轮廓配准法在MRI评估健康成年人左心室心肌细胞外容积分数中的价值

Contour-based Registration for Left Ventricular Myocardial Extracellular Volume Fraction Using T1 Mapping in Healthy Adults

【目的】探讨轮廓配准法在MRI纵向弛豫时间定量成像(T1 mapping)技术评估健康成年人左心室心肌细胞外容积分数(ECV)中的价值。【方法】纳入26名健康志愿者(男16名,女10名),进行心脏MRI检查。采用改良的Look-Locker反转恢复(MOLLI)序列采集心底部、心室中部和心尖部三个短轴位对比增强前后的图像。利用后处理软件通过常规法和配准法分别生成心肌ECV图并比较图像质量。将每个短轴位层面的图像分为形变组和对照组,比较各层面及各组的常规ECV值和配准ECV值。【结果】16位志愿者(61.5%)的T1 mapping图像出现左心室形变。左心室3个短轴位配准ECV图像质量均比常规ECV图高。左心室总体、心室中部和心尖部的常规ECV值[(26.81±2.78)%、(25.38±3.05)%、(28.66±4.10)%]与配准ECV值[(25.75±2.42)%、(24.30±2.45)%、(27.22±3.38)%]均有统计学差异(P值分别为0.001、0.016、0.010)。三个形变组配准ECV值均低于常规ECV值(P值分别为0.038、0.012、0.016)且标准差较小。心尖部ECV值高于其他层面以及总体ECV值(常规:F=4.799,P=0.004;配准:F=4.822,P=0.003)且标准差较大。【结论】T1 mapping图像出现左心室形变时轮廓配准技术可改善ECV图像质量,提高ECV值定量精确度。

【Objective】We aimed to evaluate the role of contour-based registration in quantification of myocardial extracellular volume fraction(ECV)based on T1 mapping technique.【Methods】T1 mapping images of the basal,midcavity and apical short axis slices of 26 healthy volunteers(16 males and 10 females)were obtained before and after administration of contrast agent using Modified Look-Locker Inversion Recovery(MOLLI)on a 3.0T Magnetic Resonance Imaging system. ECV was calculated by routine method and registered method. For each slice,subjects were divided into the deformation subgroup and the control subgroup. Routine ECVs and registered ECVs were compared for each subgroup.【Results】Left ventricular deformation among T1 mapping images occurred in 16 subjects(61.5%). In all three slices,registered ECV maps had better image quality and more highest grade images than routine ECV maps. Routine ECVs[(26.81 ± 2.78)%,(25.38 ± 3.05)%,(28.66 ± 4.10)%]were statistically different from registered ECVs in global[(25.75±2.42)%,P=0.001],mid-cavity[(24.30±2.45)%,P=0.016]and apical slices[(27.22±3.38)%,P=0.010].In deformation subgroups of all the 3 slices,registered ECVs were lower than routine ECVs with smaller SDs(P=0.038,0.012,0.016). The apical ECVs were higher with larger SDs than the other two slices and the global ECVs(routine:F=4.799,P=0.004;registered:F=4.822,P=0.003).【Conclusions】Contour-based registration can improve the image quality and precision of ECV quantification in cases with ventricular shape deformation among source images.