Andrew F. Hall

Can Transmitral Doppler E-Waves Differentiate Hypertensive Hearts From Normal?

Physiological models of transmitral flow predict E-wave contour alteration in response to variation of model parameters (stiffness, relaxation, mass) reflecting the physiology of hypertension. Accordingly, analysis of only the E-wave (rather than the E-to-A ratio) should be able to differentiate between hypertensive subjects and control subjects. Conventional versus model-based image processing methods have never been compared in their ability to differentiate E-waves of hypertensive subjects with respect to age-matched control subjects. Digitally acquired transmitral Doppler flow images were analyzed by an automated model-based image processing method. Model-derived indexes were compared with conventional E-wave indexes in 22 subjects: 11 with hypertension and echocardiographically verified ventricular hypertrophy and 11 age-matched nonhypertensive control subjects. Conventional E-wave indexes included peak E, E, and acceleration and deceleration times. Model-based image processing-derived indexes included acceleration and deceleration times, potential energy index, and damping and kinematic constants. Intergroup comparison yielded lower probability values for model-based compared with conventional indexes. In the subjects studied, Doppler E-wave images analyzed by this automated method (which eliminates the need for hand-digitizing contours or the manual placement of cursors) demonstrate diastolic function alteration secondary to hypertension made discernible by model-based indexes. The method uses the entire E-wave contour, quantitatively differentiates between hypertensive subjects and control subjects, and has potential for automated noninvasive diastolic function evaluation in large patient populations, such as hypertension and other transmitral flow velocity-altering pathophysiological states.

Relationship of the Third Heart Sound to Transmitral Flow Velocity Deceleration

BACKGROUND The third heart sound (S3) occurs shortly after the early (E-wave) peak of the transmitral diastolic Doppler velocity profile (DVP). It is thought to be due to cardiohemic vibrations powered by rapid deceleration of transmitral blood flow. Although the presence, timing, and clinical correlates of the S3 have been extensively characterized, derivation and validation of a causal, mathematical relation between transmitral flow velocity and the S3 are lacking. METHODS AND RESULTS To characterize the kinematics and physiological mechanisms of S3 production, we modeled the cardiohemic system as a forced, damped, nonlinear harmonic oscillator. The forcing term used a closed-form mathematical expression for the deceleration portion of the DVP. We tested the hypothesis that our models predictions for amplitude, timing, and frequency of S3 accurately predict the transthoracic phonocardiogram, using the simultaneously recorded transmitral Doppler E wave as input, in three subject groups: those with audible pathological S3, those with audible physiological S3, and those with inaudible S3. CONCLUSIONS We found excellent agreement between model prediction and the observed data for all three subject groups. We conclude that, in the presence of a normal mitral valve, the kinematics of filling requires that all hearts have oscillations of the cardiohemic system during E-wave deceleration. However, the oscillations may not have high enough amplitude or frequency to be heard as an S3 unless there is sufficiently rapid fluid deceleration (of the Doppler E-wave contour) with sufficient cardiohemic coupling.

Beat averaging alternatives for transmitral Doppler flow velocity images

To characterize diastolic function from transmitral Doppler data, the images maximum velocity envelope (MVE) is fit by a model for flow velocity. To reduce the physiologic beat-to-beat variability of best-fit determined model parameters, averaging of multiple cardiac cycles is indicated. To assess variability mathematically, we modeled physiologic noise as a random (normally-distributed) process and evaluated three methods of averaging (1, averaging model parameters from single images; 2, averaging images; and 3, averaging MVEs) using clinical datasets (50 continuous beats from 5 subjects). Method 2 generates a positive bias because low-velocity beats will not contribute to the composite MVE. The difference between Methods 3 and 1 is less than 2.0 E-5 (m/s)2 for uncorrelated model parameters. Input having 10% beat-to-beat variation yields a bias of <4% for model parameter mean. Hence, Method 1 was, in general, more robust than Method 3.

Evaluation of model-based processing algorithms for averaged transmitral spectral Doppler images.

In an effort to characterize more fully diastolic function using Doppler echocardiography, we have previously developed an automated method of model-based image processing for spectral Doppler images of transmitral blood flow. In this method, maximum velocity envelopes (MVEs) extracted from individual Doppler images are aligned and averaged over several cardiac cycles. The averaged waveform is fit by the solution of a kinematic model of diastolic filling. The results are estimates of the model parameters. As expected, the mean and standard deviation of the model parameter estimates depend on many factors such as noise, the number of cardiac cycles averaged, beat-to-beat variation, waveform shape, observation time and the processing methods used, among others. A comprehensive evaluation of these effects has not been performed to date. A simulation was developed to evaluate the performance of three automated processing methods and to measure the influence of noise, beat-to-beat variation and observation time on the model parameter estimates. The simulations design and a description and analysis of the three automated processing methods are presented. Of the three methods evaluated, using the inflection point in the acceleration portion of the velocity contour as the first data point to be fit was found to be the most robust method for processing averaged E-wave MVE waveforms. Using this method under nominal conditions, the average bias was measured to be < 3% for each of the model parameters. As expected, the biases and standard deviations of the estimates increased as a result of increased noise levels, increased beat-to-beat variation and decreased observation time. Another important finding was that the effects of noise, beat-to-beat variation and waveform observation time on the parameter estimates are dependent on the location in model parameter space.

