摘要
目的应用多变量分析方法建立胆囊结石合并胆总管结石的预测模型,并评估其诊断价值。方法2007年1-12月西安交通大学第一附属医院肝胆外科选择626例胆囊结石病人的各项指标建立数据库,筛选有鉴别意义的变量用于建立Logistic回归模型,通过χ2检验来比较Logistic回归模型与核磁共振胰胆管成像(MRCP)、B超检查诊断的灵敏度、特异度和准确度。结果构建的Logistic回归模型为P=1/1+EXP[-(-5.092+3.101X1+1.754X2+1.647X3+1.306X4+1.175X5+0.963X6)],模型的灵敏度为69.5%,特异度为97.1%,准确度为92.5%。MRCP和Logistic回归模型的灵敏度和准确度均好于B超(P<0.05),MRCP的灵敏度好于Logistic回归模型(P<0.05),而准确度与Logistic回归模型比较差异无统计学意义(P>0.05)。三种诊断方法的特异度差异无统计学意义(P>0.05)。结论Logistic回归模型诊断的灵敏度和准确度好于B超诊断,能为临床医生诊断和治疗决策提供参考依据。
Objective To develop a predictive equation with the factors previously identified by multivariate Logistic regression that were predictive for cholecystolithiasis combined with common bile duct stones ,Evaluate the diagnostic value of predictive equation,ultrasound and MRCP.Methods Make database with the index of 626 patients and select the variables that considered statistically significant for the Logistic equation.The sensitivity, specificity and accuracy of Logistic regression equation, ultrasound and MRCP are compared by x2 test.Results P=1/1＋EXP[-（-5.092＋3.101X1＋1.754X2＋1.647X3＋1.306X4＋1.175X5＋0.963X6）],The equation has a sensitivity of 69.5% , specificity of 97.1% , and accuracy of 92.5%.The sensitivity and accuracy of MRCP and Logistic regression equation are superior than ultrasound （P〈0.05）, The sensitivity of MRCP are superior than Logistic regression equation （P〈0.05）. The accuracy of MRCP and Logistic regression equation have no difference（P〉0.05）.The specificity of three methods have no difference（P〉0.05）. Conclusion The specificity and accuracy of Logistic regression equation is superior than ultrasound, it could provide reference for clinical diagnosis and treatment decision making for doctors.
出处
《中国实用外科杂志》
CSCD
北大核心
2010年第3期214-216,共3页
Chinese Journal of Practical Surgery
