人工智能定量分析不同近视程度儿童眼底豹纹状改变的差异性OACSTPCD
The difference of leopard spot density in the fundus of different myopia degrees children analyzed by artificial intelligence
目的 应用人工智能技术对近视豹纹状眼底定量分析,对比不同近视程度儿童豹纹状眼底改变的差异性.方法 收集2023年1月至2023年10月于我院眼科门诊就诊的6~13岁儿童327例(右眼327眼)验光结果、眼轴、角膜曲率和眼底照片,分为正常对照组76例、近视轻度86例、中度85例、高度80例.基于人工智能技术对近视豹纹状眼底定量,计算眼底每单位面积的平均脉络膜裸露面积,获得眼底豹纹斑密度(FTD),比较各组FTD差异.结果 高度近视组和正常对照组、轻度近视组、中度近视组的总FTD、以黄斑为中心3mm直径区域鼻上、鼻下、颞上、颞下和以黄斑为中心6mm直径区域鼻上、鼻下、颞上、颞下区域的FTD参数等相比,差异有统计学意义(P<0.05);中度近视组和正常对照组间以黄斑为中心6mm直径区域鼻下区域的差异也有统计学意义(P<0.05).结论 以黄斑为中心的鼻下区域是近视进展、眼轴增长过程中眼底最先受累部位;眼底豹纹状改变在近视早期变化不明显,但借助人工智能对其定量测量,获得FTD指数,可以帮助发现豹纹状眼底细微变化,有助于更及时、早期、全面精确的监控近视进展.
Objective Quantitative analysis of tessellated fundus by artificial intelligence techniques to compare the variability of tessellated fundus changes in children with different levels of myopia.Methods Optometry results,axial length,corneal curvature and fundus photographs of 327 right eyes of 327 cases of children aged 6~13 years attending our ophthalmology outpatient clinic from January 2023 to October 2023 were collected and categorized into 76 cases of normal control,86 cases of mild myopia,85 cases of moderate myopia and 80 cases of high myopia.Based on artificial intelli-gence techniques to quantify the myopic tessellated fundus pattern,the average choroidal exposed area per unit area of the fundus was calculated to obtain the fundus tessellated density(FTD),and the variability of the FTD was compared between the groups.Results Compared with the normal control group,the mild myopia group and the moderate myopia group,the total leopard spot density,the leopard spot density parameters of the 3mm diameter area centered on the macula above the nose,under the nose,supratemporal,infratemporal and the 6mm diameter area centered on the macula were statistically significant(P<0.05).The subnasal region with a diameter of 6mm centered on macula was also significantly different between the moderate myopia group and the normal control group(P<0.05).Conclusions The subnasal region,centered on the macula,is the initial area of the fundus to be affected by myopia and the growth of the eye axis.While tessellated fundus changes are not apparent in the early stages of myopia,the quantitative measurement of fundus tessellated density index,aided by the use of artificial intelligence,can assist in the detection of subtle changes in the tessellated fundus,thereby facilitating a more timely,comprehensive,and accurate monitoring of the progression of myopia.
刘满军;李兆生;郭雅楠;王茜;凌赛广;李莉
100045 首都医科大学附属北京儿童医院眼科100045 首都医科大学附属北京儿童医院眼科100045 首都医科大学附属北京儿童医院眼科依未科技(北京)有限公司依未科技(北京)有限公司100045 首都医科大学附属北京儿童医院眼科
人工智能近视眼底豹纹斑密度
Artificial intelligenceMyopiaFundusFundus tessellated density
《中国斜视与小儿眼科杂志》 2024 (3)
7-10,后插6,5
国家自然科学基金(81970844,82371093)
评论