Quantitative Vascular Elastography: Stiffness and Stress Estimation for Identifying Rupture-Prone Plaques

Presented by: Steven J. Huntzicker

Supervised by: Professor Marvin Doyley

Computer Studies Building Room 426

Every 40 seconds, someone in the United States dies of cardiovascular disease, with many of these deaths occurring from the rupture of atherosclerotic plaques in the carotid and coronary arteries. These events cause thombi to form in the arteries, which prevent blood flow. Clots in the carotid artery may lead to stroke, while clots in the coronary artery may lead to myocardial infarction. Therefore, there is a need for an imaging technique that can identify rupture-prone plaques. One proposed method for this is ultrasound-based elastography, which maps the stiffness of the vessel based on the extent of observed tissue motion. Plaques have been shown to rupture when the internal stresses exceed 300 kPa, and conventional elastography measures strain, which is proportional to stress. Absolute values of stress are needed to determine the likelihood of plaque rupture. To obtain these stresses, one must measure Young’s modulus and all components of strain. Computing these using ultrasound is challenging due to the poor lateral resolution. The objective of this thesis was to investigate the feasibility of using a model- based approach to reconstruct the modulus and stresses within arteries using clinically available ultrasound systems. To achieve this goal the following objectives had to be satisfied: (1) Develop accurate inversion schemes for estimating the modulus distribution within vessels; (2) assess the impact of microvessels on the performance of the techniques developed in (1) and develop methods to overcome its limitations; (3) investigate the effects of material nonlinearity on the ability to visualize the stress distribution within vascular tissues; (4) assess the feasibility of producing clinically useful images. The results of these studies showed that the proposed reconstruction method could accurately compute modulus and stress elastograms. In all simulation studies, peak stresses could be recovered with under 15% error using the proposed reconstruction method. Phantom and in vivo studies were also able to accurately reconstruct these parameters despite the presence of measurement noise, attenuation, and sub-resolution features. This research may be useful in the further advancement of clinical vascular imaging.