4705
Improved TWIST Imaging using k-Space Deep Learning1KAIST, Daejeon, Korea, Republic of, 2Gachon University Gil Medical Center, Incheon, Korea, Republic of
Synopsis
Time-resolved angiography with interleaved stochastic trajectories(TWIST) has been widely used for dynamic contrast enhanced (DCE) MRI. To achieve highly accelerated acquisitions for improved temporal and spatial resolution, the high frequency region is randomly sub-sampled at each time frame. Therefore, the periphery of the k-space data from multiple time frames are combined to obtain the uniformly sub-sampled k-space data so that the temporal resolution of TWIST is limited. The purpose of this research is to improve the temporal resolution of TWIST by reducing the view-sharing. Furthermore, we proposed the algorithm that can reconstruct the imagesat various number of view sharing using k-space deep learning.
Introduction
Time-resolved angiography with interleaved stochastic trajectories(TWIST)1 has been widely used for dynamic contrast enhanced (DCE) MRI. To achieve highly accelerated acquisitions for improved temporal and spatial resolution, the high frequency region is randomly sub-sampled at each time frame. Therefore, the periphery of the k-space data from multiple time frames are combined to obtain the uniformly subsampled k-space data so that the temporal resolution of TWIST is limited as shown in Fig. 1(a). The purpose of this research is to improve the temporal resolution of TWIST by reducing the view-sharing. Furthermore, we proposed the algorithm that can reconstruct the images at various number of view sharing using k-space deep learning.Methods
It was shown that the missing k-space data for parallel MRI can be interpolated by exploiting the redundancies along the coils and images. Moreover, a recent mathematical discovery reveals that the Hankel matrix decomposition is closely related to a convolutional neural network (CNN). By synergistically combining these findings, we propose a multi-coil k-space deep learning approach that immediately interpolates the missing k-space data using a deep convolutional neural network.
As shown in Fig. 1(b), the reduced view sharing results in irregular sampling pattern which cannot be reconstructed using the existing GRAPPA algorithm. We demonstrated that the missing elements in k-space data from two frames can be interpolated using annihilating fi lter based low-rank Hankel matrix approach (ALOHA)2, but its computational cost is too expensive. It was shown that the missing k-space data for parallel MRI can be interpolated by exploiting the redundancies along the coils and images. Moreover, a recent mathematical discovery reveals that the Hankel matrix decomposition is closely related to a convolutional neural network (CNN)3. By synergistically combining these findings, we propose a multi-coil k-space deep learning approach that immediately interpolates the missing k-space data using a deep convolutional neural network. The neural network is constructed in the k-space domain by stacking multi-coil k-space data along the channel direction of the network as shown in Fig. 1(c). To address the lack of ground-truth data and to provide the flexible reconstruction for various number of view sharing, our neural network learned the k-space interpolation relationship between the minimum number of k-space samples and the fully sampled k-space data from GRAPPA reconstruction. The proposed network was implemented in Python using TensorFlow library.
We used four sets of in vivo 3D DCE data for carotid vessel scan obtained with TWIST sequence using Siemens 3T Verio scanners. The acquisition parameters for two sets were as following: 159×640×80 matrix, 16 coils, and 30 temporal frames. For other two sets, the acquisition parameters were same as above, except for 37 temporal frames.The down sampling rate was 3 and 2 along kx and ky direction, respectively. Due to partial Fourier, only 63% of data was acquired. Among four patients, two patients data were used for training, and one patient data was used for validation. We used the remaining patient data to test the trained network.
Result
Fig. 2 showed the subtracted maximum intensity projection (MIP) images for test data. We selected the temporal frames to illustrate the propagation of the contrast agent. The reconstructed images at various number of view sharing (VS)=2, 3, and 5 were provided using same neural network. The raw data in Fig. 2 was obtained by directly applying inverse Fourier Fast Transform (FFT) to the k-space data without view sharing, which has the true temporal resolution. As shown in Fig. 2, the contrast agent in the reconstructed images using GRAPPA was suddenly propagated from the T = 11 frame the T = 12 frame. This degradation of temporal dynamics could be found frequently as the number of view-sharing increased. In the reconstructed images with VS=2 using the proposed network, the flow of the contrast agent and the detail of dynamics were captured correctly , as shown in Fig. 2. The results of proposed method with VS=5, which is same to the conventional method, provided very closer spatial and temporal resolution to the GRAPPA reconstruction. This implies that one trained network can generate multiple reconstructed results with various spatial and temporal resolution by simply changing the number of view sharing at the inference stage.Discussion&Conclusion
The proposed k-space deep learning for reconstruction with reduced view-sharing significantly improved the image quality and diagnostic validity of TWIST imaging without changing the current acquisition protocol. Furthermore, the proposed network can produce the results only in 0.029s, which is several order of magnitude faster than ALOHA and GRAPPA reconstruction.Acknowledgements
This study was supported by Korea Science and Engineering Foundation under Grant NRF-2016R1A2B3008104.References
1. Gerhard L and Randall K. syngo TWIST for dynamic time-resolved MR angiography. Magnetom Flash. 2006;34(3):92-95.
2. Kyoung Hwan J, Dong-wook L, and Jong Chul Y. A General Framework for Compressed Sensing and Parallel MRI Using Annihilating Filter Based Low-Rank Hankel Matrix. IEEE Trans. on Computational Imaging.2016;2(4):480-495.
3. Jong Chul Y, Yoseob H, and Eunju C. Deep convolutional framelets: A general deep learning framework for inverse problems. SIAM Journal on Imaging Sciences 11.2 (2018): 991-1048.
Figures
