General Departmental Seminar Series
A Nonparametric Method for Estimating Arterial Wall Shear Stress
John Carew, Graduate Student, Departments of Statistics & Biostatistics and Medical Informatics, University of Wisconsin - Madison
Friday, February 6, 2004, 12-1 p.m.
132 WARF Building, 610 Walnut St.
Arterial wall shear stress (WSS) is proportional to the blood velocity gradient at the arterial wall. Studies have shown that atherosclerotic plaques preferentially form in regions such as vessel bifurcations where WSS is relatively low. A non-invasive form of magnetic resonance imaging (MRI) that uses phase contrast is capable of measuring blood velocity within a vessel. These velocity measurements can be used to estimate WSS.
This talk describes a nonparametric method for estimating WSS from phase contrast MRI data. The approach is based on penalized least squares estimation in a reproducing kernel Hilbert space that is generated by the Matern family of radial basis functions. This approach does not require an a priori model for blood flow or assumptions about the vessel wall geometry. Estimation, model selection, and inference will be presented and illustrated with data. Some of the challenges, limitations, and future work will also be discussed.
This is joint work with Grace Wahba (Statistics), Reema Dalal (Biomedical Engineering), and Sean Fain (Medical Physics).
Return to seminar list