Detección de isquemia de miocardio y estudio de la respuesta autónoma asociada mediante procesamiento de la señal de EGG

Myocardial ischemia, consequence of coronary artery disease, is the major health problem leading to myocardial infarction (MI), arrhythmias and death. Autonomic nervous system (ANS) regulates several physiological functions including the heart rate. Characterization of the complexity of the autonomo...

Descripción completa

Detalles Bibliográficos
Autor: Magrans Nicieza, Rudys
Tipo de recurso: tesis doctoral
Fecha de publicación:2016
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:español
OAI Identifier:oai:upcommons.upc.edu:2117/106278
Acceso en línea:https://hdl.handle.net/2117/106278
https://dx.doi.org/10.5821/dissertation-2117-106278
Access Level:acceso abierto
Palabra clave:Infart de miocardi
Enginyeria biomèdica
Àrees temàtiques de la UPC::Enginyeria biomèdica
Descripción
Sumario:Myocardial ischemia, consequence of coronary artery disease, is the major health problem leading to myocardial infarction (MI), arrhythmias and death. Autonomic nervous system (ANS) regulates several physiological functions including the heart rate. Characterization of the complexity of the autonomous nervous system (ANS) response by means of multifractal techniques over RR interval time series, and the analysis of such intervals before and during induced myocardial ischemia is valuable to understand the cardiac autonomic control mechanisms. To date, however, the use of such multifractal measures has been practically negligible for the clinic diagnosis of diseases such as myocardial ischemia assessment and others. In the same way, some indices derived from both the ventricular repolarization and depolarization intervals in the electrocardiogram (ECG) signal have been commonly used to detect ischemic heart diseases. This thesis has three main aims. Firstly, to characterize the nonlinear response of the ANS during induced myocardial ischemia by analysing the nonlinear dynamics involved in the fluctuations of the short-term RR interval time series. For that, fractal (mono and multifractal) methods and surrogate data techniques are going to be used. Secondly, to analysis ventricular repolarization and depolarization indices from the high-resolution ECG signal to identify which of these indices describe better myocardial ischemia events. The analysis of the indices is going to be based in terms of the individual statistical discriminant power, the effect size measured, and the interaction degree collectively. Thirdly, building a highly robust prediction model of ischemia and myocardial infarction (MI) by using machine learning techniques based on the previously identified measures. The database used in this thesis contains ECG signals from patients underwent elective coronary angioplasty to open obstructed coronary arteries. The symptoms occurred during a complete coronary occlusion by the angioplasty balloon inflation are similar to those found in patients suffering MI. Therefore, the coronary angioplasty represents an excellent model to study myocardial ischemia and infarction changes. The procedure of angioplasty practiced here is unique in the sense that the coronary occlusion duration was longer than usual. This fact has allowed studying transient myocardial ischemic events as well as the early phase of a MI. In a general way, results show an increase in both the multifractal complexity and the nonlinearity of the ANS response during induced myocardial ischemia. It has been interpreted as a beneficial adaptive mechanism for increasing coronary blood flow to the damaged zones of myocardium. The research done from the standpoint of the assessment of the autonomic changes by means of short-term RR interval time series analysis represents a novel approach for studying the database used here. All this, jointly with the particularity of the procedure of angioplasty practiced, provides a special relevance to the research developed here. On the other hand, it has been observed that several ventricular repolarization and depolarization indices are closely interrelated, providing in consequence little different value than others that have proven to be most significant in identifying patients with myocardial ischemia at risk of suffering heart attack. Finally, the methodology carefully planned to build the different prediction models has allowed the model built with the most important measures shows a better performance in comparison with that observed for models previously developed.