使用深度学习算法鉴别青光眼性视神经病变与非青光眼性视神经病变
本横断面研究训练、验证并外部测试了一个深度学习系统,用2183张彩色眼底照片将视盘分类为正常、青光眼性视神经病变(GON)或非青光眼性视神经病变(NGON)。单中心数据集中DenseNet121表现最佳,敏感度95.36%、精确率95.35%、特异度92.19%、F1分数95.40%;外部验证中区分GON与NGON的敏感度和特异度分别为85.53%和89.02%,均高于青光眼专科医师的71.05%和82.21%。
要点
- 深度学习系统可将视盘分类为正常、GON或NGON,单中心数据集最佳模型为DenseNet121。
- 外部验证中网络区分GON与NGON的敏感度85.53%、特异度89.02%。
- 青光眼专科医师盲法诊断的敏感度71.05%、特异度82.21%,低于算法。
核心结论
- 所提出的鉴别GON与NGON的算法敏感度高于青光眼专科医师,其在未见数据上的应用因此极具前景。
研究设计与样本
- 横断面研究,训练、验证并外部测试深度学习系统以将视盘分类为正常、GON或NGON,使用2183张数字彩色眼底照片。
- 单中心数据集1822张图像(NGON 660张、GON 676张、正常视盘486张)用于训练和验证。
- 来自4个不同数据集的361张照片用于外部测试。
主要发现
- 单中心数据集分类表现最佳的算法为DenseNet121,敏感度95.36%、精确率95.35%、特异度92.19%、F1分数95.40%。
- 外部验证数据中,网络区分GON与NGON的敏感度和特异度分别为85.53%和89.02%。
- 盲法诊断这些病例的青光眼专科医师敏感度为71.05%、特异度82.21%。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
PURPOSE: A deep learning framework to differentiate glaucomatous optic disc changes due to glaucomatous optic neuropathy (GON) from non-glaucomatous optic disc changes due to non-glaucomatous optic neuropathies (NGONs). DESIGN: Cross-sectional study. METHOD: A deep-learning system was trained, validated, and externally tested to classify optic discs as normal, GON, or NGON, using 2183 digital color fundus photographs. A Single-Center data set of 1822 images (660 images of NGON, 676 images of GON, and 486 images of normal optic discs) was used for training and validation, whereas 361 photographs from 4 different data sets were used for external testing. Our algorithm removed the redundant information from the images using an optic disc segmentation (OD-SEG) network, after which we performed transfer learning with various pre-trained networks. Finally, we calculated sensitivity, specificity, F1-score, and precision to show the performance of the discrimination network in the validation and independent external data set. RESULTS: For classification, the algorithm with the best performance for the Single-Center data set was DenseNet121, with a sensitivity of 95.36%, precision of 95.35%, specificity of 92.19%, and F1 score of 95.40%. For the external validation data, the sensitivity and specificity of our network for differentiating GON from NGON were 85.53% and 89.02%, respectively. The glaucoma specialist who diagnosed those cases in masked fashion had a sensitivity of 71.05% and a specificity of 82.21%. CONCLUSIONS: The proposed algorithm for the differentiation of GON from NGON yields results that have a higher sensitivity than those of a glaucoma specialist, and its application for unseen data thus is extremely promising.
引用
- Vali M, Mohammadi M, Zarei N, et al. Differentiating Glaucomatous Optic Neuropathy From Non-glaucomatous Optic Neuropathies Using Deep Learning Algorithms. Am J Ophthalmol. 2023. PMID: 36868341. doi:10.1016/j.ajo.2023.02.016
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。