Research Of Electrical Network Modeling Of Cardiovascular System And Diagnosis Of Arteriosclerosis And Arteriostenosis

The cardiovascular system is an important system which maintains the normalfunctioning of human body, and is also a high-incidence area of disease. Cardiovasculardisease has become one of the most dangerous threats to human life. About a third ofpeople around the world die of cardiovascular disease. Early diagnosis of cardiovasculardisease is very important to reduce the mortality and economic burden caused bycardiovascular disease. This paper mainly focused on four aspects: electric networkmodeling and simulation of cardiovascular system, diagnosis modeling ofarteriosclerosis and stenosis for early detecting cardiovascular disease, equipment R&Dof cardiovascular dynamics parameters and clinical trials.By comparing arterial system with electrical network system, one refined andsummarized the characteristics of lumped electrical network model. An electric networkmodel of upper limb arteries was established. And numerical method and MatlabSimulink/Simpowersystem module model were used to solve the model. The results ofsimulation are in good agreement with actual data. However, the model couldn’teffectively simulate pulse wave propagation and reflection.Based on the distributed electric network model of55human arterial tree, a novelcalculation method of the model was proposed with data lists and recursive algorithm,which can automatically complete the computation of model. The model and algorithmwere applied in the simulation of pulse wave propagation, blood pressure and flowwaveform of normal arterial tree, and the results rendered in3D figure which directlydisplay the properties of pulse wave propagation and reflection. The effects of differentphysiological parameters: height, heart rate, stroke output, arterial diameter and wallthickness on blood pressure and flow waveform were analyzed and discussed. We alsodiscussed the effects of different physiological parameters: arterial compliance,peripheral resistance, arterial length, arterial diameter and wall thickness of the arterialtree on input impedance. The results showed that different factors affect the bloodpressure and flow waveforms and input impedance in quite different way showing theirunique characteristics. The model provided us with an important auxiliary reference forphysiological and pathological diagnosis of human arterial tree.The distributed electric network model was applied to simulate atherosclerosis andarterial stenosis. The results showed that the model can accurately simulate: PWV changes with the degree of atherosclerosis; augmentation index of systolic bloodpressure changes with pulse wave reflection; blood pressure and flow waveforms wereaffected by arteriosclerosis parameters; ABI, blood pressure and flow waveforms wereaffected by the location, size and degree of artery stenosis; transfer function caneffectively reflect the location, size and degree of artery stenosis.A prediction model of artery stenosis was established by support vector machineusing input impedance as the characteristic parameters of arterial stenosis. Predictionresults showed that overall accuracies were more than82%under different stenosisdegrees. And, the accuracy increases with the degree of stenosis increasing. When thedegree of stenosis reaches to60%, the accuracy rises to more than90%. Additionally,the nearer the artery stenosis was to heart, the higher the accuracy was. The accuracy ofaortic arch stenosis approached to95%under the degree of stenosis60%. But, theaccuracy of femoral artery stenosis dropped down to about50%. It means the methodhas some limitations for the prediction of moderate stenosis far from heart. When thedegree of arterial stenosis increased to90%and above, the limitation disappeared forfemoral artery stenosis whose prediction accuracy increased to91.6%.A prediction model of artery stenosis was built using support vector machine withtransfer function as the characteristic parameters of arterial stenosis. Prediction resultsshowed that: the10cross-validation average accuracy of SVM reached to97.8%for thedegree of stenosis90%. For moderate and severe (50%and90%) artery stenosis,transfer function can predict accurately the artery stenosis situating between two pointsof transfer function, whose accuracy were more than87%and99%, respectively. But,the prediction total accuracy Q of carotid stenosis were62.1%and91.4%, sensitivity QPwere25.8%and18%, respectively. It means the stenosis was hard to predict for thestenosis outside transfer function.A new method of the positioning of arterial stenosis was proposed using transferfunction and the multi-classification theory of support vector machine. The positioningresults showed that: total accuracy Q was82.0%,80%,81.4%,83.7%and91.5%forartery stenosis50%,60%,70%,80%and90%, respectively. It means the artery stenosisabove90%can be positioned easily. But the positioning accuracies of carotid andsubclavian artery stenosis were85.7%and44.4%, which indicated the proposed methodhas some limitations for the artery stenosis outside two points of transfer function.An arteriosclerosis and arterial stenosis detection apparatus YF/XGYD-2000B wasdeveloped on LabVIEW development platform. It can detect a series of parameters and indicators of PWV, ABI, ASI, C1and C2. The comparison between YF/XGYD-2000Band noninvasive blood pressure simulator verified that YF/XGYD-2000B can measureblood pressure accuracy, and have good repeatability. Clinical comparative studydemonstrated that the results of YF/XGYD-2000B were in good agreement with thoseresults of Omron atherosclerosis apparatus. Comparison of itself parameters ofYF/XGYD-2000B indicated a good correlation between the parameters. Clinical trial ofcerebral infarction patients was implemented, and results showed PWV had a goodcorrelation with age, SBP, and ABI, which demonstrated the application value ofYF/XGYD-2000B.