基于神经网络拟合的腹部化学交换饱和转移成像

[an error occurred while processing this directive]

王志超¹,张记磊²,赵羽³,华婷⁴,汤光宇⁴,李建奇^1,*(

)

CEST Imaging of the Abdomen with Neural Network Fitting

Zhi-chao WANG¹,Ji-lei ZHANG²,Yu ZHAO³,Ting HUA⁴,Guang-yu TANG⁴,Jian-qi LI^1,*(

)

图3. 双回波梯度回波（GRE）序列采集（第二行）和基于神经网络拟合（第三行）得到的一名健康志愿者腹部主磁场B₀图和对应CEST非对称分析结果.(a)未施加CEST饱和脉冲得到的参考模图；(b)~(d)双回波GRE序列得到的主磁场B₀图、B₀图矫正后的MTR_asym图，以及对应于图(a)橙色圆环中组织的Z谱分析；(e)~(g)通过神经网络预测得到水质子共振频率偏移量生成的主磁场B₀图、以此B₀图矫正后的MTR_asym图，以及对应于图(a)橙色圆环中组织的Z谱分析

Fig.3. Abdominal B₀ maps and the corresponding CEST results of a healthy volunteer obtained by gradient recalled echo (GRE) sequence (the 2^nd row) and neural network based method (the 3^rd row). (a) The reference image while the CEST saturation radio frequency pulse was not applied; (b) & (e) B₀ maps obtained by GRE sequence and neural network fitting, respectively; (c) & (f) MTR_asym maps based on GRE sequence and neural network fitting, respectively; (d) & (g) The Z-spectrum analysis of tissue located in the orange circle in the left kidney based on GRE sequence and neural network fitting, respectively