A deep learning system for predicting time to progression of diabetic retinopathy.

Diabetic retinopathy (DR) is the leading cause of preventable blindness worldwide. The risk of DR progression is highly variable among different individuals, making it difficult to predict risk and personalize screening intervals. We developed and validated a deep learning system (DeepDR Plus) to predict time to DR progression within 5 years solely from fundus images. First, we used 717,308 fundus images from 179,327 participants with diabetes to pretrain the system. Subsequently, we trained and validated the system with a multiethnic dataset comprising 118,868 images from 29,868 participants with diabetes. For predicting time to DR progression, the system achieved concordance indexes of 0.754-0.846 and integrated Brier scores of 0.153-0.241 for all times up to 5 years. Furthermore, we validated the system in real-world cohorts of participants with diabetes. The integration with clinical workflow could potentially extend the mean screening interval from 12 months to 31.97 months, and the percentage of participants recommended to be screened at 1-5 years was 30.62%, 20.00%, 19.63%, 11.85% and 17.89%, respectively, while delayed detection of progression to vision-threatening DR was 0.18%. Altogether, the DeepDR Plus system could predict individualized risk and time to DR progression over 5 years, potentially allowing personalized screening intervals.

Anatomical and Functional Outcomes in Delayed Onset versus Concurrent Retinal Detachment in Endophthalmitis.

Purpose: To determine the functional and anatomical outcomes of patients with endophthalmitis with concurrent or delayed onset retinal detachment (RD), and compare the preoperative, intraoperative and postoperative features. Patients and Methods: This was a retrospective review of 121 eyes in 121 patients presenting with endophthalmitis and RD. Subjects were categorized into two groups: endophthalmitis with delayed onset RD (group 1, N=76) and endophthalmitis with concurrent RD (group 2, N=45). Results: The mean age of patients in groups 1 and 2 was 38.21±21.60 and 46.78±24.42 years, respectively (P=0.047). Exogenous endophthalmitis was common in both groups 1 and 2 (86.84% and 84.44%, respectively). No significant differences were found between the groups in the type of RD, retinal breaks, number of quadrants involved or proliferative vitreoretinopathy grade. In the overall cohort, visual acuity improved post-surgery in one-third of the patients who were in the near or total blindness category at presentation. We found good anatomical success rates of an attached retina in both groups 1 and 2 (84.3% and 77.7%, P=0.376). Conclusion: Our study presents the results of patients with endophthalmitis and delayed onset RD or concurrent RD. It shows a few differences in presentation between the groups, but the anatomical and functional outcomes were almost the same.

Morphological and functional retinal changes in neovascular age-related macular degeneration treated with intravitreal bevacizumab.

Purpose: This study was conducted to determine the morphological and functional retinal changes in patients with neovascular age-related macular degeneration (nAMD) treated with intravitreal bevacizumab 1.25 mg. Methods: This was a prospective, nonrandomized, interventional study. Eighteen eyes of 18 subjects with nAMD were treated with intravitreal bevacizumab (1.25 mg) injection. Subjects underwent complete ophthalmic evaluation which included visual acuity, slitlamp examination, tonometry, binocular ophthalmoscopy, optical coherence tomography (OCT), and MP1 microperimetry before the intravitreal injection and the follow-up at 1 and 3 months. Test of significance such as Chi-squared test, paired ttest and oneway analysis of variance (ANOVA) linear trend were used to compare the pre- and post-anti-VEGF outcomes. Intraclass correlation was done to assess the intra observer variability. Results: Mean retinal sensitivity had increased from 3.77 ± 3.13 dB at baseline to 4.93 ± 2.42 dB at 3 months (P = 0.05). Visual acuity improved from 0.62 ± 0.36 at baseline to 0.52 ± 0.36 at 1 month and 0.48 ± 0.34 at 3-month followup, but overall change was not significant (P = 0.40). There was a significant reduction in central foveal thickness (CFT) from 274.61 ± 117.95 at baseline to 179.83 ± 84.18 at 1 month and 179.00 ± 126.55 at 3-month follow-up (P = 0.013). Conclusion: Intravitreal bevacizumab (1.25 mg) injection in nAMD improves retinal function, quantified by retinal sensitivity, scotoma characteristics, fixation stability by MP 1 microperimetry and morphological parameters quantified by CFT in SDOCT. These changes show the effectiveness of treatment with intravitreal bevacizumab in nAMD.

Influence of Different Types of Retinal Cameras on the Performance of Deep Learning Algorithms in Diabetic Retinopathy Screening.

