The sonographic stenosis index: computer simulation of a novel method for detecting and quantifying arterial
Sherwin Chan1, Michael Felden McNeeley, Thomas Xuan Le
1Department of Radiology, University of Washington, Seattle, WA 98195, USA. sschan@uw.edu
Purpose:
We introduce a new technique for quantifying arterial stenoses, stenosis index (SI), using frequency analysis of the spectral Doppler waveform.
Methods:
Spectral Doppler waveforms obtained distal to arterial stenoses were enveloped, and the envelope was transformed into the frequency domain via Fourier transform. Stenosis index represents the ratio of the power in high-frequency harmonics to the power in the fundamental frequency.Stenosis index was compared with resistive index (RI), pulsatility index (PI), and acceleration time using 1000 computer simulations of waveforms from arteries distal to transplanted renal artery stenoses using Kendall tau correlation coefficients and receiver operating characteristic curve analyses.
Results:
The mean r's using Kendall tau correlation for SI, RI, and PI were 0.58, 0.31, and 0.45, respectively. The 95% confidence intervals for increase in r's for SI versus RI and PI to the degree of stenosis (none, mild, and severe) in the 1000 computer simulations were 0.22 to 0.34 and 0.07 to 0.19. The difference between the areas under the curves for SI versus RI and PI for normal artery versus mild stenosis was 0.26 to 0.44 and 0.078 to 0.24. Stenosis index and acceleration time were not significantly different.
Conclusions:
In computational models, SI outperformed RI and PI in correlation and receiver operating characteristic curve analysis. Further testing of SI is warranted.
