Sources and characterization of clutter in cardiac B-mode images

摘要

In echocardiography, clutter is one of the most problematic image artifacts, often obscuring ventricular borders and introducing stationary noise in blood flow measurements. Clutter in transthoracic cardiac images is widely understood to originate from reverberations and off-axis echoes. The objective of this work is to investigate the sources of clutter in cardiac images and their relative contributions. Real-time 3D raw echo data was acquired at a volumetric frame rate of 1 kHz and speckle tracking was applied to resulting images to determine the motion characteristics of clutter and adjacent myocardium. When clutter adjacent to the myocardial wall was tracked, the clutter and adjacent myocardium had similar displacements. When clutter farther from the myocardial wall was tracked, displacements were temporally and spatially complex and did not correlate well with any portion of the myocardium. In addition, principal component analysis (PCA) was applied to the raw echo data and resulting eigenvectors were used to isolate various motion patterns in the cardiac data. Results support the hypothesis that echoes from stationary structures, such as the ribcage and chest wall, are contributors to stationary clutter noise, while the myocardium is a dominant source of nonstationary clutter.