基于水脂分离自参考相位差模型的磁共振测温方法研究
发布时间：2018-12-27 20:42
【摘要】：磁共振图像引导高强度聚焦超声的无创热疗方法是肿瘤治疗的一种重要手段。在热疗过程中,磁共振仪的作用是提供目标组织断层图像和靶区空间温度图像,以引导聚焦超声焦点移动和监控靶区体素热剂量。目前磁共振技术能够提供良好的人体组织断层图像;但在测温准确性与时空分辨率方面,磁共振空间温度定量测量技术尚未达到热疗关于温度监控的临床要求。在各种磁共振测温方法中,质子共振频率(PRF)法应用最为广泛;然而,该方法在实际应用中易受磁共振仪主磁场漂移、被测体运动及脂类质子的影响而导致测温不准确。本文提出一种基于水脂分离自参考相位差的PRF测温模型,以解决上述PRF法的测温问题。通过分析被测体空间温度信息的磁共振获取技术,掌握质子共振频率测温方法并对该方法作具体的测温因子分析。针对主磁场漂移以及被测体帧间运动导致的测温误差,建立自参考相位差测温模型予以校正。该模型考虑主磁场即时强度变化和帧间运动的影响,将实时帧图像中未加热区域的体素相位作为参考相位,并以此生成拟合平面作为靶区PRF相位相减测温法的计算参考相位,该计算参考相位只与实时帧信息相关而消除帧间运动和磁场变化等时变干扰因素的影响。针对自参考相位差测温模型因脂类质子温度不敏感所带来的测温误差,拟对所获体素信号进行水脂分离以获取与温度具有线性关系的水质子相位。模型获得的水质子相位包括参考区及靶区水质子相位,参考区水质子相位用于拟合靶区温度计算的最后参考相位,靶区水质子相位用于和最后参考相位作差得到模型相位差值。为节约模型的数据采集和参数计算时间,提出根据参考区相位变化的阈值判据作断续水脂分离,拟提出理想模型下的矢量运算计算两次水脂分离间各帧图像各体素点的水质子相位,所得水相位将被用于各帧图像的自参考相位差模型测温。针对MR测温数据复数三角运算导致相位的2π截断现象即相位缠绕问题,将测温过程由检测相位复现出真实相位的相位解缠绕,分解为空间域和时域上的相位求真解问题。提出基于马尔可夫随机场最大后验的相位解缠绕算法用于空间域的静态相位求解,与基于温度-相位关系的先验模型用于时域相位的解算。为评估新模型的测温效果,使用Bstar 150T超导磁共振仪对不同水脂比仿体进行测温实验,并将测温结果与光纤测温结果相比对。实验结果显示,新模型对于水脂比为5:5的测试体的测温结果与光纤测温结果的最大差值为1.12℃,且测温过程无温度奇异点,说明相位解缠绕有效。本研究工作为磁共振仪进一步应用于临床热疗作了前期技术准备。
[Abstract]:Noninvasive hyperthermia guided by high intensity focused ultrasound (HIFU) is an important method for tumor treatment. In the process of hyperthermia, the purpose of Mr is to provide the target tissue tomograph and target spatial temperature image to guide the focus of focused ultrasound and monitor the thermal dose of voxel in the target area. At present, magnetic resonance technology can provide good image of human tissue, but in terms of temperature measurement accuracy and space-time resolution, magnetic resonance spatial temperature quantitative measurement technology has not met the clinical requirements of hyperthermia temperature monitoring. The proton resonance frequency (PRF) method is the most widely used in all kinds of magnetic resonance temperature measurement methods, however, it is easy to be affected by the drift of the main magnetic field of the magnetic resonance instrument and the influence of the body motion and the lipid proton in the practical application, which leads to the inaccuracy of the temperature measurement. In this paper, a PRF temperature measurement model based on the self-reference phase difference of water-lipid separation is proposed to solve the temperature measurement problem of the PRF method mentioned above. The method of proton resonance frequency temperature measurement is grasped by magnetic resonance (Mr) acquisition technique of the temperature information of the body measured in space, and the temperature factor of the method is analyzed in detail. In view of the temperature measurement error caused by the drift of the main magnetic field and the motion of the measured body between frames, the self-reference phase difference temperature measurement model is established to correct the temperature measurement. The model takes into account the influence of real time intensity change of the main magnetic field and the motion between frames. The voxel phase of the unheated region in the real time frame image is taken as the reference phase, and the fitting plane is used as the target PRF phase subtraction method to calculate the reference phase. The reference phase is only correlated with the real-time frame information and eliminates the influence of time-varying interference factors such as inter-frame motion and magnetic field change. In view of the temperature measurement error caused by the temperature insensitivity of lipid protons in the self-reference phase difference temperature measurement model, the obtained voxel signal is proposed to be separated from water and lipid to obtain the sub-phase of water quality which has a linear relationship with temperature. The water quality sub-phase obtained by the model includes the reference area and the target water quality sub-phase, the reference water quality sub-phase is used to fit the final reference phase of the temperature calculation of the target area, and the model phase difference value is obtained by the difference between the target water quality sub-phase and the final reference phase. In order to save the time of data acquisition and parameter calculation of the model, the threshold criterion of phase change in the reference region is proposed to separate water and lipid intermittently. In this paper, a vector operation based on ideal model is proposed to calculate the water quality sub-phase of each voxel point of each frame image between two water and lipid separations. The resulting water phase will be used to measure the temperature of each frame image using the self-reference phase difference model. Aiming at the phenomenon of 2 蟺 truncation of phase caused by complex triangulation of MR temperature measurement data, namely phase winding problem, the phase unwinding of the real phase is repeated from the detection phase, which is decomposed into the phase truth-seeking problem in spatial domain and time domain. A phase unwinding algorithm based on the maximum a posteriori of Markov random field is proposed to solve the static phase in spatial domain and a prior model based on temperature-phase relation to solve the phase in time domain. In order to evaluate the effect of the new model, the Bstar 150T superconducting magnetic resonance instrument was used to measure the temperature of different water-lipids, and the results were compared with those of optical fiber. The experimental results show that the maximum difference between the temperature measurement results of the new model and the optical fiber temperature measurement results is 1.12 鈩，
