J Endod:AI对根尖周损伤CBCT影像CAD的影响

2020-06-01 lishiting MedSci原创

这篇研究的目的是为了评估一种深度学习(DL)运算方法对CBCT影像的自动分割以及根尖周损伤检测的影响。

这篇研究的目的是为了评估一种深度学习(DL)运算方法对CBCT影像的自动分割以及根尖周损伤检测的影响。

研究纳入了包含61个有或没有损伤牙根的小视野CBCT容积(n = 20),比较依赖于临床医师与依赖于U-Net 结构体系的DL法分割之间的差异。分割标记每个体素作为5个中的1个范畴:"损伤"(根尖周损伤)、"牙齿结构"、"骨"、"修复材料"和"背景"。通过深度学习分割法(DLS)并基于5倍交叉确认后重复拆分所有影像为一个训练组和一个验证组,随后结果取平均值。通过对分损伤的准确性检测评估DLS与依赖于临床医师的分割差异,并采用DICE指数评估每个标记的敏感性、特异性、阳性预测值、阴性预测值和体素匹配的准确性。

结果显示,DLS损伤的准确性为0.93,特异性为0.88、阳性预测值为0.87以及阴性预测值为0.93。个体标记的整个DICE指数中,损伤= 0.52, 牙齿结构= 0.74, 骨= 0.78, 修复材料= 0.58以及背景= 0.95。所有真实存在损伤的累积DICE指数为0.67。

结论:在一个有限的CBCT环境下,DL经过演练对于损伤检测的准确性展现出极好的结果。整体的体素匹配准确性可能会通过增强的AI版本而得到提升。

原始出处

Frank C Setzer, Katherine J Shi, et al. Artificial Intelligence for the Computer-aided Detection of Periapical Lesions in Cone-beam Computed Tomographic Images. J Endod., 2020 May 8;S0099-2399(20)30235-1. doi: 10.1016/j.joen.2020.03.025.

 

