动态三维血管建模法分析Stanford B型胸主动脉夹层四维相位对比MRI

Dynamic three-dimensional blood vessel modeling for analyzing 4 dimensional phase contrast MRI of Stanford B type aortic dissection

目的由动态三维（3D）血管建模法分析四维（4D）相位对比MRI（PC-MRI）所测血流参数，判定正常主动脉及StanfordB型胸主动脉夹层患者血流动力学特征。方法在带有时间分辨率的3DPC-MRI所获数据中，获取动态3D血管模型。定位任意一层血管参考截面作为参考像，通过在参考像的同一截面提取变形轮廓信息，在同一位点判定不同时相的靶血管脉动性，并应用时间描点法获取脉动性数据。应用该后处理方法，分析比较采用4DPC—MRI获取的19名健康成年志愿者及8例StanfordB型主动夹层患者的血流动力学差异。结果健康人的主动脉内血流流型绝大部分为层流，仅见极少量螺旋流，绝无涡流，主动脉全长范围内壁面切应力变化小且均匀。假腔内逆向血流出现频率高，且峰值流速到达时间提前，对应真腔内正向血流比例高，无峰值流速提前到达表现。真假腔射血量比较：真腔[中位数（M）=54．3mL；四分位间距（IQR）=43．2～64．8mL]较假腔（M=31．6mL；IQR=19．8～47．6mL）高（P〈0．01）；真腔内血流方向以前向为主（M=91．4％，IQR：90．0％～94．2％），假腔内逆向血流所占比例高（M=40．3％，IQR=23．2％～53．3％；P〈0．01）；假腔内平均流速（M=7．1cm／s，IQR=4．9～9．8cm／s）低于真腔（M=18．0cm／s，IQR=13．9～20．6cm／s；P〈0．01）；假腔内峰值流速于R波后166．0ms（IQR=132．8～210．0ms）到达，明显早于真腔215．0ms（IQR=196．3～249．0ms；P％0．01）。螺旋流通常出现在收缩早期，R波后158ms（IQR=145～249ms），持续306ms（IQR=217～537ms），最大旋转角度为820°／搏。结论血管动态3D模型能够高效分析4DPC-MRI，并且提供定量的血流参数信息，血流方向、流速峰值到达时间及螺旋流的出现或改变可能参与主动脉夹层假腔的发生和发展。

Objective To apply dynamic three-dimensional blood vessel modeling for analyzing the blood flow parameters obtained by 4 dimensional phase contrast MRI （4D PC-MRI） of Stanford B type aortic dissection, so as to demonstrate the blood flow characteristics of both healthy controls and patients with Stanford B type aortic dissection. Methods Dynamic 3D models of blood vessel were captured from 3D PC-MRI with temporal resolution. A reference vascular cross-sectional plane was defined, and the displacement contour information at the same plane was all used to determine the pulsatility of the target vascular cross-sectional planes at multiple time points. The pulsatility parameters of target vascular cross-sectional planes were obtained by temporal tracking. The hemodynamic differences between healthy adults （n= 19） and patients （n= 8） with Stanford B type aortic dissection were analyzed by comparing 4D PC-MRI data of the two groups. Results Qualitative blood flow visualization showed laminar flow in the aorta of healthy volunteers, without turbulences or vortex formation, with slight helical flow pattern found in the ascending aorta; there were little changes in the wall shear stress in the entire thoracic aorta. The blood flow in the false lumen was multidirectional and complex, with a high incidence of reverse flow, and the true lumens were dominated by aortic flow direction. The peak velocity of blood flow arrived earlier in the false lumen, not in the true lumen. The stroke volume was greater in the true lumen （media [M]54.3 mL, interquartile range [IQR] 43.2-64.8 mL） compared with the false one （M31.6 mL, IQR 19. 8-47. 6 mL, P%0. 01）. The majority of the flow in the true lumen was forward flow （M 91.4%, IQR 90.0%-94.2 % ）, whereas the false lumen had a high proportion of backward flow （M 40.3 %, IQR 23.2 %-53.3%）. The average velocity of blood flow in the false lumen （M 7.1 cm/s,IQR 4.9-9.8 cm/s） was lower than that in the true lumen （M 18.0 cm/s, IQR 13.9-20.6 era/s, P〈0.01）. The maximum velocity occurred earlier in the false lumen during the cardiac cycle 166.0 ms after the R-wave （IQR 132.8-210.0 ms） compared with that in the true lumen （M 215.0 ms, IQR 196.3-249.0 ms,P〈0.01）. Helical flow mainly occurred at early-systole stage, at 158 ms （IQR 145-249 ms） after the R- wave, and lasted for 310 ms （IQR 217-537 ms）, with the maximum rotation being 820° per cardiac cycle. Conclusion Dynamic 3D modeling method can effectively analyze the flow parameters obtained from 4D PC-MRI and can provide qualitative blood flow information. Flow direction, time to peak velocity, and development and changes of helical flow may be involved in the pathology of aortic dissection.