Department of Gastroenterology, Nippon Medical School Hospital, 1-1-5, Sendagi, Bunkyo-ku, Tokyo, 113-8603, Japan.
Tada Tomohiro Institute of Gastroenterology and Proctology, Saitama, Japan.
BMC Gastroenterol. 2022 May 12;22(1):237. doi: 10.1186/s12876-022-02312-y.
Endocytoscopy (ECS) aids early gastric cancer (EGC) diagnosis by visualization of cells. However, it is difficult for non-experts to accurately diagnose EGC using ECS. In this study, we developed and evaluated a convolutional neural network (CNN)-based system for ECS-aided EGC diagnosis.
We constructed a CNN based on a residual neural network with a training dataset comprising 906 images from 61 EGC cases and 717 images from 65 noncancerous gastric mucosa (NGM) cases. To evaluate diagnostic ability, we used an independent test dataset comprising 313 images from 39 EGC cases and 235 images from 33 NGM cases. The test dataset was further evaluated by three endoscopists, and their findings were compared with CNN-based results.
The trained CNN required 7.0 s to analyze the test dataset. The area under the curve of the total ECS images was 0.93. The CNN produced 18 false positives from 7 NGM lesions and 74 false negatives from 28 EGC lesions. In the per-image analysis, the accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 83.2%, 76.4%, 92.3%, 93.0%, and 74.6%, respectively, with the CNN and 76.8%, 73.4%, 81.3%, 83.9%, and 69.6%, respectively, for the endoscopist-derived values. The CNN-based findings had significantly higher specificity than the findings determined by all endoscopists. In the per-lesion analysis, the accuracy, sensitivity, specificity, PPV, and NPV of the CNN-based findings were 86.1%, 82.1%, 90.9%, 91.4%, and 81.1%, respectively, and those of the results calculated by the endoscopists were 82.4%, 79.5%, 85.9%, 86.9%, and 78.0%, respectively.
Compared with three endoscopists, our CNN for ECS demonstrated higher specificity for EGC diagnosis. Using the CNN in ECS-based EGC diagnosis may improve the diagnostic performance of endoscopists.
内镜下细胞学检查（ECS）通过观察细胞辅助早期胃癌（EGC）的诊断。然而，非专家很难通过 ECS 准确诊断 EGC。在这项研究中，我们开发并评估了一种基于卷积神经网络（CNN）的 ECS 辅助 EGC 诊断系统。
我们构建了一个基于残差神经网络的 CNN，训练数据集包括 61 例 EGC 病例的 906 张图像和 65 例非癌性胃黏膜（NGM）病例的 717 张图像。为了评估诊断能力，我们使用了一个独立的测试数据集，该数据集包含 39 例 EGC 病例的 313 张图像和 33 例 NGM 病例的 235 张图像。测试数据集由三名内镜医生进一步评估，并将他们的发现与基于 CNN 的结果进行比较。
经过训练的 CNN 分析测试数据集需要 7.0 秒。总 ECS 图像的曲线下面积为 0.93。CNN 从 7 个 NGM 病变中产生了 18 个假阳性，从 28 个 EGC 病变中产生了 74 个假阴性。在逐张图像分析中，CNN 的准确性、敏感性、特异性、阳性预测值（PPV）和阴性预测值（NPV）分别为 83.2%、76.4%、92.3%、93.0%和 74.6%，而内镜医生得出的值分别为 76.8%、73.4%、81.3%、83.9%和 69.6%。CNN 的特异性明显高于所有内镜医生的发现。在逐病变分析中，CNN 发现的准确性、敏感性、特异性、PPV 和 NPV 分别为 86.1%、82.1%、90.9%、91.4%和 81.1%，而内镜医生计算的结果分别为 82.4%、79.5%、85.9%、86.9%和 78.0%。
与三名内镜医生相比，我们的 ECS 专用 CNN 对 EGC 诊断具有更高的特异性。在 ECS 辅助 EGC 诊断中使用 CNN 可能会提高内镜医生的诊断性能。
