基于心音时频特征参数的肥厚型心肌病分析研究

发布时间：2018-04-28 08:58

本文选题：肥厚型心肌病 + 心音　；参考：《西华大学》2017年硕士论文

【摘要】：随着当今社会医疗知识的普及,越来越多的病症已经被人们所熟悉。肥厚型心肌病(Hypertrophic Cardiomyopathy,HCM)作为一种高遗传性、高发病率、高危险性的病症,是诱发猝死、心力衰竭和房颤的重要因素,多年来受到了国内外的广泛关注,越来越多的医疗工作者和科研工作者投入到HCM的诊断与治疗研究中。由于受医疗水平的限制,我国HCM患者的筛查率并不高,绝大多数患者只有在发病时才会到医院就诊,错过早期干预的黄金期,导致病情加重,为家庭和社会造成了沉重负担。随着病情加重,患者会出现胸闷、气短、晕厥甚至猝死等症状。因此,建立健全HCM的早期诊断体系对于患者病情的控制具有十分重要的现实意义。心音包含了心脏瓣膜、各器官功能状态的基本信息,已经被广泛应用于心血管疾病的诊断与治疗研究当中。经过临床听诊显示,左室流出道梗阻的HCM患者可能出现第二心音分裂,更为严重者会出现因左室流出道湍流和二尖瓣返流等原因引起的收缩期杂音。因此,将心音的分析研究运用到HCM的诊断与治疗具有重要的临床价值,并且对HCM的筛查甚至人体健康普查具有重要意义。本文在研究室和第四军医大学第一附属医院西京医院肥厚型心肌病诊治与遗传咨询中心前期研究工作成果的基础上,将心音的检测分析研究应用于HCM的诊断研究当中,主要完成了以下工作:(1)HCM心音信号采集利用美国BIOPAC公司生产的16导生理记录仪在可视化环境下,实时采集并同步显示心音信号,不仅操作简单而且数据采集的质量较高。通过此系统采集了年龄在20~70岁的24名正常人(男15名,女9名)和30名HCM患者(男17名,女13名)三尖瓣、二尖瓣和主动脉瓣第二听诊区三个部位的心音数据,并通过筛选剔除低质量的心音,最终选择57例正常心音数据(24名正常人)和81例HCM心音数据(30名HCM患者)用于本研究中。(2)HCM心音信号预处理临床HCM心音数据采集过程中不可避免的会受到仪器内部结构和外部环境的影响,从而产生额外噪声。因此,本文首先分析了心音信号中噪声组成部分及主要来源,然后描述了心音信号的降噪方法和过程,根据各类噪声的性质,利用50Hz工频陷波器、数字带通滤波器和自适应小波阈值降噪法,消除心音中的工频干扰和额外噪声,以提高心音信号的信噪比,获得了较好的降噪效果。(3)HCM心音信号特征提取本文通过时域和频域对HCM心音信号进行特征提取,首先讨论了基于希尔伯特变换(HT)、归一化平均香农能量(NASE)、单自由度模型(SDOF)、变频同态滤波(FMH)四种方法对本文研究数据的特征波形提取效果,通过对比分析,FMH算法具有更好的特征波形提取效果。在特征波形提取后,利用局部峰值检测和自适应双阈值门限对HCM心音信号进行分段定位,定义了信号时域时间特征参数、时域能量特征参数;最后结合小波分析,利用小波包提取频域频带能量特征参数,研究了正常心音和HCM心音在VLF、LF、MF和HF频段上的能量分布,为后续临床HCM心音的特征研究奠定了基础。(4)临床HCM心音时频特征分析研究利用前文中的降噪方法和特征波形提取方法对筛选后的HCM心音提取特征参数,在时域和频域通过统计分析研究了HCM的心音特征。通过正常人与HCM的心音特征参数对比分析,讨论了HCM第一心音、第二心音的特征;HCM心音中杂音出现的部位、时间、强度以及频带范围;通过对静息状态下梗阻和非梗阻性HCM心音的时域能量和频域小波包能量频带参数对比分析,可得杂音的性质,从而初步判断左室流出道梗阻情况。这对于HCM的诊断有十分重要的临床价值。综上所述:本文在临床HCM心音数据采集、处理及分析上进行了全面深入的研究,通过临床HCM心音特征分析,初步描述了HCM患者第一心音、第二心音特征及杂音特征,同时结合临床症状进行了初步分析,为基于心音分析的HCM诊断研究奠定了临床基础。但由于数据量相对较少,只是进行初步分析。因此,今后的工作重点将是建立临床HCM心音数据库,提出更加有效的HCM心音分析方法,为建立完善的临床HCM诊断体系而努力。
[Abstract]:With the popularization of medical knowledge in today's society, more and more diseases have become familiar. Hypertrophic Cardiomyopathy (HCM), as a high hereditary, high incidence and high risk, is an important factor in inducing sudden death, heart failure and atrial fibrillation, and has been widely concerned at home and abroad for many years. The more medical workers and researchers have put into the research of the diagnosis and treatment of HCM. Because of the limitation of medical level, the screening rate of HCM patients in China is not high. The overwhelming majority of patients only go to the hospital when they are sick, miss the golden period of early intervention, cause the aggravation of the disease and cause a heavy burden to the family and the society. As the condition worsens, the patient will have symptoms such as chest tightness, shortness of breath, syncope and even sudden death. Therefore, it is of great practical significance to establish and improve the early diagnosis system of HCM for the control of the patient's condition. The heart sound contains the basic information of the heart valve and the functional state of each organ, which has been widely used in the diagnosis and treatment of cardiovascular disease. In the clinical study, the HCM patients with left ventricular outflow tract obstruction may have second heart sound mites after clinical auscultation, and the more serious are the systolic murmur caused by the left ventricular outflow tract turbulence and mitral regurgitation. Therefore, the analysis and study of the heart sound are of important clinical value in the diagnosis and treatment of HCM. And it is of great significance for the screening of HCM and the general survey of human health. On the basis of the results of the previous research work of the diagnosis and treatment and genetic counseling center of the hypertrophic cardiomyopathy in Xijing Hospital, the First Affiliated Hospital of The Fourth Military Medical University, this paper applies the detection and analysis of heart sound to the diagnosis and research of HCM. Work: (1) HCM heart sound signal acquisition and use of the 16 guide physiological recorder produced by American BIOPAC company in visual environment, real-time collection and synchronization display of heart sound signals, not only simple operation but also the high quality of data acquisition. Through this system, 24 normal