基于T2弛豫定量的快速磁共振髓鞘水成像方法研究

Myelin water imaging is a non-invasive technique to detect Myelin sheath using quantative T2-relaxometry. However, the relatively long scanning time impedes its clinical applications. This project devotes to accelerating quantitative T2-relaxometry based fast myelin water imaging by improving the three key factors of the whole imaging process, namely, “imaging sequence, image reconstruction and parameter mapping". Specifically, this work will investigate the effect of undersampling in k-t domain on the recovery of Hankel matrixes in r-t domain，the effect of T1 and B1 errors on T2 estimation with undersampled data and the intrinsic relationship between T2 relaxometry signal and its Hankel variant. To fulfill the investigation, we will design double inverse recovery based MESE undersampling sequence, study generating function based T2 relaxometry imaging method with T2 and proton density sparsity, T2 weighted image reconstruction with model driven constraint and phase compensation based unbiased T2 estimation using complex data. It is expected that this project could enhance the clinical value of Myelin water imaging and therefore assist more efficient and effective diagnosis and prognosis of neurological and psychiatric disorders.

髓鞘水成像是一种非侵入式的采用T2弛豫定量的髓鞘检测技术。它在临床的应用主要受限于过长的扫描时间。本项目拟从“成像序列——图像重建——参数拟合”这一技术全链条入手，开展基于T2弛豫定量的快速磁共振髓鞘水成像方法研究。具体围绕：k-t空间的欠采样模式对r-t空间所形成的Hankel矩阵恢复的影响机制，母函数模型中T1值和B1场的误差对欠采样时T2弛豫时间估计的影响机制，T2弛豫信号与Hankel矩阵奇异值和酉矩阵的内在联系，这三个科学问题，开发基于双翻转恢复的MESE欠采样序列，研究基于母函数模型的T2弛豫时间和质子密度图稀疏的成像方法，基于模型拟合约束的T2加权图像重建以及基于复数相位补偿的无偏估计方法。项目的顺利实施有望为神经和精神障碍性疾病的诊断和预后提供临床可用的有效标识手段，进而提高髓鞘水成像技术在临床的应用能力和价值。

髓鞘水成像是一种非侵入式的采用T2弛豫定量的髓鞘检测技术。它在临床的应用主要受限于过长的扫描时间。本项目拟从“成像序列—图像重建—参数拟合”这一技术全链条入手，开展基于T2等参数的快速磁共振成像方法研究。具体包括：基于子空间集思路，提出高分辨率的磁共振波谱成像方法，并首次实现人脑高分辨磁敏感度和磁共振波谱的同时采集的序列开发；母函数驱动的结构与参数双正则化参数定量估计，在可接受的定量估计质量下，成功实现高倍率加速。基于拟合约束模型以及复数补偿性矩阵加性分解模型，提出高质量的T2等参数成像方法成功实现高速的全脑定量扫描进程。本项目已发表期刊论文9篇，其中SCI论文7篇，申请发明专利5项，培养硕士研究生3名。