Blindness prevention has been an important national policy in China. Previous strategies, such as deploying experienced cataract surgeons to rural areas and assisting in building local ophthalmology centers, had successfully decreased the prevalence of visual impairment and blindness. However, new challenges arise with the aging population and the shift of the disease spectrum towards age-related eye diseases and myopia. With the constant technological boom, digital healthcare innovations in ophthalmology could immensely enhance screening and diagnosing capabilities. Artifcial intelligence (AI) and telemedicine have been proven valuable in clinical ophthalmology settings. Moreover, the integration of cutting-edge communication technology and AI in mobile clinics and remote surgeries is on the horizon, potentially revolutionizing blindness prevention and ophthalmic healthcare. The future of blindness prevention in China is poised to undergo signifcant transformation, driven by emerging challenges and new opportunities.

Blindness prevention has been an important national policy in China. Previous strategies, such as deploying experienced cataract surgeons to rural areas and assisting in building local ophthalmology centers, had successfully decreased the prevalence of visual impairment and blindness. However, new challenges arise with the aging population and the shift of the disease spectrum towards age-related eye diseases and myopia. With the constant technological boom, digital healthcare innovations in ophthalmology could immensely enhance screening and diagnosing capabilities. Artifcial intelligence (AI) and telemedicine have been proven valuable in clinical ophthalmology settings. Moreover, the integration of cutting-edge communication technology and AI in mobile clinics and remote surgeries is on the horizon, potentially revolutionizing blindness prevention and ophthalmic healthcare. The future of blindness prevention in China is poised to undergo signifcant transformation, driven by emerging challenges and new opportunities.

目的：建立和验证一个涉及多级临床场景的白内障协作通用的人工智能(artificial intelligence,AI)管理平台，探索基于AI的医疗转诊模式，以提高协作效率和资源覆盖率。方法：训练和验证的数据集来自中国AI医学联盟，涵盖多级医疗机构和采集模式。使用三步策略对数据集进行标记: 1)识别采集模式；2)白内障诊断包括正常晶体眼、白内障眼或白内障术后眼；3)从病因和严重程度检测需转诊的白内障患者。此外，将白内障AI系统与真实世界中的居家自我监测、初级医疗保健机构和专科医院等多级转诊模式相结合。结果：通用AI平台和多级协作模式在三步任务中表现出可靠的诊断性能: 1)识别采集模式的受试者操作特征(receiver operating characteristic curve,ROC)曲线下面积(area under the curve,AUC)为99.28%~99.71%)；2)白内障诊断对正常晶体眼、白内障或术后眼，在散瞳-裂隙灯模式下的AUC分别为99.82%、99.96%和99.93%，其他采集模式的AUC均 > 99%；3)需转诊白内障的检测(在所有测试中AUC >91%)。在真实世界的三级转诊模式中，该系统建议30.3%的人转诊，与传统模式相比，眼科医生与人群服务比率大幅提高了10.2倍。结论：通用AI平台和多级协作模式显示了准确的白内障诊断性能和有效的白内障转诊服务。建议AI的医疗转诊模式扩展应用到其他常见疾病和资源密集型情景当中。

Objective: To establish and validate a universal artificial intelligence (AI) platform for collaborative management of cataracts involving multilevel clinical scenarios and explored an AI-based medical referral pattern to improve collaborative efficiency and resource coverage. Methods: The training and validation datasets were derived from the Chinese Medical Alliance for Artificial Intelligence, covering multilevel healthcare facilities and capture modes. The datasets were labelled using a three step strategy: (1)capture mode recognition; (2) cataract diagnosis as a normal lens, cataract or a postoperative eye and (3) detection of referable cataracts with respect to aetiology and severity. Moreover, we integrated the cataract AI agent with a real-world multilevel referral pattern involving self-monitoring at home, primary healthcare and specialised hospital services. Results: The universal AI platform and multilevel collaborative pattern showed robust diagnostic performance in three-step tasks: (1) capture mode recognition (area under the curve (AUC) 99.28%–99.71%), (2) cataract diagnosis (normal lens, cataract or postoperative eye with AUCs of 99.82%, 99.96% and 99.93% for mydriatic-slit lamp mode and AUCs >99% for other capture modes) and (3)detection of referable cataracts (AUCs >91% in all tests). In the real-world tertiary referral pattern, the agent suggested 30.3%  of people be ’referred’, substantially increasing the ophthalmologist-to-population service ratio by 10.2-fold compared with the traditional pattern. Conclusions: The universal AI platform and multilevel collaborative pattern showed robust diagnostic performance and effective service for cataracts. The context of our AI-based medical referral pattern will be extended to other common disease conditions and resource-intensive situations.

