Image quality assessment of retinal fundus photographs for diabetic retinopathy in the machine learning era: a review.

This study aimed to evaluate the image quality assessment (IQA) and quality criteria employed in publicly available datasets for diabetic retinopathy (DR). A literature search strategy was used to identify relevant datasets, and 20 datasets were included in the analysis. Out of these, 12 datasets mentioned performing IQA, but only eight specified the quality criteria used. The reported quality criteria varied widely across datasets, and accessing the information was often challenging. The findings highlight the importance of IQA for AI model development while emphasizing the need for clear and accessible reporting of IQA information. The study suggests that automated quality assessments can be a valid alternative to manual labeling and emphasizes the importance of establishing quality standards based on population characteristics, clinical use, and research purposes. In conclusion, image quality assessment is important for AI model development; however, strict data quality standards must not limit data sharing. Given the importance of IQA for developing, validating, and implementing deep learning (DL) algorithms, it's recommended that this information be reported in a clear, specific, and accessible way whenever possible. Automated quality assessments are a valid alternative to the traditional manual labeling process, and quality standards should be determined according to population characteristics, clinical use, and research purpose.

Clinician-Driven AI: Code-Free Self-Training on Public Data for Diabetic Retinopathy Referral.

Importance: Democratizing artificial intelligence (AI) enables model development by clinicians with a lack of coding expertise, powerful computing resources, and large, well-labeled data sets. Objective: To determine whether resource-constrained clinicians can use self-training via automated machine learning (ML) and public data sets to design high-performing diabetic retinopathy classification models. Design, Setting, and Participants: This diagnostic quality improvement study was conducted from January 1, 2021, to December 31, 2021. A self-training method without coding was used on 2 public data sets with retinal images from patients in France (Messidor-2 [n = 1748]) and the UK and US (EyePACS [n = 58 689]) and externally validated on 1 data set with retinal images from patients of a private Egyptian medical retina clinic (Egypt [n = 210]). An AI model was trained to classify referable diabetic retinopathy as an exemplar use case. Messidor-2 images were assigned adjudicated labels available on Kaggle; 4 images were deemed ungradable and excluded, leaving 1744 images. A total of 300 images randomly selected from the EyePACS data set were independently relabeled by 3 blinded retina specialists using the International Classification of Diabetic Retinopathy protocol for diabetic retinopathy grade and diabetic macular edema presence; 19 images were deemed ungradable, leaving 281 images. Data analysis was performed from February 1 to February 28, 2021. Exposures: Using public data sets, a teacher model was trained with labeled images using supervised learning. Next, the resulting predictions, termed pseudolabels, were used on an unlabeled public data set. Finally, a student model was trained with the existing labeled images and the additional pseudolabeled images. Main Outcomes and Measures: The analyzed metrics for the models included the area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, specificity, and F1 score. The Fisher exact test was performed, and 2-tailed P values were calculated for failure case analysis. Results: For the internal validation data sets, AUROC values for performance ranged from 0.886 to 0.939 for the teacher model and from 0.916 to 0.951 for the student model. For external validation of automated ML model performance, AUROC values and accuracy were 0.964 and 93.3% for the teacher model, 0.950 and 96.7% for the student model, and 0.890 and 94.3% for the manually coded bespoke model, respectively. Conclusions and Relevance: These findings suggest that self-training using automated ML is an effective method to increase both model performance and generalizability while decreasing the need for costly expert labeling. This approach advances the democratization of AI by enabling clinicians without coding expertise or access to large, well-labeled private data sets to develop their own AI models.

Diabetic retinopathy classification for supervised machine learning algorithms.

BACKGROUND: Artificial intelligence and automated technology were first reported more than 70 years ago and nowadays provide unprecedented diagnostic accuracy, screening capacity, risk stratification, and workflow optimization. Diabetic retinopathy is an important cause of preventable blindness worldwide, and artificial intelligence technology provides precocious diagnosis, monitoring, and guide treatment. High-quality exams are fundamental in supervised artificial intelligence algorithms, but the lack of ground truth standards in retinal exams datasets is a problem. MAIN BODY: In this article, ETDRS, NHS, ICDR, SDGS diabetic retinopathy grading, and manual annotation are described and compared in publicly available datasets. The various DR labeling systems generate a fundamental problem for AI datasets. Possible solutions are standardization of DR classification and direct retinal-finding identifications. CONCLUSION: Reliable labeling methods also need to be considered in datasets with more trustworthy labeling.

