Department of Ophthalmology, Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine, Shanghai, China.
Department of Ophthalmology, Shanghai Children's Hospital, Shanghai Jiaotong University, Shanghai, China.
Transl Vis Sci Technol. 2021 Jan 27;10(1):33. doi: 10.1167/tvst.10.1.33. eCollection 2021 Jan.
This study implements and demonstrates a deep learning (DL) approach for screening referable horizontal strabismus based on primary gaze photographs using clinical assessments as a reference. The purpose of this study was to develop and evaluate deep learning algorithms that screen referable horizontal strabismus in children's primary gaze photographs.
DL algorithms were developed and trained using primary gaze photographs from two tertiary hospitals of children with primary horizontal strabismus who underwent surgery as well as orthotropic children who underwent routine refractive tests. A total of 7026 images (3829 non-strabismus from 3021 orthoptics [healthy] subjects and 3197 strabismus images from 2772 subjects) were used to develop the DL algorithms. The DL model was evaluated by 5-fold cross-validation and tested on an independent validation data set of 277 images. The diagnostic performance of the DL algorithm was assessed by calculating the accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC).
Using 5-fold cross-validation during training, the average AUCs of the DL models were approximately 0.99. In the external validation data set, the DL algorithm achieved an AUC of 0.99 with a sensitivity of 94.0% and a specificity of 99.3%. The DL algorithm's performance (with an accuracy of 0.95) in diagnosing referable horizontal strabismus was better than that of the resident ophthalmologists (with accuracy ranging from 0.81 to 0.85).
We developed and evaluated a DL model to automatically identify referable horizontal strabismus using primary gaze photographs. The diagnostic performance of the DL model is comparable to or better than that of ophthalmologists.
DL methods that automate the detection of referable horizontal strabismus can facilitate clinical assessment and screening for children at risk of strabismus.
本研究通过临床评估作为参考，基于第一眼注视照片，实施并演示一种深度学习（DL）方法来筛查可转诊的水平斜视。本研究旨在开发和评估用于筛查儿童第一眼注视照片中可转诊的水平斜视的深度学习算法。
使用来自两家三级医院的第一眼注视照片，对患有原发性水平斜视并接受手术的儿童和接受常规屈光检查的正位儿童的 DL 算法进行开发和训练。共使用 7026 张图像（3829 张来自 3021 名视轴矫正术 [健康]受试者的非斜视图像和 2772 名受试者中的 3197 张斜视图像）来开发 DL 算法。使用 5 折交叉验证对 DL 模型进行评估，并在 277 张独立验证数据集中进行测试。通过计算准确性、敏感性、特异性和接收者操作特征曲线（AUC）下的面积来评估 DL 算法的诊断性能。
在训练过程中使用 5 折交叉验证，DL 模型的平均 AUC 约为 0.99。在外部验证数据集，DL 算法的 AUC 为 0.99，灵敏度为 94.0%，特异性为 99.3%。DL 算法在诊断可转诊的水平斜视方面的性能（准确率为 0.95）优于住院医师（准确率范围为 0.81 至 0.85）。
我们开发并评估了一种使用第一眼注视照片自动识别可转诊水平斜视的 DL 模型。DL 模型的诊断性能与眼科医生相当或优于眼科医生。
筱竹
