深度学习模型检测需转诊糖尿病视网膜病变和青光眼的可视化
该横断面研究系统可视化了两个经验证的深度学习模型(检测需转诊糖尿病视网膜病变(DR)和青光性视神经病变(GON))的卷积神经网络。使用自适应核可视化技术处理眼底图像生成热图,由一名验光师分类。结果显示GON真阳性中90%热图局限于视盘区域,DR真阳性中96%识别出典型病变区域,提示该方法可突出传统诊断区域,支持模型有效性并可能促进临床 adoption。
要点
- GON真阳性中90%热图位于视盘区域,支持模型关注传统区域。
- DR真阳性中96%热图识别出渗出、出血或血管异常等典型病变。
- 假阳性病例热图常显示非传统区域,提示模型可能误判。
核心结论
- 该可视化方法可突出疾病诊断的传统区域,证实所研究深度学习模型的有效性,并可能促进其临床 adoption。
研究设计与样本
- 横断面研究,从GON和DR验证数据集中随机选取100例真阳性照片及所有假阳性病例。
- 使用自适应核可视化技术,滑动窗口28×28像素、步长3像素生成特征图,调整阈值生成热图,由一名验光师分类。
主要发现
- GON数据集:100例真阳性中90例(90%;95%CI,82%-95%)热图仅位于视盘区域。
- GON数据集:22例假阳性中15例(68%;95%CI,45%-86%)热图仅位于视盘区域。
- DR数据集:100例真阳性中96例(96%;95%CI,90%-99%)热图识别出渗出、出血或血管异常为最重要预后区域。
- DR数据集:46例假阳性中39例(85%;95%CI,71%-94%)热图显示非传统眼底区域(伴或不伴视网膜小静脉)。
关键数值与推荐
- GON真阳性热图位于视盘区域的比例为90%(95%CI,82%-95%)。
- DR真阳性热图识别典型病变的比例为96%(95%CI,90%-99%)。
临床怎么用
- 临床医生可利用热图可视化理解深度学习模型输出依据,增强对模型决策的信任。
- 该技术可能促进深度学习模型在眼科临床中的 adoption。
局限性
- 研究仅使用一名验光师进行图像分类,可能存在主观偏倚。
- 样本来自特定数据库,外部推广性需进一步验证。
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: Convolutional neural networks have recently been applied to ophthalmic diseases; however, the rationale for the outputs generated by these systems is inscrutable to clinicians. A visualization tool is needed that would enable clinicians to understand important exposure variables in real time. OBJECTIVE: To systematically visualize the convolutional neural networks of 2 validated deep learning models for the detection of referable diabetic retinopathy (DR) and glaucomatous optic neuropathy (GON). DESIGN, SETTING, AND PARTICIPANTS: The GON and referable DR algorithms were previously developed and validated (holdout method) using 48 116 and 66 790 retinal photographs, respectively, derived from a third-party database (LabelMe) of deidentified photographs from various clinical settings in China. In the present cross-sectional study, a random sample of 100 true-positive photographs and all false-positive cases from each of the GON and DR validation data sets were selected. All data were collected from March to June 2017. The original color fundus images were processed using an adaptive kernel visualization technique. The images were preprocessed by applying a sliding window with a size of 28 × 28 pixels and a stride of 3 pixels to crop images into smaller subimages to produce a feature map. Threshold scales were adjusted to optimal levels for each model to generate heat maps highlighting localized landmarks on the input image. A single optometrist allocated each image to predefined categories based on the generated heat map. MAIN OUTCOMES AND MEASURES: Visualization regions of the fundus. RESULTS: In the GON data set, 90 of 100 true-positive cases (90%; 95% CI, 82%-95%) and 15 of 22 false-positive cases (68%; 95% CI, 45%-86%) displayed heat map visualization within regions of the optic nerve head only. Lesions typically seen in cases of referable DR (exudate, hemorrhage, or vessel abnormality) were identified as the most important prognostic regions in 96 of 100 true-positive DR cases (96%; 95% CI, 90%-99%). In 39 of 46 false-positive DR cases (85%; 95% CI, 71%-94%), the heat map displayed visualization of nontraditional fundus regions with or without retinal venules. CONCLUSIONS AND RELEVANCE: These findings suggest that this visualization method can highlight traditional regions in disease diagnosis, substantiating the validity of the deep learning models investigated. This visualization technique may promote the clinical adoption of these models.
引用
- Keel S, Wu J, Lee PY, et al. Visualizing Deep Learning Models for the Detection of Referable Diabetic Retinopathy and Glaucoma. JAMA Ophthalmol. 2019. PMID: 30570648. doi:10.1001/jamaophthalmol.2018.6035
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。