Early Detection of Myocardial Infarction in Low-Quality Echocardiography

TL;DR

This paper presents a novel three-phase deep learning-based method for early myocardial infarction detection from low-quality echocardiography, including a new public dataset and high-accuracy segmentation and detection results.

Contribution

It introduces a highly accurate LV wall segmentation method, a pseudo labeling approach for ground truth creation, and the first public dataset for MI detection in echocardiography.

Findings

Achieved 95.72% sensitivity and 99.58% specificity in LV wall segmentation.

Achieved 85.97% sensitivity and 74.03% specificity in MI detection.

Publicly shared the HMC-QU dataset for future research.

Abstract

Myocardial infarction (MI), or commonly known as heart attack, is a life-threatening health problem worldwide from which 32.4 million people suffer each year. Early diagnosis and treatment of MI are crucial to prevent further heart tissue damages or death. The earliest and most reliable sign of ischemia is regional wall motion abnormality (RWMA) of the affected part of the ventricular muscle. Echocardiography can easily, inexpensively, and non-invasively exhibit the RWMA. In this article, we introduce a three-phase approach for early MI detection in low-quality echocardiography: 1) segmentation of the entire left ventricle (LV) wall using a state-of-the-art deep learning model, 2) analysis of the segmented LV wall by feature engineering, and 3) early MI detection. The main contributions of this study are highly accurate segmentation of the LV wall from low-quality echocardiography, pseudo labeling approach for ground-truth formation of the unannotated LV wall, and the first public echocardiographic dataset (HMC-QU)* for MI detection. Furthermore, the outputs of the proposed approach can significantly help cardiologists for a better assessment of the LV wall characteristics. The proposed approach has achieved 95.72% sensitivity and 99.58% specificity for the LV wall segmentation, and 85.97% sensitivity, 74.03% specificity, and 86.85% precision for MI detection on the HMC-QU dataset. *The benchmark HMC-QU dataset is publicly shared at the repository https://www.kaggle.com/aysendegerli/hmcqu-dataset