Aortic valve morphology and the double envelope phenomenon in continuous Doppler echocardiography of aortic stenosis: a computerized model

Abstract Introduction Visualization of a double envelope (DE) is a relatively common issue in aortic transvalvular continuous-wave Doppler (CW) traces. In 1997, Song et al proposed that velocity time integral (VTI) measured at the edge of the inner envelope (IE) might equal that of the left ventricular outflow tract (LVOT). This surrogate measurement would be especially interesting in patients with irregular heart rhythms, as both LVOT and valvular VTIs could be obtained from the same beat, thus avoiding inter-beat variability when assessing aortic valve area (AVA) with continuity equation. In contrast, based on clinical data and 3D-printed models, they raised concern that DE method in dome-shaped valves could lead to significant overestimation of LVOT velocities (about 40%). Later studies, including a Partner trial subanalysis, question the validity of this measurement. Purpose We created a computerized aortic valve model to reproduce the CW tracing, and analysed how valve morphology may lead to a valid, well defined inner envelope or, by contrast, produce overestimation of LVOT velocities. Methods The simulation code was created using Python (v3.9.6) along with NumPy and Matplotlib libraries. For each simulation, an aortic valve scheme was created (LVOT diameter: 2 cm; TSVI length: 1 cm; LVOT peak velocity 1.1 m/s; varying morphologies and lengths of the aortic cusps). The longitudinal axis between LVOT base and valve tip was divided in sections or levels (at least 300 per emulation). To recreate the CW trace, for each level, luminal diameter and peak velocity (according to continuity equation) were calculated, and an array of points was generated representing the aortic waveform during the entire systole. Slight random dispersion (standard deviation 0.05 times original values) was added to produce a more realistic image. Data points were represented as scatter plot over black background with transparency, simulating CW tracings (Figure 1) Results Figure 2 represents the plots of various aortic valve configurations (1st row AVA 1.6 cm2; 2nd: 1.2 cm2; 3rd: 0.8 cm2). Note that the innermost portion of the trace results in hollow space, as these velocities are produced in portions of ventricle and aorta which are not included in the simulation. Pink dashed line represents actual LVOT velocities. In a short-length flat valve (A), the IE is well defined and shows good correlation with LVOT velocity. In a cone-shaped 6mm length valve (B), there is a progressive brightness reduction, with no recognizable contour. With long (6 mm) dome-shaped valves, an ill-defined IE is produced at higher velocities than LVOT, which could result in overestimation. Conclusion DE method could accurately assess LVOT velocities when a sharp-defined IE is present and a dome-shaped valve can be excluded (e.g. rheumatic or some bicuspid valves). Measurements in a blurred DE or in a dome-shaped valve would produce overestimation of LVOT velocities and AVA. Funding Acknowledgement Type of funding sources: None.