QUANTITATIVE ELASTICITY IMAGING AND SOFT TISSUE CHARACTERIZATION USING TAGGED MAGNETIC RESONANCE IMAGING

Abstract

Magnetic Resonance (MR) elastography is a promising non-invasive method to quantitatively determine the mechanical properties of biological tissue. A novel quasi-static MR elastography imaging system, including new image acquisition system, tagged MR image processing software and inverse elasticity algorithms, is presented in this thesis. The new image acquisition system uses a MR compatible actuator to produce a tissue indentation sequence according to simulated electrocardiograph (ECG) signal. Conventional harmonic phase (HARP) image processing technique is extended for large motion tracking during indentation. Adaptive windowed harmonic phase (AWHARP) is proposed for rapid and accurate strain calculation from tagged MR images. A probabilistic approach was introduced to approximate heterogeneity of liver tissue from diverting stress-strain data, and a novel material stiffness reconstruction method was developed assuming hyperelastic liver tissue. Efficiency and accuracy of this quasi-static approach for quantitative elasticity imaging was demonstrated via extensive imaging experiments.