期刊文献+

两种尘肺病计算机智能诊断系统的性能评价 被引量:1

Two Computer-aided Diagnosis Systems for Pneumoconiosis
下载PDF
导出
摘要 目的 对两种尘肺病计算机智能诊断系统性能进行评价,探讨不同的计算机智能诊断技术在尘肺病筛查中的应用价值。方法 选取国内五家尘肺病研究机构提供的500例胸部DR图像作为测试集,测试集由3名尘肺病诊断专家依据GBZ 70-2015集体诊断,尘肺病阳性/阴性=298/202。两种尘肺病智能诊断系统(分别来源于T科技和J科技)对测试集胸片是否有尘肺病进行独立诊断。利用ROC曲线、敏感性、特异性等指标对两种系统的诊断性能进行评价。结果 T科技尘肺病智能诊断系统诊断效能高于J科技,ROC曲线下面积分别为0.99(95%CI:0.98~0.99)和0.81(95%CI:0.77~0.85),P<0.05;诊断敏感性T科技高于J科技,诊断特异性两者无明显差异。T科技的尘肺病智能诊断系统诊断敏感性高于3名高年资诊断专家,与2名专家无统计学差异,诊断特异性高于2名高年资诊断专家,与3名专家无统计学差异。结论 不同技术的尘肺病智能诊断系统诊断性能存在差异,诊断敏感性和特异性不低于尘肺病高年资诊断专家。 Objective To evaluate the performance of two pneumoconiosis computeraided diagnosis systems,and to explore the application value of computer-aided diagnosis in pneumoconiosis.Methods 500 cases of chest DR images provided by five pneumoconiosis research institutions were selected as the test set.The test set was diagnosed by five senior diagnostic experts.Pneumoconiosis positive/negative=298/202.Two computer-aided diagnosis systems for pneumoconiosis(T Tech and J Tech)independently diagnose the test set.Use the area under the ROC curve(AUC),sensitivity and specificity to evaluate the two systems.Results 1.The AUC value of T Tech is higher than that of J Tech.The AUC value is 0.99(95%CI:0.98~0.99)and 0.81(95%CI:0.77~0.85),P<0.05,The sensitivity of T technology is higher than that of J technology,P<0.05;the specificity of diagnosis is not significantly.The sensitivity and specificity of computer-aided diagnosis by T Tech are not less than five senior diagnostic experts.Conclusion T tech's computer-aided diagnosis system for pneumoconiosis has higher performance than J tech’s,and it can be used as a second diagnosis for pneumoconiosis screening.
作者 刘瑞珍 王峥 钱青俊 多彩虹 卫晓鹏 LIU Ruizhen;WANG Zheng;QIAN Qingjun;DUO Caihong;WEI Xiaopeng(National Center for Occupational Safety and Health,NHC,Beijing 102308)
出处 《智慧健康》 2023年第7期1-5,共5页 Smart Healthcare
基金 国家卫生健康委员会职业健康司项目《尘肺病人工智能诊断技术试点应用》(项目编号:20190705)。
关键词 人工智能 尘肺病 计算机辅助诊断 深度卷积神经网络 深度残差网络 Artificial intelligence Pneumoconiosis Computer-aided diagnosis Deep convolutional neural network Deep residual network
  • 相关文献

参考文献4

二级参考文献13

共引文献47

同被引文献15

引证文献1

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部