RESUMO
Oxidative stress represents one of the main factors driving the pathophysiology of multiple ophthalmic conditions including presbyopia, cataracts, dry eye disease (DED), glaucoma, age-related macular degeneration (AMD) and diabetic retinopathy (DR). Currently, different studies have demonstrated the role of orally administered nutraceuticals in these diseases. For instance, they have demonstrated to improve lens accommodation in presbyopia, reduce protein aggregation in cataracts, ameliorate tear film stability, break up time, and tear production in dry eye, and participate in the avoidance of retinal neuronal damage and a decrease in intraocular pressure in glaucoma, contribute to the delayed progression of AMD, or in the prevention or treatment of neuronal death in diabetic retinopathy. In this review, we summarized the nutraceuticals which have presented a positive impact in ocular disorders, emphasizing the clinical assays. The characteristics of the different types of nutraceuticals are specified along with the nutraceutical concentration used to achieve a therapeutic outcome in ocular diseases.
Assuntos
Catarata , Retinopatia Diabética , Síndromes do Olho Seco , Glaucoma , Degeneração Macular , Humanos , Catarata/metabolismo , Olho/metabolismo , Degeneração Macular/tratamento farmacológico , Glaucoma/tratamento farmacológicoRESUMO
Introdução: A inteligência artificial (IA) está revolucionando a área da saúde. Na oftalmologia, esta tecnologia pode possibilitar diagnósticos mais rápidos e precisos, impedindo a progressão das alterações na visão. Médicos e algoritimosalgoritmos podem ser mais eficientes quando trabalham juntos. Desenvolver um software de IA com alta especificidade e sensibilidade para apoio no diagnóstico de algumas patologias oftalmológicas. Métodos: O software de deep learning foi construído através de redes neurais valendo de duas bases computacionais MobileNet e Inception. Para o treinamento do banco de dados foram utilizadas 2.520 imagens de glaucoma, retinopatia diabética, toxoplasmose ocular, papiledema, descolamento de retina e retina normal. Para a validação foi utilizado 428 imagens patológicas e normais para os cálculos de sensibilidade e de especificidade. Todas as imagens foram cedidas da Sociedade Americana de Especialistas da Retina. Resultados: Os resultados de sensibilidade e especificidade foram no MobileNet de 91% (IC 95%, 89-92%) e 98,5% (IC 95%, 98-99%); no Inception, de 91,4% (IC 95%, 89-93,5%) de 98,4% (IC 95%, 98-98,8%), respectivamente. Não houve diferença significativa entre os dois métodos utilizados. Conclusão: O software apresentou resultados promissores na distinção das condições oftalmológicas pesquisadas. (AU)
Introduction: Artificial intelligence (AI) is revolutionizing health care. In ophthalmology, this technology can enable faster and more accurate diagnoses, preventing the progression of vision Physicians and algorithms are most effective when working together. To develop an AI software with high specificity and sensitivity to support the diagnosis of some ophthalmic diseases. Methods: A deep learning software was built through neural networks using two computational bases, MobileNet and Inception. For training the database, 2520 images of glaucoma, diabetic retinopathy, ocular toxoplasmosis, papilledema, retinal detachment, and normal retina were used. For validation, 428 pathological and normal images were used for calculations of sensitivity and specificity. All images were obtained from the American Society of Retina Specialists. Results: The results of sensitivity and specificity were 91% (95% confidence interval [CI], 89-92%) and 98.5% (95% CI, 98-99%) on MobileNet, and 91.4% (95% CI, 89-93.5%) and 98.4% (95% CI, 98-98.8%) on Inception, respectively. There was no significant difference between the two methods. Conclusion: The software showed promising results in distinguishing the main ophthalmic conditions surveyed. (AU)