Volumetric coronary plaque composition using intravascular ultrasound: three-dimensional segmentation and spectral analysis

Abstract

Intravascular ultrasound provides precise tomographic assessment of coronary artery disease, allowing unique potential for analysis of both plaque geometry and composition, two critical factors related to the likelihood of plaque rupture. A novel three-dimensional segmentation technique and spectral analysis are used to create a unique tool for volumetric assessment of plaque composition. The semi-automated 3D segmentation technique was used to identify luminal and medial-adventitial borders in ECG-gated images created from radiofrequency (RF) IVUS data acquired during automated pullbacks in patients. Spectral analysis was applied to the RF data within the segmented plaque. Color-coded pseudo-histology images were created from these plaque component predictions using statistical classification trees. Quantitative analysis and visualization techniques were used to assess volumetric plaque composition and provide a unique tool for evaluation of plaque vulnerability.