基于AI技术的癌症风险动态预警系统的应用研究  

AI-technology-based cancer risk dynamic early warning system and its application

在线阅读下载全文

作  者:雷锦誌 张晗[2] 李永志 康熙雄[4] LEI Jin-zhi;ZHANG Han;LI Yong-zhi;KANG Xi-xiong(Zhou Pei-Yuan Center for Applied Mathematics,Tsinghua University,Beijing 100084,China;不详)

机构地区:[1]清华大学周培源应用数学研究中心,北京100084 [2]北京航天总医院健康管理中心,北京100076 [3]北京航天总医院信息处,北京100076 [4]首都医科大学附属北京天坛医院,北京100070

出  处:《中国卫生检验杂志》2020年第5期524-528,共5页Chinese Journal of Health Laboratory Technology

摘  要:目的癌症是严重危害人们健康的复杂疾病,建立便捷有效的癌症风险预警系统以进行癌症的预防和早期诊断是目前在癌症研究中所面临的重大挑战性问题。方法本文根据某医院的常规血检验数据(包括非癌症人群和癌症患者),通过非线性数理统计与分析方法提出基于检查数据的癌症风险指标。结果通过该风险指标可以有效区分癌症高风险人群,在多家医院数据的测试中特异度达到96%,灵敏度为91%,ROC曲线的AUC达到0.96以上,高于常用肿瘤标志物的指标。结论根据所建立的癌症风险指标建立人工智能(AI)技术的癌症风险动态预警系统,对普通人群提供便捷动态的癌症风险预警服务。Objective Cancer is a complex disease that seriously endangers people’s health.In cancer researches,it is a major challenge to establish a convenient and effective cancer risk early warning system,which is important for cancer prevention and early diagnosis.Methods In this paper,based on the routine blood test data from a hospital(including data from non-cancer population and cancer patients),we proposed a cancer risk index through methods of nonlinear analysis.Results The proposed cancer risk index is able to effectively distinguish the high-risk population of cancer from non-cancer populations.We further test the index with data from different hospitals,the specificity is 96%,and the sensitivity is 91%,with the AUC of ROC curve larger than 0.96,which is higher than most methods based on common tumor biomarkers.Conclusion According to the established cancer risk index,we established an artificial intelligence(AI) based system of cancer risk dynamic early warning system,which can provide convenient dynamic cancer risk early warning service for the general population.

关 键 词:癌症 大数据分析 血常规 风险评估 人工智能 

分 类 号:R73[医药卫生—肿瘤]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象