Tuning Multidomain Hemodynamic Simulations to Match Physiological Measurements

Abstract

In recent years, considerable progress has been made in creating more realistic models of the cardiovascular system, often based on patient-specific anatomic data, whereas comparatively little progress has been made on incorporating measured physiological data. We have developed a method to systematically adjust the parameters of three-element windkessel outlet boundary conditions of three-dimensional blood flow models such that desired features of pressure and flow waveforms are achieved. This tuning method was formulated as the solution of a nonlinear system of equations and employed a quasi-Newton method that was informed by a reduced-order model. The three-dimensional hemodynamic models were solved using a stabilized finite-element method incorporating deformable vessel walls. The tuning method was applied to an idealized common carotid artery, an idealized iliac arterial bifurcation, and a patient-specific abdominal aorta. The objectives for the abdominal aortic model were values of the maximum and minimum of the pressure waveform, an indicator of the pressure waveform’s shape, and the mean, amplitude, and diastolic mean of the flow waveform for an infrarenal measurement plane. The hemodynamic models were automatically generated and tuned by custom software with minimal user input. This approach enables efficient development of cardiovascular models for applications including detailed evaluation of cardiovascular mechanics, simulation-based design of medical devices, and patient-specific treatment planning.