Zhang Qiang, Su Pan, Chen Zhensen, Liao Ying, Chen Shuo, Guo Rui, Qi Haikun, Li Xuesong, Zhang Xue, Hu Zhangxuan, Lu Hanzhang, Chen Huijun
Center for Biomedical Imaging Research, Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China.
The Russell H. Morgan, Department of Radiology, Johns Hopkins University School of Medicine, Baltimore, Maryland.
Magn Reson Med. 2020 Aug;84(2):1024-1034. doi: 10.1002/mrm.28166. Epub 2020 Feb 4.
To develop a reproducible and fast method to reconstruct MR fingerprinting arterial spin labeling (MRF-ASL) perfusion maps using deep learning.
A fully connected neural network, denoted as DeepMARS, was trained using simulation data and added Gaussian noise. Two MRF-ASL models were used to generate the simulation data, specifically a single-compartment model with 4 unknowns parameters and a two-compartment model with 7 unknown parameters. The DeepMARS method was evaluated using MRF-ASL data from healthy subjects (N = 7) and patients with Moymoya disease (N = 3). Computation time, coefficient of determination (R ), and intraclass correlation coefficient (ICC) were compared between DeepMARS and conventional dictionary matching (DM). The relationship between DeepMARS and Look-Locker PASL was evaluated by a linear mixed model.
Computation time per voxel was <0.5 ms for DeepMARS and >4 seconds for DM in the single-compartment model. Compared with DM, the DeepMARS showed higher R and significantly improved ICC for single-compartment derived bolus arrival time (BAT) and two-compartment derived cerebral blood flow (CBF) and higher or similar R /ICC for other parameters. In addition, the DeepMARS was significantly correlated with Look-Locker PASL for BAT (single-compartment) and CBF (two-compartment). Moreover, for Moyamoya patients, the location of diminished CBF and prolonged BAT shown in DeepMARS was consistent with the position of occluded arteries shown in time-of-flight MR angiography.
Reconstruction of MRF-ASL with DeepMARS is faster and more reproducible than DM.
开发一种可重复且快速的方法，利用深度学习重建磁共振指纹动脉自旋标记（MRF - ASL）灌注图。
使用模拟数据并添加高斯噪声训练一个全连接神经网络，称为DeepMARS。使用两个MRF - ASL模型生成模拟数据，具体为一个具有4个未知参数的单室模型和一个具有7个未知参数的双室模型。使用来自健康受试者（N = 7）和烟雾病患者（N = 3）的MRF - ASL数据评估DeepMARS方法。比较了DeepMARS与传统字典匹配（DM）之间的计算时间、决定系数（R）和组内相关系数（ICC）。通过线性混合模型评估DeepMARS与Look - Locker PASL之间的关系。
在单室模型中，DeepMARS每体素的计算时间<0.5毫秒，而DM则>4秒。与DM相比，DeepMARS在单室模型导出的团注到达时间（BAT）以及双室模型导出的脑血流量（CBF）方面显示出更高的R和显著改善的ICC，在其他参数方面R/ICC更高或相似。此外，DeepMARS在BAT（单室）和CBF（双室）方面与Look - Locker PASL显著相关。而且，对于烟雾病患者，DeepMARS中显示的CBF降低和BAT延长的位置与飞行时间磁共振血管造影中显示的闭塞动脉位置一致。
与DM相比，使用DeepMARS重建MRF - ASL更快且更具可重复性。
