Department of Radiology, Neuroradiology Section, Stanford, CA, USA.
ETH Zürich Institute for Biomedical Engineering, Zürich, Switzerland.
Int J Stroke. 2020 Apr;15(3):332-342. doi: 10.1177/1747493019873515. Epub 2019 Sep 3.
Intravoxel incoherent motion is a diffusion-weighted imaging magnetic resonance imaging technique that measures microvascular perfusion from a multi-b value sequence. Intravoxel incoherent motion microvascular perfusion has not been directly compared to conventional dynamic susceptibility contrast perfusion-weighted imaging in the context of acute ischemic stroke. We determined the degree of correlation between perfusion-weighted imaging and intravoxel incoherent motion parameter maps in patients with acute ischemic stroke.
We performed a retrospective cohort study of acute ischemic stroke patients undergoing thrombectomy treatment triage by magnetic resonance imaging. Intravoxel incoherent motion perfusion fraction maps were derived using two-step voxel-by-voxel post-processing. Ischemic core, penumbra, non-ischemia, and contralateral hemisphere were delineated based upon diffusion-weighted imaging and perfusion-weighted imaging using a Tmax >6 s threshold. Signal intensity within different brain compartments were measured on intravoxel incoherent motion (IVIM , IVIM D*, IVIM D*) parametric maps and compared the differences using one-way ANOVA. Ischemic volumes were measured on perfusion-weighted imaging and intravoxel incoherent motion parametric maps. Bland-Altman analysis and voxel-based volumetric comparison were used to determine the agreements among ischemic volumes of perfusion-weighted imaging and intravoxel incoherent motion perfusion parameters. Inter-rater reliability on intravoxel incoherent motion maps was also assessed. Significance level was set at α < 0.05.
Twenty patients (11 males, 55%; mean age 67.1 ± 13.8 years) were included. Vessel occlusions involved the internal carotid artery (6 patients, 30%) and M1 segment of the middle cerebral artery (14, 70%). Mean pre-treatment core infarct volume was 19.07 ± 23.56 ml. Mean pre-treatment ischemic volumes on perfusion-weighted imaging were 10.90 ± 13.33 ml (CBV), 24.83 ± 23.08 ml (CBF), 58.87 ± 37.85 ml (MTT), and 47.53 ± 26.78 ml (Tmax). Mean pre-treatment ischemic volumes on corresponding IVIM parameters were 23.20 ± 25.63 ml (IVIM ), 14.01 ± 16.81 ml (IVIM D*), and 27.41 ± 40.01 ml (IVIM D*). IVIM , D, and D* demonstrated significant differences ( < 0.001). The best agreement in term of ischemic volumes and voxel-based overlap was between IVIM D* and CBF with mean volume difference of 0.5 ml and mean dice similarity coefficient (DSC) of 0.630 ± 0.136.
There are moderate differences in brain perfusion assessment between intravoxel incoherent motion and perfusion-weighted imaging parametric maps, and IVIM D* and perfusion-weighted imaging CBF show excellent agreement. Intravoxel incoherent motion is promising for cerebral perfusion assessment in acute ischemic stroke patients.
体素内不相干运动是一种扩散加权成像磁共振成像技术,可从多 b 值序列测量微血管灌注。体素内不相干运动微血管灌注尚未在急性缺血性卒中的背景下直接与传统的动态对比增强灌注加权成像进行比较。我们确定了急性缺血性卒中患者灌注加权成像和体素内不相干运动参数图之间的相关性程度。
我们对接受磁共振成像检查进行血栓切除术治疗分诊的急性缺血性卒中患者进行了回顾性队列研究。使用两步逐像素后处理方法从体素内不相干运动灌注分数图中得出。根据弥散加权成像和灌注加权成像,使用 Tmax>6s 阈值,划定缺血核心、半影、无缺血和对侧半球。在体素内不相干运动(IVIM、IVIM D*、IVIM D*)参数图上测量不同脑区的信号强度,并使用单因素方差分析比较差异。在灌注加权成像和体素内不相干运动参数图上测量缺血体积。使用 Bland-Altman 分析和基于体素的容积比较来确定灌注加权成像和体素内不相干运动灌注参数的缺血体积之间的一致性。还评估了体素内不相干运动图的组内可靠性。显著性水平设定为 α<0.05。
共纳入 20 例患者(11 例男性,55%;平均年龄 67.1±13.8 岁)。血管闭塞累及颈内动脉(6 例,30%)和大脑中动脉 M1 段(14 例,70%)。治疗前核心梗死体积平均为 19.07±23.56ml。治疗前灌注加权成像的平均缺血体积为 10.90±13.33ml(CBV)、24.83±23.08ml(CBF)、58.87±37.85ml(MTT)和 47.53±26.78ml(Tmax)。相应 IVIM 参数的平均缺血体积为 23.20±25.63ml(IVIM)、14.01±16.81ml(IVIM D*)和 27.41±40.01ml(IVIM D*)。IVIM、D 和 D* 之间存在显著差异( <0.001)。IVIM D*与 CBF 的缺血体积和基于体素的重叠具有最佳一致性,平均体积差异为 0.5ml,平均 Dice 相似系数(DSC)为 0.630±0.136。
体素内不相干运动和灌注加权成像参数图之间的脑灌注评估存在中度差异,IVIM D*与灌注加权成像 CBF 具有良好的一致性。体素内不相干运动有望用于急性缺血性卒中患者的脑灌注评估。