Quantification of new structural features of coronary plaques by computational post-hoc analysis of virtual histology-intravascular ultrasound images

Cardiovascular disease and complications are often mediated by the development and rupture of atherosclerotic plaques. Plaque composition is a major factor that determines plaque vulnerability. Intravascular ultrasound (IVUS) and spectral analysis of the radio frequency signal provide an in vivo tissue characterisation of atherosclerotic plaques, known as virtual histology (VH–IVUS). In VH–IVUS analysis, four histological tissue components are classified: fibrous, fibro/fatty, necrotic core and calcium. Existing technology determines only the area of each component within the plaque. Quantitative, objective characterisation of other plaque components' patterns within the plaque is lacking. The aim of this study was to determine new compositional and structural indices which indicate spatial distribution, heterogeneity and dispersity of each VH–IVUS-derived component within the plaque area and also with respect to the plaque–lumen border. We developed an automated computational system in Java for the analysis of both single cross-sectional segments and the whole length of the examined plaque (volumetric analysis). The following parameters were computed: the number of different solid segments and the area of the largest solid segment of each component within the plaque, the per cent of the lumen border that is surrounded by each component, the number of different solid segments and the largest area of a solid segment of each component that adjoins the lumen border. Especially components' localisation in relation to the lumen border may significantly influence plaque vulnerability and plaque–stent interaction, which should be investigated in future clinical studies.