VALUE OF CLINICAL, LABORATORY PARAMETERS AND ANALYSIS OF RETINAL VASCULARIZATION IN PEDIATRIC PATIENTS WITH PORTAL HYPERTENSION.

BACKGROUND: Portal hypertension is one of the complications of cirrhosis and is associated with numerous systemic manifestations, including renal, brain, pulmonary, cardiac and vascular changes. In routine ophthalmological examinations performed at our service, we observed that some children diagnosed with portal hypertension had increased retinal vascular tortuosity. OBJECTIVE: 1. To evaluate the presence of retinal vascular abnormalities (vascular tortuosity) in children diagnosed with portal hypertension; 2. To investigate the association between retinal vascular tortuosity and the presence of gastroesophageal varices in these children; 3. To evaluate the use of clinical and laboratory parameters to predict the presence of gastroesophageal varices in children with portal hypertension. METHODS: This was a cross-sectional and observational study that included patients aged <18 years with a diagnosis of portal hypertension. The participants included were submitted to dilated fundus examination and fundus photography with Visucam (Carl Zeiss Meditec AG) device. Besides, clinical and laboratorial data were collected from the patients' medical records. RESULTS: A total of 72 patients were included in this study, and 36% of them had an increase in retinal vascular tortuosity. Platelet count (P=0.001), bilirubin dosage (P=0.013) and aspartate transaminase dosage (AST) (P=0.042) were associated with the presence of gastroesophageal varices in digestive endoscopy. There was no association between retinal vascular tortuosity and the presence of gastroesophageal varices (P=0.498). CONCLUSION: The results of this study suggest that platelet count, bilirubin dosage, and aspartate transaminase dosage were associated with the presence of gastroesophageal varices in digestive endoscopy. Regarding the retinal findings, we found that there was an increase in retinal vascular tortuosity in 36% of pediatric patients, but no association was found with the presence of gastroesophageal varices.

Valor de parâmetros clínicos, laboratoriais e análise da vascularização da retina em pacientes pediátricos com hipertensão portal / Value of clinical, laboratory parameters and analysis of retinal vascularization in pediatric patients with portal hypertension

ABSTRACT BACKGROUND: Portal hypertension is one of the complications of cirrhosis and is associated with numerous systemic manifestations, including renal, brain, pulmonary, cardiac and vascular changes. In routine ophthalmological examinations performed at our service, we observed that some children diagnosed with portal hypertension had increased retinal vascular tortuosity. OBJECTIVE: 1. To evaluate the presence of retinal vascular abnormalities (vascular tortuosity) in children diagnosed with portal hypertension; 2. To investigate the association between retinal vascular tortuosity and the presence of gastroesophageal varices in these children; 3. To evaluate the use of clinical and laboratory parameters to predict the presence of gastroesophageal varices in children with portal hypertension. METHODS: This was a cross-sectional and observational study that included patients aged <18 years with a diagnosis of portal hypertension. The participants included were submitted to dilated fundus examination and fundus photography with Visucam (Carl Zeiss Meditec AG) device. Besides, clinical and laboratorial data were collected from the patients' medical records. RESULTS: A total of 72 patients were included in this study, and 36% of them had an increase in retinal vascular tortuosity. Platelet count (P=0.001), bilirubin dosage (P=0.013) and aspartate transaminase dosage (AST) (P=0.042) were associated with the presence of gastroesophageal varices in digestive endoscopy. There was no association between retinal vascular tortuosity and the presence of gastroesophageal varices (P=0.498). CONCLUSION: The results of this study suggest that platelet count, bilirubin dosage, and aspartate transaminase dosage were associated with the presence of gastroesophageal varices in digestive endoscopy. Regarding the retinal findings, we found that there was an increase in retinal vascular tortuosity in 36% of pediatric patients, but no association was found with the presence of gastroesophageal varices.

