Prediction Model of Left Ventricle Myocardial Infarction Based on Echo Image Data

• 作 者: Longling Fan Jing Yao Chun Yang Zheyang Wu Di Xu Dalin Tang
• 机 构: Department of Mathematics,Southeast University,Nanjing,210096,China
• 出 处: 2015年中国生物医学工程联合学术年会
• 发表日期: 2015年11期
• 关键词: Tissue material property prediction ventricle model ventricle mechanics left ven

　　Determining ventricle tissue material properties and myocardial infarction(MI)noninvasively based on in vivo image data is of great important in clinical applications.Echo data were obtained form 10 patients.The patients were divided into Group 1(n=5,with infarct)and Group 2(n=5,without infarct).Echo-based patient-specific computational LV models were constructed to quantify LV material properties and identify predictors for presence of infarction.Systolic and diastolic material parameter values were adjusted to match echo volume data.The equivalent Youngs modulus(YM)values were obtained for each material stress-strain curve for easy comparison.LV wall thickness,volume,ejection fraction,diameter,height,material stiffness parameter values,circumferential and longitudinal curvatures,stress and strain values were collected for analysis.Logistic regression analysis was used to identify the best parameters for infract prediction.The LV stiffness in fiber direction at end-systole was the best single predictor among the 12 individual parameters with an area under the ROC of 0.9841.Computational modeling and material stiffness parameters may be used as a potential tool to suggest if a patient had infarction based on echo data.Large-scale clinical studies are needed to validate these preliminary findings.