目的：建立并评估儿童Ⅱ期人工晶状体(intraocular lens,IOL)植入术后青光眼相关不良事件(glaucoma-related adverse events,GRAEs)的预测模型。方法：选取于中山大学中山眼科中心行Ⅱ期IOL植入术的无晶状体眼患儿205例(356眼)，并在术后对其随访3年。采用Cox比例风险模型确定GRAEs的预测因子，并建立列线图预测模型。采用随时间变化的受试者工作特征(receiver operating characteristic,ROC)曲线、决策曲线分析、Kaplan-Meier曲线评估模型性能，并通过Bootstrapping的C指数和校准图进行内部验证。结果：行Ⅱ期IOL植入术时年龄较大(HR=1.50, 95% CI: 1.03 ~2.19)、术后一过性高眼压(HR=9.06, 95% CI: 2.97~27.67)和IOL睫状沟植入术(HR=14.55, 95% CI: 2.11~100.57)是GRAEs的危险因素(均P<0.05)，并据此建立了两个列线图预测模型。在术后1、2、3年，模型1的ROC曲线下面积(area under curve,AUC)分别为0.747(95% CI: 0.776 ~0.935)、0.765 (95% CI: 0.804 ~0.936)和0.748 (95% CI: 0.736~0.918)，模型2的AUC分别为0.881 (95% CI: 0.836 ~0.926)、0.895 (95% CI: 0.852 ~0.938)和0.848 (95% CI: 0.752~0.945)。在内部验证和评价中，两种模型均表现出良好的性能和临床净效益。Kaplan-Meier曲线显示两个不同的风险组在两个模型中都能被显著且稳健地区分。此外，本研究也构建了在线风险计算器。结论：两种列线图均能灵敏、准确地识别Ⅱ期IOL植入术后GRAEs的高危患儿，有助对其进行早期识别和及时干预。

Aims: To establish and evaluate predictive models for glaucoma-related adverse events (GRAEs) following secondary intraocular lens (IOL) implantation in paediatric eyes. Methods: 205 children (356 aphakic eyes) receiving secondary IOL implantation at Zhongshan Ophthalmic Center with a 3-year follow-up were enrolled. Cox proportional hazard model was used to identify predictors of GRAEs and developed nomograms. Model performance was evaluated with time-dependent receiver operating characteristic (ROC) curves, decision curve analysis, Kaplan-Meier curves and validated internally through C-statistics and calibration plot of the bootstrap samples. Results: Older age at secondary IOL implantation (HR=1.5, 95% CI: 1.03 to 2.19), transient intraocular hypertension (HR=9.06, 95% CI: 2.97 to  27.67) and ciliary sulcus implantation (HR=14.55, 95% CI: 2.11 to 100.57) were identified as risk factors for GRAEs (all p<0.05). Two nomograms were established. At postoperatively 1, 2 and 3 years, model 1 achieved area under the ROC curves (AUCs) of 0.747 (95% CI: 0.776 to 0.935), 0.765 (95% CI: 0.804 to 0.936) and 0.748 (95% CI: 0.736 to 0.918), and the AUCs of model 2 were 0.881 (95% CI: 0.836 to 0.926), 0.895 (95% CI: 0.852 to 0.938) and 0.848 (95% CI: 0.752 to 0.945). Both models demonstrated fine clinical net benefit and performance in the interval validation. The Kaplan-Meier curves showing two distinct risk groups were well discriminated and robust in both models. An online risk calculator was constructed. Conclusions: Two nomograms could sensitively and accurately identify children at high risk of GRAEs after secondary IOL implantation to help early identification and timely intervention.

Background: Research innovations inoculardisease screening, diagnosis, and management have been boosted by deep learning (DL) in the last decade. To assess historical research trends and current advances, we conducted an artifcial intelligence (AI)–human hybrid analysis of publications on DL in ophthalmology.

Methods: All DL-related articles in ophthalmology, which were published between 2012 and 2022 from Web of Science, were included. 500 high-impact articles annotated with key research information were used to fne-tune alarge language models (LLM) for reviewing medical literature and extracting information. After verifying the LLM's accuracy in extracting diseases and imaging modalities, we analyzed trend of DL in ophthalmology with 2 535 articles. 

Results: Researchers using LLM for literature analysis were 70% (p= 0.000 1) faster than those who did not, while achieving comparable accuracy (97% versus 98%, p = 0.768 1). The field of DL in ophthalmology has grown 116% annually, paralleling trends of the broader DL domain. The publications focused mainly on diabetic retinopathy (p = 0.000 3), glaucoma (p = 0.001 1), and age-related macular diseases (p = 0.000 1) using retinal fundus photographs (FP, p = 0.001 5) and optical coherence tomography (OCT, p = 0.000 1). DL studies utilizing multimodal images have been growing, with FP and OCT combined being the most frequent. Among the 500 high-impact articles, laboratory studies constituted the majority at 65.3%. Notably, a discernible decline in model accuracy was observed when categorizing by study design, notwithstanding its statistical insignificance. Furthermore, 43 publicly available ocular image datasets were summarized. 

Conclusion: This study has characterized the landscape of publications on DL in ophthalmology, by identifying the trends and breakthroughs among research topics and the fast-growing areas. This study provides an efcient framework for combined AI–human analysis to comprehensively assess the current status and future trends in the feld. 

Background: Research innovations inoculardisease screening, diagnosis, and management have been boosted by deep learning (DL) in the last decade. To assess historical research trends and current advances, we conducted an artifcial intelligence (AI)–human hybrid analysis of publications on DL in ophthalmology.

