Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas, Texas, USA.
Max Planck Institute for Biological Cybernetics, Tübingen, Germany.
Magn Reson Med. 2022 Oct;88(4):1500-1515. doi: 10.1002/mrm.29314. Epub 2022 Jun 3.
To further develop MultiNet GRAPPA, a neural-network-based reconstruction, for lower SNR proton MRSI ( H MRSI) data using adapted undersampling schemes and improved training sets.
H FID-MRSI data and an anatomical image for GRAPPA reconstruction were acquired in two slices in the human brain (n = 6) at 7T. MRSI data were retrospectively undersampled for a 4×, 6×, and 7× acceleration rate. Signal-to-noise, relative error (RE) between accelerated and fully sampled metabolic maps, RMS of the lipid artifacts, and fitting reliability were compared across acceleration rates, to the fully sampled data, and with different kinds and amounts of training images.
Training with semi-synthetic images resulted in higher SNR and lower lipid RMS relative to training with acquired images from one or several subjects. SNR increased with the number of semi-synthetic training images and the 4× accelerated data retains ∼30% more SNR than other accelerated data. Spectra reconstructed with 20 semi-synthetic averages retained ∼100% more SNR and had ∼5% lower lipid RMS than those reconstructed with the center k-space points of one image as was originally proposed for very high SNR MRSI data and had higher fitting reliability. The metabolite RE was lowest when training with 20-semi-synthetic training images and highest when training with the center k-space points of one image.
MultiNet GRAPPA is feasible with lower SNR H MRSI data if 20-semi-synthetic training images are used at a 4× acceleration rate. This acceleration rate provided the best trade-off between scan time and spectral SNR.
进一步开发基于神经网络的重建方法 MultiNet GRAPPA，用于使用自适应欠采样方案和改进的训练集对低 SNR 质子 MRSI（H MRSI）数据进行重建。
在 7T 下对人脑的两个切片采集 H FID-MRSI 数据和用于 GRAPPA 重建的解剖图像。对 MRSI 数据进行回顾性欠采样，以实现 4×、6×和 7×的加速率。比较了在不同加速率下的信号噪声、加速代谢图与完全采样图之间的相对误差（RE）、脂质伪影的均方根（RMS）和拟合可靠性，与完全采样数据以及不同种类和数量的训练图像进行了比较。
与从一个或几个受试者采集的图像进行训练相比，使用半合成图像进行训练可获得更高的 SNR 和更低的脂质 RMS。SNR 随半合成训练图像数量的增加而增加，4×加速数据比其他加速数据保留约 30%的 SNR。与最初提出的非常高 SNR MRSI 数据相比，使用 20 个半合成平均值重建的谱保留了约 100%的 SNR，并且脂质 RMS 降低了约 5%，并且具有更高的拟合可靠性。使用 20 个半合成训练图像进行训练时，代谢物的 RE 最低，而使用一个图像的中心 k 空间点进行训练时，RE 最高。
如果在 4×加速率下使用 20 个半合成训练图像，则 MultiNet GRAPPA 可用于低 SNR 的 H MRSI 数据。该加速率在扫描时间和光谱 SNR 之间提供了最佳的折衷。
