LABORATORY INVESTIGATION TISSUE CHARACTERIZATION Ultrasonic tissue characterization: detection of acute myocardial ischemia in dogs

Ultrasonic tissue characterization is a new area of investigation in the field of cardiac ultrasound. The amplitude and frequency of the ultrasound signal are normally altered as the signal penetrates through tissue. It is assumed that the amplitude distribution and frequency shift of diseased or edematous tissue are different than those of normal tissue. A statistical approach to the analysis of the unprocessed radiofrequency signal in the amplitude domain was used to study the effect of acute myocardial ischemia on the parameter mean amplitude/standard deviation of the amplitude (MSR). Ten dogs were anesthetized and underwent left lateral thoracotomy. Baseline mean MSR from the interventricular septum was 1.99 + 0.05, but increased by 30 min after coronary artery occlusion and started to plateau at 1 hr (mean 2.24 + 0.06). Reproducibility in noninfarcted myocardium (left ventricular inferoposterior wall) was good, with a mean MSR of 2.00 + 0.05 at baseline and 1.98 0.04 3 to 4 hr later. There was no difference in mean MSR when data were obtained through chest wall and when they were obtained directly from the surface of the heart. We conclude that statistical analysis in the amplitude domain of the unprocessed radiofrequency signal can detect acute myocardial ischemia within 30 min after coronary artery occlusion, provides reproducible measurements, and is unaffected by chest wall filtering. Circulation 72, No. 1, 193-199, 1985. DETECTION OF early acute myocardial ischemia in man is currently suboptimal. The electrocardiogram may not always reveal acute ischemia because it may demonstrate only nonspecific changes. The technetium pyrophosphate scintigram usually requires a delay of 24 hr before it can be interpreted as positive for myocardial injury. The information obtained from serum creatine kinase-MB levels is also delayed. A history of angina pectoris may be the only evidence of ongoing myocardial infarction in these patients. It would be highly desirable to have a noninvasive tool available for diagnosis and localization of early acute myocardial injury. Echocardiography has traditionally been used as an imaging device. Recent work by several investigators suggests that interactions between the ultrasound energy and tissue can be analyzed to characterize the histologic state of tissue, both in vitro1' 2 and in vivo.36 From the Cardiology Division, Stanford University School of Medicine, Stanford, CA. Supported in part by contract No. 15534-01-00 from Hewlett-Packard, Andover, MA. Address for correspondence: Richard L. Popp, M.D., Cardiology Division, Stanford University Medical Center, Stanford, CA 94305. Received Sept. 24, 1984; revision accepted April 18, 1985. Vol. 72, No. 1, July 1985 However, several quite different approaches to ultrasonic tissue characterization have been pursued. Miller, Mimbs, and Cohen and their colleagues 2, 4 have measured quantitated integrated backscatter from the unprocessed radiofrequency (rf) signal. In vivo, lack of an absolute calibration standard and the presence of the chest wall, with its variable acoustic characteristics, have been problems. Skorton et al.3' have analyzed the rf signal after it has been processed and displayed on the traditional image, applying various texture algorithms to extract information about the tissue. The potential disadvantage of image analysis is its dependence on the transducer and machine used to collect and process data and on the position of the region of interest within the scan field.7 We have previously worked with a statistical approach to the study of the unprocessed rf signal8 in which we analyzed the amplitude domain.9"10 We have shown that the measured parameter is independent of the measurement site when tested in a tissue-mimicking phantom. In the phantom there was no statistically significant difference in the parameter mean amplitude/standard deviation of the amplitude (MSR) whether it was sampled at a depth of 1 to 2, 5 to 6, or 9 to 10 cm.9 This 193 by gest on July 5, 2017 http://ciajournals.org/ D ow nladed from