Department of Radiology, Amiens University Hospital, 1 Rond-Point du Professeur Christian Cabrol, F-80054 Amiens Cedex 01, France.
Medical Image Processing Unit, Amiens University Hospital, Amiens, France.
AJR Am J Roentgenol. 2020 Oct;215(4):807-817. doi: 10.2214/AJR.19.22726. Epub 2020 Aug 5.
The objective of our study was to assess the accuracy and reproducibility of right ventricular volumes and function measurements in patients with repaired tetralogy of Fallot using two k-adaptive-t autocalibrating reconstruction for cartesian sampling (kat-ARC) accelerated sequences: a morphologic 3D cine sequence and a functional free-breathing 4D flow sequence. Seventeen patients who underwent cardiac MRI with gadolinium injection as part of follow-up of repaired tetralogy of Fallot from March 2017 to July 2018 were prospectively included in the initial study population; the final study cohort was composed of 15 of the 17 patients. Ventricle volume measurements were performed on a 3D cine kat-ARC sequence, a 4D flow kat-ARC sequence, and a 2D cine balanced steady-state free precession (bSSFP) sequence. The 2D cine bSSFP sequence was the reference standard in cardiac assessment. Intertechnique and interobserver analyses were performed. Bland-Altman analysis and correlation tests were used to compare quantitative measurements. Ventricular end-diastolic volume (EDV) and end-systolic volume (ESV) values were well correlated in the right ventricle ( = 0.94-0.98) for both 3D and 4D sequences. Ejection fraction (EF) also showed good correlation for both 3D and 4D sequences ( = 0.79 and = 0.92). Bland-Altman analysis showed good agreement between right and left ventricular volumes, with narrower limits of agreement in the left ventricle, and an intraclass correlation coefficient (ICC) of greater than 0.80. For the 4D flow sequence, ventricular volumes were overestimated, which led to underestimation of the EF (bias for EDV = -10.2 mL, bias for ESV = -22.7 mL, bias for EF = 6.4%). Interobserver agreement was excellent for the ventricular volumes (ICC, 0.93-0.99) and fair to excellent for the EFs (ICC, 0.59-0.83). Compared with the reference standard 2D bSSFP sequence, the 3D cine kat-ARC accelerated sequence showed good accuracy and reproducibility for right ventricular measurements in patients with repaired tetralogy of Fallot. The short duration of the 4D flow kat-ARC sequence appears promising for performing volumetric measurements.
我们的研究目的是评估两种自适应 k-自动校准重建笛卡尔采样(kat-ARC)加速序列(形态学 3D 电影序列和功能自由呼吸 4D 流序列)在法洛四联症修复患者右心室容积和功能测量中的准确性和可重复性。2017 年 3 月至 2018 年 7 月,17 名接受心脏 MRI 检查并注射钆造影剂的法洛四联症修复患者前瞻性纳入初始研究人群;最终研究队列由 17 名患者中的 15 名组成。在 3D 电影 kat-ARC 序列、4D 流 kat-ARC 序列和 2D 电影平衡稳态自由进动(bSSFP)序列上进行心室容积测量。2D 电影 bSSFP 序列是心脏评估的参考标准。进行了技术间和观察者间分析。采用 Bland-Altman 分析和相关检验比较定量测量值。右心室舒张末期容积(EDV)和收缩末期容积(ESV)值在 3D 和 4D 序列中均具有很好的相关性(r = 0.94-0.98)。射血分数(EF)在 3D 和 4D 序列中也具有良好的相关性(r = 0.79 和 r = 0.92)。Bland-Altman 分析显示左右心室容积之间具有良好的一致性,左心室的一致性界限更窄,组内相关系数(ICC)大于 0.80。对于 4D 流序列,心室容积被高估,导致 EF 被低估(EDV 的偏差=-10.2mL,ESV 的偏差=-22.7mL,EF 的偏差=6.4%)。心室容积的观察者间一致性极好(ICC,0.93-0.99),EF 的观察者间一致性为良好至极好(ICC,0.59-0.83)。与参考标准 2D bSSFP 序列相比,3D 电影 kat-ARC 加速序列在法洛四联症修复患者的右心室测量中具有良好的准确性和可重复性。4D 流 kat-ARC 序列的短采集时间似乎有望进行容积测量。