Automatic Segment-Level Assessment of Regional Wall Motion Abnormality From Echocardiography Images

Regional wall motion assessment is critical for the diagnosis of coronary artery diseases, and is commonly performed using echocardiography images in clinical practice. However, manual assessment of regional wall motion is time-consuming and requires expertise. Currently, various works have been proposed to detect the existence of regional wall motion abnormality, whereas they do not provide a segment-level assessment which is essential for detailed diagnosis and treatment. In this paper, we propose a deep learning-based framework for automatic segment-level assessment of regional wall motion abnormality. We collected a dataset consisting of 198 patients each with three views in three modes. Experimental results show that our framework can detect segment-level abnormality with an excellent performance of sensitivity, specificity, and accuracy rates of 93.85%, 99.99%, and 99.73%, respectively, which has demonstrated the potential in clinical application. The dataset and code used in our study are released to the public [1].