摘要
目的 探讨左心耳(LAA)相关测量参数在基于心脏计算机断层扫描成像(cardiac computed tomography angiography,CCTA)的自动分割测量与手动测量方法 的一致性.方法 回顾性收集 2016 年 12 月 1 日～2020 年 3 月 31 日在解放军总医院心血管医学部接受CCTA检查的非瓣膜性房颤患者 168 例.以专业医师的手动测量结果 作为金标准,使用Bland-Atman统计方法 分别对使用Anythink β软件自动测量的左心耳开口平面周长、最长径、最短径,左心耳容积,开口平面与轴断面、矢状面、冠状面的夹角等参数进行一致性分析.结果 左心耳自动测量与手动测量中开口的周长、最长径、最短径的一致性评价组内相关系数(ICC)值分别为0.93、0.854、0.935(P<0.01).Bland-Altman分析显示左心耳自动测量开口平面周长、最长径、最短径,左心耳容积,开口平面与轴断面、矢状面、冠状面的夹角与手动测量值的差值超出 95%一致性限范围的点数均<10%,一致性良好.结论 基于CCTA的左心耳自动分割测量方法 可以自动完成LAA形态的识别和相关参数的测量,高效准确,可重复性高,可将其作为一种左心耳封堵术(LAAO)术前评估的测量工具.
Abstract
Objective To explore the consistency of left atrial appendage(LAA)measurement parameters between cardiac computer tomography angiography(CCTA)imaging-based automatic segmentation and manual measurement method.Methods Clinical and imaging data of non-valvular atrial fibrillation patients who underwent CCTA at the department of cardiovascular medicine,General Hospital of the PLA from December 1,2016,to March 31,2020,were collected retrospectively.Using manual measurement by professional physicians as the gold standard,Bland-Altman statistical method were used to analyze the consistency of LAA orifice plane circumference,longest diameter,shortest diameter,LAA volume,and the angle between the orifice plane and the axial,sagittal,and coronal planes,respectively,measured automatically using Anythinkβ software.Results The ICC values for the consistency evaluation of the perimeter,longest diameter,and shortest diameter of the left atrial appendage between automatic and manual measurements were 0.93,0.854,and 0.935,respectively(P<0.01).Bland-Altman analysis showed that the number of points where the difference between the automatic measurement values of the LAA orifice plane circumference,longest diameter,shortest diameter,LAA volume,and the angle between the orifice plane and the axial,sagittal,and coronal planes and the manual measurement values exceeded the 95%limits of agreement was less than 10%.Conclusion The CCTA-based automatic segmentation measurement method of the LAA efficiently identifies LAA morphology and measure related parameters with high accuracy and reproducibility.It can be used as a measurement tool for preoperative evaluation of LAAO.
基金项目
首都卫生发展科研专项(首发 2024-2-5071)
NETL
NSTL
万方数据