Background: The aim of this study was to assess the performance of regional graders and artificial intelligence algorithms across retinal cameras with different specifications in classifying an image as gradable and ungradable. Methods: Study subjects were included from a community-based nationwide diabetic retinopathy screening program in Thailand. Various non-mydriatic fundus cameras were used for image acquisition, including Kowa Nonmyd, Kowa Nonmyd α-DⅢ, Kowa Nonmyd 7, Kowa Nonmyd WX, Kowa VX 10 α, Kowa VX 20 and Nidek AFC 210. All retinal photographs were graded by deep learning algorithms and human graders and compared with a standard reference. Results: Images were divided into two categories as gradable and ungradable images. Four thousand eight hundred fifty-two participants with 19,408 fundus images were included, of which 15,351 (79.09%) were gradable images and the remaining 4057 (20.90%) were ungradable images. Conclusions: The deep learning (DL) algorithm demonstrated better sensitivity, specificity and kappa than the human graders for all eight types of non-mydriatic fundus cameras. The deep learning system showed, more consistent diagnostic performance than the human graders across images of varying quality and camera types.

Distribution of macular pigments in macular telangiectasia type 2 and correlation with optical coherence tomography characteristics and visual acuity.

BACKGROUND: To estimate macular pigment values in macular telangiectasia (MacTel) Type 2 in comparison with healthy subjects in the South Indian population across different spatial profiles and to quantify the regional differences of macular pigment optical density (MPOD) in MacTel Type 2. METHODS: In this prospective cross-sectional study, healthy controls and patients diagnosed with various stages of MacTel Type 2 underwent MPOD measurement using dual-wavelength autofluorescence technique with Spectralis HRA + OCT. RESULTS: Sixty eyes of 31 healthy subjects and 41 eyes of 22 MacTel type 2 patients were included. We found an overall decrease in MPOD values in MacTel type 2 patients (-0.109, -0.11, -0.001) in comparison with healthy subjects (0.38, 0.23, 0.06) at 1°, 2° & 6° foveal eccentricities (P < 0.001). In various stages of MacTel type 2, the mean MPOD was found to be higher in the peripheral region compared to the central region. We found a significantly lower mean MPOD in the central region in association with specific optical coherence tomography (OCT) parameters like inner retinal cavities (P = 0.035) and ellipsoid zone disruption (P = 0.034). CONCLUSIONS: In MacTel type 2, MPOD distribution varies in different spatial profiles with higher MPOD levels in the peripheral region compared to the central region. The macular pigment levels are associated with inner retinal cavities and ellipsoid zone disruption seen on OCT.

Prevalence, risk factors and association with glycemic levels of presbyopia in South Indian population.

PURPOSE: To determine the prevalence of presbyopia and its association with elevated glycemic levels in subjects ≥40 years of age in the South Indian population of Chennai. METHODS: This was a retrospective study. Subjects were included from the Sankara Nethralaya Diabetic Retinopathy Epidemiology And Molecular genetics Study (SN-DREAMS 1). Demographic data, detailed medical and ocular history, comprehensive eye examination, and biochemical investigations were performed. Glycosylated hemoglobin results were categorized as controls (4%-5.6%), prediabetic (5.7%-6.4%), and diabetic (≥6.5%) groups. The given presbyopic correction was divided into two groups as within and outside donders limit. Prevalence rates and mean values were determined and compared among the three glycemic groups. The Student t test, the Chi-square test, and multivariate logistic regression analyses were performed. RESULTS: The overall prevalence of presbyopia from our previously conducted SN-DREAMS 1 population of 1414 patients was 79.77% (95% CI: 0.775-0.818). In total, 1128 participants were included for our current secondary analysis with a mean age of 54.40 years (range: 40-83). The number of subjects within and outside donders limit was 1044 (92.55%) and 84 (7.44%), respectively. In each age group (40-49, 50-59, ≥60) regardless of being within or outside donders limits, an increasing trend in the prevalence of presbyopia was noted based on increasing glycemic levels. CONCLUSION: Our study demonstrated a high prevalence of presbyopia in the South Indian population of Chennai. Findings show that the prevalence of presbyopia in different age groups increases with worsening diabetes status.

Visual outcomes following cataract surgery with intraocular lens implantation in vitrectomized eyes among children.

