高场磁共振人体组织电特性断层成像核心技术研究

The technique of high-field magnetic resonance electrical properties tomography (MR EPT) of human tissues is an emerging MR imaging technique, which can non-invasively provide imaging of electrical properties of tissues at arbitrary tomographic slices in vivo. The mechanism of MR EPT is completely different with the mechanism which underlies the current MRI based on the transversal and longitudinal relaxation time. The electrical properties of tissues characterize the biological and pathological status of tissues, varied with different tissues. Studies have proved that the electrical properties between healthy and malignant tissues are very different. Therefore, MR EPT is a potential technique for the early detection of tumors at clinic. However, the key algorithm of current MR EPT assumes that the gradient of electrical properties in the small local region of interest is zero, conflicting with the fact that the distribution of electrical properties of human tissues is inhomogeneous, resulting in poor imaging quality. Aiming at the existed substantial deficiency of current MR EPT technique, this project proposed a novel MR EPT algorithm named finite difference time domain (FDTD) with adjustable electrical properties. Employing the proposed core equations of the new algorithm, the tomographic distribution of electrical properties were calculated using the FDTD method with adjustable electrical properties, without the unreasonable assumption adopted in the current algorithm. The preliminary experiment proved that the new algorithm exhibits excellent convergence, promising to greatly improve the imaging quality. The proposed novel algorithm is completely different with current method, promising to substantially improve the MR EPT technique. The related research may become a great contribution to the high-field MR imaging technique.

磁共振人体组织电特性断层成像（MR EPT）是新兴高场MR成像技术，其成像机制基于组织在MR射频电磁场中表现出来的电特性，与传统MRI基于组织纵向和横向弛豫时间的成像机制不同。MR EPT能够无创提供活体组织电特性断层图像，活体组织电特性能够反映其生理和病理状态，通常不同组织电特性各异，且组织癌变后电特性改变较大，因此MR EPT技术在肿瘤早期诊断等方面具有重要潜在价值。现有MR EPT技术核心算法需要假设较小成像计算区域内组织电特性梯度为零，而人体活体组织电特性分布客观上不均匀，导致成像结果误差较大。针对现有算法这一原理性缺陷，本课题提出了创新的变电特性参数时域有限差分（FDTD）MR EPT算法理论，新算法不受现有算法假设条件限制，在成像区域全局迭代求解组织电特性断层分布，预实验证实新算法收敛性较好，具可行性。本课题开辟了MR EPT全新技术路线，有望实质性提升MR EPT技术。

磁共振人体组织电特性断层成像（MR EPT）是一种“无创的人体活体组织电特性（EPs）断层成像”的新兴MR成像技术，在肿瘤早期诊断等方面具有重要潜在价值。针对以往MR EPT技术重建算法需要假设“较小成像计算区域局部组织EPs为常数”这一重大缺陷，本项目提出了不受以往算法假设条件限制的“变电特性参数时域有限差分（FDTD）” MR EPT算法理论，实现在成像区域全局迭代求解组织电特性断层分布。整体研究内容包括：（1）射频发射场的测量及鲁棒性研究；（2）射频电磁场FDTD方程组的建立；（3）变电特性参数FDTD新算法研究；（4）体模实验、动物实验和人体组织无创EPs断层成像临床测试。在项目实施期间，本研究获得以下重大进展：首先比较了双角度法（DAM）和快速预饱和法（satTFL）两种B1映像技术的鲁棒性，结果发现在低介电特性负载时，两种技术有着相同的适用性；在高介电特性负载时，satTFL较DAM适用性更佳。该结论可为不同介电特性组织选取合适的B1 mapping技术提供参考，为推动MR EPT技术的实用化提供基础研究支持。再者，根据麦克斯韦方程组构建组织EPs与RF场之间关系的FDTD核心算法方程组，实现了在方程组中反映EPs非均匀组织的σ(r)和ϵ(r)信息；实现采用变电特性参数迭代逼近准则求解未知计算区域内的σ(r)和ϵ(r)，获得组织EPs分布；并结合BSS B1 mapping技术研究优化Eps非均匀人体组织适用的MR EPT新算法；最终，分别用体模实验、动物实验和临床志愿者实验成功验证该算法的可行性。以上研究实质性地推动MR EPT技术进一步发展，为早日将MR EPT技术应用到肿瘤早期诊断中做出了贡献。项目执行期间，针对B1 mapping技术、EPT算法关键技术方面，分别在Physics in Medicine & Biology、Electromagnetic Biology and Medicine、Magnetic Resonance in Medical Sciences、NMR in Biomedicine等多个国内外期刊杂志共发表学术论文10篇，在申关于B1 mapping技术发明专利一项，进行学术报告2次。