Ultrasound Transit-time-based Pulse Wave Velocity Estimation For Common Carotid Artery

Local pulse wave velocity(PWV)is a gold standard of quantitative evaluation of carotid arterial stiffness.Non-invasive quantifications of the arterial stiffness for the early diagnosis and effective prevention of cardiovascular events and related diseases are clinically important.The ultrasound transit time(TT)-based methods determine the local PWV from the time delay between the pulse waves(PWs)at two sites over a known distance along a segment of a common carotid artery(CCA).Owing to a simple principle without measuring the vessel diameter,thickness and elasticity,and the development of the ultrafast ultrasound imaging based on the plane wave technique,the transit time-based method is preferable for the detection of the local PWV of a CCA.However,the true parameter values relate to the PWV estimation are not known in clinics,and the parameter configurations in the ultrasound scanning system are inflexible.Hence,flexible and realistic phantoms or simulation models are required to objectively and systematically investigate and evaluate the influence factors and algorithm performance.In the TT-based method,the PWV of carotid artery is obtained based on the time fiduciary points(TFP),such as the 'maximum second derivative'(M2D),'maximum first derivative'(M1D).'intersecting tangent'(IT),and '20%threshold'(TH20).However,the reflected waves(RWs)come from the carotid sinus is liable to change with the arterial elasticity,which leads to various PW deformations at different beam positions along the arterial segment.This obscures the time delays between PWs,and worsens the accuracy of the TT-based PWV estimation.In addition,the PWs estimated from ultrasonic RF echo signals of the pulsatile vessel wall are usually affected by different interferences,such as speckle noise caused by backscattered echoes from inhomogeneous soft tissues,artificial movements generated by breathing,and quantization errors produced by discrete scanning,different algorithms for a filter may affect the accuracy of the time delay estimation,and then the PWV estimation,to different degrees.According to relative literatures and clinical practice,the key problems of the effects of the PW distortions caused by the RWs on the PWV estimation are studied.Therefore,the aim of this paper is first to build a realistic model of PW propagation along the carotid artery,and then improve the accuracy of ultrasound TT-based PWV estimation algorithms.The major studies and innovations are as follows:1.A PW propagation model incorporating incident and reflected wave propagations is proposed to provide a realistic data source for validation of TT-based local PWV estimation algorithms.First,the incident wave(IW)and RW at a certain position over a CCA are separated with the simultaneously measured PW and blood flow velocity from the ultrasound radio-frequency echo signals.Then,the propagating PWs along the CCA are modelled with the synthesizations of the estimated IWs and RWs.whose occurrences are delayed in opposite sequences according to a preset PWV.Results show that the model can not only characterize the effect of the wave reflection on the local PW propagation as its clinical manifestation,but also be used for the systematic evaluation of the performance of TT-based carotid artery PWV estimation.2.An ultrasound simulation model of PW propagation is established to provide a realistic data source for the investigation of the influence of ultrasound parameter settings and quantization errors on the PWV estimation.First,a geometric model of a CCA is built,a scatter phantom is built by putting scatterers into a geometry model of a CCA randomly and uniformly.According to the wall distensions in the established P W propagation model,the dynamic phantom is yielded by updating the scatterer positions according to the scanning order.Finally,the ultrasound RF echo signals for the PW propagation within a cardiac cycle are simulated based on Field II.Results show that the ultrasound simulation model can be effective as a data source for the validation of TT-based local PWV estimation algorithms,particularly the effects of RWs on the estimation performance.3.To solve the sensitivity of TFPs location in the upstrokes of PWs caused by RWs for TT estimation,a regional upstroke tracking(RUT)approach by finding the most similar TFP-centered region in the upstrokes is proposed to detect the time delay for improving the local PWV estimation.The process of the proposed approach is as follows:first,a TFP in the upstroke of the PW at the first beam location(reference PW)is identified.Then the TFP-centered region in the upstroke of the reference PW is considered as the template to track the most similar one in the upstrokes of the PWs at other locations(i.e.,comparison PWs).And then the sum of the square difference(SSD)between the template and the tracked region at each temporal shift is calculated to measure their similarity.Finally,the time delay between the reference and comparison PWs is defined as the temporal shift with the smallest SSD.The results show that the RUT algorithms with only three tracking points provide higher accuracy,precision,and reproducibility for the local PWV estimation.4.An empirical wavelet transform(EWT)-based adaptive filtering algorithm for rejecting the quantization error and RW in PW is proposed to improve the accuracy of the ultrasound TT-based PWV estimation.First,the upstroke of PW is decomposed into a set of empirical mode functions(EMFs)using an EWT algorithm.Then the frequency and energy spectrums of each EMF is calculated,respectively,and the noise signals(the quantization error or RW)are sifted out from these EMFs based on the maximum frequency and energy threshold method sequentially.Finally,the EMF that don't meet the sifting principle(either the maximum frequency or energy)and its previous EMFs are reconstructed,then the reconstructed signal is decomposed again and sifted out until the EMFs meet the sifting criterion.Results show that the proposed adaptive filtering algorithm can not only obtain higher accuracy of PWV estimation,but also have the unique and invariant filter parameters which is useful for clinical real-time signal detection and analysis.In conclusion,a PW propagation ultrasound simulation model incorporating incident and reflected wave propagations is firstly proposed in the present study to provide a realistic data source for validation of TT-based local PWV estimation algorithms.Then,a TFP-centered RUT method and an EWT-based adaptive filtering algorithm for rejecting the quantization error and RW in PW are proposed to improve the accuracy of the ultrasound TT-based PWV estimation.The research results can provide useful suggestions for improvements of the clinical real-time processing method of PWV estimation in carotid artery by ultrasound and the accuracy of PWV estimation with practical value.