RESUMO CONTEXTO: A hipertensão portal é uma das complicações da cirrose e está associada a inúmeras manifestações sistêmicas, incluindo alterações renais, cerebrais, pulmonares, cardíacas e vasculares. Nos exames oftalmológicos de rotina realizados em nosso serviço, observamos que algumas crianças diagnosticadas com hipertensão portal apresentaram aumento da tortuosidade vascular da retina. OBJETIVO: 1. Avaliar a presença de anormalidades vasculares da retina (tortuosidade vascular) em crianças diagnosticadas com hipertensão portal; 2. Investigar a associação entre tortuosidade vascular da retina e presença de varizes gastroesofágicas nessas crianças; 3. Avaliar o uso de parâmetros clínicos e laboratoriais para prever a presença de varizes gastroesofágicas em crianças com hipertensão portal. MÉTODOS: Estudo transversal e observacional, que incluiu pacientes com idade <18 anos com diagnóstico de hipertensão portal. Os participantes incluídos foram submetidos ao exame de fundo de olho dilatado e fotografia de fundo com dispositivo Visucam (Carl Zeiss Meditec AG). Além disso, foram coletados dados clínicos e laboratoriais dos prontuários dos pacientes. RESULTADOS: Um total de 72 pacientes foi incluído neste estudo e 36% deles apresentaram aumento da tortuosidade vascular da retina. Contagem de plaquetas (P=0,001), dosagem de bilirrubina (P=0,013) e dosagem de aspartato transaminase (AST) (P=0,042) foram associados à presença de varizes gastroesofágicas na endoscopia digestiva. Não houve associação entre tortuosidade vascular da retina e presença de varizes gastroesofágicas (P=0,498). CONCLUSÃO: Os resultados deste estudo sugerem que a contagem de plaquetas, a dosagem de bilirrubina e a aspartato transaminase foram associadas à presença de varizes gastroesofágicas na endoscopia digestiva. Em relação aos achados da retina, descobrimos que houve um aumento na tortuosidade vascular da retina em 36% dos pacientes pediátricos, mas nenhuma associação foi encontrada com a presença de varizes gastroesofágicas.

INTRAVITREAL DEXAMETHASONE IMPLANT MIGRATION INTO THE ANTERIOR CHAMBER: A Multicenter Study From the Pan-American Collaborative Retina Study Group.

PURPOSE: To establish the prevalence and risk factors for intravitreal dexamethasone implant migration into the anterior chamber in eyes with macular edema. METHODS: This was a multicenter, retrospective, observational chart review of data that included patients with macular edema who had been treated with at least one intravitreal dexamethasone injection. Patients with incomplete chart information during the follow-up period were excluded. RESULTS: The prevalence of implant migration in 468 patients, considering the number of injections, was 1.6%, with significant associations between implant migration and cataract surgery (P = 0.043) and intraocular lens status (P = 0.005) and a trend toward statistical significance (P = 0.057) with vitrectomy. A higher rate of implant migration into the anterior chamber was observed in vitrectomized eyes (4.8%) when compared with patients who did not undergo a vitrectomy (1.6%). The implants that migrated were removed with forceps with/without viscoelastic expression or with 20-gauge cannulas connected to the vitreous cutter machine. CONCLUSION: The risk of implant migration into the anterior chamber was 1.6%. Risk factors were a history of cataract surgery or vitrectomy and aphakia. When anterior migration occurs, rapid removal is advised, especially if corneal edema is present.

New perspectives on macular hole surgery at three years of follow-up / Novas perspectivas em cirurgia de buraco macular com de 3 anos de seguimento

