Clinical value of two compounding techniques for IVUS Palpography

Abstract

This study aims at the in-vivo assessment of two compounding techniques for IVUS Palpography. To increase the elastographic contrast-to-noise ratio, Doyley et. al. proposed a compounding scheme for IVUS Palpography [1]. The neighboring IVUS frames acquired at diastole are paired to compute the luminal strain profiles or partial palpograms. Subsequently, the obtained strain maps are averaged to form a final compounded strain profile for a given cross-section of a coronary artery. This first scheme is further referred to as the classical compounding. The second scheme explicitly takes into account that the measured strains are only partially available. It attempts at reconstructing the missing elasticity values by using the available strain information in its direct vicinity using the Normalized Convolution method. The improved partial strain profiles are subsequently averaged in the same manner as in classical compounding. This scheme was coined as reconstructive compounding. Eight in-vivo IVUS pullbacks were used for the comparative analysis. The percentage of valid strains was 28.6±13.7% for the scheme without compounding, 94.3±4.4% and 99.7 ± 0.2% for the classical and reconstructive methods, respectively. Implementation of the compounding schemes significantly boosts the diagnostic information coming out of IVUS Palpography with the reconstructive scheme being the best.