In vivo detection of carotid plaque thrombus by ultrasonic tissue characterization

The purpose of the study was to determine whether ultrasonic tissue characterization could detect carotid plaque thrombus in vivo. Patients undergoing carotid endarterectomy were examined preoperatively and the ultrasonic tissue characterization findings were compared to those of optical microscopy of the removed plaque specimens. Ten of 15 patients studied had plaque thrombus. Ultra-ultrasonic tissue characterization entailed an analysis of parameters obtained from the power spectrum of backscattered ultrasound signals. Data were obtained with a nominal 10 MHz sector scanning transducer with an effective bandwidth of 3 to 13 MHz. The parameters were the slope and intercept derived from the linear regression of the normalized spectrum and total power (log of the integrated power of the normalized spectrum over the effective bandwidth). The combined effect of the three parameters was determined by discriminant function analysis and showed a significant difference (P < 0.05) between nonthrombus and plaque thrombus in a small sample of patients with advance carotid atherosclerosis. These parameters applied singly could not provide such a distinction. Correct classification of carotid plaque thrombus using the multiple-parameter analysis revealed a sensitivity of 90%, specificity of 80%, and accuracy of 86.7%. This study demonstrates that analysis utilizing a combination of multiple spectral parameters was able to detect carotid plaque thrombus in vivo.