van Amerom Joshua F P, Kellenberger Christian J, Yoo Shi-Joon, Macgowan Christopher K
Department of Medical Imaging, University of Toronto, and Department of Diagnostic Imaging, Hospital for Sick Children, Toronto, Ontario, Canada.
Magn Reson Imaging. 2009 Jan;27(1):38-47. doi: 10.1016/j.mri.2008.05.016. Epub 2008 Jul 23.
An automated method was evaluated to detect blood flow in small pulmonary arteries and classify each as artery or vein, based on a temporal correlation analysis of their blood-flow velocity patterns. The method was evaluated using velocity-sensitive phase-contrast magnetic resonance data collected in vitro with a pulsatile flow phantom and in vivo in 11 human volunteers. The accuracy of the method was validated in vitro, which showed relative velocity errors of 12% at low spatial resolution (four voxels per diameter), but was reduced to 5% at increased spatial resolution (16 voxels per diameter). The performance of the method was evaluated in vivo according to its reproducibility and agreement with manual velocity measurements by an experienced radiologist. In all volunteers, the correlation analysis was able to detect and segment peripheral pulmonary vessels and distinguish arterial from venous velocity patterns. The intrasubject variability of repeated measurements was approximately 10% of peak velocity, or 2.8 cm/s root-mean-variance, demonstrating the high reproducibility of the method. Excellent agreement was obtained between the correlation analysis and radiologist measurements of pulmonary velocities, with a correlation of R2=0.98 (P<.001) and a slope of 0.99+/-0.01.
基于对小肺动脉血流速度模式的时间相关性分析，评估了一种自动方法来检测小肺动脉中的血流并将每条血管分类为动脉或静脉。该方法使用在体外通过脉动流模体收集的速度敏感相位对比磁共振数据以及在11名人类志愿者体内收集的数据进行评估。该方法的准确性在体外得到验证，结果表明在低空间分辨率（每直径四个体素）下相对速度误差为12%，但在提高空间分辨率（每直径16个体素）时降低至5%。根据该方法的可重复性以及与经验丰富的放射科医生手动速度测量的一致性，在体内对该方法的性能进行了评估。在所有志愿者中，相关性分析能够检测并分割外周肺血管，并区分动脉和静脉的速度模式。重复测量的受试者内变异性约为峰值速度的10%，或2.8 cm/s的均方根，证明了该方法具有高度可重复性。相关性分析与放射科医生对肺速度的测量之间取得了极好的一致性，相关系数R2 = 0.98（P <.001），斜率为0.99±0.01。
