Automated method for characterization of diastolic transmitral Doppler velocity contours: Late atrial filling

• Published in: Ultrasound in medicine & biology Vol. 20; no. 9; pp. 859 - 869
• Main Authors: Hall, Andrew F., Aronovitz, Joseph A., Nudelman, Scott P., Kovács, Sándor J.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Inc 1994; Elsevier
• Subjects: Atrial Function; Atrial systole; Biological and medical sciences; Cardiovascular system; Diastole; Echocardiography, Doppler, Pulsed - methods; Humans; Image processing; Inverse problem; Investigative techniques, diagnostic techniques (general aspects); Left ventricular function; Mathematical modeling; Medical sciences; Parameter estimation; Pulsed Doppler echocardiography; Signal Processing, Computer-Assisted; Ultrasonic investigative techniques; Parameter estimation; Diastole; Image processing; Atrial systole; Pulsed Doppler echocardiography; Mathematical modeling; Left ventricular function; Inverse problem; Doppler ultrasound study; Echocardiography; Systole; Automatic analysis; Left ventricle performance; Mathematical model
• ISSN: 0301-5629; 1879-291X
• DOI: 10.1016/0301-5629(94)90046-9

We develop an automated method of characterizing the late atrial filling phase of diastole by fitting a kinematic model for diastolic filling to the clinical Doppler A-wave contour. The result is a set of model parameters which completely characterizes the contour. We have previously derived a parameterized diastolic filling (PDF) model, which predicts the time-dependent transmitral blood flow velocity obtained by Doppler echocardiography. An automated method to determine the PDF model parameters for early rapid filling from the clinical Doppler E-wave has also been developed and validated. The method consists of digitizing the acoustic Doppler waveform, recreating the Doppler velocity profile, extracting the maximum velocity envelope, and fitting the PDF model for early filling to the envelope. In the current work, we apply the same general approach for PDF parameter determination for the late atrial filling phase of diastole. To assess the presence and significance of near-degeneracies in the model parameter set, numerical experiments (consisting of fitting the model to a model-generated contour to which Gaussian noise was added) were performed. These revealed a two-dimensional degeneracy in four-dimensional parameter space which could be removed by using two kinematic simplifications: critical damping and resonant forcing. We show that these degeneracy-eliminating approximations do not limit the ability of the model to predict clinical A-wave contours.