Department of Biomedical Engineering, School of Biomedical Engineering and Imaging Sciences, King's College London, United Kingdom.
King's College London & Guy's and St Thomas' PET Centre, St Thomas' Hospital, London, United Kingdom.
Magn Reson Med. 2019 Mar;81(3):2120-2134. doi: 10.1002/mrm.27521. Epub 2018 Oct 16.
To propose a framework for synergistic reconstruction of PET-MR and multi-contrast MR data to improve the image quality obtained from noisy PET data and from undersampled MR data.
Weighted quadratic priors were devised to preserve common boundaries between PET-MR images while reducing noise, PET Gibbs ringing, and MR undersampling artifacts. These priors are iteratively reweighted using normalized multi-modal Gaussian similarity kernels. Synergistic PET-MR reconstructions were built on the PET maximum a posteriori expectation maximization algorithm and the MR regularized sensitivity encoding method. The proposed approach was compared to conventional methods, total variation, and prior-image weighted quadratic regularization methods. Comparisons were performed on a simulated [ F]fluorodeoxyglucose-PET and T /T -weighted MR brain phantom, 2 in vivo T /T -weighted MR brain datasets, and an in vivo [ F]fluorodeoxyglucose-PET and fluid-attenuated inversion recovery/T -weighted MR brain dataset.
Simulations showed that synergistic reconstructions achieve the lowest quantification errors for all image modalities compared to conventional, total variation, and weighted quadratic methods. Whereas total variation regularization preserved modality-unique features, this method failed to recover PET details and was not able to reduce MR artifacts compared to our proposed method. For in vivo MR data, our method maintained similar image quality for 3× and 14× accelerated data. Reconstruction of the PET-MR dataset also demonstrated improved performance of our method compared to the conventional independent methods in terms of reduced Gibbs and undersampling artifacts.
The proposed methodology offers a robust multi-modal synergistic image reconstruction framework that can be readily built on existing established algorithms.
提出一种 PET-MR 和多对比度 MR 数据协同重建的框架，以提高从噪声 PET 数据和欠采样 MR 数据中获得的图像质量。
设计了加权二次先验项，以在降低噪声、PET Gibbs 振铃和 MR 欠采样伪影的同时，保持 PET-MR 图像之间的公共边界。这些先验项使用归一化多模态高斯相似核进行迭代重新加权。协同 PET-MR 重建基于 PET 最大后验期望最大化算法和 MR 正则化敏感编码方法。与传统方法、全变差和先验图像加权二次正则化方法相比，对模拟 [F]氟脱氧葡萄糖-PET 和 T/T 加权 MR 脑体模、2 个活体 T/T 加权 MR 脑数据集以及活体 [F]氟脱氧葡萄糖-PET 和液体衰减反转恢复/T 加权 MR 脑数据集进行了比较。
模拟结果表明，与传统、全变差和加权二次方法相比，协同重建在所有图像模态下均能获得最低的定量误差。尽管全变差正则化保留了模态特有的特征，但与我们提出的方法相比，该方法无法恢复 PET 细节，也无法减少 MR 伪影。对于活体 MR 数据，我们的方法对 3×和 14×加速数据保持了相似的图像质量。与传统的独立方法相比，PET-MR 数据集的重建也证明了我们的方法在降低 Gibbs 和欠采样伪影方面的性能有所提高。
所提出的方法提供了一种稳健的多模态协同图像重建框架，可轻松构建在现有的成熟算法之上。
