Publication
Background In-vivo assessment of aortic characteristic impedance (Zao) and total arterial compliance (CT) has been hampered by the need for invasive methods to access simultaneous recordings of aortic pressure and flow, wall thickness, and cross-sectional area. In contrast, regional pulse wave velocity (PWV) measurements are noninvasive and clinically available. Given that PWV is strongly related to aortic stiffness (1), we assume that carotid-to-femoral PWV (cfPWV) and carotid-to-radial PWV (crPWV) may contain sufficient information to evaluate the elasticity of the ascending aorta. Concretely, here, we present a noninvasive regression method for estimating Zao and CT using cuf pressure, cfPWV, and crPWV.
Methods Gradient Boosting is employed for predicting Zao, and CT. The regressors are trained/tested using a pool of virtual subjects (n = 4833) generated from a previously validated in-silico model (2). The cross validation is performed using a 10-fold cross-validation (3).The population used has been previously generated (4) and reflects a wide range of hemodynamical properties and states.
Results Predictions had a high accuracy (Figure) achieving a normalized-RMSE equal to 6.24 ± 1.19% (r = 0.85, p < 0.001) for Zao, and 4.38 ± 0.36% (r = 0.97, p < 0.001) for CT, respectively. High errors were reported for high values of Zao due to the limited amount of similar data.
Conclusion The proposed approach constitutes a step forward to noninvasive screening of elastic vascular properties in human by exploiting easily obtained measurements. This study could introduce a valuable tool for assessing aortic stiffness reducing the cost and the complexity of the required measuring techniques. Clinical evaluation is required to validate the method in-vivo.