文献速递临床微生物分子诊断

深度神经网络可实现细菌性阴道病的形态学分类与诊断

Deep Neural Networks Offer Morphologic Classification and Diagnosis of Bacterial Vaginosis.
PubMed 文献 · J Clin Microbiol · 2021-01-21

该研究开发并优化卷积神经网络模型,从显微镜图像自动识别和分类Nugent评分三类,使用三家医院独立测试集评估。模型在5,815张验证图像上敏感度82.4%、特异度96.6%,在1,082张独立测试图像上三类准确率75.1%,优于技术员和产科医生,显示深度学习可自动化诊断细菌性阴道病。

要点

  • CNN模型对异常阴道菌群和BV的敏感度82.4%、特异度96.6%,优于国内顶级技术员和产科医生。
  • 独立测试集三类Nugent评分准确率75.1%,比三名医学本科技术员平均高6.6%。
  • 三名技术员对同一标本三重复检测的三类评分精确度仅54.0%,模型稳定性更优。

核心结论

  • CNN模型在三类Nugent评分诊断的准确性和稳定性上优于人类医疗从业者,深度学习有望在配套硬件支持下实现细菌性阴道病自动化诊断。

研究设计与样本

  • 开发并优化卷积神经网络模型,先用专家判定Nugent评分的显微镜图像面板建立模型,通过最小化交叉熵损失函数训练、动量优化器优化。
  • 模型包含25个卷积层、2个池化层和1个全连接层,在三家医院收集的独立测试集图像上评估。

主要发现

  • 在5,815张验证图像上,以异常阴道菌群和BV为阳性样本,模型敏感度82.4%、特异度96.6%,优于中国顶级技术员和产科医生。
  • 在1,082张独立测试图像上,模型三类Nugent评分准确率75.1%,比三名医学本科技术员平均高6.6%。

关键数值与推荐

  • 模型在5,815张验证图像上敏感度82.4%、特异度96.6%(阳性定义为异常阴道菌群和BV)。
  • 模型在1,082张独立测试图像上三类Nugent评分准确率75.1%,比三名技术员平均高6.6%。

临床怎么用

  • 该深度学习模型可在配套硬件支持下用于细菌性阴道病的自动化Nugent评分分类,辅助弥补人工判读的 labor-intensive、耗时和个体差异问题。

局限性

  • 原文未报告模型在临床常规流程中的前瞻性验证结果。
  • 原文未报告模型在不同染色批次或不同显微镜设备间的泛化性能。

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

Bacterial vaginosis (BV) is caused by the excessive and imbalanced growth of bacteria in vagina, affecting 30 to 50% of women. Gram staining followed by Nugent scoring based on bacterial morphotypes under the microscope is considered the gold standard for BV diagnosis; this method is often labor-intensive and time-consuming, and results vary from person to person. We developed and optimized a convolutional neural network (CNN) model and evaluated its ability to automatically identify and classify three categories of Nugent scores from microscope images. The CNN model was first established with a panel of microscopic images with Nugent scores determined by experts. The model was trained by minimizing the cross-entropy loss function and optimized by using a momentum optimizer. The separate test sets of images collected from three hospitals were evaluated by the CNN model. The CNN model consisted of 25 convolutional layers, 2 pooling layers, and a fully connected layer. The model obtained 82.4% sensitivity and 96.6% specificity with the 5,815 validation images when altered vaginal flora and BV were considered the positive samples, which was better than the rates achieved by top-level technologists and obstetricians in China. The capability of our model for generalization was so strong that it exhibited 75.1% accuracy in three categories of Nugent scores on the independent test set of 1,082 images, which was 6.6% higher than the average of three technologists, who are hold bachelor's degrees in medicine and are qualified to make diagnostic decisions. When three technologists ran one specimen in triplicate, the precision of three categories of Nugent scores was 54.0%. One hundred three samples diagnosed by two technologists on different days showed a repeatability of 90.3%. The CNN model outperformed human health care practitioners in terms of accuracy and stability for three categories of Nugent score diagnosis. The deep learning model may offer translational applications in automating diagnosis of bacterial vaginosis with proper supporting hardware.

来源:J Clin Microbiol,查看原文。版权归原作者所有。

引用

  1. Wang Z, Zhang L, Zhao M, et al. Deep Neural Networks Offer Morphologic Classification and Diagnosis of Bacterial Vaginosis. J Clin Microbiol. 2021. PMID: 33148709. doi:10.1128/JCM.02236-20

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。