桡骨远端骨折的检测与定位:深度学习系统与放射科医生的对比
本研究评估基于深度学习的图像分析软件在桡骨远端骨折检测与定位中的表现。系统在524张腕关节X线片(166张示骨折)上训练,并在内部(100张,42张骨折)和外部(200张,100张骨折)测试集上测试。结果显示系统在内部测试集表现优异,外部测试集性能下降,放射科医生在外部数据上表现更优(p < 0.05),但超过90%的热图峰值激活区域与医生标注一致。
要点
- 深度学习系统在内部测试集检测桡骨远端骨折的AUC为0.93–0.96,敏感度0.81–0.90,特异度0.86–1.0。
- 外部测试集上系统性能下降:AUC 0.80–0.89,敏感度0.64–0.92,特异度0.60–0.90。
- 放射科医生在外部数据上表现优于系统(敏感度0.88–0.98,特异度0.66–1.0,p < 0.05),但内部数据上两者相当。
关键影像表现/征象
- 系统输出0–1的评分和热图,热图用于定位骨折区域。
- 超过90%的病例中,热图峰值激活区域与放射科医生标注的感兴趣区一致。
诊断标准与关键数值
- 系统在内部测试集:AUC 0.93(95%CI 0.82–0.98)至0.96(0.87–1.00),敏感度0.81(0.58–0.95)至0.90(0.70–0.99),特异度0.86(0.68–0.96)至1.0(0.88–1.0)。
- 系统在外部测试集:AUC 0.80(0.71–0.88)至0.89(0.81–0.94),敏感度0.64(0.49–0.77)至0.92(0.81–0.98),特异度0.60(0.45–0.74)至0.90(0.78–0.97)。
- 放射科医生在内部数据:敏感度0.71(0.48–0.89)至0.95(0.76–1.0),特异度0.52(0.32–0.71)至0.97(0.82–1.0)。
- 放射科医生在外部数据:敏感度0.88(0.76–0.96)至0.98(0.89–1.0),特异度0.66(0.51–0.79)至1.0(0.93–1.0),p < 0.05。
- 检测性能以最佳阈值下的AUC、敏感度和特异度表示。
临床应用与阅片要点
- 深度学习系统可作为辅助工具用于腕关节X线片的骨折检测和定位,热图有助于提示骨折区域。
- 外部测试集上系统性能下降,提示在不同数据分布下需谨慎解读系统输出。
- 放射科医生在外部数据上表现优于系统,系统定位与医生标注高度一致。
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
PURPOSE: To evaluate a deep learning based image analysis software for the detection and localization of distal radius fractures.
METHOD: A deep learning system (DLS) was trained on 524 wrist radiographs (166 showing fractures). Performance was tested on internal (100 radiographs, 42 showing fractures) and external test sets (200 radiographs, 100 showing fractures). Single and combined views of the radiographs were shown to DLS and three readers. Readers were asked to indicate fracture location with regions of interest (ROI). The DLS yielded scores (range 0-1) and a heatmap. Detection performance was expressed as AUC, sensitivity and specificity at the optimal threshold and compared to radiologists' performance. Heatmaps were compared to radiologists' ROIs.
RESULTS: The DLS showed excellent performance on the internal test set (AUC 0.93 (95% confidence interval (CI) 0.82-0.98) - 0.96 (0.87-1.00), sensitivity 0.81 (0.58-0.95) - 0.90 (0.70-0.99), specificity 0.86 (0.68-0.96) - 1.0 (0.88-1.0)). DLS performance decreased on the external test set (AUC 0.80 (0.71-0.88) - 0.89 (0.81-0.94), sensitivity 0.64 (0.49-0.77) - 0.92 (0.81-0.98), specificity 0.60 (0.45-0.74) - 0.90 (0.78-0.97)). Radiologists' performance was comparable on internal data (sensitivity 0.71 (0.48-0.89) - 0.95 (0.76-1.0), specificity 0.52 (0.32-0.71) - 0.97 (0.82-1.0)) and better on external data (sensitivity 0.88 (0.76-0.96) - 0.98 (0.89-1.0), specificities 0.66 (0.51-0.79) - 1.0 (0.93-1.0), p < 0.05). In over 90%, the areas of peak activation aligned with radiologists' annotations.
CONCLUSIONS: The DLS was able to detect and localize wrist fractures with a performance comparable to radiologists, using only a small dataset for training.
引用
- Blüthgen C, Becker AS, Vittoria de Martini I, et al. Detection and localization of distal radius fractures: Deep learning system versus radiologists. Eur J Radiol. 2020. PMID: 32193036. doi:10.1016/j.ejrad.2020.108925
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。