Characterization of elastographic noise using the envelope of echo signals

A theoretical formulation characterizing the noise performance of strain estimation using envelope signals is presented for the cross-correlation based strain estimator in elastography, using a modified strain filter approach. The strain filter describes the relationship among the elastographic signal-to-noise ratio (SNRe), sensitivity, contrast-to-noise ratio and dynamic range for a given resolution in the elastogram, as determined by the cross-correlation window length and window overlap. Theoretical results indicate that the envelope strain filter noise performance (SNRe level) is about half that obtained in the ratio frequency (RF) case (fo = 7.5 MHz). Simulation results corroborate the trend predicted using the strain filter. Experimental SNRe vs. strain plots presented in this article illustrate the same trend as the theoretical results. These plots allow a quantitative comparison of the elastograms obtained with RF and envelope signal processing. For small strains, the performance obtained using RF signals is superior to that obtained for envelope signals (since jitter errors are smaller due to the utilization of phase information in RF signals). However, for large tissue strains, envelope analysis provides an accurate estimate of the tissue strain (since envelope signal decorrelation is smaller than RF signal decorrelation at large strains). An algorithm that combines the low-noise characteristics of the cross-correlation analysis using RF signals at small strains and envelope signals for estimation of large tissue strains is proposed to improve the dynamic range in the elastogram.