Is what you see in color intravascular ultrasound what you want to get?
T. Hiro
DOI: https://doi.org/10.1253/CIRCJ.CJ-10-1058
2010-11-01
Circulation Journal
Abstract:Using several sophisticated mathematical models, various novel color-coded methods of IVUS tissue characterization have been proposed because grey-scale IVUS has been widely recognized as an inadequate tool for estimating the tissue architecture of atherosclerotic plaque. There are now 3 commercially available color IVUS systems: IB-IVUS,1 VH-IVUS2 and iMap.3 The difference among these systems is how the radiofrequency (RF) signals from the plaque are processed. The time-series RF signals that are divided into many signal segments are first transformed into a frequency-domain field, from which some quantitative values are extracted according to the acoustic characteristics specific to each tissue component. These values are then color-coded to represent a plaque tissue map. The tissue components identified by these systems usually consist of 3 or 4 types of tissue, such as calcified, fibrous, lipidic or necrotic core areas, although the pathological definitions of these components differ slightly among the IVUS systems. IB-IVUS quantitatively obtains an acoustic power or energy for each segment of the RF signals to perform color-coding with 2 or 3 different thresholds in order to identify 3 or 4 different types of tissue. VH-IVUS identifies 4 types of tissue by the combination of 8 acoustic parameters obtained from the spectrum for each RF segment. The iMAP identifies 4 types of plaque tissue using similarity analysis of the spectrum by learning theory of neural network based on the hypothesis that each type tissue has a specific spectral shape. Therefore, IB-IVUS is a single-parameter system, whereas the others are multiparameter systems. Actually, there are not only these 3 systems, but other proposed methods of IVUS tissue characterization that are not yet commercially available, including angle-dependence analysis,4 attenuation-slope analysis,5 wavelet analysis,6 fractal analysis,7 self-organization mapping,8 and the modified k-nearest neighbor method.9 Each method has own merits and limitations; however, the overall accuracy of each for identifying plaque tissue components generally provides a sensitivity and specificity of approximately 80–90%. In this issue of the Journal, Kawasaki et al document their results of a validation study of a new IB-IVUS system that was recently developed on a different IVUS platform with more advanced signal processing equipment compared with previous versions, and found it had superb capability for identifying plaque tissue components.10 Interestingly, they demonstrated reasonable accuracy of the system in measuring the fibrous cap thickness by comparing it with optical coherence tomography, a result that could be considered as acceptable, especially when the cap was sufficiently thick. The trend to develop a more sophisticated IVUS modality for tissue characterization is beneficial for enhancing our understanding and clarifying of the atherosclerotic mechanisms within coronary plaque. However, we should also pay attention to the limitations as well as to ensure the proper future direction of the technology. The most fundamental limitation is that IVUS imaging is reconstructed from a reflected ultrasound wave. The ultrasound reflection is seen at the interface between 2 materials with different acoustic impedances, but not from the whole body of a single material. For example, if a material is completely homogeneous, the ultrasound image within that material is a black hole except at the border with the other material. In addition, no information can be obtained from areas with no or very weak ultrasound reflection; in other words, no reliable tissue characterization can be achieved of the area behind a complete ultrasound reflector such as a calcified mass, even if the algorithm can create some color-coding around that area. Furthermore, several acoustic principles, such as attenuation, angle-dependence, coherence, or diffraction, should be considered in ultrasound imaging, especially for very small materials of less than a couple of millimeters. It should not be forgotten that IVUS imaging is performed based on the hypothesis that the ultrasound speed is uniform within the whole plaque. The next important limitation among color IVUS systems regards the classification of the tissue being characterized. All the commercially available systems provide identification of 3 or 4 types of tissue; however, the histology of in vivo plaque has a more complicated, intermingled and graduated structure involving the 4 types. It has been reported that the pathologic diagnosis varies among pathologists as to the particular plaque that is the gold standard to validate color IVUS. Furthermore, there are many tissue components that cannot