Automatic measurement plane placement for 4D Flow MRI of the great vessels using deep learning.
Automatic measurement plane placement for 4D Flow MRI of the great vessels using deep learning.
复制标题
DOI：
10.1007/s11548-021-02475-1
复制
发表时间：
2022-01
影响因子：
3
通讯作者：
Wieben O
中科院分区：
文献类型：
--
作者：
Corrado PA;Seiter DP;Wieben O
尽管4D Flow MRI在血流动力学分析方面具有巨大的潜力和灵活性，但其主要局限性是需要耗时且依赖于用户的后处理。我们提出了一个快速的四步算法，快速，鲁棒性和可重复的流量测量的基础上自动放置的测量平面和血管分割的大血管。我们的算法通过以下方式工作：1）将3D图像子采样为3D补丁，2）通过卷积神经网络预测每个补丁包含单个血管的概率以及补丁内血管的位置/方向，3）选择每个血管概率最高的预测平面，以及4）将平面中心移动到每个平面内的最大速度。该方法在283次扫描上进行了训练，并通过使用t检验、Pearson相关性和Bland-Altman分析将算法推导的处理时间、平面位置和流量测量值与两名手动观察者（研究生）的测量值进行比较，在40次看不见的扫描上进行了评估。算法的平均处理时间（18秒）短于观察者1（362秒; P<0.001）和观察者2（317秒; P<0.001）。算法放置的平面与手动观察者放置的平面之间的距离（O 1与算法：9.0 mm，O2与算法：10.3 mm）略大于两个手动观察者放置的平面之间的距离（8.3 mm）。由算法放置的平面和由手动观察者放置的平面的流量值之间的相关性（O 1与算法：R=0.68，O2与算法：R=0.72）略低于两个手动观察者之间的流量相关性（R=0.81）。我们的方法是一种可行且准确的方法，可用于大血管4D血流MRI中的快速、可重复和自动化流量测量和可视化，与手动注释器相比，其变异性与两个手动观察者之间的变异性相似。这种方法可以应用于其他解剖区域。
Despite the great potential and flexibility of 4D flow MRI for hemodynamic analysis, a major limitation is the need for time-consuming and user-dependent post-processing. We propose a fast four-step algorithm for rapid, robust, and repeatable flow measurements in the great vessels based on automatic placement of measurement planes and vessel segmentation. Our algorithm works by 1) subsampling the 3D image into 3D patches, 2) predicting the probability of each patch containing individual vessels and location/orientation of the vessel within the patch via a convolutional neural network, 3) selecting the predicted planes with highest probabilities for each vessel, and 4) shifting the plane centers to the maximum velocity within each plane. The method was trained on 283 scans and evaluated on 40 unseen scans by comparing algorithm-derived processing times, plane locations, and flow measurements to those of two manual observers (graduate students) using t-tests, Pearson correlation, and Bland-Altman analysis. The average processing time for the algorithm (18 seconds) was shorter than observer 1 (362 seconds; P<0.001) and observer 2 (317 seconds; P<0.001). The distance between planes placed by the algorithm and those placed by manual observers was slightly greater (O1 vs. algorithm: 9.0mm, O2 vs. algorithm: 10.3mm) than the distance between planes placed by the two manual observers (8.3mm). The correlation between flow values for planes placed by the algorithm and those placed by manual observers was slightly lower (O1 vs. algorithm: R=0.68, O2 vs. algorithm: R=0.72) than the flow correlation between the two manual observers (R=0.81). Our method is a feasible and accurate approach for fast, reproducible, and automated flow measurement and visualization in 4D flow MRI of the great vessels, with similar variability compared to a manual annotator as the variability between two manual observers. This approach could be applied in other anatomical regions.
登录
查看更多内容
影响因子：
10.9
作者：
Bustamante, Mariana;Gupta, Vikas;Ebbers, Tino
通讯作者：
Ebbers, Tino
影响因子：
3.3
作者：
Stalder, A. F.;Russe, M. F.;Markl, M.
通讯作者：
Markl, M.
影响因子：
6.7
作者：
Frydrychowicz A;Wieben O;Niespodzany E;Reeder SB;Johnson KM;François CJ
通讯作者：
François CJ
DOI：
10.1186/s12968-015-0190-5
发表时间：
2015-10-05
期刊：
Journal of cardiovascular magnetic resonance : official journal of the Society for Cardiovascular Magnetic Resonance
影响因子：
--
作者：
Bustamante M;Petersson S;Eriksson J;Alehagen U;Dyverfeldt P;Carlhäll CJ;Ebbers T
通讯作者：
Ebbers T
影响因子：
4.4
作者：
Tariq, Umar;Hsiao, Albert;Alley, Marcus;Zhang, Tao;Lustig, Michael;Vasanawala, Shreyas S.
通讯作者：
Vasanawala, Shreyas S.