The Classification of Common Macular Diseases Using Deep Learning on Optical Coherence Tomography Images with and without Prior Automated Segmentation.

We compared the performance of deep learning (DL) in the classification of optical coherence tomography (OCT) images of macular diseases between automated classification alone and in combination with automated segmentation. OCT images were collected from patients with neovascular age-related macular degeneration, polypoidal choroidal vasculopathy, diabetic macular edema, retinal vein occlusion, cystoid macular edema in Irvine-Gass syndrome, and other macular diseases, along with the normal fellow eyes. A total of 14,327 OCT images were used to train DL models. Three experiments were conducted: classification alone (CA), use of automated segmentation of the OCT images by RelayNet, and the graph-cut technique before the classification (combination method 1 (CM1) and 2 (CM2), respectively). For validation of classification of the macular diseases, the sensitivity, specificity, and accuracy of CA were found at 62.55%, 95.16%, and 93.14%, respectively, whereas the sensitivity, specificity, and accuracy of CM1 were found at 72.90%, 96.20%, and 93.92%, respectively, and of CM2 at 71.36%, 96.42%, and 94.80%, respectively. The accuracy of CM2 was statistically higher than that of CA (p = 0.05878). All three methods achieved AUC at 97%. Applying DL for segmentation of OCT images prior to classification of the images by another DL model may improve the performance of the classification.

The Association Between Environmental and Social Factors and Myopia: A Review of Evidence From COVID-19 Pandemic.

Purpose: To review the association between children's behavioral changes during the restriction due to the pandemic of Coronavirus disease (COVID-19) and the development and progression of myopia. Design: A literature review. Method: We looked for relevant studies related to 1) children's behavioral changes from COVID-19 restriction and 2) children's myopia progression during COVID-19 restriction by using the following keywords. They were "Behavior," "Activity," "COVID-19," "Lockdown," "Restriction," and "Children" for the former; "Myopia," "COVID-19," "Lockdown," "Restriction" for the latter. Titles, abstracts and full texts from the retrieved studies were screened and all relevant data were summarized, analyzed, and discussed. Results: Children were less active and more sedentary during COVID-19 restriction. According to five studies from China and six studies, each from Hong Kong, Spain, Israel, South Korea, Turkey and Taiwan included in our review, all countries without myopia preventive intervention supported the association between the lockdown and myopia progression by means of negative SER change ranging from 0.05-0.6 D, more negative SER change (compared post- to pre-lockdown) ranging from 0.71-0.98 D and more negative rate of SER changes (compared post- to pre-lockdown) ranging from 0.05-0.1 D/month. The reported factor that accelerated myopia is an increase in total near work, while increased outdoor activity is a protective factor against myopia progression. Conclusion: The pandemic of COVID-19 provided an unwanted opportunity to assess the effect of the behavioral changes and myopia in the real world. There is sufficient evidence to support the association between an increase in near work from home confinement or a reduction of outdoor activities and worsening of myopia during the COVID-19 lockdown. The findings from this review of data from the real world may help better understanding of myopia development and progression, which may lead to adjustment of behaviors to prevent myopia and its progression in the future.

Capacity building in screening and treatment of diabetic retinopathy in Asia-Pacific region.

The focus of capacity building for screening and treatment of diabetic retinopathy (DR) is on health professionals who are nonophthalmologists. Both physicians and nonphysicians are recruited for screening DR. Although there is no standardization of the course syllabus for the capacity building, it is generally accepted to keep their sensitivity >80%, specificity >95%, and clinical failure rate <5% for the nonophthalmologists, if possible. A systematic literature search was performed using the PubMed database and the following search terms: diabetic retinopathy, diabetic retinopathy screening, Asia, diabetic retinopathy treatment, age-related macular degeneration, capacity building, deep learning, artificial intelligence (AI), nurse-led clinic, and intravitreal injection (IVI). AI may be a tool for improving their capacity. Capacity building on IVIs of antivascular endothelial growth factors for DR is focused on nurses. There is evidence that, after a supervision of an average of 100 initial injections, the trained nurses can do the injections effectively and safely, the rate of endophthalmitis ranges from 0.03 to 0.07%, comparable to ophthalmologists. However, laws and regulations, which are different among countries, are challenges and barriers for nonophthalmologists, particularly for nonphysicians, for both screening and treatment of DR. Even if nonphysicians or physicians who are nonophthalmologists are legally approved for these tasks, sustainability of the capacity is another important challenge, this may be achieved if the capacity building can be part of their career development. Patient acceptability is another important barrier for initiating care provided by nonophthalmologists, particularly in Asia. There are also collaborations between national eye institutes of high-income countries, nongovernment organizations, and local eye institutes to improve both the quality and quantity of ophthalmologists and retinal specialists in low-income countries in Asia. This approach may require more labor, cost, and time consuming than training nonophthalmologists.

Longitudinal Screening for Diabetic Retinopathy in a Nationwide Screening Program: Comparing Deep Learning and Human Graders.

