Fundamental mechanical limitations on the visualization of elasticity contrast in elastography

Abstract

Elastography is a new ultrasonic imaging technique that produces images (elastograms) of the elastic properties of compliant tissue. To determine the Young's modulus it is necessary to measure or estimate any five of seven relevant variables. In elastography, the measured quantity is the normal strain component in the direction of the applied load, and the three normal components of stress may be estimated using the modified Love's analytical models while assuming a value close to 0.5 (incompressible) for Poisson's ratio. The distribution of Young's moduli can thus be computed and displayed in the form of two-dimensional images called elastograms. The analytical models used for the estimation of the three normal components of stress assume that the target is semi-infinite and homogeneous in composition. The objective of this article is to determine some of the errors associated with the assumption of homogeneity of the target. Experiments using computer simulations were performed to study the efficiency with which elastograms display the contrast in the Young's modulus of a lesion or target, with respect to its background under certain conditions. It was observed (using the definition of contrast-transfer efficiency of elastography as the ratio of the elasticity contrast as measured from an elastogram, to the true contrast) that elastograms were consistently efficient in quantitatively depicting the elasticity contrast of hard lesions; however, they showed suboptimal contrast-transfer efficiency in cases of soft lesions in a hard background. In general, elastograms are efficient in displaying the elasticity contrast of hard or soft lesions which have a low contrast level with respect to the surroundings, irrespective of their size and location.