Processing parameter effects on the robustness of the solution to the "inverse problem" of diastole from doppler echocardiographic data

We have prevlously reported a metbod for automated quaa t i f lca t lon of diastollc fl l l ing parameters from Doppler echocardiographic data by fitting tbe traasmltral E-wave to n parameterlzed dIastollc filllng (PDF) model. Analysis ol the effect of two empirlcally set processlng variables, a spectral noise threshold and a vetoclty threshold, on the results of the PDF model f i t to cllnical data was performed. I t lndlcates that the mean-square error and the estimated variance of each of the model parameters, as a functlon of the processing varlables, suggest a means to objectively choose these variables. TLls makes the fitting process completely automated and objectlve. Furthermore, this process CPU also be used to Identify data (Ewaves) which are not good candidates for modeling. Therefore , ob jec t lve d e t e r m l n a t l o n of the proccsslng varlsbles lmcreases the robustness of our method and enhances Its utlllty lor direct cllnlcal appllcatloo.

Postextrasystolic left ventricular isovolumic pressure decay is not monoexponential

OBJECTIVE The relationship between the left ventricular (LV) relaxation time constant and early diastolic filling is not fully defined. This study provides additional evidence that LV isovolumic pressure fall in the normal intact heart in response to certain interventions is not adequately described by a model of monoexponential decay and that its relationship to filling is complex. METHODS AND RESULTS To gain further insight into the relationship between LV relaxation and early rapid filling we measured LV isovolumic relaxation rate, peak early filling velocity (E), LV volumes, and transmitral pressures at baseline and in the first postextrasystolic beat after a short-coupled extrasystole in 9 anesthetized dogs. Postextrasystolic isovolumic relaxation rate was slowed as measured by 3 commonly used time constants, while E was increased 32%. LV contractility and peak pressure were also increased, while LV end-systolic volume was decreased. LV minimum pressure was deceased, while the early diastolic transmitral pressure gradient was increased. Although all relaxation time constants measured over the entire isovolumic relaxation phase indicated slowed relaxation, direct measurement of isovolumic relaxation time indicated no change in relaxation rate. Calculation of the time constants and direct measurement of isovolumic relaxation time during early isovolumic pressure decay indicated slowed postextrasystolic pressure decay rate compared with baseline, while calculation of time constants and direct measurement of isovolumic relaxation time during late isovolumic relaxation indicated augmented postextrasystolic pressure decay rate versus baseline. CONCLUSIONS This non-exponential behavior of LV isovolumic pressure decay in postextrasystolic beats after short-coupled extrasystoles provides further evidence that the relationship that exists between ventricular relaxation and early filling is not simple. The results are interpreted in terms of current theoretical formulations that attribute control of myocardial relaxation to the interaction between inactivation-dependent and load-dependent mechanisms.

Doppler echocardiographic determination of mitral valvular resistance to inertiance ratio

Mltral valve resistance Is defined as the ratlo of the atrlo-ventrlcular pressure gradient to the volumetrlc l low across the valve In dlastole. The mltral valve lnertlance Is deflned as the product or the blood density and the votumetrlc flow rate. These physlologlc parameters have been usually determlned by lnvaslve methods such as cardlsc catheterization. We show that quantltatlve determlnatlon of the mitral resistance to lnertlance ratlo can be accompllshed by solution of the ‘Inverse problem’ of diastole, when the cllnlcal transmltral Doppler tradng Is analyzed using the parametrized dlastollc filling (PDF) formalism.

Comparison of audio and video data sources for quantitative analysis of echocardiographic Doppler velocity profiles

Computer based, off-line quantitative Doppler velocity profile image analysis can utilize either the audio or video formats of the echocardiography machine. The audio and video data are each corrupted, in some way, relative to the Doppler information from which they originate. To compare the suitability of each source for quantitative model-based image processing analysis, audio and video transmitral Doppler data were input to an algorithm for automated determination of diastolic filling parameters. Selected clinical examples were used to assess the sensitivity of the algorithm to variations in processing parameters in order to provide a measure of robustness for each data source. The frame-grabbed video image was less sensitive than the processed audio to variations in the algorithms spectral noise threshold setting.<<ETX>>

Model-based image processing of transmitral Doppler velocity profiles: toward automation

Transmitral Doppler ultrasound has not yet fulfilled its potential as a clinical tool for the noninvasive quantitation of left ventricular diastolic function. In the effort to achieve this goal we have developed a model-based image processing method for quantitative characterization of transmitral Doppler velocity profiles (DVP). The method fits a kinematic model for diastolic filling to the single-valued velocity contour extracted from the DVP image. Improvement of the method includes complete automation of the overall process. Two aspects of automation are examined: data set selection and beat averaging. An algorithm for E-wave (early diastolic filling interval) data selection and three methods for averaging are discussed.

Echocardiographic Doppler velocity contour computation from acoustic signal

Doppler velocity contour analysis has been hampered by uncertainty in determination of the contour. Current methods of contour determination are subjective and prone to observer dependent variations since they require manual tracing of the perceived contour. A computer-based method of Doppler velocity contour computation from the echocardiographic acoustic signal has been developed and is applied to the human transmitrai diastolic Doppler velocity profile.