Department of Ophthalmology, The University of Tokyo, Tokyo, Japan.
Department of Ophthalmology, The University of Tokyo, Tokyo, Japan.
Am J Ophthalmol. 2019 Feb;198:136-145. doi: 10.1016/j.ajo.2018.10.007. Epub 2018 Oct 12.
We sought to construct and evaluate a deep learning (DL) model to diagnose early glaucoma from spectral-domain optical coherence tomography (OCT) images.
Artificial intelligence diagnostic tool development, evaluation, and comparison.
This multi-institution study included pretraining data of 4316 OCT images (RS3000) from 1371 eyes with open angle glaucoma (OAG) regardless of the stage of glaucoma and 193 normal eyes. Training data included OCT-1000/2000 images from 94 eyes of 94 patients with early OAG (mean deviation > -5.0 dB) and 84 eyes of 84 normal subjects. Testing data included OCT-1000/2000 from 114 eyes of 114 patients with early OAG (mean deviation > -5.0 dB) and 82 eyes of 82 normal subjects. A DL (convolutional neural network) classifier was trained using a pretraining dataset, followed by a second round of training using an independent training dataset. The DL model input features were the 8 × 8 grid macular retinal nerve fiber layer thickness and ganglion cell complex layer thickness from spectral-domain OCT. Diagnostic accuracy was investigated in the testing dataset. For comparison, diagnostic accuracy was also evaluated using the random forests and support vector machine models. The primary outcome measure was the area under the receiver operating characteristic curve (AROC).
The AROC with the DL model was 93.7%. The AROC significantly decreased to between 76.6% and 78.8% without the pretraining process. Significantly smaller AROCs were obtained with random forests and support vector machine models (82.0% and 67.4%, respectively).
A DL model for glaucoma using spectral-domain OCT offers a substantive increase in diagnostic performance.
我们旨在构建并评估一种基于深度神经网络（DL）的模型，以从频域光学相干断层扫描（OCT）图像中诊断早期青光眼。
人工智能诊断工具的开发、评估和比较。
本多中心研究纳入了来自 1371 只眼的 4316 张 RS3000 型 OCT 图像（不论青光眼阶段），这些眼患有开角型青光眼（OAG）；还纳入了来自 94 只早期 OAG 眼（平均偏差>-5.0dB）和 84 只正常眼的 94 名患者的 OCT-1000/2000 图像，以及来自 114 只早期 OAG 眼（平均偏差>-5.0dB）和 82 只正常眼的 114 名患者的 OCT-1000/2000 图像作为测试数据。DL（卷积神经网络）分类器首先使用预训练数据集进行训练，然后使用独立的训练数据集进行第二轮训练。DL 模型的输入特征是来自频域 OCT 的 8×8 网格黄斑视网膜神经纤维层厚度和节细胞复合体层厚度。在测试数据集上评估诊断准确性。为了比较，还使用随机森林和支持向量机模型评估了诊断准确性。主要观察指标为接收者操作特征曲线下面积（AROC）。
DL 模型的 AROC 为 93.7%。如果没有预训练过程，AROC 显著降低至 76.6%至 78.8%。随机森林和支持向量机模型的 AROC 显著较小（分别为 82.0%和 67.4%）。
基于频域 OCT 的青光眼 DL 模型可显著提高诊断性能。
