【作者】 张孝通；

【导师】 朱善安； Bin He；

【作者基本信息】 浙江大学 ， 控制理论与控制工程， 2009， 博士

【摘要】 生物组织的电特性(电导率与介电常数)反映了组织的生理、病理信息,其分布图有助于对病变(如癌症等)组织的早期诊断。本论文在回顾了几种重要医学电磁成像方法的基础上,从物理原理以及实验技术等方面论述了当前研究热门之一的磁共振电阻抗成像技术(MREIT),原创性地提出了一种仅利用单方向磁感应强度、基于自适应网络模糊推理系统的ANFIS-MREIT算法,并在两种典型三维头模型上验证了该算法的有效性和抗噪性。即使在加入噪声或者电极位置发生偏移的情况下,该算法仍然可以十分准确地对头部组织电导率比值进行重建。通过与同类算法的结果比较发现,对于具有多层组织、电导率各向同性且分层连续体模型的阻抗重建,该算法具有较大的优越性。继加拿大、韩国和土耳其的研究小组之后,我们在国内率先开展了电流密度成像(MRCDI)实验,利用一台1.5T的MRI设备得到了成像模型内部横截面竖直方向上的电流密度分布信息,测量计算结果与实际值误差为10.62%。论文分析了引起噪声的若干因素,这其中包括来自MRI系统电路的干扰信号以及成像物体翻转过程中带来的误差；同时提出了抑制噪声、提高实验结果信噪比的若干软硬件改进方法。磁共振电特性成像技术(MREPT)是新近提出的电磁成像方法。该技术基于射频场成像技术(B1-mapping),无需外加能量注入成像体,仅通过常规磁共振扫描即可实现,克服了MREIT技术的局限性,真正实现了对人体测量的非侵入性和无损伤性,因而有着很大的发展潜力。本论文分析了该技术的物理原理,克服了现有三种MREPT重建算法的不足,原创性地提出了非迭代的Dual-excitation算法,并利用一系列二维及三维仿真实验验证了其可行性和有效性,证明了其较同类算法的优越性。同时,率先对MREPT图像重建过程中的诸多技术环节做了深入的理论分析。与MREIT技术不同,MREPT技术同时对组织电导率和介电常数两者进行重建,因此可以提供比MREIT更为丰富的组织功能结构信息。该技术同时有助于我们对高频下衡量人体因吸收电磁波生热的“比吸收率”进行计算,因此可用于衡量高场强磁共振系统发热安全问题以及手机辐射安全等领域。更多还原

【Abstract】 The electric properties (EPs) of biological tissue, i.e., the electric conductivity and permittivity, can provide important information in diagnosis of various diseases. The EPs distribution within human body has been the subject of research for over 30 years since the Electrical Impedance Tomography (EIT) was proposed. This dissertation reviews several important EPs imaging modalities, and provides theoretical and experimental studies on Magnetic Resonance Electrical Impedance Tomography (MREIT), which is popularly pursuit in recent 10 years. Using adaptive neuro-fuzzy inference system (ANFIS), a new ANFIS-MREIT algorithm has been developed, and only one component of the magnetic flux densities was utilized. Simulations were performed on sphere and realistic-geometry models to estimate head tissues conductivity with and without noise contamination, and electrodes excursion was also concerned to evaluate its effectiveness. Promising simulation results suggest the merits of ANFIS-MREIT in estimating the conductivity values of head volume conductor for piece-wise homogeneous head volume-conductor models.The first Magnetic Resonance Current Density Imaging (MRCDI) experiment was carried out in China. The feasibility of MRCDI as the tool of measurements for MREIT has been shown after a phantom experiment with a 1.5T MRI. The relative error of measured current density on a transverse plane was 10.62%, while noise induced by electronics and phantom rotation was observed to be the main cause of errors. Several effective ways to improve CDI result have been proposed according to the error analysis.Based on the measurement of the active transverse magnetic component of the applied RF field (known as B1-mapping), Magnetic Resonance Electric Properties Tomography (MREPT) has been newly developed to image EPs distributions within biological tissues. MREPT can be performed on a standard MRI system using a regular volume coil, and it differs from other noninvasive imaging techniques in that no electrode mounting is required and no external energy is introduced into the body during MRI scanning. MREPT can also help with specific absorption rate (SAR) calculation which is a major concern in high-field Magnetic Resonance Imaging (MRI) as well as in non-medical areas like wireless-telecommunications. We have proposed a novel MREPT algorithm, Dual-excitation algorithm, which uses two sets of measured B1 data, to noninvasively reconstruct the biological tissue’s electric properties. The Finite Element Method (FEM) has been utilized in three-dimensional (3D) modeling and B1 field calculation. A series of computer simulations were conducted to evaluate the feasibility and performance of the proposed method on a 3D head model within a birdcage coil and a transverse electromagnetic (TEM) coil. Compared with other B1-mapping based reconstruction algorithms, our approach provides superior performance without the need for iterative computations. The present simulation results indicate good reconstruction of electric properties from B1 mapping.更多还原