A parametric model of the relationship between EIT and total lung volume
A parametric model of the relationship between EIT and total lung volume
复制标题
DOI：
10.1088/0967-3334/26/4/006
复制
发表时间：
2005-08-01
影响因子：
3.2
通讯作者：
Guardo, R
中科院分区：
文献类型：
--
作者：
Coulombe, N;Gagnon, H;Guardo, R
利用26例健康受试者(男14例，女12例)的肺活量测定和电阻抗断层扫描(EIT)数据，建立了EIT差异图像的对比度变化与肺体积变化之间的联系模型。每个受试者以四种姿势(站立、坐着、仰卧45度、仰卧)和两种呼吸模式(安静、潮汐和深呼吸)进行8次录音，每一次S长。记录了年龄、性别和五个人体测量学变量。数据库分为四个子集。第一个子集是22名受试者(12名男性，10名女性)在深呼吸模式下记录的数据，用于创建模型。对其他子集进行验证：同样22名受试者在平静潮气呼吸期间记录的数据，以及其他4名受试者在两种呼吸模式下记录的数据。Delta V-P(肺活量计记录的肺容量变化)中的二次方程与EIT图像中的总对比度变化非常吻合。研究发现，模型系数取决于体位、性别、胸围和肩部皮肤褶皱。为了验证模型的有效性，将二次方程倒置以从EIT图像中估计肺体积的变化。然后将估计的变化与测量的体积变化进行比较。对每个数据子集进行验证，平均标准误差在9.3%到12.4%之间。提出的模型是实现EIT图像个体间比较的第一步，因为：(1)它提供了一个框架，用于纳入人体测量变量、性别和姿势的影响，以及(2)它将图像引用到可通过肺活量测量验证的物理量(体积)。
Spirometry and electrical impedance tomography (EIT) data from 26 healthy subjects (14 males, 12 females) were used to develop a model linking contrast variations in EIT difference images to lung volume changes. Eight recordings, each 64 s long, were made for each subject in four postures (standing, sitting, reclining at 45 degrees, supine) and two breathing modes (quiet tidal and deep breathing). Age, gender and five anthropometric variables were recorded. The database was divided into four subsets. The first subset, data from 22 subjects (12 males, 10 females) recorded in deep breathing mode, was used to create the model. Validation was done with the other subsets: data recorded during quiet tidal breathing in the same 22 subjects, and data recorded in both breathing modes for the other four subjects. A quadratic equation in Delta V-P (lung volume changes recorded by the spirometer) provided a very good fit to total contrast changes in the EIT images. The model coefficients were found to depend on posture, gender, thoracic circumference and scapular skin fold. To validate the model, the quadratic equation was inverted to estimate lung volume changes from the EIT images. The estimated changes were then compared to the measured volume changes. Validations with each data subset yielded mean standard errors ranging from 9.3% to 12.4%. The proposed model is a first step in enabling inter individual comparisons of EIT images since: (1) it provides a framework for incorporating the effects of anthropometric variables, gender and posture, and (2) it references the images to a physical quantity (volume) verifiable by spirometry.