Aims: To identify the characteristic retinal neurovascular changes in patients in different stages of nondiabetic chronic kidney disease (CKD) and to develop a model for the accurate diagnosis of nondiabetic CKD.
Methods: Peripapillary retinal nerve fiber layer (pRNFL) thickness and average macular ganglion cell-inner plexiform layer (GC-IPL) thickness of nondiabetic CKD patients and healthy controls (HC) were evaluated by spectral-domain optical coherence tomography (OCT). The vessel density (VD) and perfusion density (PD) of the macula were obtained from optical coherence tomography angiography (OCTA). The estimated glomerular filtration rate (eGFR) was obtained to access the kidney function of CKD patients. Multiple linear regression models were used to adjust for confounding factors in statistical analyzes. The diagnostic capabilities of the parameters were evaluated by logistic regression models.
Results: 131 nondiabetic CKD patients and 62 HC entered the study. eGFR was found significantly associated with parafoveal VD and PD (average PD: β = 0.000 4, Padjusted < 0.001) in various sectors. Thinning of pRNFL (β = -6.725, Padjusted < 0.001) and GC-IPL (β = -4.542, Padjusted < 0.001), as well as decreased VD (β = -2.107, P- adjusted < 0.001) and PD (β = -0.057, Padjusted = 0.0328) were found in CKD patients. Thinning of pRNFL and deteriorated perifoveal vasculature were found in early CKD, and the parafoveal and foveal VD significantly declined in advanced CKD. Logistic regression models were employed, and selected neurovascular parameters showed an AUC of 0.853 (95% Confidence Interval [CI]: 0.795 to 0.910) in distinguishing CKD patients from HC.
Conclusions: Distinctive retinal neurovascular characteristics could be observed in nondiabetic CKD patients of different severities. Our results suggest that retinal manifestations could be valuable in the screening, diagnosis, and follow-up evaluation of patients with CKD.
Aims: To identify the characteristic retinal neurovascular changes in patients in different stages of nondiabetic chronic kidney disease (CKD) and to develop a model for the accurate diagnosis of nondiabetic CKD.
Methods: Peripapillary retinal nerve fiber layer (pRNFL) thickness and average macular ganglion cell-inner plexiform layer (GC-IPL) thickness of nondiabetic CKD patients and healthy controls (HC) were evaluated by spectral-domain optical coherence tomography (OCT). The vessel density (VD) and perfusion density (PD) of the macula were obtained from optical coherence tomography angiography (OCTA). The estimated glomerular filtration rate (eGFR) was obtained to access the kidney function of CKD patients. Multiple linear regression models were used to adjust for confounding factors in statistical analyzes. The diagnostic capabilities of the parameters were evaluated by logistic regression models.
Results: 131 nondiabetic CKD patients and 62 HC entered the study. eGFR was found significantly associated with parafoveal VD and PD (average PD: β = 0.000 4, Padjusted < 0.001) in various sectors. Thinning of pRNFL (β = -6.725, Padjusted < 0.001) and GC-IPL (β = -4.542, Padjusted < 0.001), as well as decreased VD (β = -2.107, Padjusted < 0.001) and PD (β = -0.057, Padjusted = 0.0328) were found in CKD patients. Thinning of pRNFL and deteriorated perifoveal vasculature were found in early CKD, and the parafoveal and foveal VD significantly declined in advanced CKD. Logistic regression models were employed, and selected neurovascular parameters showed an AUC of 0.853 (95% Confidence Interval [CI]: 0.795 to 0.910) in distinguishing CKD patients from HC.
Conclusions: Distinctive retinal neurovascular characteristics could be observed in nondiabetic CKD patients of different severities. Our results suggest that retinal manifestations could be valuable in the screening, diagnosis, and follow-up evaluation of patients with CKD.
近年来随着人口老龄化的发展、人群用眼方式的改变,现有的眼科医疗资源正越来越难以满足日渐增长的医疗需求,亟需新型的诊疗模式予以补足。眼科人工智能作为眼科领域的新兴元素,在眼病的筛查诊断中发展迅速,主要表现为“眼部图像数据+人工智能”的模式。近年来,随着该模式在白内障、青光眼、糖尿病性视网膜病变(diabetic retinopathy,DR)等常见病中研究的深入,相关技术日渐成熟,表现出了较大的应用优势与应用前景,部分技术甚至成功转化并被逐渐应用于临床。眼科诊疗向智慧医学模式的过渡,有望缓解日益增长的医疗需求与紧缺的医疗资源之间的矛盾,从而提高整体的医疗服务水平。
The development of population aging and changes in the way people use their eyes over the recent years have increasingly challenged the existing ophthalmic medical resources to meet the growing medical needs, thus urgently calling for a novel diagnostic and treatment mode. Despite its status as an emerging sector in ophthalmology, ophthalmic artificial intelligence has developed rapidly in the screening and diagnosis of eye diseases, as can be seen in practices adopting the “eye imaging data + AI” mode. In recent years, with the intensified research on this mode with respect to common diseases such as cataract, glaucoma and diabetic retinopathy, relevant technologies have grown increasingly mature, presenting undeniable application superiority and prospects. Some of the relevant technical achievements have also been successfully transformed for practical usage, and are gradually being applied to clinical practices. Ophthalmic diagnosis and treatment are transitioning toward the era of intelligent medical services, which are expected to reduce the contradictions between the growing medical needs and the shortage of medical resources, as well as ultimately improve the overall experience of medical services.
