Reproducibility of intravascular ultrasound virtual histology analysis

Objective: Intravascular ultrasound with virtual histology analysis (IVUS-VH) is a novel technology which allows for the identification of discrete atherosclerotic plaque components using radiofrequency backscatter data. Whether the composition of these plaques can be monitored accurately in a longitudinal fashion remains unclear. We sought to evaluate the reproducibility of plaque composition measurements as determined by IVUS-VH in a clinical setting. Methods: Sixteen consecutive patients who underwent percutaneous coronary intervention were included in this study. Prior to and then again following intervention, IVUS images with VH processing were obtained with motorized pullback. Up to 4 frames were selected for analysis for each patient, with a total of 24 IVUS frames used. An IVUS frame was chosen with visible angiographic and ultrasound landmarks. The matching frame on the second pullback was identified using these landmarks. For each frame, vessel and lumen area as well as plaque composition by VH were determined. Results: There was a high level of agreement between the two pullback measurements for lumen area, vessel area, and plaque burden (the Spearman rank-order correlation coefficients were 0.96, 0.96, and 0.95, respectively). Similarly, for plaque components by VH, the coefficients ranged from .90 to .97 and 0.84 to 0.92 for segmental volumetric analysis. The Bland-Altman plots indicated proportional error for the differences of the four measurements between the two pullback trials and were associated with high coefficients of reproducibility. Conclusions: Discrete measurements of plaque compositional area and volume in the clinical setting appear to have reproducibility comparable to that of traditional IVUS measurements.