Impact of Leaflet Geometry on the Acoustic Spectrum in Mimicking Aortic Stenosis

The non-invasive detection of an obstruction in fluid flow systems has applications ranging widely, including determining the severity of heart diseases. A quantitative study of the effect of leaflet thickness of model aortic valves is performed through assessing characteristic acoustic frequencies and their interactions. In the model setup, the flow of blood through the aortic valve is simulated with a system using water and a pulsatile pump. Flexible leaflets modeled after the tricuspid aortic valve are 3D-printed to represent different restrictions. Dynamic similarity between the model and actual human heart flow conditions is established by matching Reynolds and Womersley numbers. The sounds produced by the flow through the leaflets are measured with a contact microphone, mimicking the function of an electronic stethoscope. The sound signals are then decomposed into energy spectra showing the energy content across frequencies. From the associated Fourier transform, the bicoherence is calculated to evaluate interactions between frequencies. These results are then compared with those of real heart sounds spanning healthy to severe cases of aortic stenosis. In both this study and the real data, enhanced high frequency content and lower coherence levels are shown to correlate with more severe restrictions.