文献速递青光眼神经眼科

深度学习用于视盘出血的局部检测

Deep Learning for Localized Detection of Optic Disc Hemorrhages.
PubMed 文献 · Am J Ophthalmol · 2023-07-23

本研究开发了基于视盘照片自动检测视盘出血存在及位置的深度学习系统,使用来自5家机构的1562只眼的597张含至少1处视盘出血图像和1075张无视盘出血图像,由专家分级后分为训练集1340张、验证集167张和测试集165张。目标检测模型在全部165张留出图像上的AUC为0.936(95%CI 0.857-0.964),显著优于图像分类模型的AUC 0.845(95%CI 0.740-0.912;P=0.006),并与2名青光眼专科医生的表现相当。

要点

  • 目标检测模型AUC 0.936,显著优于图像分类模型AUC 0.845(P=0.006)。
  • 高特异性操作点:特异性94.3%、敏感性70.0%,与专家医生差异无统计学意义(P=0.7)。
  • 高敏感性操作点:敏感性96.7%、特异性73.3%。

核心结论

  • 自主目标检测模型在检测视盘出血方面优于图像分类模型,并与2名临床医生的表现相当。

研究设计与样本

  • 研究为深度学习算法的开发与测试,使用来自5家机构的1562只眼的视盘照片,其中597张图像含至少1处视盘出血、1075张图像无任何视盘出血,由专家分级者根据视盘出血的有无进行分类。
  • 图像被分为训练集(n=1340)、验证集(n=167)和测试集(n=165),训练了基于目标级检测或图像级分类的两种先进深度学习算法。
  • 模型之间相互比较,并与2名独立的青光眼专科医生进行比较,通过受试者工作特征曲线下面积(AUC)评估模型性能,AUC比较采用Hanley-McNeil方法。

主要发现

  • 目标检测模型在全部留出图像(n=165张照片)上的AUC为0.936(95%CI=0.857-0.964),显著优于图像分类模型的AUC=0.845(95%CI=0.740-0.912;P=0.006)。
  • 在高特异性操作点,模型特异性为94.3%、敏感性为70.0%,与专家临床医生的差异无统计学意义(P=0.7)。
  • 在高敏感性操作点,模型敏感性为96.7%、特异性为73.3%。

关键数值与推荐

  • 目标检测模型AUC=0.936(95%CI=0.857-0.964),图像分类模型AUC=0.845(95%CI=0.740-0.912),两者比较P=0.006。
  • 高特异性操作点:特异性94.3%、敏感性70.0%,与专家临床医生比较P=0.7。
  • 高敏感性操作点:敏感性96.7%、特异性73.3%。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

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

PURPOSE: To develop an automated deep learning system for detecting the presence and location of disc hemorrhages in optic disc photographs. DESIGN: Development and testing of a deep learning algorithm. METHODS: Optic disc photos (597 images with at least 1 disc hemorrhage and 1075 images without any disc hemorrhage from 1562 eyes) from 5 institutions were classified by expert graders based on the presence or absence of disc hemorrhage. The images were split into training (n = 1340), validation (n = 167), and test (n = 165) datasets. Two state-of-the-art deep learning algorithms based on either object-level detection or image-level classification were trained on the dataset. These models were compared to one another and against 2 independent glaucoma specialists. We evaluated model performance by the area under the receiver operating characteristic curve (AUC). AUCs were compared with the Hanley-McNeil method. RESULTS: The object detection model achieved an AUC of 0.936 (95% CI = 0.857-0.964) across all held-out images (n = 165 photographs), which was significantly superior to the image classification model (AUC = 0.845, 95% CI = 0.740-0.912; P = .006). At an operating point selected for high specificity, the model achieved a specificity of 94.3% and a sensitivity of 70.0%, which was statistically indistinguishable from an expert clinician (P = .7). At an operating point selected for high sensitivity, the model achieves a sensitivity of 96.7% and a specificity of 73.3%. CONCLUSIONS: An autonomous object detection model is superior to an image classification model for detecting disc hemorrhages, and performed comparably to 2 clinicians.

来源:Am J Ophthalmol,查看原文。版权归原作者所有。

引用

  1. Brown A, Cousins H, Cousins C, et al. Deep Learning for Localized Detection of Optic Disc Hemorrhages. Am J Ophthalmol. 2023. PMID: 37490992. doi:10.1016/j.ajo.2023.07.007

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。