版权声明:
本网站所有内容来源注明为“梅斯医学”或“MedSci原创”的文字、图片和音视频资料,版权均属于梅斯医学所有。非经授权,任何媒体、网站或个人不得转载,授权转载时须注明来源为“梅斯医学”。其它来源的文章系转载文章,或“梅斯号”自媒体发布的文章,仅系出于传递更多信息之目的,本站仅负责审核内容合规,其内容不代表本站立场,本站不负责内容的准确性和版权。如果存在侵权、或不希望被转载的媒体或个人可与我们联系,我们将立即进行删除处理。
在此留言
评论区 (3)
#插入话题
  1. [GetPortalCommentsPageByObjectIdResponse(id=1310432, encodeId=58cb131043234, content=<a href='/topic/show?id=3e9441e65f' target=_blank style='color:#2F92EE;'>#CBCT#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=40, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=4176, encryptionId=3e9441e65f, topicName=CBCT)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=a12645, createdName=智慧医人, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1397981, encodeId=94fc139e9815f, content=<a href='/topic/show?id=8bee565436d' target=_blank style='color:#2F92EE;'>#损伤#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=35, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=56543, encryptionId=8bee565436d, topicName=损伤)], attachment=null, authenticateStatus=null, createdAvatar=https://thirdwx.qlogo.cn/mmopen/vi_32/DYAIOgq83eoNX3lUOlDKGBUEPNwZx0nibmdcFyeibnX4WMhUSCU2REpGRWtSmnT2CCs0dA5waxvJknYDxwtYnIcQ/132, createdBy=294e2500218, createdName=ms3046638856685384, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1545942, encodeId=9125154594216, content=<a href='/topic/show?id=9abf4e247a1' target=_blank style='color:#2F92EE;'>#尖周损伤#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=47247, encryptionId=9abf4e247a1, topicName=尖周损伤)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=a12645, createdName=智智灵药, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=)]
    2020-06-03 智慧医人
  2. [GetPortalCommentsPageByObjectIdResponse(id=1310432, encodeId=58cb131043234, content=<a href='/topic/show?id=3e9441e65f' target=_blank style='color:#2F92EE;'>#CBCT#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=40, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=4176, encryptionId=3e9441e65f, topicName=CBCT)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=a12645, createdName=智慧医人, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1397981, encodeId=94fc139e9815f, content=<a href='/topic/show?id=8bee565436d' target=_blank style='color:#2F92EE;'>#损伤#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=35, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=56543, encryptionId=8bee565436d, topicName=损伤)], attachment=null, authenticateStatus=null, createdAvatar=https://thirdwx.qlogo.cn/mmopen/vi_32/DYAIOgq83eoNX3lUOlDKGBUEPNwZx0nibmdcFyeibnX4WMhUSCU2REpGRWtSmnT2CCs0dA5waxvJknYDxwtYnIcQ/132, createdBy=294e2500218, createdName=ms3046638856685384, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1545942, encodeId=9125154594216, content=<a href='/topic/show?id=9abf4e247a1' target=_blank style='color:#2F92EE;'>#尖周损伤#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=47247, encryptionId=9abf4e247a1, topicName=尖周损伤)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=a12645, createdName=智智灵药, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=)]
  3. [GetPortalCommentsPageByObjectIdResponse(id=1310432, encodeId=58cb131043234, content=<a href='/topic/show?id=3e9441e65f' target=_blank style='color:#2F92EE;'>#CBCT#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=40, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=4176, encryptionId=3e9441e65f, topicName=CBCT)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=a12645, createdName=智慧医人, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1397981, encodeId=94fc139e9815f, content=<a href='/topic/show?id=8bee565436d' target=_blank style='color:#2F92EE;'>#损伤#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=35, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=56543, encryptionId=8bee565436d, topicName=损伤)], attachment=null, authenticateStatus=null, createdAvatar=https://thirdwx.qlogo.cn/mmopen/vi_32/DYAIOgq83eoNX3lUOlDKGBUEPNwZx0nibmdcFyeibnX4WMhUSCU2REpGRWtSmnT2CCs0dA5waxvJknYDxwtYnIcQ/132, createdBy=294e2500218, createdName=ms3046638856685384, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1545942, encodeId=9125154594216, content=<a href='/topic/show?id=9abf4e247a1' target=_blank style='color:#2F92EE;'>#尖周损伤#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=0, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=47247, encryptionId=9abf4e247a1, topicName=尖周损伤)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=a12645, createdName=智智灵药, createdTime=Wed Jun 03 10:49:28 CST 2020, time=2020-06-03, status=1, ipAttribution=)]

相关资讯

OCC 2020:对话人工智能:云医院和AI医生渐行渐近

由于疫情影响,今年的东方心脏病学会议(OCC2020)举办了一场独特的云上学术盛宴。新时代新技术为新医疗带来了无限可能。本届大会展开了一场人工智能和互联网医疗论坛。

J Dent Res: 通过成像和人工智能改善口腔癌的预后

早期诊断是口腔和口咽鳞状细胞癌(OPSCC)预后最重要的决定因素,然而,多数癌症是在晚期才发现的,预后较差。通常,由非专业人士(例如牙医)筛查口腔癌的风险,然后将高危患者转诊给专科医生进行活检诊断。由

李飞飞团队正在研发家用AI系统,可监测独居老人新冠症状

在新冠肺炎大流行期间,照顾老年人变得更加困难。人工智能是否在这个领域发挥作用?

疫情大考下的人工智能,交出了一份多少分的答卷?

疫情期间,北京地铁10号线牡丹园站新增了几套红外测温设备,安检人员无需接触就可以对乘客进行体温筛检。

人工智能在肿瘤放射治疗中的研究进展

据中国国家癌症中心统计,截至2016年1月,中国现有肿瘤患者750万人左右。根据WHO的数据显示,大约70%的癌症患者需要接受放射治疗。虽然目前临床放射治疗技术日趋成熟,但是仍存在诸多具有挑战性的难题

Thorax:一种预测肺结节恶性程度卷积神经网络人工智能工具的外部验证

与布鲁克模型相比,LCP-CNN评分具有更好的辨别力,可以识别更多的良性结节而不会遗漏癌症。