[Improved reproducibility of contrast echocardiography by SH U 454. Experimental studies using digital subtraction echocardiography].

The right heart chambers of 10 animals were contrasted by conventional (NaCl, CO2, H2O2, indocyanine green (ICG), haemaccel) and a newly developed echo-contrast medium (SH U 454) and studied by 2-D echocardiography. By means of digital subtraction echocardiography (DSE) endocardial borders were defined automatically and the results were compared with the manual input of endocardial borders of original and contrast echocardiograms. The area enclosed by these borders served as basis for the calculation of reproducibility (in %) and correlations. The following correlation coefficients (r) and SEE were calculated between the areas defined by the different contrast media and DSE and manually derived borders: r = 0.85, 3.98 cm2 (ICG), and 0.89, 1.00 cm2 (haemaccel). The best calculations were found using SH U 454 in concentrations between 100 and 300 mg/ml. The correlation coefficients were in the range of r = 0.95 and 0.98 with an SEE of 0.21 to 0.56 cm2 between manually and automatically derived contours. Comparing the reproducibility of data between the different evaluation methods we found the following results: manual input of endocardial borders in original echocardiograms 12.3%-16.9%; manual definition of endocardial borders in contrast echocardiograms 2.0% (SH U 454) - 15.7% (CO2); automatic contour finding in original echocardiograms 8.6%-28.9% (mean 21.6%); automatic definition of endocardium by DSE in contrast echocardiograms 7.6% (ICG) - 0.9% (SH U 454, 300 mg/ml). Our results demonstrate that digital subtraction echocardiography is a simple an safe procedure to define endocardial contours if echo contrast media lead to a uniform and homogeneous opacification of the left and right cardiac cavities.(ABSTRACT TRUNCATED AT 250 WORDS)