people (15 men, 9 women) and 30 HCM patients (male 17) aged 20~70 years old (male 17) were collected. The heart sound data of three parts of the three cusp, mitral and aortic valve in second auscultation areas, and the selection and elimination of low quality heart sound, and the final selection of 57 normal heart sound data (24 normal persons) and 81 HCM heart sound data (30 HCM patients) were used in this study. (2) HCM heart sound signal preprocessing clinical HCM heart sound data acquisition process It is unavoidable to be influenced by the internal structure of the instrument and the external environment, thus producing extra noise. Therefore, this paper first analyzes the components and main sources of the noise in the heart sound signal, and then describes the noise reduction methods and processes of the heart sound signals. According to the properties of all kinds of noise, the 50Hz power frequency trap and digital bandpass filter are used. And adaptive wavelet threshold denoising method to eliminate the power frequency interference and extra noise in the heart sound to improve the signal to noise ratio of the heart sound signal, and get a better noise reduction effect. (3) the feature extraction of HCM heart sound signal is extracted by the feature extraction of the HCM heart sound signal in the time domain and frequency domain. The first discussion is based on the Hilbert transform (HT) and the normalized average. Shannon energy (NASE), single degree of freedom model (SDOF) and frequency conversion homomorphic filter (FMH) are used to study the feature waveform extraction effect of this paper. Through comparison and analysis, the FMH algorithm has a better feature extraction effect. After the extraction of the characteristic waveform, the local peak detection and adaptive double threshold threshold are used to divide the HCM heart sound signal. Segment location, the time domain time characteristic parameters and time domain energy characteristic parameters are defined. Finally, the energy distribution of frequency band energy in frequency domain is extracted with wavelet packet, and the energy distribution of normal heart sound and HCM heart sound in VLF, LF, MF and HF frequency bands is studied. (4) clinical HCM heart. The characteristic analysis of sound time frequency characteristics is used to extract the characteristic parameters of the HCM heart sound after the noise reduction and characteristic waveform extraction. In the time and frequency domain, the heart sound characteristics of HCM are studied by the statistical analysis in the time domain and frequency domain. The characteristics of the first heart sound of HCM, the second heart sound, and the HCM heart are discussed by the comparison and analysis of the heart sound characteristic parameters of the normal people and the HCM. The location, time, intensity, and frequency range of the sound in the sound, by comparing the time domain energy of the resting state with the non obstructive HCM heart sound and the wavelet packet energy frequency band parameters in the frequency domain, the properties of the murmurs can be obtained, thus the obstruction of the left ventricular outflow tract is preliminarily judged. This is of great clinical value for the diagnosis of HCM. To sum up, this paper makes a comprehensive and in-depth study on the clinical HCM heart sound data acquisition, processing and analysis. Through the analysis of the clinical HCM heart sound characteristics, the first heart sound, second heart sound characteristics and the murmurs characteristics of the HCM patients are preliminarily described, and the preliminary analysis of the clinical symptoms is carried out, which is established for the HCM diagnosis based on the heart sound analysis. The clinical basis. But due to the relatively small amount of data, it is only a preliminary analysis. Therefore, the focus of the future work will be to establish a clinical HCM heart sound database, and put forward a more effective method of HCM heart sound analysis, so as to establish a perfect clinical HCM diagnosis system.

【学位授予单位】：西华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R542.2;TN911.7

【参考文献】