Development of LuxIA, a Cloud-Based AI Diabetic Retinopathy Screening Tool Using a Single Color Fundus Image.

Purpose: Diabetic retinopathy (DR) is the leading cause of vision impairment in working-age adults. Automated screening can increase DR detection at early stages at relatively low costs. We developed and evaluated a cloud-based screening tool that uses artificial intelligence (AI), the LuxIA algorithm, to detect DR from a single fundus image. Methods: Color fundus images that were previously graded by expert readers were collected from the Canarian Health Service (Retisalud) and used to train LuxIA, a deep-learning-based algorithm for the detection of more than mild DR. The algorithm was deployed in the Discovery cloud platform to evaluate each test set. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve were computed using a bootstrapping method to evaluate the algorithm performance and compared through different publicly available datasets. A usability test was performed to assess the integration into a clinical tool. Results: Three separate datasets, Messidor-2, APTOS, and a holdout set from Retisalud were evaluated. Mean sensitivity and specificity with 95% confidence intervals (CIs) reached for these three datasets were 0.901 (0.901-0.902) and 0.955 (0.955-0.956), 0.995 (0.995-0.995) and 0.821 (0.821-0.823), and 0.911 (0.907-0.912) and 0.880 (0.879-0.880), respectively. The usability test confirmed the successful integration of LuxIA into Discovery. Conclusions: Clinical data were used to train the deep-learning-based algorithm LuxIA to an expert-level performance. The whole process (image uploading and analysis) was integrated into the cloud-based platform Discovery, allowing more patients to have access to expert-level screening tools. Translational Relevance: Using the cloud-based LuxIA tool as part of a screening program may give diabetic patients greater access to specialist-level decisions, without the need for consultation.

The role of retinal fluid location in atrophy and fibrosis evolution of patients with neovascular age-related macular degeneration long-term treated in real world.

PURPOSE: To assess the effect of clinical factors on the development and progression of atrophy and fibrosis in patients with neovascular age-related macular degeneration (nAMD) receiving long-term treatment in the real world. METHODS: An ambispective 36-month multicentre study, involving 359 nAMD patients from 17 Spanish hospitals treated according to the Spanish Vitreoretinal Society guidelines, was designed. The influence of demographic and clinical factors, including the presence and location of retinal fluid, on best-corrected visual acuity (BCVA) and progression to atrophy and/or fibrosis were analysed. RESULTS: After 36 months of follow-up and an average of 13.8 anti-VEGF intravitreal injections, the average BCVA gain was +1.5 letters, and atrophy and/or fibrosis were present in 54.8% of nAMD patients (OR = 8.54, 95% CI = 5.85-12.47, compared to baseline). Atrophy was associated with basal intraretinal fluid (IRF) (OR = 1.87, 95% CI = 1.09-3.20), whereas basal subretinal fluid (SRF) was associated with a lower rate of atrophy (OR = 0.40, 95% CI = 0.23-0.71) and its progression (OR = 0.44, 95% CI = 0.26-0.75), leading to a slow progression rate (OR = 0.34, 95% CI = 0.14-0.83). Fibrosis development and progression were related to IRF at any visit (p < 0.001). In contrast, 36-month SRF was related to a lower rate of fibrosis (OR = 0.49, 95% CI = 0.29-0.81) and its progression (OR = 0.50, 95% CI = 0.31-0.81). CONCLUSION: Atrophy and/or fibrosis were present in 1 of 2 nAMD patients treated for 3 years. Both, especially fibrosis, lead to vision loss. Subretinal fluid (SRF) was associated with good visual outcomes and lower rates of atrophy and fibrosis, whereas IRF yields worse visual results and a higher risk of atrophy and especially fibrosis in routine clinical practice.

Use of intravitreal dexamethasone implants in the treatment of diabetic macular edema: Expert recommendations using a Delphi approach.

OBJECTIVE: The aim of this study is to develop guidance on the use of intravitreal dexamethasone implants in the treatment of diabetic macular edema. METHOD: The study was performed using the modified Delphi method to obtain a consensus among a panel of experts on management of patients with diabetic macular edema and use of intravitreal dexamethasone implants in clinical practice. Thirty-seven panel members, experts on retina, from different Spanish centers were invited to participate. Individual and anonymous opinions were asked by answering a 76-item questionnaire across 11 topic areas (two rounds were done). Level of agreement was assessed using a Likert-type scale of 9 points. RESULTS: Agreement on "consensus" was reached during the first round in 63 items. The 13 remaining items underwent a second round of voting. After the second round, agreement on "consensus" was reached on five items. Finally, eight items remained without consensus. CONCLUSION: Intravitreal dexamethasone implants are useful in the treatment of patients with diabetic macular edema with different profiles, for example, pseudophakic, poor-adherents, vitrectomized, candidates for cataract surgery, patients with high inflammatory component, and with a history of cardiovascular events. The use of intravitreal dexamethasone reduces the number of visits and facilitates compliance. Experts thought that the switch from anti-vascular endothelial growth factor therapy to intravitreal dexamethasone implants should be done preferably after three injections. Also, pro re nata treatment provides better results in diabetic macular edema patients as it helps to prevent undertreatment. Finally, experts concluded that clinical guidelines and treatment protocols for diabetic macular edema need to be updated.

