深度学习Grad-CAM与影像组学联合用于数字乳腺断层摄影中结构扭曲的自动定位与诊断
本研究纳入500例数字乳腺断层摄影(DBT)报告提示结构扭曲(AD)的病例,前292例用于训练、后208例用于测试,采用深度学习Grad-CAM自动定位病灶并生成感兴趣区(ROI),再经影像组学模型估计恶性概率并构建ROC曲线,与放射科医师手动勾画ROI比较。测试集中手动ROI的诊断AUC为0.82,深度学习生成ROI的AUC为0.84;纯AD病例中深度学习生成ROI的AUC为0.77,显著低于AD合并其他特征者的0.86。
要点
- 深度学习Grad-CAM可自动定位DBT上的结构扭曲并生成ROI,诊断效能与手动ROI相当。
- 全部病例恶性率为57%(285/500),纯AD恶性率39.7%(106/267)显著低于AD合并其他特征者的76.8%(179/233)。
- 纯AD病例诊断更难,深度学习生成ROI的AUC为0.77,低于AD合并其他特征者的0.86。
核心结论
- 深度学习可在DBT上检测结构扭曲,诊断效能与手动ROI相当;该策略对纯AD有效,但效能低于AD合并其他特征的病例。
研究设计与样本
- 研究纳入500例DBT报告提示结构扭曲的病例,前292例用于训练、后208例用于测试,以病理恶性概率为参考构建ROC曲线并用DeLong检验比较。
关键影像表现/征象
- 结构扭曲病例分为纯AD和AD合并其他特征两类,后者包括肿块、区域性高密度和钙化。
- 全部500例中恶性285例(57%),纯AD 267例中恶性106例(39.7%),AD合并其他特征233例中恶性179例(76.8%)。
诊断标准与关键数值
- 测试集中手动ROI的诊断AUC为0.82,深度学习生成ROI的AUC为0.84。
- 纯AD病例中深度学习生成ROI的AUC为0.77,AD合并其他特征者为0.86。
临床应用与阅片要点
- 深度学习Grad-CAM可自动定位结构扭曲并生成ROI,用于影像组学模型估计恶性概率,效能与放射科医师手动勾画ROI相当。
- 纯AD病例的恶性概率估计效能低于AD合并其他特征的病例,阅片时需结合是否存在肿块、区域性高密度和钙化等特征综合判断。
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)。
英文原文(摘要)· 点击展开
RATIONALE AND OBJECTIVES: Detection and diagnosis of architectural distortion (AD) on digital breast tomosynthesis (DBT) is challenging. This study applied artificial intelligence (AI) using deep learning (DL) algorithms to detect AD, followed by radiomics for classification.
MATERIALS AND METHODS: 500 cases with AD on DBT reports were identified; the earlier 292 cases for training, and the later 208 cases for testing. The DL Gradient-weighted Class Activation Mapping (Grad-CAM) was applied to automatically localize abnormalities and generate a region of interest (ROI), which was put into the radiomics model to estimate the malignancy probability for constructing ROC curves. Radiologists delineated ROI manually for comparison. Cases were categorized into pure AD and AD associated with other features, including mass, regional high-density, and calcifications. The ROC curves were compared using the DeLong test.
RESULTS: The overall malignancy rate was 57% (285/500). Of them, 267 cases were classified as pure AD, and the malignancy rate (106/267 = 39.7%) was significantly lower compared to AD cases associated with other features (179/233 = 76.8%, p < 0.01). In the testing set, the diagnostic AUC was 0.82 when using the manual ROI and 0.84 when using the DL-generated ROI. In the more challenging pure AD cases, DL-generated ROI yielded an AUC of 0.77, significantly lower than 0.86 for AD associated with other features.
CONCLUSION: DL could detect AD on DBT, and the diagnostic performance was comparable to manual ROI. The strategy worked for pure AD, but the performance was worse than that for AD with other features.
引用
- Chen X, Zhang Y, Zhou J, et al. Combination of Deep Learning Grad-CAM and Radiomics for Automatic Localization and Diagnosis of Architectural Distortion on DBT. Acad Radiol. 2024. PMID: 39496537. doi:10.1016/j.acra.2024.10.031
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。