Purpose: To assess the visual outcomes and associated factors in pediatric patients undergoing cataract surgery following pars plana vitrectomy (PPV). Methods: A total of 52 eyes of 52 children who underwent cataract surgery and intraocular lens (IOL) implantation in previously vitrectomized eyes between January 2008 and December 2017 were included in this retrospective study. Descriptive and inferential statistical measurements (Chi-square test, Fisher's exact test, odds ratio, multiple logistic regression) were done. Results: The mean age at PPV was 11.3 ± 2.9 years. The most common indication for PPV was retinal detachment in 43 eyes (82.69%) and vitreous hemorrhage in 6 eyes (11.53%). The mean preoperative and postoperative best-corrected visual acuity (BCVA) was 1.31 ± 0.48 logMAR (logarithm of the minimum angle of resolution) units (20/400) and 0.69 ± 0.45 logMAR units (20/100), respectively; P < 0.001. The final visual outcome was good in 37 eyes (71.15%) and poor in 15 eyes (28.85%). In mean follow-up of 4 years, 42.3% developed posterior capsular opacification, 13.5% developed glaucoma, and 3.8% had retinal redetachment after cataract surgery. In the presence of macular pathology, the visual outcome was poor (aOR [adjusted odds ratio] = 4.26, P = 0.002). Conclusion: Cataract surgery with IOL implantation in vitrectomized eyes among children is a safe procedure and can improve visual acuity and quality of life. However, the extent of improvement is limited by macular pathology.

Association Between Gut Microbial Abundance and Sight-Threatening Diabetic Retinopathy.

Purpose: To study the association between gut microbial abundance and sight-threatening diabetic retinopathy among patients with a history of type 2 diabetes mellitus. Methods: An observational case-control study was performed using a sample population of diabetics referred to a tertiary eye institute. Sample subjects were identified as cases if they were diagnosed with sight-threatening diabetic retinopathy and controls if they were not but had at least a 10-year history of diabetes. Fecal swabs for all patients were collected for enumeration and identification of sequenced gut microbes. Statistical analyses were performed to associate the clinically relevant Bacteroidetes to Firmicutes relative abundance ratio (B/F ratio) with sight-threatening diabetic retinopathy and an optimal cutoff value for the ratio was identified using Youden's J statistics. Results: A sample size of 58 diabetic patients was selected (37 cases, 21 controls). No statistically significant difference in the relative abundance among the predominant phyla between the groups were found. In our univariate analysis, the B/F ratio was elevated in cases compared to controls (cases, 1.45; controls, 0.94; P = 0.049). However, this statistically significant difference was not seen in our multivariate regression model. Optimal cutoff value of 1.05 for the B/F ratio was identified, and significant clustering of cases above this value was noted in beta diversity plotting. Conclusions: No difference in gut microbial abundance for any particular phylum was noted between the control and diseased population. Increased gut microbial B/F ratio can be a potential biomarker for the development of sight-threatening diabetic retinopathy among type 2 diabetic patients.

Incidence and risk factors for retinal detachment following laser-assisted in-situ keratomileusis.

Purpose: The aim of this work was to study the incidence and potential risk factors for development of retinal detachment (RD) after laser-assisted in situ keratomileusis (LASIK) surgery over a long-term follow-up. Methods: This was a retrospective interventional case series. A total of 694 eyes of 352 patients were included who had undergone LASIK surgery at a tertiary eye care institute between January 2005 and September 2015 who had a minimum follow-up of 5 years after the surgery were included. Kaplan-Meier analysis and Cox proportion regression model was used to estimate the potential risk factors and cumulative risk for the development of RD. Results: Out of the total patients, 5 eyes developed RD after a mean follow-up of 7 years. The cumulative risk of RD after 1 year was 0.4%, after 2 years was 0.5% and after 7 years was 0.7%. Increased risk of RD was not associated with age at LASIK surgery, gender, laterality, and spherical equivalent (P > 0.05). A significantly increased risk for the development of RD was observed in eyes that were given prior prophylactic laser photocoagulation for peripheral lesions in multivariate cox proportional regression analysis with a hazard ratio (HR) of 9.33 (CI- 1.554-56.094; P = 0.015). Conclusion: We emphasize the need for a regular follow-up after the LASIK procedure to ensure timely treatment of any new retinal lesions.

Accuracy of Intraocular Lens Power Calculation in Children With Vitrectomized Eyes Undergoing Cataract Surgery.