Efficacy and Safety of an Aflibercept Treat-and-Extend Regimen in Treatment-Naïve Patients with Macular Oedema Secondary to Central Retinal Vein Occlusion (CRVO): A Prospective 12-Month, Single-Arm, Multicentre Trial.

OBJECTIVES: To evaluate efficacy and safety of an aflibercept treat-and-extend (TAE) regimen in patients with macular oedema (MO) secondary to central retinal vein occlusion (CRVO). DESIGN SETTING AND PATIENTS: Phase IV, prospective, open-label, single-arm trial in 11 Spanish hospitals. Treatment-naïve patients with <6 month diagnosis of MO secondary to CRVO and best-corrected visual acuity (BCVA) of 73-24 ETDRS letters were included between 23 January 2015 and 17 March 2016. INTERVENTION: Intravitreal aflibercept 2 mg monthly (3 months) followed by proactive individualized dosing. MAIN OUTCOMES: Mean change in BCVA after 12 months. RESULTS: 24 eyes (24 patients) were included; mean (SD) age: 62.8 (15.0) years; 54.2% male; median (IQR) time since diagnosis: 7.6 (3.0, 15.2) days. Mean BCVA scores significantly improved between baseline (56.0 (16.5)) and Month 12 (74.1 (17.6)); mean (95% CI) change: 14.8 (8.2, 21.4); P=0.0001. Twelve (50.0%) patients gained ≥15 ETDRS letters. Foveal thickness improved between baseline (mean: 569.4 (216.8) µm) and Month 12 (mean 257.4 (48.4) µm); P < 0.0001. At Month 12, 8.3% patients had MO. The mean (SD) number of injections: 8.3 (3.0). No treatment-related AEs were reported. Five (20.8%) patients experienced ocular AEs. Two nonocular serious AEs were reported. CONCLUSIONS: An aflibercept TAE regimen improves visual acuity in patients with MO secondary to CRVO over 12 months with good tolerability.

Clinical Decision-Making when Treating Diabetic Macular Edema Patients with Dexamethasone Intravitreal Implants.

Diabetes mellitus (DM) is a metabolic disease frequently associated with comorbidities that include diabetic macular edema (DME). The current medical approach to treating DME involves intravitreal injections with either anti-vascular endothelial growth factors or steroids. However, the burden associated with intravitreal injections and DM-derived complications is high, underlining the need to find optimal treatment regimens. In this article we describe the considerations we apply when treating DME patients with dexamethasone intravitreal implants (Ozurdex®), particularly those that influence the clinical decision-making process during the follow-up period. These considerations are based both on the available medical literature and on our clinical experience following the use of these implants in this type of patient, the goal being to optimize the number of injections and the clinical outcome of this therapy. We also provide a general overview of the pathophysiology of DME, highlighting the inflammatory component as a rationale to use steroids in these patients.

Expression of Angiogenic MicroRNAs in Endothelial Progenitor Cells From Type 1 Diabetic Patients With and Without Diabetic Retinopathy.

PURPOSE: MicroRNA (miR) expression in endothelial progenitor cells (EPCs) in type 1 diabetes (DM1) and its relation with different stages of diabetic retinopathy (DR) have not been reported to date. Our aim was to analyze miR-222, miR-221, and miR-126 expression in EPCs from DM1 patients with and without DR. METHODS: We included 41 patients with DR, 35 without DR, and 38 controls. Blood was collected for flow cytometry and EPC culture. Total RNA was extracted and purified and real-time quantitative PCR was performed for miR expression in cultured EPCs. Relative changes in miR expression were analyzed with the 2-ΔΔCT method. RESULTS: Circulating EPCs were reduced and miR-126 expression was increased in DM1 compared to controls (0.030 [interquartile range [IQR], 0.020-0.050] vs. 0.060 [IQR, 0.030-0.110], P = 0.004; 1.740 [IQR, 0.890-4.120] vs. 0.990 [IQR, 0.487-3.015], P = 0.047 respectively) without differences between patients with and without DR. Patients with DR had higher expression of miR-221 than those without DR (1.405 [IQR, 0.820-2.867] vs. 0.915 [IQR, 0.507-1.292], P = 0.019) without differences among degrees of DR. Circulating EPCs were reduced in patients on statins (0.010 [IQR, 0.010-0.050] vs. 0.045 [IQR, 0.020-0.087], P = 0.008), and miR-221 expression increased in patients on angiotensin-converting enzyme (ACE) inhibitors/angiotensin receptor blocker (ARB) II (1.430 [IQR, 1.160-2.705] vs. 1.000 [IQR, 0.520-1.330], P = 0.021) compared to those without treatment. MicroRNA-126 expression was associated with body mass index (BMI; ρ = -0.267, P = 0.026) and diastolic blood pressure (ρ = -0.267, P = 0.034). MicroRNA-221 was associated with triglyceride concentration (ρ = 0.296, P = 0.012). CONCLUSIONS: Circulating EPCs were reduced and miR-126 expression was increased in DM1 compared to controls. Patients with DR had higher expression of miR-221 than those without DR. The identification of biomarkers of diabetic complications might be useful for monitoring disease progression and potential therapeutic targets.