ABSTRACT Purpose: The aim of this study was to determine the functional and anatomical success rates as well as the safety of sutureless combined surgery involving vitreous base removal and internal limiting membrane peeling after Brilliant Blue G (0.5 mg/mL) staining for the management of idiopathic macular holes after three years. Methods: Forty-six eyes of 46 patients with an idiopathic macular hole were enrolled in this retrospective study. The inclusion criteria were macular holes with a minimum linear diameter below 1,500 mm, 0.05 or better decimal best-corrected visual acuity and duration of symptoms less than two years. The exclusion criteria included pregnancy, optic nerve atrophy, advanced glaucoma, and other chronic ocular diseases. The surgical procedure included internal limiting membrane peeling after Brilliant Blue G (0.5 mg/mL) staining, along with C3F8 tamponade and face-down positioning for three days postoperatively. Ophthalmologic examinations and optical coherence tomography were performed at 1 and 7 days and 1, 6, 12, 24, and 36 months postoperatively. If no anatomic closure of the macular holes occurred within the first month, the area of the internal limiting membrane peeling was enlarged in a second procedure. Multiple logistic regression and chi-squared tests were used for data analyses, and p-values of <0.05 were considered significant. Results: Out of 46 eyes with a preoperative idiopathic macular hole, anatomic closure was achieved in 42 (91.3%) after one procedure and in 45 (97.8%) after an additional surgery. The median postoperative best-corrected visual acuity improvement was 0.378 (range: 0.050-0.900) decimal. None of the patients experienced macular hole reopening, surgery-related complications, or ocular complications related to the dye. Conclusion: Combined surgery including vitreous base removal and internal limiting membrane peeling after staining with Brilliant Blue G (0.5 mg/mL) for the management of idiopathic macular holes resulted in adequate staining, best-corrected visual acuity improvement, and macular hole closure with no signs of ocular toxicity at the three-year follow-up examination.

RESUMO Objetivo: Determinar, após 3 anos de seguimento, as taxas de sucesso funcional e anatômico e a segurança da cirurgia combinada sem sutura, incluindo remoção da base vítrea e da membrana limitante interna após coloração com azul brilhante (0,5 mg/ml) para o manejo de buracos maculares idiopáticos. Métodos: Quarenta e seis olhos de 46 pacientes com buraco macular idiopático foram incluídos neste estudo retrospectivo. Os critérios de inclusão foram: buraco macular com diâmetro linear mínimo menor que 1500 micrômetros, acuidade visual com melhor correção de 0,05 decimal ou melhor e tempo de sintomas menor que 2 anos. Os critérios de exclusão foram gravidez, atrofia do nervo óptico, glaucoma avançado ou outra doença ocular crônica. A técnica cirúrgica incluiu a remoção da membrana limitante interna após coloração com Azul Brilhante 0,5 mg/ml, tamponamento com C3F8 posicionamento em prona ção durante 3 dias de pós-operatório. O seguimento foi realizado por exame oftalmológico e Tomografia de Coerência Óptica no 1 e 7 dias, 1, 6, 12, 24 e 36 meses de pós-operatório. Se o fechamento anatômico do buraco macular não fosse atingido na visita de um mês, realizava-se um segundo procedimento no qual a área do peeling da membrana limitante interna era ampliada. Para análise estatística, foram utilizados testes de regressão logística múltipla e Qui-quadrado. Valores de p menores que 0.05 foram considerados estatisticamente significativos. Resultados: Dos 46 olhos com buraco macular idiopático, 42 (91,3%) obtiveram fechamento do buraco macular após um procedimento cirúrgico e 45 (97,8%) após uma cirurgia adicional. A média de melhora da acuidade visual com melhor correção no pós-operatório foi de 0.378 (0.050-0.900) decimal. Não foram observados: reabertura do buraco macular, complicações relacionadas ao procedimento cirúrgico ou complicações relacionadas ao corante. Conclusão: A cirurgia combinada sem sutura que incluiu remoção da base vítrea e remoção membrana limitante interna após coloração com Azul Brilhante (0,5 mg/ml) para o tratamento de buracos maculares idiopáticos foi realizada com adequada capacidade de coloração, melhora da acuidade visual e fechamento do buraco macular sem sinais de toxicidade ocular no seguimento de 3 anos.

New perspectives on macular hole surgery at three years of follow-up.