PURPOSE: To assess the predictability of desired postoperative refractive outcomes using the SRK-II formula for intraocular lens (IOL) power calculation in children undergoing cataract surgery in eyes with a previous pars plana vitrectomy (PPV). METHODS: In this retrospective study, 68 eyes of 66 children who underwent cataract surgery and IOL implantation in eyes that had previous vitrectomy between January 2008 and December 2017 were included. Data were collected on preoperative and postoperative characteristics. The Mann-Whitney test, Kruskal-Wallis test, and multinominal logistic regression were used for comparing the results. RESULTS: Absolute prediction error (APE) in the cohort was 1.29 ± 1.13. Desirable refractive outcome with insignificant prediction error of less than 0.50 diopters (D) was found in approximately one-third of the children. Patients were further subdivided by magnitude of APE as ≥ 0.50 to ≤ 1.00 D and > 1.00 D. Age at the time of surgery, axial length, mean keratometry, silicone oil removal, IOL position, scleral buckle, and corneal suture did not affect APE and there was no significant difference between the groups (P > .05). CONCLUSIONS: The mean APE of the SRK-II formula for IOL power calculation in pediatric eyes that had vitrectomy is comparable to that reported in the literature for routine pediatric cataract surgeries. Factors such as age, axial length, corneal power, IOL position (bag/sulcus), scleral buckle, corneal suture, and silicone oil removal done prior to surgery or along with cataract surgery did not affect the APE. [J Pediatr Ophthalmol Strabismus. 2021;58(2):126-131.].

Evaluation of Macular Pigment Optical Density in Healthy Eyes Based on Dual-Wavelength Autofluorescence Imaging in South Indian Population.

Purpose: To estimate macular pigment optical density (MPOD) values across different age groups in the South Indian population across various spatial profiles using dual-wavelength autofluorescence. Methods: Sixty eyes of 31 healthy subjects underwent MPOD measurement with Spectralis HRA+OCT. The average MPOD and macular pigment optical volume (MPOV) at 1°, 2°, and 6° radii, the mean MPOD in the classical Early Treatment Diabetic Retinopathy Study (ETDRS) grid, and the spatial profiles of two different age groups across 12 plots covering the radial sectors were recorded. Results: The mean age was 39.1 ± 12.7 years. The mean MPOD and MPOV values were 0.38 ± 0.11 and 787.95 ± 225.13 at 1° eccentricity, 0.23 ± 0.08 and 2000 ± 708.24 at 2° eccentricity, and 0.05 ± 0.02 and 4335 ± 2007.71 at 6° eccentricity, respectively. In the ETDRS grid, the mean MPOD was found to be highest in the central sector and lowest in the inferior peripheral ring. We also found that along the radial sectors the lower quadrants tended to have low MPOD as compared to the upper quadrants. Subjects 40 years of age or older had significantly higher averaged MPOD in certain areas (-15° to 15° and 75° to 105°) along the radial sectors than subjects less than 40 years of age. Conclusions: This study establishes a reference value for future studies of diseased eyes in the South Indian population. Translational Relevance: Our study is unique in that it reports MPOD among the South Indian population across different age groups, as well as the distribution of MPOD in all nine zones of the classical ETDRS grid and various spatial profiles covering the 30° radial sectors centered on the fovea.

Fluorophotometric Measurement of Corneal Autofluorescence in Diabetic and Normal Eyes.

INTRODUCTION: Corneal autofluorescence (AF) has been recommended as an indicator of diabetic retinopathy. This study was conducted to evaluate corneal autofluorescence in non-insulin dependent diabetes subjects (with and without diabetic retinopathy) and compare the changes with healthy subjects. MATERIALS AND METHODS: In this cross-sectional study, 145 eyes of diabetes mellitus subjects with and without diabetic retinopathy and 34 eyes of normal healthy subjects were included. Diabetic patients were subdivided by the severity of retinopathy based on international clinical diabetic retinopathy disease severity scale. The investigations included fluorophotometric determination of corneal autofluorescence, fasting glucose level (FBS) and glycosylated hemoglobin (HbA1c). RESULTS: The corneal autofluorescence values were significantly higher in moderate NPDR, severe NPDR and PDR when compared with healthy subjects (p<0.001). FBS values and HbA1c were significantly higher in all diabetic groups than the healthy subjects (p<0.001). In linear regression test, corneal autofluorescence seem to be related to duration of diabetes and FBS, but not with age and HbA1c. CONCLUSION: We conclude that the level of corneal autofluorescence is dependent on the severity of diabetic retinopathy. Our study suggests that the formation of autofluorescent products in cornea depend upon the glucose concentration in the aqueous humor.