Non-invasive vascular elastography based on a new 2-D strain estimator: simulation and in vitro results

Previous studies showed that knowledge about the composition of diseased arteries can improve diagnostic decision making. Accordingly, stiffness of the vascular wall and size ratio between structures composing it were proven to be useful to characterize pathology evolution. In order to make it as a routine clinical exam, the non-invasive vascular ultrasound elastography method (NIVE) was recently developed. With NIVE, the blood pressure induces radial strain in the vessel wall that is tracked by ultrasound means. In the current study, displacement fields were computed with a new 2D pixel-based estimator that uses pre-and post-motion radiofrequency (RF) ultrasound images. Linear affine transformations considering pixel intensity variations were used in this method. To better interpret the 2D strain distributions, Von Mises elastograms were computed. The robustness of the method was investigated with simulations of RF signals obtained from a homogeneous tissue contaminated with decorrelation noise, and from simulations of healthy and pathological vessel geometries. In addition, in vitro experiments were conducted on vascular phantoms to validate the method. Results show a good agreement between theory and estimated data. The experimental study allowed us to identify different tissue structures with specific strain distributions around the vessel lumen. These results proved the potential of our estimator to compute elastograms of pathological vessels.