he prospect of improved clinical outcomes and more efficient health systems has fueled a rapid rise in the development and evaluation of AI systems over the last decade. Because most AI systems within healthcare are complex interventions designed as clinical decision support systems, rather than autonomous agents, the interactions among the AI systems, their users and the implementation environments are defining components of the AI interventions' overall potential effectiveness. Therefore, bringing AI systems from mathematical performance to clinical utility needs an adapted, stepwise implementation and evaluation pathway, addressing the complexity of this collaboration between two independent forms of intelligence, beyond measures of effectiveness alone 1 . Despite indications that some AI-based algorithms now match the accuracy of human experts within preclinical in silico studies 2 , there

Diabetic retinopathy (DR) is the major ocular complication of diabetes mellitus, and is a problem with significant global health impact. Major advances in diagnostics, technology and treatment have already revolutionized how we manage DR in the early part of the 21st century. For example, the accessibility of imaging with optical coherence tomography, and the development of anti-vascular endothelial growth factor (VEGF) treatment are just some of the landmark developments that have shaped the DR landscape over the last few decades. Yet, there are still more exciting advances being made. Looking forward to 2030, many of these ongoing developments are likely to further transform the field. First, epidemiologic projections show that the global burden of DR is not only increasing, but also shifting from high-income countries towards middle- and low-income areas. Second, better understanding of disease pathophysiology is placing greater emphasis on retinal neural dysfunction and non-vascular aspects of diabetic retinal disease. Third, a wealth of information is becoming available from newer imaging modalities such as widefield imaging systems and optical coherence tomography angiography. Fourth, artificial intelligence for screening, diagnosis and prognostication of DR will become increasingly accessible and important. Fifth, new pharmacologic agents targeting other non-VEGF-driven pathways, and novel therapeutic strategies such as gene therapy are being developed for DR. Finally, the classification system for diabetic retinal disease will need to be continually updated to keep pace with new developments. In this article, we discuss these major trends in DR that we expect to see in 2030 and beyond.

Background By 2050, almost 5 billion people globally are projected to have myopia, of whom 20% are likely to have high myopia with clinically significant risk of sight-threatening complications such as myopic macular degeneration. These are diagnoses that typically require specialist assessment or measurement with multiple unconnected pieces of equipment. Artificial intelligence (AI) approaches might be effective for risk stratification and to identify individuals at highest risk of visual loss. However, unresolved challenges for AI medical studies remain, including paucity of transparency, auditability, and traceability.Methods In this retrospective multicohort study, we developed and tested retinal photograph-based deep learning algorithms for detection of myopic macular degeneration and high myopia, using a total of 226 686 retinal images. First we trained and internally validated the algorithms on datasets from Singapore, and then externally tested them on datasets from China, Taiwan, India, Russia, and the UK. We also compared the performance of the deep learning algorithms against six human experts in the grading of a randomly selected dataset of 400 images from the external datasets. As proof of concept, we used a blockchain-based AI platform to demonstrate the real-world application of secure data transfer, model transfer, and model testing across three sites in Singapore and China.Findings The deep learning algorithms showed robust diagnostic performance with areas under the receiver operating characteristic curves [AUC] of 0•969 (95% CI 0•959-0•977) or higher for myopic macular degeneration and 0•913 (0•906-0•920) or higher for high myopia across the external testing datasets with available data. In the randomly selected dataset, the deep learning algorithms outperformed all six expert graders in detection of each condition (AUC of 0•978 [0•957-0•994] for myopic macular degeneration and 0•973 [0•941-0•995] for high myopia). We also successfully used blockchain technology for data transfer, model transfer, and model testing between sites and across two countries.Interpretation Deep learning algorithms can be effective tools for risk stratification and screening of myopic macular degeneration and high myopia among the large global population with myopia. The blockchain platform developed here could potentially serve as a trusted platform for performance testing of future AI models in medicine.Funding None.