Greenleaf, J. F.

Diastolic dysfunction is characterized by the stiffening of the left-ventricular myocardium. An imaging technique capable of quantifying viscoelasticity of the left-ventricle would aid clinicians in evaluating diastolic function. Lamb wave dispersion ultrasound vibrometry (LDUV) is an ultrasound-based method for measuring viscoelasticity of plate-like materials by exciting Lamb waves and measuring change of wave velocity as a function of frequency (dispersion). Lamb wave dispersion equation is fit to the dispersion data to estimate mechanical properties. We report in vivo transthoracic studies aimed at quantifying elasticity and viscosity of a beating porcine left-ventricle using LDUV. A single ultrasound probe operating at 5 MHz was used for impulse excitation in the myocardium and to track the wave motion. A Fourier space-time analysis of the motion was used to obtain Lamb wave velocity dispersion. ECG-gated transthoracic in vivo measurements of group velocity, elasticity, and viscosity throughout a single heart cycle were obtained. Group velocity, elasticity, and viscosity in the frequency range from 50 to 500 Hz increased by at least two-fold from diastole to systole, consistent with contraction and relaxation of the myocardium.

Multi-frequency signals were transmitted using a plane-wave transducer, through a water immersed phantom consisting of two finger cots, one containing 50% isopropanol alcohol and the other containing normal saline. The scattered ultrasound was received with a small hydrophone and detected with a quadrature detection system. The resulting phase quadrature data were submitted to a reconstruction algorithm which utilized the Rytov approximation to the Helmholtz equation in order to obtain the inverse solution for the distribution of refraction index within the phantom. In one experiment pseudo-random noise was generated by loading a 256 by 8-bit memory with random numbers and then synchronously reading these values through a D/A converter into an EMI power amplifier which drove the plane-wave transducer. The data received from each of 60 angles of view were spectral analyzed and specific frequencies selected for the cw reconstruction algorithm. Multiple frequency reconstructions will be shown as will the effect of frequency on the various underlying assumptions for the approximation. [Supported in part by Grants CA 24085 and NCI and ECS 7296008 from NSF.]

Data and simulated data from circumscribing and linear sampled arrays has been used to form high resolution reflection images by two techniques. The first technique (synthetic focusing) uses the mathematical equivalent of a lense to focus the image at each point by adding either linear (uncompressed) or nonlinear (compressed) transforms of the data. Logarithmic and square root compression is studied. The lense effect is obtained by backprojecting convolved or other linear transformed versions of the data along equal time propagation surfaces in the image with appropriate corrections for refraction and frequency dependent attenuation. The choice of apodizing function and different wave forms can greatly improve image resolution. The second technique (reflection ART) is not based on a focusing model but rather on direct inversion of a matrix (linear transform) model of the object-acoustic field interaction and data acquisition process by a Kacmarz iterative inversion method. The two methods are compared.