A comparison of digital and analog methods of Doppler spectral analysis for quantifying flow.

Ultrasonic methods can be used for calculating flow when the mean Doppler frequency is representative of spatial average velocity. We have examined the capabilities of two commercially available methods of Doppler spectral analysis for providing measurements of spatial average velocity and flow. In a steady state flow model, Doppler audio spectra were recorded using a 5-MHz duplex scanner. Fast Fourier transform (FFT) spectral analysis was used to determine mean (M), mode (MO), and maximum (MAX) frequencies. An analog method (offset zero crossing detector = ZC) was used to determine root mean square (RMS) frequencies. The results of comparing Doppler flow estimates (QM, QMO, QMAX and QRMS) with direct flow measurements (n = 10; range = 128-1098 ml/min) were (1) QM = 0.67Q + 23 ml/min (SEE = 36 ml/min); (2) QMO = 0.96Q + 152 ml/min (SEE = 32 ml/min); (3) QMAX = 1.19Q + 171 ml/min (SEE = 23 ml/min); and (4) QRMS = 0.93Q + 76ml/min (SEE = 92 ml/min). Estimates of flow using M and RMS frequencies were adversely affected by experimental conditions likely to result in turbulence. We conclude that application of commercially available FFT determined M frequencies could result in significant errors in calculations of spatial average velocity and flow. Alternatively, FFT determined MO frequencies and ZC determined RMS frequencies resulted in accurate estimates of flow in this model. This study demonstrates the importance of evaluating the capabilities of commercially available methods of Doppler spectral analysis when using ultrasound for determining velocity and flow.