Envelope Approximation on Doppler Ultrasound Spectrogram for Estimating Flow Speed in Carotid Artery

Abstract

The development of Doppler ultrasound allows us to convert the distribution of reflective ultrasound energy into colored blood flow imaging. The Doppler spectrogram also reveals the change of flow velocity over time when specifying an observation window inside the vessel wall. With the information doctors can visually search for abnormalities in blood vessels or valves from irregular patterns in the Doppler spectrogram within several cardiac cycles. Currently there are very few automatic analysis that can be applied to either color flow imaging or Doppler spectrogram. We hypothesize that the dynamics of flow speed in carotid could be an indicator to abnormalities on vascular wall like arteriosclerosis and stenosis. In this work the time-varying Doppler spectrogram during a carotid ultrasound exam was recorded as a video clip, and the image series were concatenated as a long spectrogram waveform image. To assess the carotid stiffness with previously developed nonlinear analysis and measures, the waveform image was further converted to 1-D signal by several proposed methods. Among them we found one that generated the most reasonable and less noisy curve, and expressed the dynamics of flow speed with minimal distortion.