In vivo ultrasonic tissue characterization of human intracardiac masses

Interactions between an ultrasonic signal and cardiac tissue have been used to characterize the histologic state of myocardium in vitro. To assess the utility of in vivo ultrasonic tissue characterization, stochastic analysis was applied to the digitized echocardiographic signals from 15 patients with 2-dimensional echocardiograms suggesting intracardiac masses. Ten subjects with echocardiograms suggesting mural thrombi underwent subsequent surgery or necropsy, which confirmed thrombi in 6 and revealed no thrombi (designated artifact) in 4. Five other patients had intracardiac tumors. The amplitudes within the digitized ultrasonic signals were displayed as histograms, which were described by a parameter k that represented the degree to which each histogram departed from a totally random probability density function. In 5 of 6 thrombi, k = 0, but in all 4 artifacts, k greater than 0. The sixth thrombus had k = 0.5 due to the specular effect of the interface between the thrombus' 2 lobes. All 5 tumors had k greater than 0. Ultrasonic tissue characterization using a stochastic analysis of backscatter can be performed in vivo and helps differentiate thrombus from artifact and tumor in the heart.