An automatic Doppler angle analysis method for interventional blood flow velocity calibration

Abstract

Quantitative monitoring of blood flow velocity, is helpful to accurately evaluate the hemodynamic status of patients. However, the traditional blood flow velocity calibration method depends on manual operation, and it is not suitable for interventional blood flow velocity measurement. In this study, a new intravascular ultrasound (IVUS) catheter design combined with an automated Doppler velocity angle calibration method was investigated. Our ultrasound catheter will not have non-uniform rotation distortion (NURD), so the original IVUS images will be directly used to analyze Doppler angle. The deflection of ultrasound catheter is analyzed and simulated, and the formulas for estimating Doppler angle and calibrating blood flow velocity are given. The U-Net-based model segmented the lumen region from the image, and then used an ellipse to quickly fit the lumen boundary. A hemodynamic circulation circuit was constructed, and the measurements of ultrasound catheters and commercially available instruments were compared. The model achieves good segmentation performance, Dice similarity coefficient is 99.464%，Jaccard index is 98.934%, Hausdorff distance is 0.15780 mm. Statistical tests showed that the cosines of the true and estimated Doppler angles have a good correlation and consistency. The mean absolute and relative errors of the estimated Doppler angles are 1.49374° and 7.42654%, respectively. The mean relative error of calibrated blood flow velocity was 8.453925%. This study may have a good application prospect in ultrasound hemodynamic monitoring.