System for Early Detection of Myocardial Infarction
Degerli, Aysen (2019)
2019
Information Technology
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2019-05-24
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tty-201905211683
https://urn.fi/URN:NBN:fi:tty-201905211683
Tiivistelmä
Myocardial infarction (MI) is known as 'heart attack,' and it is a life-threatening worldwide health problem. Mainly, MI happens because of a blockage or plaque in one or more arteries that feed/s the heart muscle. Heart muscle consists of several layers. Mostly, the damage happens to the myocardium layer since it includes the cardiac cells. By the time, cardiac cells cannot get enough oxygen because of the blocked artery, and they start to die. The death of the cardiac cells accumulates, and myocardium tissue dies, at the end, MI happens. That is why it is critical to detect any abnormalities in the early stage so that the MI can be prevented.
Echocardiography (echo) is a non-invasive imaging technique that can monitor the heart in real time from different views using ultrasound technology. There are four sections in the human heart, and, the focus of this thesis is to detect problematic segments on the left ventricle (LV) section. Echo is used for this purpose since it can give hints about the global and regional wall motion of the LV. Cardiologists decide on the status of the patient by interpreting the LV motion in an echocardiogram. However, that causes subjective results which are prone to human error. Additionally, since all the segments on the LV are connected and move together, even an akinetic segment looks as if it somewhat moves, that adds up difficulties to the diagnosis as the outcome is not accurate enough, and the process of the diagnosis is slow. Hence, a system that can diagnose MI more quickly and more accurately is an urgent need.
In order to tackle this problem, in this thesis, a systematic approach for the early detection of MI from the wall-motion abnormalities is developed. In the analysis, 4-chamber view echos are used where the four sections of the heart are visible. The proposed algorithm first extracts the boundary of the LV muscle wall, divides the LV-wall into segments and then, measures the exact motion of each of the segment on the LV. Lastly, based on these quantitative results, the case is classified as healthy or MI. The system also enables the visualization of maximum motion displacements for each segment on the LV, a displacement curve to compare the motion of each segment and the measurement of LV Ejection-Fraction as output. These provide crucial information for the objective evaluation of the LV-wall and the early diagnosis of MI.
The results over the dataset show that the proposed method can achieve an elegant accuracy, sensitivity and, precision on MI detection and identification. Primarily, high accuracy is achieved with the echo records which have high-level of noise and poor temporal resolution, considering the failures of the state-of-the-art methods in such cases. The enhanced visual capabilities of the system can further be used to help cardiologists to improve the diagnosis of MI and, can also be used as an assistive tool to guide echo-technicians to record echos with good quality.
Echocardiography (echo) is a non-invasive imaging technique that can monitor the heart in real time from different views using ultrasound technology. There are four sections in the human heart, and, the focus of this thesis is to detect problematic segments on the left ventricle (LV) section. Echo is used for this purpose since it can give hints about the global and regional wall motion of the LV. Cardiologists decide on the status of the patient by interpreting the LV motion in an echocardiogram. However, that causes subjective results which are prone to human error. Additionally, since all the segments on the LV are connected and move together, even an akinetic segment looks as if it somewhat moves, that adds up difficulties to the diagnosis as the outcome is not accurate enough, and the process of the diagnosis is slow. Hence, a system that can diagnose MI more quickly and more accurately is an urgent need.
In order to tackle this problem, in this thesis, a systematic approach for the early detection of MI from the wall-motion abnormalities is developed. In the analysis, 4-chamber view echos are used where the four sections of the heart are visible. The proposed algorithm first extracts the boundary of the LV muscle wall, divides the LV-wall into segments and then, measures the exact motion of each of the segment on the LV. Lastly, based on these quantitative results, the case is classified as healthy or MI. The system also enables the visualization of maximum motion displacements for each segment on the LV, a displacement curve to compare the motion of each segment and the measurement of LV Ejection-Fraction as output. These provide crucial information for the objective evaluation of the LV-wall and the early diagnosis of MI.
The results over the dataset show that the proposed method can achieve an elegant accuracy, sensitivity and, precision on MI detection and identification. Primarily, high accuracy is achieved with the echo records which have high-level of noise and poor temporal resolution, considering the failures of the state-of-the-art methods in such cases. The enhanced visual capabilities of the system can further be used to help cardiologists to improve the diagnosis of MI and, can also be used as an assistive tool to guide echo-technicians to record echos with good quality.