基于超声多平面测量的胸腔积液定量分析

Quantitative analysis of pleural effusion based on ultrasonic multiplanar measurement

目的探讨2个超声多平面测量公式用于定量坐位患者胸腔积液的准确性。方法前瞻性收集我院2023年2~7月因临床需要行穿刺引流的胸腔积液患者87例(共引流96处胸腔积液),分为训练组56例,验证组31例。均取坐位进行超声检查,采用1个线性模型和2个超声多平面测量公式(柱体模型、可变柱体模型)分别计算胸腔积液量(V_(预1)、V_(预2)、V_(预3)),随后立即于超声引导下行胸腔穿刺术,记录实际胸腔积液量(V_(实))。采用线性回归分析V_(预1)、V_(预2)、V_(预3)与V_(实)的关系;采用组内相关系数(ICC)和Bland-Altman图分析V_(预1)、V_(预2)、V_(预3)与V_(实)的一致性。结果训练组中V_(预1)、V_(预2)、V_(预3)分别为(507.6±158.2)ml、(515.2±170.7)ml、(508.3±166.9)ml;验证组中V_(预1)、V_(预2)、V_(预3)、V_(实)分别为(551.0±172.7)ml、(568.1±198.4)ml、(563.2±200.6)ml、(563.2±208.6)ml。线性回归分析显示,训练组中V_(预1)、V_(预2)、V_(预3)与V_(实)均呈线性相关(R^(2)分别为0.697、0.912、0.945,均P<0.05)。训练组和验证组中V_(预1)、V_(预2)、V_(预3)与V_(实)的一致性均好(ICC分别为0.833、0.955、0.972和0.828、0.957、0.984,均P<0.05);Bland-Altman图分析显示,验证组中V_(预2)、V_(预3)与V_(实)均具有良好的一致性。结论通过超声多平面测量能准确定量坐位患者胸腔积液,其中可变柱体模型的准确率最高,柱体模型次之。

Objective To investigate the accuracy of two ultrasound multiplanar measurement formulas for quantifying pleural effusion volume in patients in the seated upright position.Methods A prospective study was conducted,collecting data from 87 patients with pleural effusion who underwent thoracentesis drainage due to clinical indications between the February and July 2023 in our hospital,with a total of 96 sites of pleural effusion drainage.These patients were divided into a training group(n=56)and a validation group(n=31).All patients underwent ultrasonography in the seated upright position.One linear model and two ultrasound multiplanar measurement formulas(cylindrical model,variable cylindrical model)were used to calculate the volume of pleural effusion(V_(forecast1),V_(forecast2),V_(forecast3)).Subsequently,thoracentesis drainage was immediately performed under ultrasound guidance,and the actual volume of pleural effusion(V_(real))was recorded.Linear regression analysis was used to analyze the relationship between V_(forecast1),V_(forecast2),V_(forecast3)and V_(real).Intra-class correlation coefficient(ICC)and Bland-Altman diagram were used to analyze the consistency between V_(forecast1),V_(forecast2),V_(forecast3)and V_(real).Results In the training group,the V_(forecast1),V_(forecast2),V_(forecast3)were(507.6±158.2)ml,(515.2±170.7)ml and(508.3±166.9)ml,respectively.In the validation group,the V_(forecast1),V_(forecast2),V_(forecast3)were(551.0±172.7)ml,(568.1±198.4)ml and(563.2±208.6)ml,respectively.Linear regression analysis showed that V_(forecast1),V_(forecast2),V_(forecast3)were linearly correlated with V_(real)(R^(2)=0.697,0.912,0.945,all P<0.05).There was good consistency between V_(forecast1),V_(forecast2),V_(forecast3)and V_(real)in both the training and validation groups(ICC=0.833,0.955,0.972,0.828,0.957,0.984,all P<0.05).Bland-Altman diagram analysis showed good consistency between V_(forecast2),V_(forecast3)and V_(real)in the validation group.Conclusion Ultrasound multiplanar measurements can accurately quantify pleural effusion volume in patients in the seated upright position,with the variable cylindrical model demonstrating the highest accuracy,followed by the cylindrical model.