Echo machine-imposed limit on transmitral spectral Doppler velocity- profile analysis

We have previously developed a kinematic model of ventricular filling. Its application to in vivo transmitral Doppler velocity profiles provides a quantitative characterization of filling. However, the model parameters computed by solving the 'inverse problem' may depend on ultrasound machine type and setting (e.g., gain, baseline filter, dynamic range). To determine machine-based effects on the computed model parameters, we performed a flow phantom study using Acuson and lip echocardiography machines at various settings. We compared maximum velocity envelopes (MVEs), as well as the model fit to these MVEs, for 3 simulated waveforms imaged by both machines. For all 3 waveforms, the machines generated comparable MVEs, fit by the model within a mean-square difference of 5E-5 (m/s)². The associated variations in model parameters for the 3 waveforms were not uniform. Two waveforms showed slight variation between machines, with model parameters varying by less than 6%. The shortest duration waveform showed model parameter variations of 10-15%. Analysis of the parameter space for this waveform showed a constant mean- square error contour that was larger than that for the other two, causing similar small variations in measured MVEs to result in larger differences in the parameter estimates for this waveform. Because this method completely eliminates inter- and intraobserver variability, we conclude that, within the limits established, the slight contour variations due to machine type and setting should not affect this method's applicability in clinical Doppler- flow analysis.