In vivo ultrasonic tissue characterization of human intracardiac masses

Clinical application of a new method that segments the region of interest into multiple layers for RF amplitude histogram analysis in the cirrhotic liver

“…Texture analysis using statistics from RF signal amplitude and histograms has been used to study myocardial tissue 23,24 and metastatic liver tumors, 25 in both of which the SNR is utilized. SNR was a useful parameter in the early diagnosis of myocardial ischemia in dogs 23 and the qualitative diagnosis of intracardiac tumors.…”

“…SNR was a useful parameter in the early diagnosis of myocardial ischemia in dogs 23 and the qualitative diagnosis of intracardiac tumors. 24 It was useless, however, for differenting metastatic liver tumors and normal liver tissue. 25 Skewness and kurtosis have not been used in any of the studies published in MEDLINE or Japana Centra Revuo Medicina between 1980 and 1999.…”

Ultrasonic Tissue Characterization in Cardiology

“…Specifically, quantitative analysis of the 2D distribution of echo am plitudes (referred to as tissue texture) has been shown capable of identifying a variety of abnormalities includ ing acute ischemic [26] and nonischemic [27] injury in dogs, and chronic cardiomyopathies in humans [28]. Analysis of the overall distribution of echo amplitudes has also proved useful in identifying the nature of intra cardiac masses (thrombi) in humans [29], An advantage of these techniques is that they may be carried out using relatively standard échocardiographie imaging systems. A major shortcoming of image/signal analysis ap proaches is that, similar to image display analysis, the techniques are dependent upon échocardiographie imag ing system adjustments, and thus may be difficult to standardize from laboratory to laboratory.…”

“…In an initial evaluation of the efficacy of échocardiographie texture analysis, Skorton et al [27] were able to identify acutely contused canine myocardium by quantitative texture analysis. Green et al [29] used statistical analysis of radiofrequency ultrasound signals, similar to that pre viously described by Joynt et al [35,36], to differentiate intracardiac thrombi and tumor from artifact in pa tients. Shaw et al [44] used color-encoded echocardio grams to evaluate regional echo amplitude characteris tics in patients with valvular, congenital (and ischemic) heart diseases.…”

Automatic textural feature selection on echocardiographic images

“…For these reasons, the segmentation of echocardiographic images continues being a problem, and several methods have been used to obtain an automatic result [3,14]. In general, two kinds of methods have been proven for segmenting echocardiographic images: methods based on textural analysis [4,5,8,9] and those based on direct analysis of the remote echo signal [6,7].…”