OBJECTIVE: To evaluate diabetic retinopathy (DR) screening via deep learning (DL) and trained human graders (HG) in a longitudinal cohort, as case spectrum shifts based on treatment referral and new-onset DR. METHODS: We randomly selected patients with diabetes screened twice, two years apart within a nationwide screening program. The reference standard was established via adjudication by retina specialists. Each patient's color fundus photographs were graded, and a patient was considered as having sight-threatening DR (STDR) if the worse eye had severe nonproliferative DR, proliferative DR, or diabetic macular edema. We compared DR screening via two modalities: DL and HG. For each modality, we simulated treatment referral by excluding patients with detected STDR from the second screening using that modality. RESULTS: There were 5,738 patients (12.3% STDR) in the first screening. DL and HG captured different numbers of STDR cases, and after simulated referral and excluding ungradable cases, 4,148 and 4,263 patients remained in the second screening, respectively. The STDR prevalence at the second screening was 5.1% and 6.8% for DL- and HG-based screening, respectively. Along with the prevalence decrease, the sensitivity for both modalities decreased from the first to the second screening (DL: from 95% to 90%, p = 0.008; HG: from 74% to 57%, p < 0.001). At both the first and second screenings, the rate of false negatives for the DL was a fifth that of HG (0.5-0.6% vs. 2.9-3.2%). CONCLUSION: On 2-year longitudinal follow-up of a DR screening cohort, STDR prevalence decreased for both DL- and HG-based screening. Follow-up screenings in longitudinal DR screening can be more difficult and induce lower sensitivity for both DL and HG, though the false negative rate was substantially lower for DL. Our data may be useful for health-economics analyses of longitudinal screening settings.

Predicting optical coherence tomography-derived diabetic macular edema grades from fundus photographs using deep learning.

Center-involved diabetic macular edema (ci-DME) is a major cause of vision loss. Although the gold standard for diagnosis involves 3D imaging, 2D imaging by fundus photography is usually used in screening settings, resulting in high false-positive and false-negative calls. To address this, we train a deep learning model to predict ci-DME from fundus photographs, with an ROC-AUC of 0.89 (95% CI: 0.87-0.91), corresponding to 85% sensitivity at 80% specificity. In comparison, retinal specialists have similar sensitivities (82-85%), but only half the specificity (45-50%, p < 0.001). Our model can also detect the presence of intraretinal fluid (AUC: 0.81; 95% CI: 0.81-0.86) and subretinal fluid (AUC 0.88; 95% CI: 0.85-0.91). Using deep learning to make predictions via simple 2D images without sophisticated 3D-imaging equipment and with better than specialist performance, has broad relevance to many other applications in medical imaging.

Deep learning versus human graders for classifying diabetic retinopathy severity in a nationwide screening program.

Deep learning algorithms have been used to detect diabetic retinopathy (DR) with specialist-level accuracy. This study aims to validate one such algorithm on a large-scale clinical population, and compare the algorithm performance with that of human graders. A total of 25,326 gradable retinal images of patients with diabetes from the community-based, nationwide screening program of DR in Thailand were analyzed for DR severity and referable diabetic macular edema (DME). Grades adjudicated by a panel of international retinal specialists served as the reference standard. Relative to human graders, for detecting referable DR (moderate NPDR or worse), the deep learning algorithm had significantly higher sensitivity (0.97 vs. 0.74, p < 0.001), and a slightly lower specificity (0.96 vs. 0.98, p < 0.001). Higher sensitivity of the algorithm was also observed for each of the categories of severe or worse NPDR, PDR, and DME (p < 0.001 for all comparisons). The quadratic-weighted kappa for determination of DR severity levels by the algorithm and human graders was 0.85 and 0.78 respectively (p < 0.001 for the difference). Across different severity levels of DR for determining referable disease, deep learning significantly reduced the false negative rate (by 23%) at the cost of slightly higher false positive rates (2%). Deep learning algorithms may serve as a valuable tool for DR screening.

Erratum: Author Correction: Deep learning versus human graders for classifying diabetic retinopathy severity in a nationwide screening program.

[This corrects the article DOI: 10.1038/s41746-019-0099-8.].

Sutureless 25-Gauge Vitrectomy for Rhegmatogenous Retinal Detachment Caused by Superior Breaks Using Air Tamponade.

PURPOSE: This study was aimed to evaluate the efficacy and safety of 25-gauge sutureless vitrectomy in repairing primary rhegmatogenous retinal detachment (RRD) with air tamponade. DESIGN: This is a prospective, clinic-based, case series. METHODS: Twenty consecutive eyes of 20 patients with primary RRD caused by superior breaks of less than a month underwent transconjunctival sutureless 25-gauge vitrectomy with intraocular air tamponade. Patients who had a follow-up of less than 6 months were excluded. Outcome measures included best corrected visual acuity (BCVA), reattachment rate by a single procedure, final reattachment rate by additional procedures, and complications. RESULTS: The mean follow-up was 10 months (range, 6-15 months). The proportion of eyes with BCVA of between 20/200 and 20/70 increased significantly from 15% at baseline to 65% on day 14 (P = 0.024). At final follow-up, 15%, 60%, and 25% had BCVA worse than 20/200, between 20/200 and 20/70, and better than 20/70, respectively. The mean BCVA was significantly better than baseline (logMAR, 1.4) by day 14 (logMAR, 0.87). The reattachment rate by a single procedure was 70%, and the final success rate was 100% after 1 additional procedure. The primary success rate increased to 77.8% after excluding 2 eyes with proliferative vitreoretinopathy grade C1. High myopia and large retinal break were 2 other conditions associated with failed primary reattachment. No postoperative complication was observed. CONCLUSIONS: Selected eyes with primary RRD may gain the benefit of early visual recovery when treated with 25-gauge vitrectomy and air tamponade.