Safety and efficacy of a flexible dosing regimen of ranibizumab in neovascular age-related macular degeneration: the SUSTAIN study.

OBJECTIVE: To evaluate the safety and efficacy of individualized ranibizumab treatment in patients with neovascular age-related macular degeneration. DESIGN: Twelve-month, phase III, multicenter, open-label, single-arm study. PARTICIPANTS: A total of 513 ranibizumab-naïve SUSTAIN patients. INTERVENTION: Three initial monthly injections of ranibizumab (0.3 mg) and thereafter pro re nata (PRN) retreatment for 9 months based on prespecified retreatment criteria. Patients switched to 0.5 mg ranibizumab after approval in Europe. MAIN OUTCOME MEASURES: Frequency of adverse events (AEs), monthly change of best-corrected visual acuity (BCVA) and central retinal thickness (CRT) from baseline, the time to first re-treatment, and the number of treatments were assessed. RESULTS: A total of 249 patients (48.5%) reported ocular AEs, and 8 (1.5%) deaths, 5 (1.2%) patients with ocular serious AEs of the study eye (retinal hemorrhage, cataract, retinal pigment epithelial tear, reduced visual acuity [VA], vitreous hemorrhage), and 19 (3.7%) patients with arteriothromboembolic events were observed. Most frequent AEs in the study eye were reduced VA (18.5%), retinal hemorrhage (7.2%), increased intraocular pressure (7.0%), and conjunctival hemorrhage (5.5%). The average number of re-treatments from months 3 to 11 was 2.7. Mean best-corrected visual acuity increased steadily from baseline to month 3 to reach +5.8 letters, decreased slightly from month 3 to 6, and remained stable from month 6 to 12, reaching +3.6 at month 12. Mean change in CRT was -101.1 µm from baseline to month 3 and -91.5 µm from baseline to month 12. CONCLUSIONS: The safety results are comparable to the favorable tolerability profile of ranibizumab observed in previous pivotal clinical studies; individualized treatment with less than monthly re-treatments shows a similar safety profile as observed in previous randomized clinical trials with monthly ranibizumab treatment. Efficacy outcomes were achieved with a low average number of re-treatments. Visual acuity in SUSTAIN patients with individualized re-treatment based on VA/optical coherence tomography assessment reached on average a maximum after the first 3 monthly injections, decreased slightly under PRN during the next 2 to 3 months, and was then sustained throughout the treatment period.

Intravitreal ranibizumab for myopic choroidal neovascularization: factors predictive of visual outcome and need for retreatment.

PURPOSE: To identify predictive factors for visual outcome and need for retreatment after treating myopic choroidal neovascularization (CNV) with ranibizumab. DESIGN: A prospective interventional case series. METHODS: Sixty-seven eyes of 67 patients with myopic CNV were treated with 3 intravitreal ranibizumab injections given monthly. Best-corrected visual acuity (BCVA) and optical coherence tomography-determined central macular thickness (CMT) were recorded monthly during follow-up. Fluorescein angiography changes and the number of injections needed were also assessed. RESULTS: Mean follow-up was 15.9 months. Mean BCVA improved by 7.8 letters after the first injection, 12.5 letters after 3 injections, and 12 letters by end follow-up. In 53 eyes (79.1%), BCVA improved; 40.3% gained more than 15 letters. No differences were detected in visual outcome between treatment-naïve and previously treated patients. Myopic CNV area and greatest linear dimension had diminished at the study end. The mean reduction in CMT was 93.6 µm. The mean number of injections given was 4.2. A total of 53.7% of eyes received only 3 injections. Through regression analysis, baseline BCVA (P = .006) and myopic CNV location (P = .026) were significantly correlated with BCVA at the end of follow-up. Myopic CNV location (P = .023) and prior treatment (P = .047) were significantly linked to the number of injections given. No major complications arose. CONCLUSION: An initial treatment regimen of 3 monthly ranibizumab injections seems effective and safe to treat myopic CNV. Baseline BCVA and myopic CNV location emerged as predictive factors for visual outcome. A need for retreatment was associated with myopic CNV location and prior treatment.