User Tools

Site Tools


10a.002.TAU_WP2 - Early Myocardial Infarction Detection by Echocardiography

Project - Summary

World Health Organization (WHO) has recently reported that coronary artery disease (CAD) is the reason for 16% of total deaths worldwide [1]. Myocardial infarction (MI), commonly known as heart attack, is the most severe manifestation of CAD that leads to irreversible necrosis of the myocardium [2]. Accordingly, MI is the leading cause of mortality and morbidity in the world from which 32.4 million people suffer each year [3]. The early detection of MI is crucial to prevent fatal damages in the myocardium muscle of the heart. Echocardiography is the fundamental imaging technique for revealing any regional wall motion abnormality (RWMA), which is the earliest signs of MI. Assessing the motion of the left ventricle (LV) wall only from a single echocardiography view may lead to missing the diagnosis of MI as the RWMA may not be visible on that specific view. Moreover, the subjective and operator-dependent diagnosis suffers from the time-consuming process and low accuracy. Hence, there is a need for an automated, robust, and accurate tool to help cardiologists as diagnosing MI. In this project, we propose to fuse apical 4-chamber (A4C) and apical 2-chamber (A2C) views in which a total of 12 myocardial segments can be analyzed for MI detection. Considering the scarcity of echocardiographic datasets for the MI detection, which is the major issue for training data-driven classification algorithms, we propose a framework for early detection of MI over multi-view echocardiography that leverages one-class classification (OCC) techniques. The OCC techniques are advantageous over scarce datasets since in the training of the model only the instances from a specific target class are used. The project investigates the performance of uni-modal and multi-modal OCC techniques using the HMC-QU dataset that includes A4C and A2C views in a total of 260 echocardiography recordings. Experimental results show that the outperforming multi-modal approach achieves a sensitivity level of 85.23% and F1-Score of 80.21%.

Project - Team

Team Member Role Email Phone Number Academic Site/IAB
Moncef Gabbouj PI 358 40 073 6613 Tampere University
Serkan Kiranyaz Co-PI 97 43 063 5600 Tampere University
Aysen Degerli Researcher 358 46 521 9737 Tampere University
Christian Sundell Project Mentor N/A TietoEVRY
Iftikhar Admad Project Mentor N/A TietoEVRY

Project - Novelty of Approach

See “Project Summary” section above.

Project - Deliverables

1 Low-quality echocardiography restoration
2 Detection of MI over restored recordings

Project - Benefits to IAB

The proposed system will be used as an assistive tool to help cardiologists and technicians to prevent subjective and operator-dependent assessments by accurately measuring the LV wall motion. Moreover, the system will provide a time-efficient diagnosis, which will be a life-saving tool in critical situations. Additionally, the proposed system can be used as a verification tool when a group of cardiologists concludes the diagnosis differently. Then, the mismatching conclusions can be re-evaluated by the proposed system. Lastly, a multi-view dataset, HMC-QU that includes A4C and A2C echocardiographic views to detect MI is publicly shared with the research community.

Project - Documents

projects/year10/10a.002.tau_wp2.txt · Last modified: 2022/10/03 14:53 by sally.johnson