目的:住院医师委托培养是我国医学教育标准化和国际化的重要举措。本研究采用客观考核和主观问卷两种方法评估我国眼科住院医师委托培养效果。方法:本研究对象为广东省深圳市政府于2012年8月至2015年7月期间委托中山大学中山眼科中心进行眼科住院医师规范化培训的9名学员。本研究对所有学员的基本信息、临床培训情况、以及考核成绩等客观指标进行统计分析,同时设 计了一份包含13个问题的调查问卷,对每位学员的培训情况进行主观评估。结果:本研究纳入9名研究对象,包括2名男性和7名女性,平均年龄为(26±3)岁,学历水平情况为学士7名和硕士2名,其中有7名毕业于国家重点医学院校。3年内人均轮转眼科亚专科超过10科,平均主管病人数及参加手术例数分别为736和1 219例,门诊工作总量人均6 274人次,所有学员均按规定至少完成综 述1篇。8名(88.89%)学员认为可以独立诊断并治疗大部分常见眼科疾病,而且能独立完成大部分眼科基本临床操作,5名(55.56%)学员可以单独完成翼状胬肉切除术、霰粒肿刮除术、前房穿刺术等。6名(66.67%)学员认为培训时间安排合理。8名(88.89%)学员对这次委托培养总体比较满意。 所有9名培训学员中,最终有7名(77.8%)顺利通过中山大学第一阶段住院医师规范化培训考核。结论:培训学员对现行的眼科住院医师规范化培训方案的接受程度较高,基本达到预期的培训效果,对常见眼病能独立进行诊治。委托培养学员的学历和学习能力的差异在一定程度上影响了最终的培训考核通过率。
Objective: The entrusted standardized training for residents is an important measure to gear the medical education in China to the international conventions. In this study, the effectiveness of standardized training for entrusted ophthalmic residents in China was evaluated both objectively and subjectively. Methods: Nine ophthalmic residents, commissioned by Shenzhen government of Guangdong Province, studied at Zhongshan Ophthalmic Center, Sun Yat-sen University during August 2012 to July 2015 were included in this study. The objective indicators of all participants were analyzed, including the basic information, clinical training, the score of examination, etc. The subjective self-assessment was also implemented thought a questionnaire including 13 designed questions.Results: All 9 participants included 2 males and 7 females, 2 medical masters and 7 bachelors, and the mean age was 26±3 years. Seven of them graduated from the national key medical universities. The mean number of rotated sub-clinical departments was 10.3, the mean number of managed inpatients and the participated operations were 736 and 1,219, respectively. The total number of managed outpatients was 6,274 in average. All participants completed at least one review article. Eight (88.89%) participants could independently diagnose and treat the most common ophthalmic diseases, they also could complete the basic clinical ophthalmic operation independently. Five (55.56%) participants could independently manage the pterygium excision, curettage of chalazion, anterior chamber penetration, etc. And 6 (66.67%) of the participants believed that the training length was reasonable. Eight (88.89%) of them were satisfied with the standardized training for residents on the whole. Finally, 7 participants successfully passed the first stage of standardized training program in Sun Yat-sen University. Conclusion: There was a high level acceptance rate of the standardized training programs for entrusted ophthalmic residents. The participants achieved the expected training effects, and could managed the diagnosis and treatment of common ophthalmic diseases independently. But the training effects and passing rate of examination were partly affected by the learning ability of the training students.
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.
目的:评估白内障人工智能辅助诊断系统在社区筛查中的应用效果。方法:采用前瞻性观察性研究方法对白内障人工辅助诊断系统的应用效果进行分析,结合远程医疗的模式,由社区卫生人员对居民进行病史采集、视力检查和裂隙灯眼前节检查等,将数据上传至云平台,由白内障人工智能辅助诊断系统和人类医生依次进行白内障评估。结果:受检人群中男性所占比例为35.7%,年龄中位数为66岁,裂隙灯眼前节照片有98.7%的图像质量合格。该白内障人工智能辅助诊断系统在外部验证集中检出重度白内障的曲线下面积为0.915。在人类医生建议转诊的病例中,有80.3%也由人工智能系统给出了相同的建议。结论:该白内障人工智能辅助诊断系统在白内障社区筛查的应用中具有较好的可行性和准确性,为开展社区筛查疾病提供了参考依据。
Objective: To evaluate the effectiveness of an artificial intelligence-assisted diagnostic system for cataract screening in community. Methods: A prospective observational study was carried out based on a telemedicine platform. Patient history, medical records and anterior ocular segment images were collected and transmitted from community healthcare centers to Zhongshan Ophthalmic Center for evaluation by both ophthalmologists and artificial intelligence-assisted cataract diagnostic system. Results: Of all enumerated subjects, 35.7% were male and the median age was 66 years old. Of all enumerated slit-lamp images, 98.7% met the requirement of acceptable quality. This artificial intelligence-assisted diagnostic system achieved an AUC of 0.915 for detection of severe cataracts in the external validation dataset. For subjects who were advised to be referred to tertiary hospitals by doctors, 80.3% of them received the same suggestion from this artificial intelligence-assisted diagnostic system.Conclusion: This artificial intelligence-assisted cataract diagnostic system showed high applicability and accuracy in community-based cataract screening and could be a potential model of care in community-based disease screening.