Flexible Prediction of CT Images From MRI Data Through Improved Neighborhood Anchored Regression for PET Attenuation Correction
Flexible Prediction of CT Images From MRI Data Through Improved Neighborhood Anchored Regression for PET Attenuation Correction
复制标题
通过改进邻域锚定回归进行 PET 衰减校正，根据 MRI 数据灵活预测 CT 图像
DOI：
10.1109/jbhi.2019.2927368
复制
发表时间：
2020-04
影响因子：
7.7
通讯作者：
Yang Wei
中科院分区：
文献类型：
--
作者：
Zhong Liming;Chen Yanlin;Zhang Xiao;Liu Shupeng;Wu Yuankui;Liu Yunbi;Lin Liyan;Feng Qianjin;Chen Wufan;Yang Wei
考虑到磁共振成像(MRI)信号与衰减值之间的复杂关系，正电子发射断层成像(PET)/MRI混合系统的衰减校正仍然是一个具有挑战性的任务。目前，现有的方法要么很耗时，要么需要足够的样本来训练模型。本文提出了一种在有限数据下从T1和T2加权MRI数据中预测伪层析(CT)图像的有效方法。该方法以改进的邻域锚定回归(INAR)为基线，预先计算投影矩阵，灵活地预测伪CT斑块。通过增加MR/CT数据集、学习MR图像的非线性描述子、分层搜索最近邻、数据驱动优化和多元回归集成等技术，提高了该方法的有效性。总共有22名健康受试者参加了这项研究。使用INAR和多元回归组合得到的伪CT图像的平均绝对误差(MAE)为92.73；内嵌公式notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；为14.86HU，峰值信噪比为29.77；内嵌公式notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；1.63分贝，皮尔逊线性相关系数为0.82；直列公式notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；为0.05，骰子相似系数为0.81；直列公式notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；与真实CT图像相比，正电子发射计算机断层扫描衰减校正的相对平均绝对误差notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；为0.20%。此外，我们提出的INAR方法，在没有任何改进策略的情况下，只需7个学科(MAE 106.89；内联公式；<；TeX-MATH notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；14.43Hu；RMAE1.51<；内联-公式>；<；TeX-MATH notation=“LaTeX”>；$\pm$<；/tex-math>；<；/inline-formula>；0.21%)就可以取得相当大的效果。实验证明，该方法比六种创新方法具有更好的性能。此外，该方法可以快速生成适合于PET衰减校正的伪CT图像。
Given the complicated relationship between the magnetic resonance imaging (MRI) signals and the attenuation values, the attenuation correction in hybrid positron emission tomography (PET)/MRI systems remains a challenging task. Currently, existing methods are either time-consuming or require sufficient samples to train the models. In this paper, an efficient approach for predicting pseudo computed tomography (CT) images from T1- and T2-weighted MRI data with limited data is proposed. The proposed approach uses improved neighborhood anchored regression (INAR) as a baseline method to pre-calculate projected matrices to flexibly predict the pseudo CT patches. Techniques, including the augmentation of the MR/CT dataset, learning of the nonlinear descriptors of MR images, hierarchical search for nearest neighbors, data-driven optimization, and multi-regressor ensemble, are adopted to improve the effectiveness of the proposed approach. In total, 22 healthy subjects were enrolled in the study. The pseudo CT images obtained using INAR with multi-regressor ensemble yielded mean absolute error (MAE) of 92.73 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 14.86 HU, peak signal-to-noise ratio of 29.77 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 1.63 dB, Pearson linear correlation coefficient of 0.82 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 0.05, dice similarity coefficient of 0.81 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 0.03, and the relative mean absolute error (rMAE) in PET attenuation correction of 1.30 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 0.20% compared with true CT images. Moreover, our proposed INAR method, without any refinement strategies, can achieve considerable results with only seven subjects (MAE 106.89 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 14.43 HU, rMAE 1.51 <inline-formula><tex-math notation="LaTeX">$\pm$</tex-math></inline-formula> 0.21%). The experiments prove the superior performance of the proposed method over the six innovative methods. Moreover, the proposed method can rapidly generate the pseudo CT images that are suitable for PET attenuation correction.
登录
查看更多内容
DOI：
--
发表时间：
2017
期刊：
--
影响因子：
--
作者：
Fang Liu;Hyungseok Jang;Richard Kijowski;T. Bradshaw;A. McMillan
通讯作者：
Fang Liu;Hyungseok Jang;Richard Kijowski;T. Bradshaw;A. McMillan
影响因子：
9.3
作者：
Delso, Gaspar;Wiesinger, Florian;Veit-Haibach, Patrick
通讯作者：
Veit-Haibach, Patrick
影响因子：
9.3
作者：
Leynes, Andrew P.;Yang, Jaewon;Larson, Peder E. Z.
通讯作者：
Larson, Peder E. Z.
影响因子：
9.3
作者：
Hofmann, Matthias;Steinke, Florian;Pichler, Bernd J.
通讯作者：
Pichler, Bernd J.
影响因子：
10.6
作者：
Burgos, Ninon;Cardoso, M. Jorge;Ourselin, Sebastien
通讯作者：
Ourselin, Sebastien