PURPOSE: The aim of this study was to determine the functional and anatomical success rates as well as the safety of sutureless combined surgery involving vitreous base removal and internal limiting membrane peeling after Brilliant Blue G (0.5 mg/mL) staining for the management of idiopathic macular holes after three years. METHODS: Forty-six eyes of 46 patients with an idiopathic macular hole were enrolled in this retrospective study. The inclusion criteria were macular holes with a minimum linear diameter below 1,500 mm, 0.05 or better decimal best-corrected visual acuity and duration of symptoms less than two years. The exclusion criteria included pregnancy, optic nerve atrophy, advanced glaucoma, and other chronic ocular diseases. The surgical procedure included internal limiting membrane peeling after Brilliant Blue G (0.5 mg/mL) staining, along with C3F8 tamponade and face-down positioning for three days postoperatively. Ophthalmologic examinations and optical coherence tomography were performed at 1 and 7 days and 1, 6, 12, 24, and 36 months postoperatively. If no anatomic closure of the macular holes occurred within the first month, the area of the internal limiting membrane peeling was enlarged in a second procedure. Multiple logistic regression and chi-squared tests were used for data analyses, and p-values of <0.05 were considered significant. RESULTS: Out of 46 eyes with a preoperative idiopathic macular hole, anatomic closure was achieved in 42 (91.3%) after one procedure and in 45 (97.8%) after an additional surgery. The median postoperative best-corrected visual acuity improvement was 0.378 (range: 0.050-0.900) decimal. None of the patients experienced macular hole reopening, surgery-related complications, or ocular complications related to the dye. CONCLUSION: Combined surgery including vitreous base removal and internal limiting membrane peeling after staining with Brilliant Blue G (0.5 mg/mL) for the management of idiopathic macular holes resulted in adequate staining, best-corrected visual acuity improvement, and macular hole closure with no signs of ocular toxicity at the three-year follow-up examination.

Analysis of Anthocyanins Extracted from the Acai Fruit (Euterpe oleracea): A Potential Novel Vital Dye for Chromovitrectomy.

PURPOSE: To classify and quantify anthocyanins in a vital dye extracted from the acai fruit (Euterpe oleracea), adjust pH and osmolarity, and perform lyophilization to develop a new chromovitrectomy dye. METHODS: Three dye concentrations 10%, 25%, and 35% (equivalent to 100, 250, and 350 mg of lyophilized acai fruit pulp extract samples) were evaluated when diluted in 1 ml of phosphate-buffered solution (pH 7 and 300 mOsm). The dye was analyzed by mass spectrometry and high-performance liquid chromatography (HPLC) to identify and quantify anthocyanins molecules. RESULTS: The pH and osmolarity correction and lyophilization were performed without damaging the anthocyanin molecular structure. Mass spectrometry confirmed the presence of five anthocyanins in the three concentrations of the dye. Cyanidin-3-O-glucoside was the major anthocyanin found. HPLC showed that the concentration of anthocyanin was similar, independent of the dye concentration tested. CONCLUSIONS: Lyophilization and the correction of pH and osmolarity (7.00 and 300 mOsm, resp.) were performed successfully. Five anthocyanins are present in the dye from the acai fruit. The major anthocyanin is cyanidin-3-O-glucoside. Independent of the dye concentration tested, the anthocyanin concentration was similar. Standardized chemical characteristics of this new dye may allow its use during chromovitrectomy in humans.

Combined Femtosecond Laser-Assisted Cataract Surgery and Small-Gauge Pars Plana Vitrectomy Using Different Devices: A New Trend for Vitreoretinal Surgery?

BACKGROUND AND OBJECTIVE: To report the efficacy, safety, and benefits of femtosecond laser-assisted cataract surgery (FLACS) combined with sutureless 23-gauge pars plana vitrectomy (PPV). PATIENTS AND METHODS: This multicenter, retrospective study evaluated patient records and videos of 43 cases with retinal pathologies and cataract who underwent the combined procedure. RESULTS: In 44.2% and 55.8% of cases, respectively, the LenSx Laser (femtosecond machine; Alcon, Fort Worth, TX) and the Constellation (vitreous cutter; Alcon, Fort Worth, TX), and the Victus (femtosecond machine; Bausch + Lomb, Rochester, NY) and Stellaris PC (vitreous cutter; Bausch + Lomb, Rochester, NY) were used. No complications developed during capsulorrhexis, even without a red fundus reflex, retrobulbar block, or scleral indentation. Foldable intraocular lenses remained stable in the capsular bag during the vitreoretinal surgeries and postoperative visits. The mean times of femtosecond phacoemulsification, vitreoretinal surgery, and total surgery were 22.9 minutes ± 4.7 minutes, 43.1 minutes ± 9.8 minutes, and 65.3 minutes ± 8.6 minutes, respectively. CONCLUSION: This emerging technology is safe and offers several potential benefits for the success of the combined procedure. [Ophthalmic Surg Lasers Imaging Retina. 2018;49:374-379.].