optical Coherence ◽  
Depth Imaging ◽  
Imaging Optical Coherence Tomography

Purpose To assess the choroidal thickness in different degrees of myopia using enhanced depth imaging optical coherence tomography (EDI-OCT) compared with healthy subjects. Methods We included 240 patients with myopia and 60 emmetropes as controls. Participants underwent full ophthalmologic examination, axial length measurement, and EDI-OCT imaging of the choroid. Choroidal thickness (CT) was measured at 5 locations, including subfoveal (SFCT), 2 mm nasal, temporal, upper, and lower to fovea. Results Choroidal thickness was significantly lower in myopic eyes compared to controls. Regardless of the degree of myopia, nasal regions showed the lowest CT with decremental pattern with advance of myopia (low myopia 279.00 ± 24.50 µm, moderate myopia 269.58 ± 20.69 µm, high myopia 189.58 ± 25.95 µm, advanced myopia 96.75 ± 24.83 µm). Highest CT was variable according to the degree of myopia with decremental pattern with advance of myopia (low myopia in subfoveal region 354.40 ± 35.14 µm, moderate myopia in temporal region 337.87 ± 35.75 µm, high myopia in lower region 312.15 ± 38.90 µm, and advanced myopia in upper region 201.25 ± 18.27 µm). Axial length showed significant negative correlation with SFCT and CT in different studied regions. Conclusions Different degrees of myopia showed thinner choroidal thickness than that of normal control eyes with decremental thinning with progress of myopia. This might be secondary to the longer axial length, which was the determining factor in some locations such as subfoveal, nasal, and upper CT.

scholarly journals Evaluation of retinal vascular density and related factors in youth myopia without maculopathy using OCTA

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Tiantian Wang ◽  
Hui Li ◽  
Rongrong Zhang ◽  
Yan Yu ◽  
Xin Xiao ◽  
...  
Keyword(s):  
Optical Coherence Tomography ◽  
Refractive Error ◽  
Axial Length ◽  
High Myopia ◽  
Optical Coherence Tomography Angiography ◽  
Flow Density ◽  
Vascular Density ◽  
Multivariate Linear Regression Analysis ◽  
Optical Coherence ◽  
Spherical Equivalent

AbstractTo evaluate the retinal vascular flow density changes of myopic eyes of young adults using optical coherence tomography angiography and the factors affecting these changes. In this cross-sectional study, 90 eyes of 45 participants were analyzed and divided into three groups: mild, moderate, and high myopia (without pathological changes). Macular and radial peripapillary capillary flow densities were measured using optical coherence tomography angiography. Their relationships with the axial length, the spherical equivalent of the refractive error, and age were analyzed using analysis of variance, Pearson’s correlation coefficient, and multivariate linear regression analysis. Superficial and deep macular vascular densities were significantly decreased in the high myopia group compared to the other groups. In the high myopia group, the nasal peripapillary flow density decreased, whereas the flow density inside the disc increased. The axial length negatively correlated with the superficial and deep macular vascular density, but positively correlated with the vascular density inside the disc. The spherical equivalent of the refractive error negatively correlated with the macular vascular density. The retinal vascular density decreased in the high myopia group. Hence, the microvascular network inside the disc may have a compensatory action in the hypoxic setting of high myopia.

Repeatability of Spectral Domain Optical Coherence Tomography Measurements in High Myopia

2016 ◽  
Vol 25 (5) ◽  
pp. e526-e530 ◽  
Author(s):  
Harsha L. Rao ◽  
Addepalli U. Kumar ◽  
Sampath R. Bonala ◽  
Kadam Yogesh ◽  
Bodduluri Lakshmi
Keyword(s):  
Optical Coherence Tomography ◽  
High Myopia ◽  
Spectral Domain ◽  
Optical Coherence
Sign in / Sign up

Export Citation Format

Share Document