Patient-Specific Echo-Based Left Ventricle Finite Element Modeling And Applications

Background: Patient-specific image-based computational models have been widely used in cardiovascular research.Accurate evaluation of left ventricle(LV)function is of vital importance for the diagnosis,prognosis and therapeutic treatment of cardiovascular disease.In this paper,patient-specific echocardiographic(Echo)-based LV models were constructed and used to obtain accurate 3D LV morphological and biomechanical information,determine myocardial material properties,simulate ventricular active contraction and cardiac blood flow conditions,evaluate ventricular function,and provide helpful information for the diagnosis,prediction,and personalized treatment for patients with possible myocardial infarction.Methods: Left ventricle Echo and arm pressure data were acquired at the First Affiliated Hospital of Nanjing Medical University with patient constent obtained.Echo-based computational mechanical models were constructed,which included LV morphology,two-layer structure with fiber direction,nonlinear myocardial material properties,active contraction and fluid-struction interactions(FSI).The nonlinear modified Mooney-Rivlin material model was used to characterize the nonlinear isotropic and anisotropic materials.Active contraction and expansion of myocardium were modeled by material stiffening and softening.The geometryfitting mesh generation technique was applied to generate mesh,and the models were solved by the finite element software ADINA.Statistical methods including student t-test analysis,correlation analysis,linear mixed effect method and logistic regression model were used in the morphological,biomechanical and fluid dynamics data analysis.Echo-based left ventricle simulation model using different zero-load geometries introduced later in this paper,methods for determining myocardial material properties and predicting myocardial infarction,and methods for performing virtual surgery to optimize pacemaker position using simulation models are all innovative.Result: First,Echo-based structure-only LV models were used to estimate in vivo LV material properties.By analyzing the differences in morphology and mechanics between patients with myocardial infarction(MI)and healthy volunteers,it was found that patients with infarcted LV have stiffer LV material properties and lower material stiffness variations during the cardiac cycle.Secondly,using the logistic regression method as the predictive model and 12 parameters including patient morphology,mechanics and bioinformatics as predictors,results from the 10 patient data indicated that material stiffness parameters may be used as a potential predictor to suggest if a patient had infarction.The third effort was new modeling approach using two different diastole and systole zero-load geometries(2G)for 15 patients to properly model active contraction and relaxation to obtain more accurate stress/strain data.By comparing 2G model and one zero-load geometry(1G)model,the 2G model can obtain more accurate stress-strain values,and the material parameters from 2G models were stiffer than that from 1G model at the minimum volume,and softer than that from 1G model at the maximum volume.The next modeling effort was the construction of FSI models for three patients to study the cardiac flow dynamics in different pathologies.By analyzing flow characteristics and mechanical parameters,it was found that infarcted LV had lower maximum flow velocity,maximum shear stress,stress and strain,higher vorticity and smaller vortex area.Finally,as a clinical application,Echobased LV FSI models with different pacing modes were constructed to perform ventricle function analysis and investigate impact of pacemaker location on cardiac outcome using data obtained from an experimental pig model.The relationship between different pacing interventions and myocardial material function was analyzed to determine the myocardial analysis corresponding to specific pacing mode.The correspondence between material stiffness and electrical signal conduction was validated by the pig data.The preliminary results indicated that pacing implantation at the posterior interventricular septum may be a better pacemaker implantation scheme.Modeling results had good agreement with experimental measurements from the pig model.Conclusion: With the development of cardiovascular imaging technology and numerical simulation technology,it is possible to establish accurate individualized ventricular mechanical models for patients.The Echo-based 3D patient-specific computational model can effectively simulate in vivo condition of patients and perform virtual surgery under different heart surgical procedures to replace high-risk clinical surgical experimentation using real patients.