本文编号：2393579
[Abstract]:Noninvasive hyperthermia guided by high intensity focused ultrasound (HIFU) is an important method for tumor treatment. In the process of hyperthermia, the purpose of Mr is to provide the target tissue tomograph and target spatial temperature image to guide the focus of focused ultrasound and monitor the thermal dose of voxel in the target area. At present, magnetic resonance technology can provide good image of human tissue, but in terms of temperature measurement accuracy and space-time resolution, magnetic resonance spatial temperature quantitative measurement technology has not met the clinical requirements of hyperthermia temperature monitoring. The proton resonance frequency (PRF) method is the most widely used in all kinds of magnetic resonance temperature measurement methods, however, it is easy to be affected by the drift of the main magnetic field of the magnetic resonance instrument and the influence of the body motion and the lipid proton in the practical application, which leads to the inaccuracy of the temperature measurement. In this paper, a PRF temperature measurement model based on the self-reference phase difference of water-lipid separation is proposed to solve the temperature measurement problem of the PRF method mentioned above. The method of proton resonance frequency temperature measurement is grasped by magnetic resonance (Mr) acquisition technique of the temperature information of the body measured in space, and the temperature factor of the method is analyzed in detail. In view of the temperature measurement error caused by the drift of the main magnetic field and the motion of the measured body between frames, the self-reference phase difference temperature measurement model is established to correct the temperature measurement. The model takes into account the influence of real time intensity change of the main magnetic field and the motion between frames. The voxel phase of the unheated region in the real time frame image is taken as the reference phase, and the fitting plane is used as the target PRF phase subtraction method to calculate the reference phase. The reference phase is only correlated with the real-time frame information and eliminates the influence of time-varying interference factors such as inter-frame motion and magnetic field change. In view of the temperature measurement error caused by the temperature insensitivity of lipid protons in the self-reference phase difference temperature measurement model, the obtained voxel signal is proposed to be separated from water and lipid to obtain the sub-phase of water quality which has a linear relationship with temperature. The water quality sub-phase obtained by the model includes the reference area and the target water quality sub-phase, the reference water quality sub-phase is used to fit the final reference phase of the temperature calculation of the target area, and the model phase difference value is obtained by the difference between the target water quality sub-phase and the final reference phase. In order to save the time of data acquisition and parameter calculation of the model, the threshold criterion of phase change in the reference region is proposed to separate water and lipid intermittently. In this paper, a vector operation based on ideal model is proposed to calculate the water quality sub-phase of each voxel point of each frame image between two water and lipid separations. The resulting water phase will be used to measure the temperature of each frame image using the self-reference phase difference model. Aiming at the phenomenon of 2 蟺 truncation of phase caused by complex triangulation of MR temperature measurement data, namely phase winding problem, the phase unwinding of the real phase is repeated from the detection phase, which is decomposed into the phase truth-seeking problem in spatial domain and time domain. A phase unwinding algorithm based on the maximum a posteriori of Markov random field is proposed to solve the static phase in spatial domain and a prior model based on temperature-phase relation to solve the phase in time domain. In order to evaluate the effect of the new model, the Bstar 150T superconducting magnetic resonance instrument was used to measure the temperature of different water-lipids, and the results were compared with those of optical fiber. The experimental results show that the maximum difference between the temperature measurement results of the new model and the optical fiber temperature measurement results is 1.12 鈩，
本文编号：2393579
本文链接：https://www.wllwen.com/yixuelunwen/zlx/2393579.html
