基于眼底照片检测青光眼性视神经病变的深度学习系统的开发与验证
该横断面研究利用来自中国青光眼研究联盟、邯郸眼研究及在线数据库的眼底图像开发了GD-CNN深度学习系统用于检测青光眼性视神经病变(GON)。在本地验证集中,GD-CNN模型AUC为0.996(95%CI,0.995-0.998),敏感度96.2%,特异度97.7%。研究提示该自动化系统对不同来源和质量的图像具有高敏感度、特异度和泛化能力,可能有助于提升现有筛查项目的成本效益和时间效率。
要点
- GD-CNN模型在本地验证集中检测GON的AUC为0.996(95%CI,0.995-0.998),敏感度96.2%,特异度97.7%。
- GD-CNN与人工判读中最常见的假阴性和假阳性原因均为病理性近视或高度近视。
- GD-CNN在不同来源和不同质量眼底图像中表现出高敏感度、特异度和泛化能力。
核心结论
- GD-CNN应用于不同场景和不同质量眼底图像检测GON时表现出高敏感度、特异度和泛化能力。
- 上述发现提示自动化深度学习系统可能以成本效益高和时间效率高的方式增强现有筛查项目。
研究设计与样本
- 该研究为横断面研究,开发了基于卷积神经网络的GD-CNN深度学习系统用于自动分类GON。
- 研究从中国青光眼研究联盟、邯郸眼研究和在线数据库获取眼底图像,选取241032张图像作为训练数据集,来自68013例患者(明确GON 29865张[12.4%]、可能GON 11046张[4.6%]、不太可能GON 200121张[83%])。
- GD-CNN模型的验证和评估使用中国青光眼研究联盟剩余的28569张图像。
- 泛化能力在多个验证数据集中测试,包括无排除条件的临床场景、基于网站获取的眼底照片的变质量图像、反映队列中青光眼患者自然分布的基于人群的研究以及具有多样民族分布的附加数据集。
- 建立了在线学习系统,将经过训练和验证的GD-CNN迁移以泛化来自新来源的眼底图像结果;并进行了预测可视化测试以识别DLS用于诊断的眼底图像区域。
主要发现
- GD-CNN模型在主要本地验证数据集中的AUC为0.996(95%CI,0.995-0.998),敏感度为96.2%,特异度为97.7%。
- GD-CNN判读中假阴性和假阳性最常见原因分别为51/119(46.3%)和191/588(32.3%),人工判读中分别为50/113(44.2%)和183/538(34.0%),原因均为病理性近视或高度近视。
关键数值与推荐
- GD-CNN模型在主要本地验证数据集中检测GON的AUC为0.996(95%CI,0.995-0.998),敏感度96.2%,特异度97.7%。
- 训练数据集共241032张图像,其中明确GON 29865张(12.4%)、可能GON 11046张(4.6%)、不太可能GON 200121张(83%),来自68013例患者。
- 验证和评估使用中国青光眼研究联盟剩余的28569张图像。
- GD-CNN判读假阴性51/119(46.3%)、假阳性191/588(32.3%)的最常见原因为病理性近视或高度近视;人工判读分别为50/113(44.2%)和183/538(34.0%)。
临床怎么用
- 原文结论指出,上述发现提示自动化深度学习系统可能以成本效益高和时间效率高的方式增强现有筛查项目。
局限性
- 原文未报告局限性。
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
IMPORTANCE: A deep learning system (DLS) that could automatically detect glaucomatous optic neuropathy (GON) with high sensitivity and specificity could expedite screening for GON. OBJECTIVE: To establish a DLS for detection of GON using retinal fundus images and glaucoma diagnosis with convoluted neural networks (GD-CNN) that has the ability to be generalized across populations. DESIGN, SETTING, AND PARTICIPANTS: In this cross-sectional study, a DLS for the classification of GON was developed for automated classification of GON using retinal fundus images obtained from the Chinese Glaucoma Study Alliance, the Handan Eye Study, and online databases. The researchers selected 241 032 images were selected as the training data set. The images were entered into the databases on June 9, 2009, obtained on July 11, 2018, and analyses were performed on December 15, 2018. The generalization of the DLS was tested in several validation data sets, which allowed assessment of the DLS in a clinical setting without exclusions, testing against variable image quality based on fundus photographs obtained from websites, evaluation in a population-based study that reflects a natural distribution of patients with glaucoma within the cohort and an additive data set that has a diverse ethnic distribution. An online learning system was established to transfer the trained and validated DLS to generalize the results with fundus images from new sources. To better understand the DLS decision-making process, a prediction visualization test was performed that identified regions of the fundus images utilized by the DLS for diagnosis. EXPOSURES: Use of a deep learning system. MAIN OUTCOMES AND MEASURES: Area under the receiver operating characteristics curve (AUC), sensitivity and specificity for DLS with reference to professional graders. RESULTS: From a total of 274 413 fundus images initially obtained from CGSA, 269 601 images passed initial image quality review and were graded for GON. A total of 241 032 images (definite GON 29 865 [12.4%], probable GON 11 046 [4.6%], unlikely GON 200 121 [83%]) from 68 013 patients were selected using random sampling to train the GD-CNN model. Validation and evaluation of the GD-CNN model was assessed using the remaining 28 569 images from CGSA. The AUC of the GD-CNN model in primary local validation data sets was 0.996 (95% CI, 0.995-0.998), with sensitivity of 96.2% and specificity of 97.7%. The most common reason for both false-negative and false-positive grading by GD-CNN (51 of 119 [46.3%] and 191 of 588 [32.3%]) and manual grading (50 of 113 [44.2%] and 183 of 538 [34.0%]) was pathologic or high myopia. CONCLUSIONS AND RELEVANCE: Application of GD-CNN to fundus images from different settings and varying image quality demonstrated a high sensitivity, specificity, and generalizability for detecting GON. These findings suggest that automated DLS could enhance current screening programs in a cost-effective and time-efficient manner.
引用
- Liu H, Li L, Wormstone IM, et al. Development and Validation of a Deep Learning System to Detect Glaucomatous Optic Neuropathy Using Fundus Photographs. JAMA Ophthalmol. 2019. PMID: 31513266. doi:10.1001/jamaophthalmol.2019.3501
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。