基于被动声信号检测的HIFU疗效评价方法研究

Evaluation of curative effect is an important problem in the application of high intensity focused ultrasound (HIFU) therapy, and its traditional methods usually need to introduce another source for monitoring besides the HlFU source. A new approach to evaluate HIFU curative effect based on the passive acoustic signal detection will be investigated in this project, which can avoid the influence of treatment source on the evaluation of curative effect, and it has important significance to achieve real-time evaluation. The separation method of harmonics and acoustic cavitation noise in the passive acoustic signal will be studied .Through analyzing the relative change of harmonic amplitude and the time-varying characteristics of harmonic components for various HIFU radiation intensity and HIFU radiation time, the size of effective treatment area and the degree of tissue damage will be assessed by using the harmonic characteristics. Through analyzing the features of broadband noise in time domain and in frequency domain, the representation mode of cavitation intensity and cavitation threshold will be explored, and the relationship between cavitation occurrence and the degree of tissue damage will be investigated. Combination of harmonic imaging and cavitation imaging , the thermal damage and cavitation damage will be evaluated at the same time. Finally, the theory support will be supplied for the established evaluation method in this project by numerical simulation. The results of this project can enrich and develop the method of HIFU curative effect evaluation, and improve the accuracy of the curative effect evaluation.

疗效评价是高强度聚焦超声治疗应用中的一个重要问题，传统方式通常需要在HlFU源之外再引入一个源用于监控。本项目研究基于被动声信号检测的HlFU疗效评价方法，避免了治疗源对疗效评价的影响，对实现HlFU疗效实时评价有重要意义。项目将研究被动声信号中谐波分量与声空化噪声的分离方法，通过分析谐波分量幅值的相对变化及各谐波分量的时变特性随HIFU辐照强度和辐照时间的变化规律，探索利用谐波特征评估有效治疗区大小及生物组织损伤程度；通过分析宽带噪声的时域和频域特征，探索空化阈值、空化强度的表达方式及空化的出现与组织损伤程度的关系；利用谐波信号进行成像和声空化信号进行成像，并将两者融合探索热损伤和空化损伤同时评价；最后通过数值仿真研究，为本项目建立的评价方法提供理论支持。本项目研究成果可丰富和发展HIFU疗效评价方法，提高疗效评价的准确性。

高强度聚焦超声（HIFU）技术已在肿瘤临床治疗中得到应用，目前其疗效评价主要依赖于B超或者磁共振成像，这些传统方式均需在HIFU源之外再引入一个激励源用于监控。本项目研究基于被动声信号检测的HIFU疗效评价方法，避免了治疗源对疗效评价的影响。主要研究内容有：1. 用不同方法对检测到的被动声信号进行去噪，针对信号特点通过改进已有算法提出了更有效的方法，同时研究了有用信号的分离和预处理方法。2. 基于熵的概念和频谱分析研究了被动声信号特征与组织损伤之间的关系。研究发现经过精细复合化处理后的信号熵，能更好地区分出正常组织与变性组织；AR模型的参数取值受组织损伤影响。3. 揭示了被动声信号中谐波分量的变化规律。当组织出现损伤后，HIFU产生的谐波分量幅值的改变量随着温度增加将减小。依据谐波分量的变化趋势，可以判定组织特性变化情况。4.研究了空化噪声及声空化效应阈值的影响因素。空化噪声的出现，与声源激励方式、空泡的存在有关。微泡振动的幅度和固有频率与其靠近的界面形状及特性有关；多频超声以适当方式组合可以降低声空化阈值。5. 利用深度学习方法研究了组织损伤的识别，提出了一种基于语义分割算法的超声图像目标区域分割方法。基于邻近治疗点对组织损伤的贡献提出了一种优化组织损伤区域的方法，研究了组织损伤的形成过程，可为被动声信号检测提供指导。项目研究成果丰富和发展了HIFU疗效评价方法，对提高疗效评价的准确性有重要意义。