2019/05/13 by Vinny Davies, Umberto Noè, Davies, Vinny +13
Engineering · Medicine · #Applications (stat.AP) #Cardiomyopathy and Myosin Studies #Cardiovascular Function and Risk Factors #Elasticity and Material Modeling #FOS: Computer and information sciences #Methodology (stat.ME)
paper · pdf · doi:10.48550/arxiv.1905.06310
openalex publication_date 2019/05/13 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
A central problem in biomechanical studies of personalised human left\nventricular (LV) modelling is estimating the material properties and\nbiophysical parameters from in-vivo clinical measurements in a time frame\nsuitable for use within a clinic. Understanding these properties can provide\ninsight into heart function or dysfunction and help inform personalised\nmedicine. However, finding a solution to the differential equations which\nmathematically describe the kinematics and dynamics of the myocardium through\nnumerical integration can be computationally expensive. To circumvent this\nissue, we use the concept of emulation to infer the myocardial properties of a\nhealthy volunteer in a viable clinical time frame using in-vivo magnetic\nresonance image (MRI) data. Emulation methods avoid computationally expensive\nsimulations from the LV model by replacing the biomechanical model, which is\ndefined in terms of explicit partial differential equations, with a surrogate\nmodel inferred from simulations generated before the arrival of a patient,\nvastly improving computational efficiency at the clinic. We compare and\ncontrast two emulation strategies: (i) emulation of the computational model\noutputs and (ii) emulation of the loss between the observed patient data and\nthe computational model outputs. These strategies are tested with two different\ninterpolation methods, as well as two different loss functions...\n