• IEEE Trans Ultrason Ferroelectr Freq Control · May 2004

    Comparative Study

    Model-based reconstructive elasticity imaging of deep venous thrombosis.

    • Salavat Aglyamov, Andrei R Skovoroda, Jonathan M Rubin, Matthew O'Donnell, and Stanislav Y Emelianov.
    • Department of Biomedical Engineering, University of Michigan, Ann Arbor, MI 48109, USA.
    • IEEE Trans Ultrason Ferroelectr Freq Control. 2004 May 1; 51 (5): 521-31.

    AbstractDeep venous thrombosis (DVT) and its sequela, pulmonary embolism, is a significant clinical problem. Once detected, DVT treatment is based on the age of the clot. There are no good noninvasive methods, however, to determine clot age. Previously, we demonstrated that imaging internal mechanical strains can identify and possibly age thrombus in a deep vein. In this study the deformation geometry for DVT elasticity imaging and its effect on Young's modulus estimates is addressed. A model-based reconstruction method is presented to estimate elasticity in which the clot-containing vessel is modeled as a layered cylinder. Compared to an unconstrained approach in reconstructive elasticity imaging, the proposed model-based approach has several advantages: only one component of the strain tensor is used; the minimization procedure is very fast; the method is highly efficient because an analytic solution of the forward elastic problem is used; and the method is not very sensitive to the details of the external load pattern--a characteristic that is important for free-hand, external, surface-applied deformation. The approach was tested theoretically using a numerical model, and experimentally on both tissue-like phantoms and an animal model of DVT. Results suggest that elasticity reconstruction may prove to be a practical adjunct to triplex scanning to detect, diagnose, and stage DVT.

      Pubmed     Copy Citation  

      Add institutional full text...

    Notes

    hide…

What will the 'Medical Journal of You' look like?

Start your free 21 day trial now.

We guarantee your privacy. Your email address will not be shared.