资源描述:
《基于bayes理论的计算机辅助诊断系统在孤立性肺结节ct诊断中的应用》由会员上传分享,免费在线阅读,更多相关内容在工程资料-天天文库。
1、基于Bayes理论的计算机辅助诊断系统在孤立性肺结节CT诊断中的应用陈伟刘进康李文政熊曾龙学颖周晖(中南大学湘雅医院放射科,湖南长沙410008)摘要:目的初步探讨基于Bayes理论的计算机辅助诊断(computer-aideddiagnosis,CAD)系统在孤立性肺结节(solitarypulmonarynodule,SPN)CT诊断中的价值。方法依据Bayes理论先从352例SPN训练集(恶性135例,良性217例)中求出恶性SPN的验前比及各临床和CT表现的似然比,再建立基于Bayes理论的CAD系统,用它计算每个SPN的恶性概率。并前瞻性地检验该系统在132例SPN测
2、试集(恶性61例,良性71例)中的诊断效能,与2位高年资和2位低年资放射科医生常规阅片的表现作比较。结果成功构建基于Bayes理论的CAD系统,它诊断训练集SPN的敏感度、特异度和符合率分别为88.9%、93.1%、91.5%,诊断测试集SPN的敏感度、特异度、符合率、阳性预测值及阴性预测值分别为88.5%、85.9%、87.1%、84.4%、89.7%,其诊断符合率与高年资甲、乙医生比较无统计学差异(P均>0.05),但高于低年资丙、丁医生(P均<0.05)。结论基于Bayes理论的CAD系统将是一个有效的诊断辅助工具,可以提高医生尤其是低年资医生对SPN良恶性质的正确判断,
3、并在指导SPN的临床决策中有一定的参考作用。关键词:硬币病变,肺;体层摄影术,X线计算机;诊断,计算机辅助;Bayes理论ApplicationofBayesiantheorybasedcomputer-aideddiagnosissystemintheCTdiagnosisofsolitarypulmonarynodulesCHENWei,LIUJin-kang,LIWen-zheng,XIONGZeng,LONGXue-ying,ZHOUHui.(DepartmentofRadiology,XiangyaHospital,CentralSouthUniversity,Cha
4、ngsha410008,China)AbstractObjectiveTopreliminarilyexplorethevalueofcomputer-aideddiagnosis(CAD)systembasedonBayesiantheoryinthediagnosisofsolitarypulmonarynodules(SPNs)withCT.MethodsThreehundredandfifty-twoconsecutiveSPNcases(malignancyn=135,benignityn=217)werecollectedretrospectivelytoformt
5、hetrainingset.AccordingtoBayesiantheory,theprioroddsofmalignantSPNsandthelikelihoodratiosofclinicalandCTfindingswerederivedfromthetrainingsetfirstly,thenaBayesiantheorybasedCADsystemwasconstructedandtheprobabilityofmalignancyineachSPNwasobtainedfromthissystem.SPswith≥50%calculatedprobability
6、werejudgedasmalignancyandthosewith<50%calculatedprobabilitywasjudgedasbenignity.Onthetestset(malignancyn=61,benignityn=71),theBayesiantheorybasedCADsystemwastestedprospectivelyforitsdiagnosticvalidation,comparedwiththeperformanceofthetwochestradiologistsandtworadiologicresidentsusingroutined
7、iagnosticmethod.ResultsTheBayesiantheorybasedCADsystemwasconstructedsuccessfully.Thesensitivity,specificity,accuracyofthissystemforthetrainingsetwere88.9%,93.1%,91.5%respectively.Onthetestset,thesensitivity,specificity,accuracy,positivepredictiveva