Department of Cardiovascular Radiology, Institut Cœur-Poumon, CHU de Lille, Boulevard du Pr Jules Leclercq, 59037 Lille Cedex, France; INSERM UMR 1011, Institut Pasteur de Lille, EGID (European Genomic Institute for Diabetes), FR3508, Univ. Lille, 59000 Lille, France.
Department of Cardiovascular Radiology, Institut Cœur-Poumon, CHU de Lille, Boulevard du Pr Jules Leclercq, 59037 Lille Cedex, France.
Diagn Interv Imaging. 2021 May;102(5):297-303. doi: 10.1016/j.diii.2020.11.012. Epub 2020 Dec 8.
To evaluate the reliability of a real-time compressed sensing (CS) cine sequence for the detection of left ventricular wall motion disorders after myocardial infarction in comparison with the reference steady-state free precession cine sequence.
One hundred consecutive adult patients referred for either initial work-up or follow-up by cardiac magnetic resonance (CMR) in the context of myocardial infarction were prospectively included. There were 77 men and 23 women with a mean age of 63.12±11.3 (SD) years (range: 29-89 years). Each patient underwent the reference segmented multi-breath-hold steady-state free precession cine sequence including one short-axis stack and both vertical and horizontal long-axis slices (SSFP) and the CS real-time single-breath-hold evaluated sequence (CS) providing the same slices. Wall motion disorders were independently and blindly assessed with both sequences by two radiologists, using the American Heart Association left ventricle segmentation. Paired Wilcoxon signed-rank test was used to search for differences in wall motion disorders conspicuity between both sequences and receiver operating characteristic curve (ROC) analysis was performed to assess the diagnosis performance of CS sequence using SSFP as the reference method.
Each patient had at least one cardiac segment with wall motion abnormality on SSFP and CS images. The 1700 segments analyzed with SSFP were classified as normokinetic (360/1700; 21.2%), hypokinetic (783/1700; 46.1%), akinetic (526/1700; 30.9%) or dyskinetic (31/1700; 1.8%). Sensitivity and specificity of the CS sequence were 99.6% (95% CI: 99.1-99.9%) and 99.7% (95% CI: 98.5-100%), respectively. Area under ROC of CS diagnosis performance was 0.997 (95% CI: 0.993-0.999).
CS real-time cine imaging significantly reduces acquisition time without compromising the conspicuity of left ventricular -wall motion disorders in the context of myocardial infarction.
评估实时压缩感知（CS）电影序列检测心肌梗死后左心室壁运动障碍的可靠性，并与参考稳态自由进动电影序列进行比较。
连续前瞻性纳入 100 例因心肌梗死初始检查或随访而接受心脏磁共振（CMR）检查的成年患者。其中男性 77 例，女性 23 例，平均年龄 63.12±11.3（SD）岁（范围：29-89 岁）。每位患者均行参考分段多屏气稳态自由进动电影序列（SSFP）检查，包括短轴堆栈和垂直及水平长轴切片，以及提供相同切片的 CS 实时单次屏气评估序列（CS）。两名放射科医师独立且盲法使用美国心脏协会左心室分段法评估两种序列的壁运动障碍。采用配对 Wilcoxon 符号秩检验比较两种序列的壁运动障碍显示差异，并采用受试者工作特征曲线（ROC）分析 CS 序列的诊断性能，以 SSFP 作为参考方法。
SSFP 和 CS 图像上每个患者至少有一个节段存在壁运动异常。分析了 1700 个 SSFP 节段，分为正常运动（360/1700；21.2%）、低运动（783/1700；46.1%）、无运动（526/1700；30.9%）或反常运动（31/1700；1.8%）。CS 序列的敏感性和特异性分别为 99.6%（95%CI：99.1-99.9%）和 99.7%（95%CI：98.5-100%）。CS 诊断性能的 ROC 曲线下面积为 0.997（95%CI：0.993-0.999）。
CS 实时电影成像可显著缩短采集时间，而不影响心肌梗死后左心室壁运动障碍的显示。
