AI-Based Estimation of End-Systolic Elastance From Arm-Pressure and Systolic Time Intervals

Development and Clinical Application of Left Ventricular-Arterial Coupling Non-Invasive Assessment Methods

The constant and dynamic interaction between ventricular function and arterial afterload, known as ventricular-arterial coupling, is key to understanding cardiovascular pathophysiology. Ventricular-arterial coupling has traditionally been assessed invasively as the ratio of effective arterial elastance over end-systolic elastance (Ea/Ees), calculated from information derived from pressure-volume loops. Over the past few decades, numerous invasive and non-invasive simplified methods to estimate the elastance ratio have been developed and applied in clinical investigation and practice. The echocardiographic assessment of left ventricular Ea/Ees, as proposed by Chen and colleagues, is the most widely used method, but novel echocardiographic approaches for ventricular-arterial evaluation such as left ventricle outflow acceleration, pulse-wave velocity, and the global longitudinal strain or global work index have arisen since the former was first published. Moreover, multimodal imaging or artificial intelligence also seems to be useful in this matter. This review depicts the progressive development of these methods along with their academic and clinical application. The left ventricular-arterial coupling assessment may help both identify patients at risk and tailor specific pharmacological or interventional treatments.

The Impact of Left Ventricular Performance and Afterload on the Evaluation of Aortic Valve Stenosis: A 1D Mathematical Modeling Approach

The transaortic valvular pressure gradient (TPG) plays a central role in decision-making for patients suffering from severe aortic stenosis. However, the flow-dependence nature of the TPG makes the diagnosis of aortic stenosis challenging since the markers of cardiac performance and afterload present high physiological interdependence and thus, isolated effects cannot be measured directly in vivo. We used a validated 1D mathematical model of the cardiovascular system, coupled with a model of aortic stenosis, to assess and quantify the independent effect of the main left ventricular performance parameters (end-systolic (Ees) and end-diastolic (Eed) elastance) and principal afterload indices (total vascular resistance (TVR) and total arterial compliance (TAC)) on the TPG for different levels of aortic stenosis. In patients with critical aortic stenosis (aortic valve area (AVA) ≤ 0.6 cm2), a 10% increase of Eed from the baseline value was associated with the most important effect on the TPG (-5.6 ± 0.5 mmHg, p < 0.001), followed by a similar increase of Ees (3.4 ± 0.1 mmHg, p < 0.001), in TAC (1.3 ±0.2 mmHg, p < 0.001) and TVR (-0.7 ± 0.04 mmHg, p < 0.001). The interdependence of the TPG left ventricular performance and afterload indices become stronger with increased aortic stenosis severity. Disregarding their effects may lead to an underestimation of stenosis severity and a potential delay in therapeutic intervention. Therefore, a comprehensive evaluation of left ventricular function and afterload should be performed, especially in cases of diagnostic challenge, since it may offer the pathophysiological mechanism that explains the mismatch between aortic severity and the TPG.

Estimation of Left Ventricular End-Systolic Elastance From Brachial Pressure Waveform via Deep Learning

Determination of left ventricular (LV) end-systolic elastance (E es ) is of utmost importance for assessing the cardiac systolic function and hemodynamical state in humans. Yet, the clinical use of E es is not established due to the invasive nature and high costs of the existing measuring techniques. The objective of this study is to introduce a method to assess cardiac contractility, using as a sole measurement an arterial blood pressure (BP) waveform. Particularly, we aim to provide evidence on the potential in using the morphology of the brachial BP waveform and its time derivative for predicting LV E es via convolution neural networks (CNNs). The requirement of a broad training dataset is addressed by the use of an in silico dataset (n = 3,748) which is generated by a validated one-dimensional mathematical model of the cardiovasculature. We evaluated two CNN configurations: 1) a one-channel CNN (CNN1) with only the raw brachial BP signal as an input, and 2) a two-channel CNN (CNN2) using as inputs both the brachial BP wave and its time derivative. Accurate predictions were yielded using both CNN configurations. For CNN1, Pearson's correlation coefficient (r) and RMSE were equal to 0.86 and 0.27 mmHg/ml, respectively. The performance was found to be greatly improved for CNN2 (r = 0.97 and RMSE = 0.13 mmHg/ml). Moreover, all absolute errors from CNN2 were found to be less than 0.5 mmHg/ml. Importantly, the brachial BP wave appeared to be a promising source of information for estimating E es . Predictions were found to be in good agreement with the reference E es values over an extensive range of LV contractility values and loading conditions. Therefore, the proposed methodology could be easily transferred to the bedside and potentially facilitate the clinical use of E es for monitoring the contractile state of the heart in the real-life setting.

Determination of Aortic Characteristic Impedance and Total Arterial Compliance From Regional Pulse Wave Velocities Using Machine Learning: An in-silico Study

In-vivo assessment of aortic characteristic impedance (Z ao ) and total arterial compliance (C T ) has been hampered by the need for either invasive or inconvenient and expensive 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 non-invasive and clinically available. In this study, we present a non-invasive method for estimating Z ao and C T using cuff pressure, carotid-femoral PWV (cfPWV), and carotid-radial PWV (crPWV). Regression analysis is employed for both Z ao and C T . The regressors are trained and tested using a pool of virtual subjects (n = 3,818) generated from a previously validated in-silico model. Predictions achieved an accuracy of 7.40%, r = 0.90, and 6.26%, r = 0.95, for Z ao , and C T , respectively. The proposed approach constitutes a step forward to non-invasive screening of elastic vascular properties in humans by exploiting easily obtained measurements. This study could introduce a valuable tool for assessing arterial stiffness reducing the cost and the complexity of the required measuring techniques. Further clinical studies are required to validate the method in-vivo.