超声心动图预测膜部瘤型室间隔缺损自然闭合  

Echocardiography in predicting spontaneous closure of aneurysm of membranous ventricular septal defect

作　　者：张昕彤 李欣洋 任卫东[1] 宋光[1] 王南 崔莉 ZHANG Xintong;LI Xinyang;REN Weidong;SONG Guang;WANG Nan(Department of Ultrasound,Shengjing Hospital of China Medical University,Shenyang 110004,China)

机构地区：[1]中国医科大学附属盛京医院超声科,辽宁沈阳110004

出　　处：《中国医学影像技术》2018年第12期1801-1804,共4页Chinese Journal of Medical Imaging Technology

基　　金：国家自然科学基金(81571686)

摘　　要：目的建立以膜部瘤型室间隔缺损(AMVSD)临床和超声心动图指标为变量的Logistic回归模型,评价其预测AMVSD自然闭合的价值。方法选取66例经我院超声科诊断为AMVSD的患儿,根据是否自然闭合将其分为自然闭合组(n=38)和未闭合组(n=28),测量间隔膜部瘤的基底宽度、膨出高度、破口宽度和破口数量,计算基底宽度/膨出高度比值、破口宽度/基底宽度比值及多破口占比。记录患儿身高及体质量,计算其体表面积,以体表面积校正基底宽度(基底宽度/体表面积)、膨出高度(膨出高度/体表面积)和破口宽度(破口宽度/体表面积)。比较2组各参数的差异,选择有统计学意义的指标为自变量,建立Logistic回归模型及其产生的Logit(P)的ROC曲线,评价Logistic模型的拟合效果。结果 2组体表面积(X_1)及破口宽度/体表面积比值(X^2)差异有统计学意义(P均<0.05)。Logistic回归方程为:Logit(P)=-10.986+21.484 X1+0.278 X2(χ~2=19.67,P<0.001);回归模型预测AMVSD自然闭合的准确率为74.24%(49/66)。ROC曲线的AUC为0.801±0.054(P<0.001),模型拟合效果良好。结论联合临床及超声心动图指标的Logistic回归模型可较精准地预测AMVSD的自然闭合情况,为临床选择治疗方案提供参考。Objective To establish Logistic regression model according to clinical and echocardiographic features of aneurysm of membranous ventricular septal defect(AMVSD),and to evaluate the predictive effect on spontaneous closure of AMVSD.Methods Totally 66 infants with AMVSD diagnosed with ultrasound were enrolled and divided into naturally close group(n=38)and unclosed group(n=28).The basal width,bulging height,width and numbers of the hole of aneurysm were measured,and the ratio of basal width/bulging height,width of the hole/basal width and multi-hole of each group were calculated.The height and weight of infants were recorded,and the body surface area(BSA)was calculated.The basal width(basal width/BSA),bulging height(bulging height/BSA)and width of the hole(width of the hole/BSA)of aneurysm were adjusted with BSA.Then the above parameters were compared between the two groups.Logistic regression models were established with parameters which were significantly different between 2 groups.ROC curves were constructed using the new parameters Logit(P)for assessing the imitative effect of the Logistic model.Results BSA(X1)and width of the hole/body surface area had significant differences between the 2 groups(all P<0.05).The predictive multivariate model was established as Logit(P)=-10.986+21.484 X1+0.278 X2(χ2=19.67,P<0.001).The accuracy of prediction was 74.24%(49/66),and AUC was 0.801±0.054(P<0.001),indicating that the imitative effect was good.Conclusion According to echocardiographic and clinical parameters,the predictive effect of the Logistic regression model is accurate,which may provide references for clinical treatment.

关 键 词：婴儿 室间隔缺损 超声心动描记术 LOGISTIC模型 

分 类 号：R725.4[医药卫生—儿科] R540.45[医药卫生—临床医学]