Variability in tissue characterization of atherosclerotic plaque by intravascular ultrasound: a comparison of four intravascular ultrasound systems.

Abstract

Different intravascular ultrasound (IVUS) systems vary in their image presentation. The purpose of this study was to compare four IVUS systems in vitro to determine the accuracy of tissue characterization of atherosclerotic plaque compared with histology. Ninety-eight plaque segments from 23 formalin-fixed human iliac arteries were imaged in saline at room temperature with four different IVUS systems. To assess the accuracy of IVUS in describing plaque, three types of analysis were performed: (1) the ability to identify the presence and extent of lumen or plaque boundary; (2) sensitivity, specificity, and interobserver variability of IVUS in qualitatively identifying plaque components compared with histology; and (3) quantification of calcification. The synthetic aperture device had a lower sensitivity in identifying lumen and plaque boundaries (87%, 38% respectively) compared with other machines (96%-100%, 95%-100%). All three mechanically rotating systems had fair to good sensitivities for identifying calcification (57%-73%) or lipid filled areas (50%-83%). The sensitivity of discriminating fibrous tissue from fatty areas was low (39%-52%). The synthetic aperture system had a significantly lower sensitivity for identifying all three tissue types (4%-21%). There was significant interobserver variability (kappa value = 0.47-0.68) as well as machine to machine variability (kappa value = 0.52) for tissue characterization. Calcified areas were underestimated by System 1 (p < .05) and System 4 (p < .01) because of weaker echo reflections or poor image quality. There are significant differences in image representation among these four IVUS systems in the diagnosis of tissue components of complex atherosclerotic plaque. These variabilities should be considered when interpreting studies performed with different machines.