Department of Medical Imaging, University Health Network, 1 PMB-273, 585 University Avenue, Toronto, Ontario, M5G 2N2, Canada.
Department of Medical Imaging, University of Toronto, Toronto, Ontario, Canada.
Eur Radiol. 2019 Jul;29(7):3658-3668. doi: 10.1007/s00330-019-06019-9. Epub 2019 Feb 15.
To validate deformable registration algorithms (DRAs) for cine balanced steady-state free precession (bSSFP) assessment of global longitudinal strain (GLS) and global circumferential strain (GCS) using harmonic phase (HARP) cardiovascular magnetic resonance as standard of reference (SoR).
Seventeen patients and 17 volunteers underwent short axis stack and 2-/4-chamber cine bSSFP imaging with matching slice long-axis and mid-ventricular spatial modulation of magnetization (SPAMM) myocardial tagging. Inverse DRA was applied on bSSFP data for assessment of GLS and GCS while myocardial tagging was processed using HARP. Intra- and inter-observer variability assessment was based on repeated analysis by a single observer and analysis by a second observer, respectively. Standard semi-automated short axis stack segmentation was performed for analysis of left ventricular (LV) volumes and ejection fraction (EF).
DRA demonstrated strong relationships to HARP for myocardial GLS (R = 0.75; p < 0.0001) and endocardial GLS (R = 0.61; p < 0.0001). GCS result comparison also demonstrated significant relationships between DRA and HARP for myocardial strain (R = 0.61; p < 0.0001) and endocardial strain (R = 0.51; p < 0.0001). Both methods demonstrated small systematic errors for intra- and inter-observer variability but DRA demonstrated consistently lower CV. Global LVEF was significantly lower (p = 0.0099) in patients (53.7%; IQR 43.9/64.0%) than in healthy volunteers (62.6%; IQR 61.1/66.2%). DRA and HARP strain data demonstrated significant relationships to LVEF.
Non-rigid deformation method-based DRA provides a reliable measure of peak systolic GCS and GLS based on cine bSSFP with superior intra- and inter-observer reproducibility compared to HARP.
• Myocardial strain can be reliably analyzed using inverse deformable registration algorithms (DRAs) on cine CMR. • Inverse DRA-derived strain shows higher reproducibility than tagged CMR. • DRA and tagged CMR-based myocardial strain demonstrate strong relationships to global left ventricular function.
使用谐波相位(HARP)心血管磁共振作为参考标准(SoR),验证用于电影平衡稳态自由进动(bSSFP)评估整体纵向应变(GLS)和整体周向应变(GCS)的变形配准算法(DRA)。
17 名患者和 17 名志愿者接受短轴堆栈和 2/4 腔电影 bSSFP 成像,匹配片层长轴和中间心室磁化空间调制(SPAMM)心肌标记。在 bSSFP 数据上应用反向 DRA 以评估 GLS 和 GCS,而心肌标记使用 HARP 进行处理。观察者内和观察者间可重复性评估分别基于单个观察者的重复分析和第二个观察者的分析。标准半自动短轴堆栈分割用于分析左心室(LV)容积和射血分数(EF)。
DRA 与 HARP 用于心肌 GLS(R=0.75;p<0.0001)和心内膜 GLS(R=0.61;p<0.0001)具有很强的相关性。心肌应变和心内膜应变的 GCS 结果比较也表明 DRA 和 HARP 之间存在显著关系(R=0.61;p<0.0001)。两种方法均显示观察者内和观察者间变异性的系统误差较小,但 DRA 的 CV 始终较低。患者的整体 LVEF 明显低于健康志愿者(p=0.0099)(53.7%;IQR 43.9/64.0%)(62.6%;IQR 61.1/66.2%)。DRA 和 HARP 应变数据与 LVEF 存在显著关系。
基于电影 bSSFP 的非刚性变形方法 DRA 提供了一种可靠的峰值收缩期 GCS 和 GLS 测量方法,与 HARP 相比,具有更好的观察者内和观察者间可重复性。
• 心肌应变可以使用电影 CMR 上的逆变形配准算法(DRA)可靠地进行分析。• 基于 DRA 的应变具有更高的可重复性。• DRA 和基于标记的 CMR 的心肌应变与整体左心室功能有很强的关系。