Wang Zengshuo, Zou Haohan, Guo Yin, Sun Minghe, Zhao Xin, Wang Yan, Sun Mingzhu
Nankai University Eye Institute, Nankai University, Tianjin, China.
National Key Laboratory of Intelligent Tracking and Forecasting for Infectious Diseases, Engineering Research Center of Trusted Behavior Intelligence, Ministry of Education, Tianjin Key Laboratory of Intelligent Robotics, Institute of Robotics and Automatic Information System, Nankai University, Tianjin, China.
Transl Vis Sci Technol. 2025 Aug 1;14(8):38. doi: 10.1167/tvst.14.8.38.
Identifying and monitoring the onset and progression of myopia (myopia onset and progression [MOP]) based on the changes in anatomical structures in fundus retinal images has significant clinical application prospects. For this purpose, we tested the performance of deep neural networks.
We established a deep neural network, called Myopic-Net, to detect anatomical changes owing to the MOP from a pair of retinal images collected during different fundoscopies. Myopic-Net was developed using 3964 fundus image pairs without MOP and 2380 fundus image pairs with MOP. Five indicators-accuracy, precision, recall, specificity, and F1-score-were evaluated on the internal test set and the independent external test set. In addition, we use a deep network visualization method to explore the factors driving Myopic-Net to predict.
On the internal test set, Myopic-Net achieved an accuracy of 87.3%; the precision, recall, and specificity were 86.2%, 80.1%, and 91.9% respectively, while the identification accuracy of two ophthalmologists is only 66.1% and 73.5%, respectively. Even on the external test set, Myopic-Net still achieved an accuracy of 84.1%. In addition, we found that the factors driving Myopic-Net to predict are mainly anatomical changes in the optic disc and surrounding areas.
Myopic-Net has been shown to be able to identify the MOP from fundus image pairs using anatomical changes in optic disc and surrounding areas. And Myopic-Net has good accuracy, reliability, and generalization ability. These factors show that deep neural networks have strong potential in monitoring and final diagnosing the MOP based on fundus image analysis.
With the development of fundus imaging technology based on intelligent mobile terminals, embedding the program based on Myopic-Net has great potential to achieve convenient and fast personalized monitoring of myopia.
基于眼底视网膜图像解剖结构的变化来识别和监测近视的发生与发展（近视发生与发展[MOP]）具有重要的临床应用前景。为此，我们测试了深度神经网络的性能。
我们建立了一个名为Myopic-Net的深度神经网络，用于从不同眼底检查时采集的一对视网膜图像中检测因MOP引起的解剖结构变化。Myopic-Net是使用3964对无MOP的眼底图像和2380对有MOP的眼底图像开发的。在内部测试集和独立外部测试集上评估了五个指标——准确率、精确率、召回率、特异性和F1分数。此外，我们使用深度网络可视化方法来探索驱动Myopic-Net进行预测的因素。
在内部测试集上，Myopic-Net的准确率达到87.3%；精确率、召回率和特异性分别为86.2%、80.1%和91.9%，而两位眼科医生的识别准确率分别仅为66.1%和73.5%。即使在外部测试集上，Myopic-Net仍达到了84.1%的准确率。此外，我们发现驱动Myopic-Net进行预测的因素主要是视盘及其周围区域的解剖结构变化。
已证明Myopic-Net能够利用视盘及其周围区域的解剖结构变化从眼底图像对中识别MOP。并且Myopic-Net具有良好的准确性、可靠性和泛化能力。这些因素表明深度神经网络在基于眼底图像分析监测和最终诊断MOP方面具有强大潜力。
随着基于智能移动终端的眼底成像技术的发展，嵌入基于Myopic-Net的程序在实现便捷、快速的个性化近视监测方面具有巨大潜力。
