2H-1 In Vivo 3D Cardiac and Skeletal Muscle Strain Estimation

Abstract

In this study, BiPlane imaging was adapted for measuring strain in actively deforming tissue in three orthogonal directions. BiPlane imaging assures a sufficient frame rate (75-120 Hz) for accurate strain estimation. A coarse-to-fine iterative 2D strain algorithm using spatial correction and local stretching was implemented. Considering the huge amount of generated data, a fast interpolation scheme was implemented for measuring sub-sample and sub-line displacements. Assuming a 2D parabolic shape of the cross-correlation function, a straightforward and direct calculation of the displacements is possible. The strain estimation method was validated by means of a simulation study and phantom experiments. Rf-data were acquired with a 3D X4 matrix array transducer (Philips Sonos 7500) in BiPlane mode. In vivo verification in human skeletal muscle was performed. Furthermore, cardiac strain imaging was conducted using cardiac BiPlane data of dogs. In a pilot animal study, beagles with an induced valvular aortic stenosis were monitored. The Field II simulation was used for determining the accuracy and detectibility of the algorithm and revealed excellent correlation between applied and measured axial strain (SNR = 43 dB) for a window of 0.60 mm. Obviously, a lower SNR was found in lateral and elevational direction. The in vivo verification experiment in the skeletal muscles revealed similar cumulative axial strain curves (up to 8%) in both the azimuth and elevational direction. The shape of the strain curve matched perfectly with the curve of the measured force. The lateral strain values parallel to the direction of the muscle fibers matched the axial strain curves, whereas the shape of the lateral strain in the perpendicular plane differed due to anisotropy. Finally, strain images of the beagles were calculated. The beagle with the most excessive pressure gradient revealed a decrease of the radial strain. Furthermore, an elongated plateau in the radial strain indicated hypertrophy. In conclusion, 3D cardiac and strain estimation is feasible using a real-time 3D scanner. Additional validation studies of full 3D imaging modes are required to fully validate the technique