A spatiotemporal strategy for automated analysis of vasodilator response

Currently there is considerable interest in the use of ultrasound imaging to quantify flow-mediated vasodilatation as a metric of endothelial function. Continuous mid-sagittal imaging of brachial arteries after a flow stimulus provides an opportunity to use automated boundary detection strategies to fully characterize these vasodilator responses. The near and far wall medial-advential interfaces are the most reliable anatomic boundaries to use for monitoring change in vessel diameter. However, in any given frame it may be difficult to distinguish this boundary from other similarly oriented anatomic boundaries (e.g. blood-intimal interface) or image artifacts. To resolve these ambiguities, a temporal domain algorithm was developed to select among various candidate boundaries in any given frame to produce the most spatially continuous boundary surface in time