脑出血的磁感应电阻抗图像监测关键技术研究

Cerebral hemorrhage is a kind of high incidence of disease with sudden pathogenesis, rapid development and high mortality rate. The key of treatment is early detection as soon as possible. The current clinical technologies are not able to continuous monitor the process of cerebral hemorrhage. So, early detection of cerebral hemorrhage cannot be achieved and new technology is urgently expected to solve this question. The electrical impedance of human blood is very different from that of the other brain tissues. The status of cerebral hemorrhage can be reflected by detecting the electrical impedance in brain. Magnetic induction tomography is a kind of contactless electrical impedance tomography. The advantages of MIT are time sensitive, safe, invasive and very convenient to use because of non-contact. Based on the research since ten years ago, this project will focus on the key methods of MIT, such as phase detection method with high precise and high stability and reconstruction algorithm adaptive for in-vivo cerebral hemorrhage. In order to improve the quality of the measurement data, we want to apply the high precise excitation source, drift inhibition by temperature feedback, digital demodulation by decreasing the frequency and data preprocess method. In order to improve the performance of the reconstruction algorithm, we will adopt the optimal regularization parameter, time correlative algorithm using multiple data frames and prior conductivity distribution. After the classical imaging just using single or two data frame, we plan to study the continuous imaging monitor during a long time. Physical phantom simulated the reality brain and in-vivo animal model are used to verify the above methods and obtain the series measurement data. By analyze the series data and series images, early detection and process monitoring of cerebral hemorrhage by MIT are expected. In future, this project may provide new technology for clinical alert of cerebral hemorrhage.

脑出血是一种临床高发疾病，发病突然，发展迅速，死亡率高。治疗脑出血的关键是早发现。现有临床技术不能连续监测脑出血过程，无法早发现，迫切需要新技术。人体血液电阻抗与脑部其他组织差异显著，检测电阻抗可反映脑出血状态。磁感应断层成像是非接触电阻抗成像方法，它既有时间敏感性、安全、无创等特点，又因非接触在应用上极为便利。在多年研究基础上，重点针对高精度高稳定性相位检测、活体脑出血成像算法等关键技术开展深入系统的研究，具体拟采用高精度激励源、温度反馈漂移抑制、降频数字解调、数据预处理等方法改善测量数据质量，采用优化正则化参数、多帧测量数据的时间相关性成像算法、先验电导率分布信息等方法提升成像算法的性能。在单次成像基础上，研究长时间连续图像监测，并通过高仿真物理模型和活体动物模型实验，分析序列数据和图像，推进磁感应断层成像在活体脑出血早期发现和病程监护方面的进展，为将来临床脑出血预警提供新的技术手段。

脑出血是临床高发疾病，发病突然，发展迅速，死亡率高。治疗脑出血的关键是早发现。现有临床技术不能连续监测脑出血过程，无法早发现，迫切需要新技术。人体血液电阻抗与脑部其他组织差异显著，检测电阻抗可反映脑出血状态。磁感应断层成像是非接触电阻抗成像方法，它既有时间敏感性、安全、无创等特点，又因非接触在应用上极为便利。.本项目从数据采集系统、时间序列化成像算法和模型成像实验等三个方面对磁感应断层成像连续监测技术进行研究。在数据采集系统方面，研制了高稳定信号源，优化了系统结构，形成了高精度的新型数据采集系统，整体数据采集精度优于0.3%。在成像算法方面，重点研究了利用多帧连续测量数据的线性卡尔曼滤波（LKF）算法和时间序列化一步高斯牛顿（TGN）算法等两种时间序列化MIT成像算法，并评估了TGN算法中各参数对重建结果的影响，然后对多种模拟出血情况进行仿真成像，结果表明TGN算法能够良好反映连续时间的电导率分布状态的变化过程，且目标定位准确、尺寸相当、电导率比例一致，是有效的算法。在上述基础上，编程实现了软件系统，完成了软硬件的联调，能够实现长时间的连续数据采集并进行实时差分成像，每帧数据采集最快为3s，从而构建完成了MIT实验样机。利用该样机开展了系列模型实验。物理模型实验的成像目标分别选择了空气域内的鸡蛋或油桃等生物组织，以及盐溶液中小体积扰动目标，均能获得MIT图像，且目标定位和大小基本准确，其中盐溶液中的5ml扰动目标可成像。动物模型实验对象为家兔，在自建的滑动式实验平台上，先对麻醉后的家兔腹部逐次注射3ml生理盐水后进行成像，结果可见电阻抗分布改变，位置准确，且重建值随注射剂量的变化曲线近似呈线性关系；再对家兔腹部注射2ml自体血液，结果可见电阻抗分布的改变。.总之，本项目系统研究了脑出血的磁感应电阻抗图像监测中的多项关键技术，为将来的临床脑出血预警提供了一种可行的新技术。