Methods: All DL-related articles in ophthalmology, which were published between 2012 and 2022 from Web of Science, were included. 500 high-impact articles annotated with key research information were used to fne-tune alarge language models (LLM) for reviewing medical literature and extracting information. After verifying the LLM's accuracy in extracting diseases and imaging modalities, we analyzed trend of DL in ophthalmology with 2 535 articles. 

Results: Researchers using LLM for literature analysis were 70% (p = 0.000 1) faster than those who did not, while achieving comparable accuracy (97% versus 98%, p = 0.768 1). The field of DL in ophthalmology has grown 116% annually, paralleling trends of the broader DL domain. The publications focused mainly on diabetic retinopathy (p = 0.000 3), glaucoma (p = 0.001 1), and age-related macular diseases (p = 0.000 1) using retinal fundus photographs (FP, p = 0.001 5) and optical coherence tomography (OCT, p = 0.000 1). DL studies utilizing multimodal images have been growing, with FP and OCT combined being the most frequent. Among the 500 high-impact articles, laboratory studies constituted the majority at 65.3%. Notably, a discernible decline in model accuracy was observed when categorizing by study design, notwithstanding its statistical insignificance. Furthermore, 43 publicly available ocular image datasets were summarized. 

Conclusion: This study has characterized the landscape of publications on DL in ophthalmology, by identifying the trends and breakthroughs among research topics and the fast-growing areas. This study provides an efcient framework for combined AI–human analysis to comprehensively assess the current status and future trends in the feld. 

Artificial intelligence (AI) is about simulating and expanding human intelligence. AI based on deep learning (DL) can analyze images well by using their inherent features, such as outlines, frames and so on. As researchers generally diagnoses ocular fundus diseases by images, it makes sense to apply AI to fundus examination. In ophthalmology, AI has achieved doctor-like performance in detecting multiple ocular fundus diseases through optical coherence tomography (OCT) images, fundus photographs, and ultra-wide-field (UWF) images. It has also been widely used in disease progression prediction. Nonetheless, there are also some potential challenges with AI application in ophthalmology, one of which is the black-box problem. Researchers are devoted to developing more interpretable deep learning systems (DLS) and confirming their clinical feasibility. This review describes a summary of the state-of-the-art AI application in the most popular ocular fundus diseases, potential challenges and the path forward.

Artificial intelligence (AI) is about simulating and expanding human intelligence. AI based on deep learning (DL) can analyze images well by using their inherent features, such as outlines, frames and so on. As researchers generally diagnoses ocular fundus diseases by images, it makes sense to apply AI to fundus examination. In ophthalmology, AI has achieved doctor-like performance in detecting multiple ocular fundus diseases through optical coherence tomography (OCT) images, fundus photographs, and ultra-wide-field (UWF) images. It has also been widely used in disease progression prediction. Nonetheless, there are also some potential challenges with AI application in ophthalmology, one of which is the black-box problem. Researchers are devoted to developing more interpretable deep learning systems (DLS) and confirming their clinical feasibility. This review describes a summary of the state-of-the-art AI application in the most popular ocular fundus diseases, potential challenges and the path forward.

Background: The “flipped classroom” is a learner-centered approach that centers on delivering videos,podcasts or slide-based material to learners prior to a lecture or class session. The class session is then dedicated to discussion, analysis, and problem-solving activities. The aim of this study was to investigate whether the flipped classroom could be adapted to medical (ophthalmology) students learning about ocular trauma and to assess the impact of the flipped classroom on those students’ performance and attitudes.
Methods: Questionnaires (using a 4-point scale) were distributed to 93 fifth-year medical students at Sun Yat-sen University, and the data showed that the majority of students preferred the flipped classroom approach to the traditional lecture method.
Results: The results of pre- and post-test scores were 14.35±3.404 and 20.37±4.356, which showed a significant improvement in students’ performance after the flipped classroom was introduced (P<0.05).
Conclusions: Student response to the flipped classroom strategy was largely positive, indicating that the strategy received a high level of approval in an ophthalmology clerkship course taken by medical students in China.

Background: The “flipped classroom” is a learner-centered approach that centers on delivering videos,podcasts or slide-based material to learners prior to a lecture or class session. The class session is then dedicated to discussion, analysis, and problem-solving activities. The aim of this study was to investigate whether the flipped classroom could be adapted to medical (ophthalmology) students learning about ocular trauma and to assess the impact of the flipped classroom on those students’ performance and attitudes.
Methods: Questionnaires (using a 4-point scale) were distributed to 93 fifth-year medical students at Sun Yat-sen University, and the data showed that the majority of students preferred the flipped classroom approach to the traditional lecture method.
Results: The results of pre- and post-test scores were 14.35±3.404 and 20.37±4.356, which showed a significant improvement in students’ performance after the flipped classroom was introduced (P<0.05).
Conclusions: Student response to the flipped classroom strategy was largely positive, indicating that the strategy received a high level of approval in an ophthalmology clerkship course taken by medical students in China.