Estimating Left Ventricular Elastance from Aortic Flow Waveform, Ventricular Ejection Fraction, and Brachial Pressure: An In Silico Study

Although left ventricular end-systolic elastance (E-es) serves as a major index of cardiac contractility, a widely-accepted noninvasive estimation of E-es does not exist. To overcome this limitation, we developed a two-step inverse method that allows for its noninvasive estimation from measurements of aortic flow and brachial pressure using a previously validated one-dimensional model of the cardiovascular system. In a first step, aortic flow is set as the model input and the output brachial pressure is compared with the real values. Subsequently, the basic properties of the arterial tree are tuned according to an optimization algorithm. In a second step, the same optimization method is used to estimate the elastance parameters that produce an aortic flow waveform that matches the real one. Additional knowledge of the ejection fraction can allow for the accurate estimation of the entire P-V loop, including end-diastolic elastance. The method was tested on a database of 50 different in silico hemodynamic cases generated after varying cardiac and arterial model parameters. Implementation of the method yielded good agreement (r=0.99) and accuracy (n-RMSE=4%) between real and estimated values of E-es. Furthermore, a sensitivity analysis revealed that errors due to poor arterial adjustment and measurements are small (8% for E-es).